<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Blog | Ayush Mangal</title><link>https://ayushtues.github.io/blog/</link><atom:link href="https://ayushtues.github.io/blog/index.xml" rel="self" type="application/rss+xml"/><description>Blog</description><generator>Hugo Blox Builder (https://hugoblox.com)</generator><language>en-us</language><lastBuildDate>Thu, 09 Jul 2026 00:00:00 +0000</lastBuildDate><image><url>https://ayushtues.github.io/media/icon_hu7729264130191091259.png</url><title>Blog</title><link>https://ayushtues.github.io/blog/</link></image><item><title>Do it for the plot</title><link>https://ayushtues.github.io/blog/do-it-for-the-plot/</link><pubDate>Thu, 09 Jul 2026 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/do-it-for-the-plot/</guid><description>&lt;p>(This turned out to be a very weird rant. I am writing a blog after 1 year, you can skip this)&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/do-it-for-the-plot/image-1.jpg"
alt="Well my bank now says I can’t be free of debt anytime soon">&lt;figcaption>
&lt;p>Well my bank now says I can’t be free of debt anytime soon&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Yours truly has been 25 for a while and is &lt;em>almost&lt;/em> about to turn 26. At this point, I feel like my life has taken a form which I never even imagined to be possible, for better or worse. I have also started seeing signs of a quarter crisis in my behaviour. Randomly trying to do a startup, going to Japan solo randomly, deciding to go for a master&amp;rsquo;s after a random conversation with a senior, it&amp;rsquo;s been a weird year. If I look back at the last year and try to figure out my rationale for doing the things that I did, I find no real reason other than what one of my dear friends might say — &lt;em>Just do it for the plot&lt;/em>. ( I used an em dash; I DID, not ChatGPT, you judgemental prick).&lt;/p>
&lt;p>On paper, and in reality as well, my life as a 25 or so year old in BLR couldn’t have been better. I was employed in one of the hottest AI startups in India, getting paid more than I knew what to do with, I had a fairly active social life, and my health had never been better. I also had enough free time to pursue whatever hobbies I wanted to and also study whatever I wanted while getting paid. I was getting better and better job opportunities, both in India and then increasingly abroad, and I was starting to explore the founder/startup ecosystem in BLR and attracting interest from some really cool people.&lt;/p>
&lt;p>Why, then, oh why, did I decide to throw it all away? Why did I decide to move to a different country, zeroing my savings and racking up massive debt and dealing with job uncertainty, away from all my friends and family and everyone I care about and love?&lt;/p>
&lt;p>So much, just for a plot twist, eh?&lt;/p>
&lt;p>Even though I am &lt;em>almost&lt;/em> sure I have aphantasia ( the lack of visual mental imagery, comically manifested by my total inability to imagine a mental abacus while being put in an abacus class in my childhood for 3 years straight), I kinda have this mental image where I had the realisation that I need a plot twist in my life. The image is of me sitting in a pretty cafe in South Delhi (or in Indiranagar/HSR in BLR or Jubilee Hills in HYD) and sipping my hot chocolate (Blue Tokai ftw for this) and reading a book on my Kindle, without any care in the world. Adjusted to my own needs (which are not really bourgeois), money has sort of lost its importance; I can go on international trips, I can buy whatever I want, eat wherever I want. Any form of art is made to inspire you to ask questions, and the only question I can think of when I ponder this particular imagery is — &lt;em>Is this it? Is this all I want to do in life?&lt;/em> — Cue an immediate episode of a quarter-life crisis.&lt;/p>
&lt;p>So for a while, I did what any self-respecting 24–25-year-old techie would do when faced with a quarter-life crisis. I tried to find meaning in new hobbies. I never thought I could sing, so I joined a singing class and also learnt guitar, keyboard and cajon. Although it ended up with me not being able to do any one of these even semi-decently, I now know I am at least capable of doing any of these if I just put in effort. I never thought I could do a pull-up, so I joined a callisthenics class and broke my front tooth, but I can now easily do one or 10 (okay, maybe not 10 &lt;em>that easily)&lt;/em>. Oh, I have also been practising doing handstands for the last 2 years and can kinda get away with a decent one about 20% of the time.&lt;/p>
&lt;p>I got bored with my 9–5 job, (although it was more like a 10–10 job 🙈 ), even though it was full of exciting problems ( my task for a day once was to recreate the one piece Wano trailer using a video model, or to listen to ASMR audios to teach our models to whisper or to try to replicate the voice of some of the most famous people in India or to figure out if we can do GRPO on TTS models to improve their pronounciations) and insane amounts of ownership ( I directly reported to the CEO, and had practically infinite freedom to approach problems however the fuck I wanted and a semi-decent amount of GPUs to play around with). I got approached by recruiters daily and had to explicitly turn down many lucrative offers regularly (my peak moment was saying F you to a big HFT firm in India cause they refused to pay me for a take-home assignment ).&lt;/p>
&lt;p>Oh and I also travelled a bit. I was totally blessed to have found an awesome friend group at MSFT, and have been to Vietnam, Pondicherry and Ladakh with them over the last 2 years. I also ended up living the otaku dream on my trip to Japan, &lt;em>alone.&lt;/em> Needless to say, all the travel needed money, which my job provided for, and I also was lucky enough to have employers who respected holidays ( I once took a holiday on a day which turned out to be one of the worst messes in our company’s history, and to their credit, they didn’t call me even once )&lt;/p>
&lt;p>Why oh why, did I leave this life? Answer 1 — Cause I am an idiot. Answer 2 — Just for the plot?&lt;/p>
&lt;p>The above 2 answers are probably the real answers. But if someone asks me to give an answer which sounds better and more motivating, I think I have one.&lt;/p>
&lt;p>I realised at 25 that no matter how badly I fuck up my life, I can always come back to India, take some tech job and earn my daily expenses since India is fucking cheap, (unless you have a health issue or want to start a family). I have zero family responsibilities being the youngest kid; nobody in the whole world depends on me, and I really don’t have many materialistic needs ( after seeing my apartment, a friend commented that I live like someone earning 20x less than what I do ). My parents are middle-class but well off, and can comfortably provide me a bed in their home whenever I want them to. They are insanely loving, supportive and will fight against the world for me if needed.&lt;/p>
&lt;p>And what do you do when you have someone who has your back? You bungee jump (10/10 experience btw, totally recommended, I remember feeling like, if I end up doing this, no other risk would feel big in the future, and that has been partially true). You jump without fear ( but a lot of adrenaline ) from a cliff, and let gravity and the inertia of your motion guide you, oh, and you enjoy the view. Once you are safe at the bottom and safe again, all you want to do is try again.&lt;/p>
&lt;p>I think my master represents exactly that. I am trying to do the most moonshot thing I can imagine other than opening my own startup ( for which I don’t have a cofounder or an idea I want to spend 3 years on ), and it&amp;rsquo;s also something I have regretted not doing a lot of times over the last 4 years. Is it the best financial decision? Definitely not. The best career decision? Probably not. A decision my parents are happy about? Nopes. A decision I am happy about? Kinda.&lt;/p>
&lt;p>The kid inside me has always wanted to study in a nice college in the US. The kid inside me also finds robots to be cool and wants to be a mechanical or electrical engineer. The kid inside me loves AI, but not LLMs. The kid inside me is also obsessed with the US, having watched unhealthy amounts of History TV 18 in my childhood. The kid inside me also just really loves to study and do research. The kid inside me also loves college and wants to escape from adulting. Master&amp;rsquo;s in Robotics at CMU is the kid taking over the reins for a little while, kicking the startup tech bro Ayush outside for a while.&lt;/p>
&lt;p>I guess I will turn 28 whether or not I go for a master&amp;rsquo;s. My life will have some plot regardless, and knowing myself, it won’t be straightforward. Might as well try to increase the entropy of the universe while I am at it. If I have one guiding rule over the years, let it be — Just do it for the plot 😛&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://ayushtues.medium.com/do-it-for-the-plot-c86fe11efb55" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>ADHD on a train ride</title><link>https://ayushtues.github.io/blog/adhd-on-a-train-ride/</link><pubDate>Thu, 13 Mar 2025 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/adhd-on-a-train-ride/</guid><description>&lt;p>So here I am, on a train ride, visiting my home after a few months, for Holi. I am on official leave from my work and its supposed to be a chill af day, with a lot of relaxation, and mental peace.&lt;/p>
&lt;p>Well yeah, tough luck.&lt;/p>
&lt;p>Now I don’t want to self diagnose myself as ADHD, even though a lot of people have commented me to be so, but we don’t really do diagnosis in India (cause well, mental health is a myth, and waking up early is the only thing you need to solve all your problems, guess what I already wake up at 5 fucking AM, and I still have issues, so fuck that theory at the very least.)&lt;/p>
&lt;p>Anyway, as you can see from my other posts and also this post as well, I find it super hard to focus my attention on anything for a long time, a long time here just means 2 mins, which is super weird, because while I have the attention span of a goldfish, I also like to read books a lot, and not tiny, smutty YA romance (even though that &lt;em>is&lt;/em> my favorite ) but also dense fiction or hard core sci-fi ( Asimov’s Foundation trilogy has become a recent obsession, only to be devoured in its entirety in a few days, with nothing left to fill the void ). And reading books like this needs attention, a LOT of attention, and somehow I am able to spend an entire day reading books, but not able to spend like 10 mins alone with myself, in a fucking train ride to home.&lt;/p>
&lt;p>In the last 40 mins of sitting in this train, I have, read parts of a book from my kindle, did some work from office ( on my leave day, by myself, cause well ), and then due to the slow internet speed retired myself to pouring my thoughts on this website, to fucking calm myself down a bit. All the while also twirling around with my phone and changing songs like every 2 mins.&lt;/p>
&lt;p>Yeah, ADHD much&lt;/p>
&lt;p>It kinda makes it hard to be around other people, when you just can’t sit by and just &lt;em>exist.&lt;/em> I am pretty sure I am coming off as a bit of crackpot to the people around me, but oh well, hello strangers, I’ll never meet again.&lt;/p>
&lt;p>I sometimes feel I should get this diagnosed, it doesn&amp;rsquo;t feel normal, and idk how much it affects my normal life, my mental peace, my personal and professional life and well everything else. It might be why I have found it terribly hard to focus inside any classroom, why I can’t learn from any live teacher, and why I have a tendency of becoming addicted to my mobile phone if I start using it for longer ( I actively supress it a lot ) .&lt;/p>
&lt;p>It feels a bit hard to deal with this, while also managing a lot of things in your life, but also sometimes it helps, because it develops the ability to multi-task a lot more than the average Joe, but then there is the whole debate about deep work and multi-tasking, so not really sure what the tradeoff is there as well.&lt;/p>
&lt;p>Well anyway, enough ranting in a train, should go back to civilization.&lt;/p>
&lt;p>Until the next one !&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://ayushtues.medium.com/adhd-on-a-train-ride-a46b290e0ce3" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>A long walk to work</title><link>https://ayushtues.github.io/blog/a-long-walk-to-work/</link><pubDate>Sun, 16 Feb 2025 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/a-long-walk-to-work/</guid><description>&lt;p>I am currently 24 years old, with a professional lifespan of ~2.5–3 years , so talking about something spanning 40 years of my life, while I haven’t even lived 25 of my own in the first place, feels a little weird, but what else do you do on a nice lazy Sunday morning? Also out of the 24 years lived, I just remember the ones from around the 16th year, and so I really feel more like 8 year old¹, the first 16 years all seem coalesced into this one giant panorama of blurry snapshots of past memories².&lt;/p>
&lt;p>While no one can predict how long a career would last, owing to the unpredictability of both death, and health³ , if all goes well, it would at least be for 40 odd years, if not more, owing to the improvements in healthcare, given that society sets an average retirement age of 60ish?&lt;/p>
&lt;p>Even though I can’t really imagine sitting on a beach⁴ somewhere, enjoying my retirement reading books for the rest of my life after that, there would be some element of work, in some form. Given the fact that my ability to create impact would probably be the maximum at that age, given the connections I have hopefully developed by that time⁵, and the free time that old age would provide, should more or less offset the depreciation in productivity that a declining health will lead to.&lt;/p>
&lt;p>Thinking about this long journey ahead, actually feels quite relaxing, it takes the pressure off my shoulders to reach certain “milestones” earlier in life, since there is quite a bit of time to do whatever I want to do. It makes it okay to spend a few more months in the city you have come to love⁶, in a job that you don’t love anymore, or spend a bit more time with the people you have come to think of as family, before leaving your nest once again. It makes it okay to take that job that excites you, but doesn’t pay that well. It makes it okay to stay with your family for a bit longer, and put aside the plans for exploring the countries you want to explore for sometime. It makes it okay to spend some more time learning the ropes of your field, to become a student again and pursue your intellectual curiosities a bit more.&lt;/p>
&lt;p>Taking this long view, also makes me appreciate the value of the people around me quite a bit. It only makes sense, that other people will have their own 40 year career around me, and will achieve things which are incredibly hard to predict right now. You never know who will become what, or who you’ll get close to, and it just makes it super important to treat everyone well, and also looking back at the brief career I already have had, it was the people who defined the life I was having⁷.&lt;/p>
&lt;p>Another important thing to take care of is quite obviously my health. I am already kind of a fitness enthusiast of sorts⁸, but there is a lot in the mental space that needs to be worked upon. It becomes evident on the days when I am sick, or even slightly under the weather, how futile everything else suddenly seems, and all things come to halt, until my body is back to normal, so if this machine needs to be working for 40 more years, it will only do so under proper maintenance and under acceptable workloads, its possible to overclock it a bit at times, but not always⁹.&lt;/p>
&lt;p>Also, the work I do will definitely change over time, I will probably not be an IC for 40 years, and my work will not be defined by what I do, but what I enable other people to do¹⁰, and thus its increasingly more important to get better at people skills, not just my technical skills. Also, the field which I work in, which is currently AI, might change, and I might find an alternative career more fulfilling, or just be forced into it due to any number of reasons¹¹, this puts things in perspective, when it sometimes feel like the work I am doing right now, is the only sort of work I’ll be doing forever.&lt;/p>
&lt;p>So much for a 40+ year long career, so many things to explore, people to meet, places to be and…&lt;/p>
&lt;p>Work to be done.&lt;/p>
&lt;p>&lt;strong>Tangents&lt;/strong>&lt;/p>
&lt;ol>
&lt;li>My emotional age, might actually be somewhere around that lol&lt;/li>
&lt;li>This reminds me of the ending of the Death’s end, when SPOILER ALERT, the entirety of humanity is converted from a 3D world, to a 2D image of sorts, only in my case, its like time is flattened into one place for the first score of my life, so it would be more like a 3D cube.&lt;/li>
&lt;li>Increasingly started to think about &lt;em>health-span&lt;/em>, instead of just lifespan after reading David Sinclair’s book Lifespan: Why We Age — and Why We Don’t Have To&lt;/li>
&lt;li>Actually, make that a library, my old bod might need the beach sun, but the brain would crave library, OR actually, I’ll take my Kindle equivalent to the beach lol&lt;/li>
&lt;li>I currently suck at making professional ( or even for that matter personal ) connections, but that will hopefully improve over time, or I’ll shamelessly latch on the connections of the few people I know, similar to how my strategy for making friends as a kind of introvert is to just befriend one very extroverted person lol ( cue to my super extrovert pseudo-sister here )&lt;/li>
&lt;li>Hyderabad — home, hardest city to leave.&lt;/li>
&lt;li>There is this view of not making friends at work, I disagree with it, maybe its just yet to be broken naivety, but its impossible for me to not connect with the people I work on a more personal level, when I spend almost the entirety of 5/7 of my days with them&lt;/li>
&lt;li>Someone recently pointed out, that this is in fact not the case, I am just good at controlling my impulses for junk food and I just like to work out lol&lt;/li>
&lt;li>Which is a bit hard while working in a rapidly growing startup&lt;/li>
&lt;li>While I’ll always want to be an IC in some capacity, I do understand that I am not super talented enough to keep growing in that role, and will have to leverage other people’s talents to make up for my own&lt;/li>
&lt;li>Plain old market forces, or a war, AI automation, hostile AGI takeover, something like the cultural revolution in China, climate change, health, family commitments, aliens, super-intelligent animal takeover, or &lt;INSERT FAVORITE POST APOCYLPITIC TROPE>&lt;/li>
&lt;/ol>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://ayushtues.medium.com/a-long-walk-to-work-0ac2e42dc600" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Book Notes : Foundation By Issac Asimov</title><link>https://ayushtues.github.io/blog/book-notes-foundation-by-issac-asimov/</link><pubDate>Sun, 16 Feb 2025 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/book-notes-foundation-by-issac-asimov/</guid><description>&lt;ul>
&lt;li>Another of the cult sci-fi books by Asimov&lt;/li>
&lt;/ul>
&lt;blockquote>
&lt;p>&lt;em>The premise of the stories is that in the waning days of a future&lt;/em> &lt;a href="https://en.wikipedia.org/wiki/Galactic_Empire_%28Isaac_Asimov%29" target="_blank" rel="noopener">&lt;em>Galactic Empire&lt;/em>&lt;/a>&lt;em>, the mathematician&lt;/em> &lt;a href="https://en.wikipedia.org/wiki/Hari_Seldon" target="_blank" rel="noopener">&lt;em>Hari Seldon&lt;/em>&lt;/a> &lt;em>devises the theory of&lt;/em> &lt;a href="https://en.wikipedia.org/wiki/Psychohistory_%28fictional%29" target="_blank" rel="noopener">&lt;em>psychohistory&lt;/em>&lt;/a>&lt;em>, a new and effective&lt;/em> &lt;a href="https://en.wikipedia.org/wiki/Mathematics" target="_blank" rel="noopener">&lt;em>mathematics&lt;/em>&lt;/a> &lt;em>of sociology. Using statistical laws of&lt;/em> &lt;a href="https://en.wikipedia.org/wiki/Mass_action_%28sociology%29" target="_blank" rel="noopener">&lt;em>mass action&lt;/em>&lt;/a>&lt;em>, it can predict the future of large populations. Seldon foresees the imminent fall of the Empire, which encompasses the entire&lt;/em> &lt;a href="https://en.wikipedia.org/wiki/Milky_Way" target="_blank" rel="noopener">&lt;em>Milky Way&lt;/em>&lt;/a>&lt;em>, and a&lt;/em> &lt;a href="https://en.wikipedia.org/wiki/Dark_Ages_%28historiography%29" target="_blank" rel="noopener">&lt;em>dark age&lt;/em>&lt;/a> &lt;em>lasting 30,000 years before a second empire arises. Although the momentum of the Empire’s fall is too great to stop, Seldon devises a plan by which “the onrushing mass of events must be deflected just a little” to eventually limit this&lt;/em> &lt;a href="https://en.wikipedia.org/wiki/Interregnum" target="_blank" rel="noopener">&lt;em>interregnum&lt;/em>&lt;/a> &lt;em>to just one thousand years. The books describe some of the dramatic events of those years as they are shaped by the underlying political and social mechanics of Seldon’s Plan&lt;/em>&lt;/p>
&lt;/blockquote>
&lt;ul>
&lt;li>Starts with Gaal Dornick, the biographer of Hari Seldon&lt;/li>
&lt;li>Talks about imperial coronation, so probably a monarchy?&lt;/li>
&lt;li>The Jumps → Wormholes, for instant travel&lt;/li>
&lt;/ul>
&lt;blockquote>
&lt;p>&lt;em>Childishness comes almost as naturally to a man as to a child&lt;/em>&lt;/p>
&lt;/blockquote>
&lt;p>Well, talk about relatable 😛&lt;/p>
&lt;p>They seem to have flying cars, for one, and also tunnels. So I think the car moves through tunnels, like walls which have holes in them and each hole is a tunnel.&lt;/p>
&lt;p>It seems a bit post-apocalyptic. Actually, it is. It is not post-apocalyptic, it is pre-apocalyptic. Because it predicts a crumble in the future. But yeah, it seems like hyper-loop’s inspiration.&lt;/p>
&lt;p>The story is based in Trantor, which is the capital of the imperial government, which is like an intergalactic, multi-planetary, galactic empire. It has a population of 40 billion which is a lot, but all of these people work in administration because I think it’s a play on how mundane bureaucracy can get so that 40 billion people are in it just for the bureaucracy and all of them are dependent on other colonies of sorts for their daily needs agricultural things and stuff so they say that this makes it very vulnerable to I’ll show some forwarding to the fact that this actually happened and they were rewarded and stuff. And I think this is a play on… Maybe England and its colonies and how it fall later how it fell later or maybe even uh u.s who knows&lt;/p>
&lt;p>There is no greenery, no soil, no life other than man, apart from the imperial palace which is in 100 square miles of natural soil and has greenery. So yeah, definitely apocalyptic. Or maybe not.&lt;/p>
&lt;p>Actually, now that I think of it, the imperial nature might be for China because of the play on population and the economical destruction. I think since it is written from an American perspective, this might be referring to China. Because why would they be dissing on their own country after all? Or maybe they would be. I don’t know. I need to look this up.&lt;/p>
&lt;p>They use tunnels, actually not tunnels, they have tunneled the ground so that they’re almost a mile below the ground level and they use the temperature difference, I think the geothermal energy for all the power needs and they just say they used to use atomic energy but now they just use this.&lt;/p>
&lt;p>Seldon’s a chad&lt;/p>
&lt;p>So they are trying to build a very giant encyclopaedia which captures all the information possible so that when the eventual fall of civilisation occurs humanity has access to all the knowledge it had in the past so that it can bootstrap it faster and the amount of suffering is greatly reduced probably like a thousand years instead of 30,000 years of suffering so that is what its plan and the rest of the book is how they go about doing that i think so now in the next chapter they are preparing to build that giant encyclopaedia on a planet called Terminus but there’s this one guy, I think some general or something or some king who wants to occupy Terminus because it has strategical importance and there’s some back and forth&lt;/p>
&lt;p>I can just feel the generic distaste for anything other than science in the author’s writing. He’s like, science is the biggest thing and rest all these… Be it beuaracracy, civil force, military force, it doesn’t understand anything. It is, I can see why this inspired the hacker’s culture.&lt;/p>
&lt;p>I think this new society which is called Foundation is having an internal revolt of sorts because initially it was just… And they want to create a normal society there. The scientists are against it. And there’s also the matter of protecting the foundation from outside attackers.&lt;/p>
&lt;p>And I think Sheldon has already overlooked or predicted… .&lt;/p>
&lt;p>Well, turns out the encyclopedia thing was a scam. And now they are just trying to fend off… Having a standoff based on that and later they established like a diplomatic solution in which they give their technical expertise to the kingdoms around and are able to survive for now.&lt;/p>
&lt;blockquote>
&lt;p>&lt;em>Violence is the last refuge for the incompetent&lt;/em>&lt;/p>
&lt;/blockquote>
&lt;p>It pays to be obvious, especially if you have a reputation for subtelty&lt;/p>
&lt;p>One overarching theme in this book is that of, I think, statistical predictions. Anyway, so what they’re proposing is that it is possible to predict the future based on a theory of mass behaviour, how masses act, but only if the masses are not aware that they are being observed.&lt;/p>
&lt;p>So there’s this one guy in this who is sort of trained in psychology, he sometimes tries to analyze things. Things which are going on and other than that also, he said that in the moment of crisis There would be only one possible solution And one obvious way to go, which is a way to reduce any deviation from the predicted path.&lt;/p>
&lt;p>So, he keeps on letting things go by until there is exactly one possible… solution or one possible plan of action in the face of crisis. Pretty cool.&lt;/p>
&lt;p>This book reminds me so much of the three body problem because it has these awesome moments where shit just changes in like a page. And I mean in this section the guy Haradin was he’s also almost chad like Seldon. I don’t know if seldon is the inspiration for Sheldon. Sheldon and Sheldon sound very similar. But anyway, he did an UNO reverse on the guys who were invading their planets. Basically, he gave them his technology, but he gave them under the guise of a religion. He created a religion from the technology.&lt;/p>
&lt;p>and like he indoctrinated and created a dogmatic religion around it and once he had that cult following he just It was like a soft takeover where he overtook the military by through And it is interesting to note how much that is related to Christianity.&lt;/p>
&lt;p>They say because Seldon did not know the individual heroic actions that anyone would do. He betted on more macroeconomic factors, so he betted on the macroeconomic forces that will emerge and solve the crises. In the first case, it was some sort of technological advantage.&lt;/p>
&lt;p>In the second case, it was religion. They converted the other, the enemy, to the missionary approach and the third, I think it is trade. They will… Survive by offering trade to the partners. So yeah, it’s pretty cool. So the Person who planned this all planned not on the individual heroics of people but on Statistically predictable economical forces.&lt;/p>
&lt;p>Pretty cool.&lt;/p>
&lt;p>So yeah, the book ended with the third crisis, which was to be solved using trade. Their daily technology like the ovens and stuff would fail and they wouldn’t want to do that and it turned out to be true It is actually A bit similar to the first one, the first one was also technology related and I think the last one was also technology related.&lt;/p>
&lt;p>They just wrapped it more in the sense of trade instead of technology. The first one was purely technology based. There was no trade there. The second one was religion-based. The third one is trade-based, but that trade is only enabled due to the technology they have.&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://ayushtues.medium.com/book-notes-foundation-by-issac-asimov-c4e0e23aa86a" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Book Notes: Anxious Generation</title><link>https://ayushtues.github.io/blog/book-notes-anxious-generation/</link><pubDate>Sun, 16 Feb 2025 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/book-notes-anxious-generation/</guid><description>&lt;p>Some connection between the social media companies and the tobacco, vape and even the oil industry. Which harm children, but still have a propaganda around it.&lt;/p>
&lt;p>Talks about the lack of or the slow decline of play children have in their childhood and how play Becoming more used to just sitting on the computer and this is so much true for me myself&lt;/p>
&lt;p>Another point is made on the lack of commitment required in virtual worlds. You physically talk to someone and there was a high bar for entry and exit and that made you more amiable and like you develop better relationship skills. Whereas in the virtual world, you don’t even need to use your own name and you can easily block someone and like living in such a world where you can easily cut off connections It’s very different from how the real world is.&lt;/p>
&lt;blockquote>
&lt;p>&lt;em>If your body was turned over to just anyone, you would doubtless take exception. Why aren’t you ashamed that you have made your mind vulnerable to anyone who happens to criticize you so that it automatically becomes confused and upset.&lt;/em>&lt;/p>
&lt;/blockquote>
&lt;p>Epictetus&lt;/p>
&lt;p>Most parents don’t want their children to have a phone-based childhood, but somehow the world has reconfigured itself so that any parent who resists is condemning their child to social isolation.&lt;/p>
&lt;p>Makes a counterpoint to the argument that isn’t Gen Z right to be anxious and depressed because 21st century is Having so many disasters, financial crisis, 9–11, covid and whatnot. But it said that if you actually map the data of those events to the increase in Depression rates, there is not a strong correlation&lt;/p>
&lt;p>Also the gender bias that it affected teenage girls more than teenage boys is not explainable by economic events.&lt;/p>
&lt;p>Also they say the generation which grows up under a stronger political cause like a war, actually shows higher degrees of cooperation and trust&lt;/p>
&lt;p>Forever elsewhere&lt;/p>
&lt;p>Children’s brains grow to 90% of full size by age 5 but then take a long time to configure themselves, this is slow-growth for cultural learning&lt;/p>
&lt;p>Animals are either designed to be in defend mode or discover mode, based on where they are in the food chain. Those in the discover mode are happier&lt;/p>
&lt;p>Online world is much more dangerous, because small mistakes can bring enormous costs, like vitality, large scale public humiliation&lt;/p>
&lt;blockquote>
&lt;p>&lt;em>The cost of a thing is the amount of life required to be exchanged for it, immediately or in the long run.&lt;/em>&lt;/p>
&lt;/blockquote>
&lt;p>Kurt Voneger book mentioned Harrison Bergeron, is set in a modern America, where by constitutional amendment, nobody is allowed to be smarter, better looking or physically more capable then anyone else, so if someone has a higher IQ, they have an earpiece, just interrupting them every 20 secs to drop their IQ lol&lt;/p>
&lt;blockquote>
&lt;p>&lt;em>Once I was a scuba diver in the sea of words, Now I zip along the surface like a guy on a jet ski&lt;/em>&lt;/p>
&lt;/blockquote>
&lt;p>As jobs involve lesser physical skill and more people skill, boys are being left behind?&lt;/p>
&lt;p>Thoda went red pill lol&lt;/p>
&lt;p>Also says that boys used to learn a lot from risky behaviours, which are being reduced?&lt;/p>
&lt;blockquote>
&lt;p>&lt;em>Do not judge, so that you may not be judged. For the judgment you give will be the judgment you get. The measure you give will be the measure you get.&lt;/em>&lt;/p>
&lt;/blockquote>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://ayushtues.medium.com/book-notes-anxious-generation-6dd5d426f44d" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Clarity</title><link>https://ayushtues.github.io/blog/clarity/</link><pubDate>Mon, 13 Jan 2025 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/clarity/</guid><description>&lt;p>Writing this just after trying to smudge my glasses clear, it’s safe to say I can really use some clarity in my life at the moment.&lt;/p>
&lt;p>While I don’t like normalising disorders, by claiming to have them, since I am not diagnosed with any of them, &lt;em>yet.&lt;/em> I still feel I might be borderline ADD, since I can’t seem to focus on anything, and my mental clarity is pretty much worse than my visual clarity, and they don’t even make glasses for that, do they?&lt;/p>
&lt;p>If there is one type of people, I envy the most, it’s the fucking &lt;em>specialists&lt;/em>. The ones who seem to be born with a purpose in mind, who have absolute clarity in their thoughts, and are perfectly fine in executing that with almost psychopathic single-mindedness.&lt;/p>
&lt;p>In a world with so many choices, how do people get this zen like clarity about what they want in their life, and how do they commit to something, without being worried about the opportunity costs it has, and all the other shiny things they can be chasing, instead of doubling down on one single thing?&lt;/p>
&lt;p>I find it extremely hard to commit to any single thing, hell, with the year starting, I ended up opening almost 10 books in parallel, not being able to decide which one to read, and ended up reading all of them a bit by bit ( although eventually settling to finish a few of them, but still wasting a lot of time in the exploration).&lt;/p>
&lt;p>The same goes for my professional life, when I was still an undergrad, I used to spend my time chasing every new field I could find, and trying to learn the basics of it, but which lead to a very superficial understanding of things, and always made me look with awe at people, who dedicated their effort in a single field, and became fucking good at it.&lt;/p>
&lt;p>I feel the world is designed to reward people who show clarity of thought, I think it is because it is hard to achieve that. Because it is hard to have the conviction and the confidence to believe that what you have chosen is the right thing, and the universe rewards you for that conviction.&lt;/p>
&lt;p>One problem for this lack of clarity is the seeming impossibility of predicting anything in an adult life. Most of the impactful events of the past have been unplanned, and life just happened to smack me in my face, one sudden day, when it chose so, how are you supposed to have any clarity in things while living in such an uncertain world?&lt;/p>
&lt;p>It is also however true, that the moments when I have felt at peace with myself, and had absolute clarity of thoughts, are clearly somehow superior to the usual brain fog that inhabits my brain. I want to maximise moments such as those, but they are few to find and small to last, and disappear into thin air, like fog on glass.&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://ayushtues.medium.com/clarity-9b634ad591f6" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Wake up to reality</title><link>https://ayushtues.github.io/blog/wake-up-to-reality/</link><pubDate>Mon, 13 Jan 2025 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/wake-up-to-reality/</guid><description>&lt;p>Given the (debatable) fact that we come into this earth tabula rasa, or as a clean slate, its is very interesting to observe it being slowly written on, and erased, time and time again, becoming more of a used slate, with dust traces from the past settling as biases withhold (there is a word for exactly this, which my dumb self will later use google to figure out — EDIT — Palimpsest), and new spaces made for me to be sensitive about issues which I was till now unaware, or simply ignorant of.&lt;/p>
&lt;p>Yours truly, has never been much of a politically inclined person, social issues were never a concern for an upbringing, which was middle class enough to allow glimpses into the fact that something was potentially wrong with the world around, but also privileged enough to allow the bliss that is ignorance.&lt;/p>
&lt;p>Well, shit changes…..sometimes, when you least expect it.&lt;/p>
&lt;p>While reading books is usually seen has a harmless hobby, a mere pastime of the intelligentsia, a thing people do on lazy Sundays, with a well brewed cup of coffee in hand and the sun shining through the window, (totally not describing my own Sundays.. ahem) sometimes, they are not so harmless, and might change the entirety of your being, in just a few pages. Afterall, most of society, is driven by a few religious books, and most political philosphies, were first written as a book or a manifesto or a treatise.&lt;/p>
&lt;p>While I started my reading journey, with typical young adult, coming of age books ( the harry potters, Percy Jackson, Colleen Hoover? ) , over the last 1 year or so, there is a whole new genere which is starting to attract me. And with it, came a whole new phase of discovering a new sensitivity to the society around, which I inhabit.&lt;/p>
&lt;p>This is not a political piece however, I still don’t have a political philosphy of my own, and as someone who hasn’t even voted yet, ever, I am the last person to say anything about any social issue, but I will do so anyway, cause…it’s my blog lol&lt;/p>
&lt;p>When talking about social issues, does it make sense to only talk about the country you live in? or are you supposed to talk about the issues which are faced in the world in general. Should I talk about racism in the US, or castesim in India? Is a global problem like climate change one that demands attention, or the threats to democracy in the world’s most populous country the issue? With so many things wrong in a world, which pretends to be working absolutely fine like clockwork, what problems do you claim as your own, what problems you chose to ignore as somehow less worthy of care, how do people develop their own socio-political agenda? I have no clue.&lt;/p>
&lt;p>Are you supposed to follow a bottom-up approach, trying to make small improvements in the world immediately around you, or do you take a top-down approach, and campaign the fix the largest problems we face as a society. I personally feel more attracted towards the top-down approach, since I feel powerless against the tyranny of the masses, and the pestilence like nature of the inertia to change, renders a bottom up approach inefficient and a Sisyphean attempt at best.&lt;/p>
&lt;p>I have never felt much attached to the nation I was born in, I didn’t chose to be born in this weird as hell world, with more problems than solutions, I didn’t chose the religion my family adopts, nor the cultural surroundings which have tainted my view of the world, in their view. I believe that the first 22 years of my life, were a blip, with barely any identity being formed, or maybe I have just dissolved whatever identity I had created in the past.&lt;/p>
&lt;p>For the future, I hope I find issues that I think are worth solving, people I want to support, movements which feel like my own, identities I adopt, and maybe again discard as the time goes by.&lt;/p>
&lt;p>Whatever the case be, its finally time for me, to wake up to reality —&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://ayushtues.medium.com/wake-up-to-reality-215dbabaf208" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Being the human behind an AI</title><link>https://ayushtues.github.io/blog/being-the-human-behind-an-ai/</link><pubDate>Sat, 23 Nov 2024 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/being-the-human-behind-an-ai/</guid><description>&lt;p>Over the last 8 months or so, in my new (now &lt;em>slightly&lt;/em> oldi&lt;em>sh&lt;/em>) job as an AI Engineer at a startup, I have been one of the humans behind some of the AI products we offer ( mostly text to speech, and other speech stuff ), and I have also spent a lot of time in the past 5 years trying to get something meaningful out of these models. And its, fun, &lt;em>mostly.&lt;/em>&lt;/p>
&lt;p>Andrej Karpathy, famously spent a lot of time manually labelling the Imagenet dataset, looking a just a shit ton of images, and trying to figure out what they represent, and that effort led to the recent AI boom almost as much as anything else. But, how did it feel for him, and other researchers at Stanford, and at other academic labs, to be the ones putting in the manual grind to painfully, and slowly label these datasets, and get these models to work, when they, just fucking wont. Probably fun, &lt;em>mostly.&lt;/em>&lt;/p>
&lt;p>I am in no way even remotely close to the level of work done my these now almost mythical forefathers of our fields, but, hell, I have spent a LOT of time manually labelling datasets, looking at way too many images of travel destinations I had no hope of actually travelling during my stint in the travel team at MSFT, and an inhuman amount of time listening to different pronunciations of names and different tonalities and accents of humans in my current job. And I think I have become a lot more sensitive to human voices and sounds due to it, &lt;em>mostly.&lt;/em>&lt;/p>
&lt;p>A few of my colleagues work in lip-sync, where you have to create accurate AI generated lip-motions for an accompanying AI generated audio for a human speaker, and I can safely say that I suck at identifying what makes a good lip-sync. My teammates attribute this to my habit of consuming a plethora of anime, that too in sub, so I am more used to looking down at the subtitles than at the face of the characters to the point where I didn’t even realise that most animes have minimal lip-motion and at one point I almost sweared to someone that animes have diverse lip movements only to be shown multiple videos characters having just simple vertical or horizontal movements lol. My teammates working on lip-sync are however a different case, and have almost a superhuman ability to detect even a minor offset between lip movements and audios, while I am just sitting there finding no difference at all, &lt;em>mostly.&lt;/em>&lt;/p>
&lt;p>I wrote about occupational hazards in a previous post, which was mostly a negative take on how the tech industry is affecting my personality, but there is also something really transforming about being the human behind an AI. It has really made me more sensitive to things which we often take for granted as humans, and has probably changed me fundamentally in how I reason about things. A simple example of this is, text to speech, after working on it for 8 months now, I mentally play out any text I see or any name I see and think about how easy or hard would it be for an AI to read it out, and what mistakes it might make. Also since I have seen how hard it is for an AI to speak properly, even when trained on hundreds of thousands of hours, it is hard not to appreciate the easy at which we are able to just speak, and communicate with each other so well, &lt;em>mostly.&lt;/em>&lt;/p>
&lt;p>Another interesting effect of becoming more cognisant of this effect is the important of choosing important problems to work on, which I am actually interested in ( and quite frankly, I wasn’t interested in text-to-speech at all previously, now I do find it quite interesting ), since those are the areas of the human experience I will become more knowledgeable about and probably develop and unfair advantage in, well…. &lt;em>mostly :P&lt;/em>&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://ayushtues.medium.com/being-the-human-behind-an-ai-d256f282898f" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Occupational Hazards</title><link>https://ayushtues.github.io/blog/occupational-hazards/</link><pubDate>Sun, 06 Oct 2024 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/occupational-hazards/</guid><description>&lt;p>The term occupational hazard is usually used to describe “physical” hazards, which one is exposed to in the process of pursuing their occupation/work. Like being exposed to anthrax as a wool rearer, dangerous radiations for nuclear factory workers or just plain old back pain in the tech industry (I am turning 24, and my back is already starting to show signs of rebel, even if I am fairly physically active).&lt;/p>
&lt;p>What stroke my fancy in one of my ever present parallel thought plane of mental existence, was the “personality” hazards which your career might expose you to.&lt;/p>
&lt;p>I have been working in the tech industry for a little over 2 years now, and a little introspection makes it clear that my personality has been affected by my choice of occupation, if its been a hazard or not in this regard, is upto anyone’s guess.&lt;/p>
&lt;p>I think my current mental utility function, the force that drives me, assuming I am a rational agent (which I certainly am not), is heavily conditionally dependent on my occupation. A lot of things which are important to me as a person, might not have hold any value for me if I was in a different career.&lt;/p>
&lt;p>I am a deeply individualistic person (having already use the phrase “I” 9 times since beginning this article), and I don’t really care about anyone other than myself except maybe a very small handful of people, which sounds like something you’d say to sound dramatic, but I mean that in all seriousness. Sometimes I wonder how much of it is my innate nature, and how deeply I am nurtured by the occupation I pursue. Tech as an industry is individualistic, fascinated by powerful individual role models who created a path beyond all odds &lt;em>and despite all costs.&lt;/em> With leaders using wartime terminology in boardrooms, somehow glorifying the horrific destruction that is a war, to the chants of doing “whatever it takes” to win the arms race to technical supremacy.&lt;/p>
&lt;p>I am not looking to question the culture,I myself am a child of, but I wonder, what would happen if I took upon a different career path. What would be my values. Currently I only care about my own success, have little regard to whatever the hell happens to society or even the people around me, I am deeply meritocratic, almost certainly sapiosexual and I work in an industry which has seen an influx of the most talented (and often ruthless) brains of the generation, and is capitalism incarnate it its most glorious ( and scary) manifestation in the 21st century. People (including yours truly) are driven by profits, by academic/professional glory, relationships are put on the sidelines and the most unlikely of alliances are forged for a chance to get a chance to make an “impact” on the world, and transform our society, often forgetting, that that’s not the only way to make an impact or even the most “impactful” one.&lt;/p>
&lt;p>In contrast to this individualism, I wondered, what if I went into maybe something like the military, what would that make of me. I don’t have any idea of what goes inside a military life, but surely it involves a level of sacrifice of the individual identity? Atleast while watching a military parade, with people clad in the same uniforms (maybe decorated with varying level of paraphernalia indicating their rank), marching in groups to the same beat, their must be a strong sense of belonging to a larger group and of surrendering yourself to a cause, for the greater good. I don’t think when my job is to literally protect other people, I would care a lot about what petty concerns make of my daily headache in my current life, and obviously that life comes with its very own (and MUCH more physical) occupational hazards, but I was just wondering if I was drawn to the tech world because it suits my personality, or my personality has just been moulded to like the work I do.&lt;/p>
&lt;p>Or what if I was a doctor? It feels like such a noble profession, though I know there is a LOT of stress and pain and occupational hazards that go into that lifestyle, but again, would that have made me a “better” person, would my personality be more caring, more sensitive and empathetic to the suffering of others, or would it have made me numb to it all, due to the constant overexposure to it?&lt;/p>
&lt;p>I also often wonder about the total absence of social responsibility I carry, bordering almost to an apathy to whatever socio-political issues in the world around me, always having a feeling that all these problems are happening somewhere else in the world, and are not my problem. But what if I was working in an NGO, would I have been able to turn away my eye and pretend not to see anything wrong while staring right at the problems that are omnipresent around us. As an example, I am finding myself unable to look street beggars in the eye, because it somehow makes me very uncomfortable, and I can’t do any better than just pretend they don’t exist or I don’t see them, and that is so fucking brutal, because its only a coin flip which decides which side of the line I was born in, and it doesn’t take a lot of misfortune to even make me cross lines from one end to another in the future, and be on the receiving end of being pushed around in a world, that pretends to not see you, exist.&lt;/p>
&lt;p>Maybe it&amp;rsquo;s just the nascent emotional maturity developing inside me, as I am growing slightly older, although still being VERY young emotionally to be sensitive to a lot of issues, that my brain is making me think of all these things. It has affected the content I am consuming, the books I am reading (I am very much interested in history, political books, even though I didn’t like them before). I don’t know if its just a phase of life, where my brain is trying to figure out what its identity is, and how much of it, is just a result of occupational hazards in a mad industry?&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://ayushtues.medium.com/occupational-hazards-5758b553782a" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Stuck</title><link>https://ayushtues.github.io/blog/stuck/</link><pubDate>Sun, 25 Aug 2024 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/stuck/</guid><description>&lt;p>Is 3 months left until becoming 24 years old too early for having a quarter life crisis?&lt;/p>
&lt;p>Its funny how I used to go around saying to everyone that I am just a kid, how people treated me like a kid, and how that made it okay for me to be a kid. As time passes by, I feel less and less like a kid, and really, just plain old.&lt;/p>
&lt;p>It’s easy to feel the world moving around at lightspeed, leaving you behind, in the dust, when you’re sitting alone in your room, working in a field (AI) which moves at speeds invoking relativistic effects, and in a city full of dust (Noida).&lt;/p>
&lt;p>Is it normal, or is it just me, who spends most of his time, feeling stuck, with no fucking clue about what to do with his future (while people tell him, he’s &lt;em>sorted&lt;/em>, what am I, an array?).&lt;/p>
&lt;p>I am at that weird point in my career, where the things I can do are mostly individualistic in nature, and by definition not scalable, leading to a feeling of almost minimal impact, which was fine, when I was a student, and could always wait to enter the “real” world to do something impactful. While with the years stacking in experience, there is this growing desire to do something which I can be proud of (my current proudest professional acheivement till now is breaking something on the world’s 2nd most used search engine, ah…while I was employed there :P ). This leads to feeling stuck in my career as well, with no clear path ahead, the age old dilemma of being an IC vs a manager in the future looming closer and closer in the already big list of dilemmas inside my mind.&lt;/p>
&lt;p>I have developed so many interests, and have tried picking up so many skills, but atleast at the moment, I am mediocre in all of them, and that adds to the feeling of being stuck, always waiting for somehow being able to crack open the box, (maybe by repeatedly banging your head against it lol).&lt;/p>
&lt;p>I don’t know what the future holds for me, I don’t even know if I have a future, or even how many days it consists of, but I hope I am able to unstuck myself a bit, or else it might end up suffocating me and be my end.&lt;/p>
&lt;p>But ah well —&lt;/p>
&lt;blockquote>
&lt;p>The best case is I’m a rat who might be able to look up at the clouds once in a while&lt;br>
&lt;em>- I have no clue who said this, neither does google :P&lt;/em>&lt;/p>
&lt;/blockquote>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://ayushtues.medium.com/stuck-89363349ac2d" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Ye shalt shut up and doth ye work</title><link>https://ayushtues.github.io/blog/ye-shalt-shut-up-and-doth-ye-work/</link><pubDate>Fri, 17 May 2024 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/ye-shalt-shut-up-and-doth-ye-work/</guid><description>&lt;figure>&lt;img src="https://ayushtues.github.io/blog/ye-shalt-shut-up-and-doth-ye-work/image-1.jpeg"
alt="Gotta love DALLE-3 lol">&lt;figcaption>
&lt;p>Gotta love DALLE-3 lol&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>While the title suggests that I should go back and do my work, we shall do just the opposite and rant the hell out of our thoughts in this lost corner of the virtual world, while we sit in a room lit up with fairy lights, blasting emo music (I like the way you kiss me by Artemas), only missing a cup of coffee to make things better (which on second thought I should make). So much for being a hypocrite eh?&lt;/p>
&lt;p>Engaging in a favourite past time of mine, ie. overthinking about random things which have little to nothing to do with my future, while reading a book which is supposed to make me think (The Clockwork Universe), but only serves as an alternative to scrolling random shit on my phone owing to my attention span of a goldfish. I started contemplating the meaning of life, universe and everything, until I had a major realisation — I should just the shut the fuck up and do my work. &lt;em>Its that simple!&lt;/em> Btw the shut up part comes from the fact that I spend most of my time talking to myself in my room in full volume, like the total crackpot I am.&lt;/p>
&lt;p>I have a bad habit of over-analyzing my own actions, criticising them to the point of being often rendered paralyzed by the sheer brtuality my brain subject unto itself ( I think, therefore I am? Yeah Descartes, I think I have better odds if I don’t think :P ). I can observe my own actions from a detached perspective and figure out what I should have done in the moment, and why I didn’t do it, and what I should do to fix it….but do we fix it, nah. Instead we just rant about how unfair life is and yeah, just rant a lot.&lt;/p>
&lt;p>I wonder if instead of being stuck in the past or trying to make sense of the present or predicting the future, it would be better if I just, shut the fuck up and did my work. Like really, just work and let things happen the way they are going to happen. I spend so much time trying to figure out what might be the best way to go doing about things or exploring random hacks that might make me better at something I want to skill up in, it might be instead better to just shut up, and put in some hours and get shit done?&lt;/p>
&lt;p>If I look back at the past, then the moments when I just get lost in the work I am doing or the skills I am trying to develop, the times when I just tune out everything and work the hell out of my capacity, are the moments which end up making all the difference &lt;em>bit by bit&lt;/em>. They don’t lead to a sudden metamorphosis, but instead slowly set up a chain reaction of events that can only be made sense of &lt;em>after&lt;/em> the fact. In the moment it feels like life doesn’t make any sense, it feels frustrating, and I feel like ranting out to no-one in particular (hence this blog), but really I should just shut the fuck up and get my work done.&lt;/p>
&lt;p>So we’ll just do that, shall we?&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/ye-shalt-shut-up-and-doth-ye-work-c62deb9f1710" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Fluid/Self</title><link>https://ayushtues.github.io/blog/fluid/self/</link><pubDate>Sun, 07 Apr 2024 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/fluid/self/</guid><description>&lt;p>Its a new year, and while I know its an artificial construct created as a by-product of us following a particular calendar AND we could easily be still stuck in 2023 if we just followed something else, imagine choosing a different calendar, so that you’re stuck in 2020 lol and still 20 years old, and ahem…in 2020? Doesn’t sound like a good idea anymore, does it. (EDIT — This part was written in January, hence the “its a new year”)&lt;/p>
&lt;p>The last year has been a blur, while I am mentally still starting out my life in Hyderabad in my first job, I am actually already leaving the city and moving to a new job. (UPDATE — I have shifted to Noida, its been one month, and it sucks, HYD&amp;raquo;&amp;raquo;&amp;gt;NCR)&lt;/p>
&lt;p>Amongst all the madness of the past year, I have started to lose track of any sense of self image that I had. Whatever I used to think that I can do or the values I thought are set in stone for me, are actually turning out to be fuzzy and quite fluid.&lt;/p>
&lt;p>Its been an year of, Oh I don’t do that, I have never done that before, I am not that type of guy. TO — oh well, let me give it a shot, what could happen, TO — Oh, I CAN do this, I actually enjoy doing this, why didn’t anyone tell me about this before!!!? Often this does lead to me having a mental breakdown because I see so many possibilities in the future about things that I can do and so possible MEs I can be, that I almost freeze about what I should be?&lt;/p>
&lt;p>I don’t remember my childhood a lot, since it feels very static, life was very static, I WAS VERY STATIC through the years ( On a side note, whenever I start thinking about my childhood in these blogs, it makes me sound like I am giving a therapy session to myself lol and psycho-analyzing the shit out of myself). My personality didn’t change a lot. I had a particular personality, shaped through my experiences (OR a lack of them actually in retrorespect). Just go to school, study ( or pretend to study? I kinda always resorted to self-study more ), come home, play video games/watch cartoons, maybe study some more. I didn’t have much or really &lt;em>any&lt;/em> hobbies to shape my personality, and what’s weirder is that I didn’t even think about trying to expand my mono dimensional personality into something more vibrant and easier to sustain in the longer term. And this is something that I have observed quite a bit around me nowadays, and it honestly scares me, knowing how easy it is to fall into this trap of….just remaining as is.&lt;/p>
&lt;p>Now not to say that you shouldn’t remain static, there are some things which define who you are, your core values, your constants, things that make you who you are. Unless you find something very wrong about them, I feel it’s okay, and probably even helpful to not change a lot on these, since you would probably just be changing a big part of yourself based on some minor, fluctuating, ephemeral feeling of being influenced by someone who you think is cool, or thought patterns/processes which seem cool in the moment, but in the long term, they don’t really align with you and will only confuse you more than anything.&lt;/p>
&lt;p>Apart from these core values though, I have come to believe that you should be up and ready to change anything and everything about yourself. I have found it very liberating and honestly just fun, to know that I can be whatever I want to be, and am not limited by the person I am or was in the past. It’s fun to know that I can change my life in any manner however I want to change it. There are sooooo many new things to learn, so many ideas to get exposed to, and just so many things to think and process and believe and un-believe that having a fluid-self is both rewarding and self-preserving.&lt;/p>
&lt;p>To end this post, I feel like every time I write such posts, and then come back to them years later, I am surprised by the thoughts that used to occupy my mind at the time of writing, and often end up with me being like “ah this kiddo has no idea what is going to happen with him”. But that is the point ain’t it, of having a fluid/self?&lt;/p>
&lt;p>Until the next one — So long!&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/fluid-self-05b6683dd7dc" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Smart enough?</title><link>https://ayushtues.github.io/blog/smart-enough/</link><pubDate>Sun, 07 Apr 2024 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/smart-enough/</guid><description>&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/smart-enough/image-1_hu609046021417295643.webp 400w,
/blog/smart-enough/image-1_hu1142474448697356221.webp 760w,
/blog/smart-enough/image-1_hu3903915476634724484.webp 1200w"
src="https://ayushtues.github.io/blog/smart-enough/image-1_hu609046021417295643.webp"
width="760"
height="283"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>I was reading a blog somewhere online, at Less Wrong ( &lt;a href="https://www.lesswrong.com/posts/kXSETKZ3X9oidMozA/the-level-above-mine" target="_blank" rel="noopener">The Level above mine&lt;/a>), one of those rare online gems that you happen to know about after waaaaaay too long in your life, that we can’t identify intelligence at more than one standard deviation away from our own level and anything over that just blends into well “genius”. That made me thinking that where do I stand and what level of intelligence am I sensitive to, and what am I oblivious to?&lt;/p>
&lt;p>I have been considered “smart” by the people around me ever since I was a child, it started from being the topper in the class in school, to being pushed into a hyper competitive race during my high school and then being a part of the “elitist” society of tier-1 college grads in my country. I think this assessment lays from the fact that I &lt;em>might&lt;/em> have a slightly higher than average IQ, while still being at a level at which normal people remain sensitive to changes in intelligence. It also was an easy explanation to two facts 1) being my total lack of aptitude in anything other than acads and giving people a way to justify that by giving me the “oh he’s a genius, they tend to be weirdos anyway” pass 2) it served as a relatively easier explanation to my academic success than the others simple. fact that I just worked hard lol&lt;/p>
&lt;p>This uni-dimensional analysis of my personality, where the only redeeming factor of myself as a human was the fact that I am “smart” led to a bunch of problems for me as a functioning member of human society, which in fact requires you not just to be smart but also, a bunch of other things like I don’t know, kind, gentle, humble, wise, charismatic/ a potential sociopath who mocks all these qualities while actually planning a serial murder? How was I supposed to suddenly learn all these other qualities which are now more important than me being smart? How do I fit myself into society, how do I keep up with the dumb as hell questions of my relatives at family functions without sounding like a total uninterested asshole or worse, a smartass?&lt;/p>
&lt;p>Another problem, which occurred slightly late in life, is the fact that being “smart” is relative, to the people surrounding you. I was “smart” in my school, based on my peer group then. But ever since I entered college, joining the so-called elite club of the engineers of the country, the average IQ of the people surrounding me shot above, probably (most likely) beyond my own ( ALSO, I know IQ is not a good measure, and its not the only measure of intelligence, and there are other forms of intelligence, and blah blah, but let’s tackle the simplest case shall we, I am fucked in EQ and SQ anyway :P ). When you are used to being called “smart” and take pride in it being something that distinguishes yourself from others, what happens when everyone around is just as smart, if not more? Well, shit… shit happens.&lt;/p>
&lt;p>Suddenly being thrown into the lower rungs of the intellectual scale, while being used to occupying it upper echelons, warps the space around you and makes you lose sense of your (mental) direction, and you’re left, lost, questioning to yourself and to the people around you…how the hell do I reach back to the top? Anything and everything you do is accompanied by the dreadful question of, am I smart enough for this? Am I doing the right thing? Is someone more suitable for this than me? Does any of this matter, do I matter? Imposter syndrome runs amok your headspace rent free like that latest anime song (BLING-BANG-BANG-BORN, who the best, I am the best!!)&lt;/p>
&lt;p>Not to say that, its not incredibly humbling as well to be sooooo outmatched by your peers in terms of their sheer intelligence that it gives you a deeper appreciation for the importance of the &lt;em>other&lt;/em> things that we mentioned previously. You appreciate that while you can’t be smarter, you can be kinder, gentler, and overall more understanding of the people around you and not be a smartass all the time. You appreciate the importance of working hard, even if you know you are outmatched intellectually, still trying anyway, to give it your best shot. Whatever you aim to do, has probably already been OR can be done 10x better and faster by some much much smarter than you, but that doesn’t mean you can’t give it a shot. The world exists at all scales of existence, you will eventually find your own plane, and maybe if you aim high enough, surrounded by these being occupying a higher plane of existence, you might get a peek at their lives and model yours around it, a bit, as much as possible.&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/smart-enough-07040e2f6675" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>FLARE — “Advanced” RAG implemented from scratch</title><link>https://ayushtues.github.io/blog/flare-advanced-rag-implemented-from-scratch/</link><pubDate>Thu, 16 Nov 2023 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/flare-advanced-rag-implemented-from-scratch/</guid><description>&lt;p>In an effort to learn more about LLMs, it was inevitable that I’d come across RAG (or Retrieval Augmented Generation). After watching a few online videos about different architectures for RAG and some advanced techniques, I was looking for something which I could implement.&lt;/p>
&lt;p>Now, I can’t train any LLMs, limited by compute, but nonetheless, I found that there are some cool approaches which enhance RAG, while treating the LLM as a black box ( okay, maybe more like a grey box, more on this later). Enter FLARE or &lt;strong>F&lt;/strong>orward &lt;strong>L&lt;/strong>ooking &lt;strong>A&lt;/strong>ctive &lt;strong>RE&lt;/strong>trieval Augmented Generation (do they decide on the name first or the acronym nowadays?).&lt;/p>
&lt;p>But we’re getting ahead; first things first, what the hell is RAG?&lt;/p>
&lt;h4 id="retrieval-augmented-generation">Retrieval Augmented Generation&lt;/h4>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/flare-advanced-rag-implemented-from-scratch/image-1.png"
alt="https://cs.stanford.edu/~myasu/blog/racm3/">&lt;figcaption>
&lt;p>&lt;a href="https://cs.stanford.edu/~myasu/blog/racm3/" target="_blank" rel="noopener">https://cs.stanford.edu/~myasu/blog/racm3/&lt;/a>&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>There are a ton of resources on RAG, so I’ll keep it quick. So one of the problems with LLMs is that they only have information about the data they were trained on, cause well, duh, how would they know about anything that they were not shown during training?&lt;/p>
&lt;p>Now, this creates a problem for use-cases where you want to use your LLM on different data. Maybe you want to know who won the latest FIFA match and want to query the internet for it, OR you want to connect your LLM to your personal Notion data to make your own personal assistant.&lt;/p>
&lt;p>Basically, you want to &lt;strong>Augment your LLM with an external knowledge base&lt;/strong>. So the idea goes something like this:&lt;/p>
&lt;ul>
&lt;li>You have a question or a &lt;em>query,&lt;/em> and you have your knowledge base. You then &lt;strong>retrieve&lt;/strong> relevant information for that query, from the knowledge base. This is usually done via embedding the query into a vector and doing some sort of nearest neighbours with your knowledge base, which might be a vector database, having all information represented by a vector.&lt;/li>
&lt;li>We then &lt;strong>augment&lt;/strong> the query with the retrieved info, making the job easier for the LLM&lt;/li>
&lt;li>The LLM then &lt;strong>generates&lt;/strong> the answer based on the augmented query.&lt;/li>
&lt;/ul>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/flare-advanced-rag-implemented-from-scratch/image-2.jpeg"
alt="https://python.langchain.com/docs/use_cases/question_answering/">&lt;figcaption>
&lt;p>&lt;a href="https://python.langchain.com/docs/use_cases/question_answering/" target="_blank" rel="noopener">https://python.langchain.com/docs/use_cases/question_answering/&lt;/a>&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;h4 id="why-flare">Why FLARE?&lt;/h4>
&lt;p>Okay, so if RAG works, why do we need FLARE? The authors claim that most RAG pipelines currently, only invoke the retrieval step once, that is, using just the input. This works fine for short-form generation but is problematic for long-form generation, which needs complex information that is &lt;em>not always evident from the input alone.&lt;/em> They claim that, similar to humans, gradually acquiring information about a topic in multiple steps, &lt;em>long-form generation with LMs would require gathering multiple pieces of knowledge throughout the generation process.&lt;/em>&lt;/p>
&lt;p>Following this, they ask the question — C&lt;em>an we create a simple and generic retrieval augmented LM that actively decides when and what to retrieve throughout the generation process?&lt;/em>&lt;/p>
&lt;p>Well, yes, and that is what FLARE does.&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/flare-advanced-rag-implemented-from-scratch/image-3_hu4280828639515023817.webp 400w,
/blog/flare-advanced-rag-implemented-from-scratch/image-3_hu8086262284540949645.webp 760w,
/blog/flare-advanced-rag-implemented-from-scratch/image-3_hu15601990942820347597.webp 1200w"
src="https://ayushtues.github.io/blog/flare-advanced-rag-implemented-from-scratch/image-3_hu4280828639515023817.webp"
width="760"
height="540"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>FLARE on a high level works as follows&lt;/p>
&lt;ul>
&lt;li>Deciding &lt;em>when&lt;/em> to retrieve — They claim that LLMs should retrieve only when they are uncertain in their prediction. Assuming LLMs are well calibrated, meaning that the probability they predict, &lt;em>is actually&lt;/em> how unsure they are about the prediction, they adopt an active retrieval strategy to retrieve only when the generated tokens have a low probability.&lt;/li>
&lt;li>&lt;em>What&lt;/em> to retrieve — They say it is important to consider what LMs intend to generate in the future, as the goal of active retrieval is to benefit future generations. Therefore, they propose anticipating the future by generating a temporary next sentence, using it as a query to retrieve relevant documents, and then regenerating the next sentence conditioning on the retrieved documents.&lt;/li>
&lt;/ul>
&lt;p>That is it on a high level, I always find that things become clearer when you see the code, so let’s dive into a simple, minimal implementation of FLARE, (&lt;a href="https://github.com/ayushtues/FLARE_from_scratch" target="_blank" rel="noopener">ayushtues/FLARE_from_scratch: Implementing Forward-Looking Active REtrieval augmented generation (FLARE) from scratch (github.com)&lt;/a>).&lt;/p>
&lt;h4 id="the-vectordb">The vector DB&lt;/h4>
&lt;p>First, we’ll create our Vector Database, which is the external knowledge source we want the LLM to know about. In my implementation, I am looking to load all the transcripts of the videos on my &lt;a href="https://www.youtube.com/channel/UCGEWHZv8Gn10Lk56PzCeFMA" target="_blank" rel="noopener">YT channel&lt;/a> into the database.&lt;/p>
&lt;p>Now, the transcripts are just a bunch of text files; how do we know what part of the text to use for our LLM? In an ideal scenario, we’ll just pass all the text to our LLM as context, but that doesn’t work, because a) LLM’s can handle limited context length b) It will be very expensive. So instead, we find out the most relevant chunk of text based on our query.&lt;/p>
&lt;p>First, we split the text files into chunks of size 200 each, using RecursiveCharacterTextSplitter from Langchain. Now, we need to convert the plain text into something on which we can do a similarity search with our query. We use a sentence transformer for this.&lt;/p>
&lt;p>I used FAISS as my vector index, which calculates the sentence embedding of all the chunks from the transcripts and stores it in a vector index, allowing for efficient search for most similar chunks given a query ( embedded using the same sentence transformer ).&lt;/p>
&lt;p>Also, I cache the embeddings, to make stuff faster, using LocalFileStore and CacheBackedEmbeddings from Langchain.&lt;/p>
&lt;div class="highlight">&lt;pre tabindex="0" class="chroma">&lt;code class="language-python" data-lang="python">&lt;span class="line">&lt;span class="cl">&lt;span class="kn">from&lt;/span> &lt;span class="nn">langchain.document_loaders&lt;/span> &lt;span class="kn">import&lt;/span> &lt;span class="n">DirectoryLoader&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="kn">from&lt;/span> &lt;span class="nn">langchain.text_splitter&lt;/span> &lt;span class="kn">import&lt;/span> &lt;span class="n">RecursiveCharacterTextSplitter&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="kn">from&lt;/span> &lt;span class="nn">langchain.vectorstores&lt;/span> &lt;span class="kn">import&lt;/span> &lt;span class="n">FAISS&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="kn">from&lt;/span> &lt;span class="nn">langchain.storage&lt;/span> &lt;span class="kn">import&lt;/span> &lt;span class="p">(&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">LocalFileStore&lt;/span>&lt;span class="p">,&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="kn">from&lt;/span> &lt;span class="nn">langchain.embeddings&lt;/span> &lt;span class="kn">import&lt;/span> &lt;span class="n">CacheBackedEmbeddings&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">HuggingFaceEmbeddings&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="kn">from&lt;/span> &lt;span class="nn">transformers&lt;/span> &lt;span class="kn">import&lt;/span> &lt;span class="n">T5Tokenizer&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">T5ForConditionalGeneration&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="kn">import&lt;/span> &lt;span class="nn">torch&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="c1"># create a vector db from the transcripts&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="k">def&lt;/span> &lt;span class="nf">create_vector_db&lt;/span>&lt;span class="p">():&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="c1"># we use sentence transformer to get the vector embeddings for the database&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">model_name&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="s2">&amp;#34;sentence-transformers/all-mpnet-base-v2&amp;#34;&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">model_kwargs&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="p">{&lt;/span>&lt;span class="s1">&amp;#39;device&amp;#39;&lt;/span>&lt;span class="p">:&lt;/span> &lt;span class="s1">&amp;#39;cpu&amp;#39;&lt;/span>&lt;span class="p">}&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">encode_kwargs&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="p">{&lt;/span>&lt;span class="s1">&amp;#39;normalize_embeddings&amp;#39;&lt;/span>&lt;span class="p">:&lt;/span> &lt;span class="kc">False&lt;/span>&lt;span class="p">}&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">hf&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">HuggingFaceEmbeddings&lt;/span>&lt;span class="p">(&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">model_name&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="n">model_name&lt;/span>&lt;span class="p">,&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">model_kwargs&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="n">model_kwargs&lt;/span>&lt;span class="p">,&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">encode_kwargs&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="n">encode_kwargs&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="c1"># load all the transcripts stored in the data folder&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">loader&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">DirectoryLoader&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="s1">&amp;#39;data_test/&amp;#39;&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">glob&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="s2">&amp;#34;**/*.txt&amp;#34;&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">show_progress&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="kc">True&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">docs&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">loader&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">load&lt;/span>&lt;span class="p">()&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">text_splitter&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">RecursiveCharacterTextSplitter&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">chunk_size&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="mi">200&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">chunk_overlap&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="mi">0&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">documents&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">text_splitter&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">split_documents&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">docs&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="c1"># cache the embeddings for faster loadup&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">fs&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">LocalFileStore&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="s2">&amp;#34;./cache/&amp;#34;&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">cached_embedder&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">CacheBackedEmbeddings&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">from_bytes_store&lt;/span>&lt;span class="p">(&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">hf&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">fs&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">namespace&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="s2">&amp;#34;sentence&amp;#34;&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="c1"># create the vector db&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">db&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">FAISS&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">from_documents&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">documents&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">cached_embedder&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">return&lt;/span> &lt;span class="n">db&lt;/span>
&lt;/span>&lt;/span>&lt;/code>&lt;/pre>&lt;/div>&lt;h4 id="the-llm">The LLM&lt;/h4>
&lt;p>While the authors use GPT 3.5, not wanting to spend any money, I instead resorted to using an Open-Source LLM, Flan-T5 from Google &lt;a href="https://huggingface.co/t5-large" target="_blank" rel="noopener">https://huggingface.co/t5-large&lt;/a>.&lt;/p>
&lt;p>We’ll also define our input question, to which we want the answer from our LLM. Note that the LLM doesn’t have knowledge about this, since it&amp;rsquo;s not in its training data, and it will have to figure out how to use the external knowledge from the vector db to answer this question.&lt;/p>
&lt;div class="highlight">&lt;pre tabindex="0" class="chroma">&lt;code class="language-python" data-lang="python">&lt;span class="line">&lt;span class="cl">&lt;span class="c1"># initialize the LLM and its tokenizer, we are using Flan T5 Large for this&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="n">tokenizer&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">T5Tokenizer&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">from_pretrained&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="s2">&amp;#34;google/flan-t5-large&amp;#34;&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="n">model&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">T5ForConditionalGeneration&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">from_pretrained&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="s2">&amp;#34;google/flan-t5-large&amp;#34;&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="c1"># function to get the prediction and scores from the LLM, given a prompt&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="k">def&lt;/span> &lt;span class="nf">get_prediction_and_scores&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">prompt&lt;/span>&lt;span class="p">):&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">input_ids&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">tokenizer&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">prompt&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">return_tensors&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="s2">&amp;#34;pt&amp;#34;&lt;/span>&lt;span class="p">)&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">input_ids&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">outputs&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">model&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">generate&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">input_ids&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">output_scores&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="kc">True&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">return_dict_in_generate&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="kc">True&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">max_length&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="mi">100&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">generated_sequence&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">outputs&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">sequences&lt;/span>&lt;span class="p">[&lt;/span>&lt;span class="mi">0&lt;/span>&lt;span class="p">]&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="c1"># get the probability scores for each generated token&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">transition_scores&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">torch&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">exp&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">model&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">compute_transition_scores&lt;/span>&lt;span class="p">(&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">outputs&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">sequences&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">outputs&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">scores&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">normalize_logits&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="kc">True&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="p">)[&lt;/span>&lt;span class="mi">0&lt;/span>&lt;span class="p">])&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">return&lt;/span> &lt;span class="n">tokenizer&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">decode&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">generated_sequence&lt;/span>&lt;span class="p">),&lt;/span> &lt;span class="n">generated_sequence&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">transition_scores&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="c1"># the input prompt for the LLM containing the question we want to ask&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="n">input_text&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="s2">&amp;#34;Q: What are streaming LLMs in the context of Large Language Models? Give a brief overview of the paper.&lt;/span>&lt;span class="se">\n&lt;/span>&lt;span class="s2">A:&amp;#34;&lt;/span>
&lt;/span>&lt;/span>&lt;/code>&lt;/pre>&lt;/div>&lt;p>We also return the probability for each generated token, since that will come in use later when deciding about whether to do a retrieval step or not.&lt;/p>
&lt;h4 id="the-flarestep">The FLARE step&lt;/h4>
&lt;div class="highlight">&lt;pre tabindex="0" class="chroma">&lt;code class="language-python" data-lang="python">&lt;span class="line">&lt;span class="cl">&lt;span class="c1"># keep generating tokens until we get a &amp;lt;/s&amp;gt; token&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="k">while&lt;/span> &lt;span class="kc">True&lt;/span>&lt;span class="p">:&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="c1"># get the prediction and scores from the LLM, given the input and all the tokens generated so far&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">generated_sequence&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">tokens&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">scores&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">get_prediction_and_scores&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">input_text&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="c1"># if any token is low in confidence, then do a RAG step&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">if&lt;/span> &lt;span class="n">torch&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">min&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">scores&lt;/span>&lt;span class="p">)&lt;/span> &lt;span class="o">&amp;lt;&lt;/span> &lt;span class="mf">0.1&lt;/span>&lt;span class="p">:&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="c1"># extract all tokens with high confidence as query&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">high_confidence_tokens&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">tokens&lt;/span>&lt;span class="p">[&lt;/span>&lt;span class="n">torch&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">where&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">scores&lt;/span> &lt;span class="o">&amp;gt;&lt;/span> &lt;span class="mf">0.1&lt;/span>&lt;span class="p">)]&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">query&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">tokenizer&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">decode&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">high_confidence_tokens&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="c1"># get the context from the vector db&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">docs&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">db&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">similarity_search&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">query&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">context&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="s2">&amp;#34;&lt;/span>&lt;span class="se">\n&lt;/span>&lt;span class="s2">&amp;#34;&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">join&lt;/span>&lt;span class="p">([&lt;/span>&lt;span class="n">doc&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">page_content&lt;/span> &lt;span class="k">for&lt;/span> &lt;span class="n">doc&lt;/span> &lt;span class="ow">in&lt;/span> &lt;span class="n">docs&lt;/span>&lt;span class="p">])&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">new_input_text&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="sa">f&lt;/span>&lt;span class="s2">&amp;#34;Given the below context:&lt;/span>&lt;span class="se">\n&lt;/span>&lt;span class="si">{&lt;/span>&lt;span class="n">context&lt;/span>&lt;span class="si">}&lt;/span>&lt;span class="se">\n\n&lt;/span>&lt;span class="s2"> Answer the following &lt;/span>&lt;span class="se">\n&lt;/span>&lt;span class="si">{&lt;/span>&lt;span class="n">input_text&lt;/span>&lt;span class="si">}&lt;/span>&lt;span class="se">\n&lt;/span>&lt;span class="s2">&amp;#34;&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="c1"># get the prediction and scores from the LLM, given the new input&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">generated_sequence&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">_&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">_&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">get_prediction_and_scores&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">new_input_text&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">input_text&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="sa">f&lt;/span>&lt;span class="s2">&amp;#34;&lt;/span>&lt;span class="si">{&lt;/span>&lt;span class="n">input_text&lt;/span>&lt;span class="si">}&lt;/span>&lt;span class="s2"> &lt;/span>&lt;span class="si">{&lt;/span>&lt;span class="n">generated_sequence&lt;/span>&lt;span class="si">}&lt;/span>&lt;span class="s2">&amp;#34;&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">if&lt;/span> &lt;span class="s2">&amp;#34;&amp;lt;/s&amp;gt;&amp;#34;&lt;/span> &lt;span class="ow">in&lt;/span> &lt;span class="n">input_text&lt;/span>&lt;span class="p">:&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">break&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">else&lt;/span>&lt;span class="p">:&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="c1"># if all tokens are high in confidence, then just add the generated tokens to the input&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">input_text&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="sa">f&lt;/span>&lt;span class="s2">&amp;#34;&lt;/span>&lt;span class="si">{&lt;/span>&lt;span class="n">input_text&lt;/span>&lt;span class="si">}&lt;/span>&lt;span class="s2"> &lt;/span>&lt;span class="si">{&lt;/span>&lt;span class="n">generated_sequence&lt;/span>&lt;span class="si">}&lt;/span>&lt;span class="s2">&amp;#34;&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">if&lt;/span> &lt;span class="s2">&amp;#34;&amp;lt;/s&amp;gt;&amp;#34;&lt;/span> &lt;span class="ow">in&lt;/span> &lt;span class="n">input_text&lt;/span>&lt;span class="p">:&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">break&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="c1"># print the final output&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="nb">print&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="s2">&amp;#34;Final output:&amp;#34;&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">input_text&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;/code>&lt;/pre>&lt;/div>&lt;p>Going step by step&lt;/p>
&lt;ul>
&lt;li>We start with our input query, and give it to the LLM and get the answer from the LLM. ( In later steps, this will also contain all the stuff generated from the LLM yet)&lt;/li>
&lt;li>Now, if the probability of &lt;em>any&lt;/em> token in the generated output is lower than a particular threshold. We trigger a RAG step. This is because the LLM is not confident about its answer and we want to augment it with extra context. If all tokens were high confidence, we take the output of the LLM as it is and repeat.&lt;/li>
&lt;li>For the RAG step, our query is all the tokens in the generated sequence, which have a high probability. We don’t take the tokens with low probability, since they can be noisy.&lt;/li>
&lt;li>We then get the closest chunks from our DB using a vector search, and append this to our input as context.&lt;/li>
&lt;li>We then again query the LLM, but this time with the additional context, and get our final generated sentence and append it to our input.&lt;/li>
&lt;/ul>
&lt;p>As a dry run from a sample input, we get the below response :&lt;/p>
&lt;div class="highlight">&lt;pre tabindex="0" class="chroma">&lt;code class="language-fallback" data-lang="fallback">&lt;span class="line">&lt;span class="cl">input_text: Q: What are streaming LLMs in the context of Large Language Models? Give a brief overview of the paper.
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">generated_sequence: &amp;lt;pad&amp;gt; The paper presents a new approach to streaming LLMs, based on the use of a Streaming LLM model.&amp;lt;/s&amp;gt;
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">----RAG step----
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">query: &amp;lt;pad&amp;gt;The paper presents a new approach to streaming LLMs, based on use of Streaming LLM model.
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">new_input_text: Given the below context:
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">say that deploying LMS in streaming applications such as multi-round dialogue where long interactions are expected is is urgently needed but has some challenges so this approach is for streaming
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">say 4,000 tokens you don&amp;#39;t want to stop after 4,000 tokens you want to keep on going on and on and that is called streaming uh so they develop an approach for it which enables llms which were trained
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">input the streaming llm but only summarize the concluding paragraphs which might not be very uh insightful so
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">infinite length inputs without sacrificing efficiency or performance and if you see here the left you have so this approach is also called streaming llm so without llm on the left without streaming
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Answer the following
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Q: What are streaming LLMs in the context of Large Language Models? Give a brief overview of the paper.
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">A:
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">generated_sequence after RAG: &amp;lt;pad&amp;gt; The paper describes a streaming approach to large language models for multi-round dialogue where long interactions are expected.&amp;lt;/s&amp;gt;
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">------------
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Final output: Q: What are streaming LLMs in the context of Large Language Models? Give a brief overview of the paper.
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">A: &amp;lt;pad&amp;gt; The paper describes a streaming approach to large language models for multi-round dialogue where long interactions are expected.&amp;lt;/s&amp;gt;
&lt;/span>&lt;/span>&lt;/code>&lt;/pre>&lt;/div>&lt;p>So there you have it, a minimal implementation of FLARE. There are A LOT of things which I have glossed over since this was just a quick hack I tried out. Some of them are&lt;/p>
&lt;ul>
&lt;li>Prompt Engineering for the specific tasks&lt;/li>
&lt;li>Attempting the different datasets/tasks in the original paper&lt;/li>
&lt;li>In the query generation, I used what they call the Masked sentences approach, but they also have another approach in which they use another LLM to help generate a query, for the low-confidence tokens.&lt;/li>
&lt;/ul>
&lt;p>Maybe we’ll do this sometime if I get time. That’s all for this post!&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/flare-advanced-rag-implemented-from-scratch-07ca75c89800" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>A hacky minimal Tree of Thoughts implementation</title><link>https://ayushtues.github.io/blog/a-hacky-minimal-tree-of-thoughts-implementation/</link><pubDate>Wed, 08 Nov 2023 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/a-hacky-minimal-tree-of-thoughts-implementation/</guid><description>&lt;p>Wanting to learn more about LLMs, I stumbled upon a particular approach called Tree of Thoughts prompting or ToT. The idea seemed interesting, but also relatively straightforward to implement, and so I thought, why not code up something myself? Hence the post.&lt;/p>
&lt;p>Now I spend &amp;lt; 2 hrs coding this up, and it involves little to no prompt engineering. So it doesn’t really solve the problem it was supposed to solve, but it does demonstrate the concept well enough.&lt;/p>
&lt;h4 id="what-istot">What is ToT?&lt;/h4>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/a-hacky-minimal-tree-of-thoughts-implementation/image-1_hu2800935882799567865.webp 400w,
/blog/a-hacky-minimal-tree-of-thoughts-implementation/image-1_hu2361567517851420480.webp 760w,
/blog/a-hacky-minimal-tree-of-thoughts-implementation/image-1_hu15925731139548966104.webp 1200w"
src="https://ayushtues.github.io/blog/a-hacky-minimal-tree-of-thoughts-implementation/image-1_hu2800935882799567865.webp"
width="760"
height="359"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>In normal IO Prompting, you take in a prompt &amp;amp; feed it into the LLM to get the output.&lt;/p>
&lt;p>CoT is similar, but you ask the model to “think step by step”, so the LLM (hopefully) breaks down its output to different steps/thoughts to reason more explicitly and structurally about the problem. COT-SC is just running multiple CoT runs and then averaging out the result.&lt;/p>
&lt;p>Now, these approaches have a problem; from the paper itself, two key shortcomings of existing approaches are: 1) Locally, they do not explore different possibilities at each state i.e. they don’t have the branches of the tree. 2) Globally, they do not incorporate any type of planning, lookahead, or backtracking to help evaluate these different options– the kind of heuristic-guided search that seems characteristic of human problem-solving.&lt;/p>
&lt;p>Enter ToT, which allows LLMs to keep track of multiple thought paths in a tree of thoughts/states. It essentially is a tree search over states, where each state is an intermediate thought generated by an LLM, to solve the problem.&lt;/p>
&lt;p>Anyway, we’ll get to it, once we go into the code, its much easier to understand from code.&lt;/p>
&lt;p>But first, what’s the problem we’re trying to solve? Well…Game of 24&lt;/p>
&lt;p>Game of 24 is a mathematical reasoning challenge, where the goal is to use 4 numbers and basic arithmetic operations (+-*/) to obtain 24. For example, given input “4 9 10 13”, a solution output could be “(10- 4) * (13- 9) = 24”.&lt;/p>
&lt;p>Okay that’s the game, but we need an LLM for the task. I used the Llama-2–7b-chat model by Replicate — &lt;a href="https://replicate.com/meta/llama-2-7b-chat" target="_blank" rel="noopener">https://replicate.com/meta/llama-2-7b-chat&lt;/a>, since I have no OpenAI credits lol.&lt;/p>
&lt;div class="highlight">&lt;pre tabindex="0" class="chroma">&lt;code class="language-python" data-lang="python">&lt;span class="line">&lt;span class="cl">&lt;span class="c1"># Calls the LLM with the given prompt and returns the output&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="k">def&lt;/span> &lt;span class="nf">run_llm&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">prompt&lt;/span>&lt;span class="p">):&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">output&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">replicate&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">run&lt;/span>&lt;span class="p">(&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="s2">&amp;#34;meta/llama-2-7b-chat:13c3cdee13ee059ab779f0291d29054dab00a47dad8261375654de5540165fb0&amp;#34;&lt;/span>&lt;span class="p">,&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="nb">input&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="p">{&lt;/span>&lt;span class="s2">&amp;#34;prompt&amp;#34;&lt;/span>&lt;span class="p">:&lt;/span> &lt;span class="n">prompt&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="s2">&amp;#34;max_new_tokens&amp;#34;&lt;/span>&lt;span class="p">:&lt;/span> &lt;span class="mi">500&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="s2">&amp;#34;temperature&amp;#34;&lt;/span>&lt;span class="p">:&lt;/span> &lt;span class="mf">0.55&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="s2">&amp;#34;top_p1&amp;#34;&lt;/span>&lt;span class="p">:&lt;/span> &lt;span class="mf">1.0&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="s2">&amp;#34;repetition_penalty&amp;#34;&lt;/span>&lt;span class="p">:&lt;/span> &lt;span class="mf">1.0&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="s2">&amp;#34;system_prompt&amp;#34;&lt;/span>&lt;span class="p">:&lt;/span> &lt;span class="s2">&amp;#34;&amp;#34;&lt;/span>&lt;span class="p">}&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">output_all&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="s2">&amp;#34;&amp;#34;&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">for&lt;/span> &lt;span class="n">item&lt;/span> &lt;span class="ow">in&lt;/span> &lt;span class="n">output&lt;/span>&lt;span class="p">:&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">output_all&lt;/span> &lt;span class="o">+=&lt;/span> &lt;span class="n">item&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">return&lt;/span> &lt;span class="n">output_all&lt;/span>
&lt;/span>&lt;/span>&lt;/code>&lt;/pre>&lt;/div>&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/a-hacky-minimal-tree-of-thoughts-implementation/image-2_hu2501849489082210360.webp 400w,
/blog/a-hacky-minimal-tree-of-thoughts-implementation/image-2_hu14577108568510312199.webp 760w,
/blog/a-hacky-minimal-tree-of-thoughts-implementation/image-2_hu582332091459852538.webp 1200w"
src="https://ayushtues.github.io/blog/a-hacky-minimal-tree-of-thoughts-implementation/image-2_hu2501849489082210360.webp"
width="760"
height="225"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>Next, we need a “&lt;strong>thought generator&lt;/strong>”, which takes a state and generates possible next steps from it.&lt;/p>
&lt;p>For our game of 24, I used the below prompt, in which we first state the overall task. Then ask the LLM to generate possible next steps, also providing a few examples in context to make it a bit easier.&lt;/p>
&lt;div class="highlight">&lt;pre tabindex="0" class="chroma">&lt;code class="language-fallback" data-lang="fallback">&lt;span class="line">&lt;span class="cl">Your goal is to use the given numbers and the basic arithmetic operations
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">(+, -, *, /) to obtain the number 24. You can use each number only once,
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">but you can use the operations in any order and as many times as you want.
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">This task will take multiple steps.
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">For the current step, you choose two numbers and perform an arithmetic operation on them.
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Examples
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Input: 4 9 10 13
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Possible next steps:
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Output1: 4 + 9 = 13 (left 10 13 13)
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Output2: 10 - 4 = 6 (left: 6 9 13)
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Output3: 10/9 = 1 (left: 4 1 13)
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Output4: 9*13 = 117 (left: 4 10 117)
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Input: 4 10 12 1
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Possible next steps:
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Output1: 12 - 10 = 2 (left: 4 2 1)
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Output2: 4 * 10 = 40 (left: 40 12 1)
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Output3: 12 + 1 = 13 (left: 4 10 13)
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Output4 12/10 = 1 (left: 4 1 1)
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Now for the below input
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Input: {state}
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Possible next steps:
&lt;/span>&lt;/span>&lt;/code>&lt;/pre>&lt;/div>&lt;p>This will give an output like&lt;/p>
&lt;div class="highlight">&lt;pre tabindex="0" class="chroma">&lt;code class="language-fallback" data-lang="fallback">&lt;span class="line">&lt;span class="cl">Great, let&amp;#39;s get started! For the input numbers 5, 10, 12, and 11, we have several options for how to proceed:
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Output1: 10 - 5 = 5 (left: 10 12 11)
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Output2: 12 / 5 = 2.4 (left: 10 12 11)
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Output3: 11 + 5 = 16 (left: 10 12 16)
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Output4: 12 *
&lt;/span>&lt;/span>&lt;/code>&lt;/pre>&lt;/div>&lt;p>We just want the part following “left:”, so I wrote a quick parser for that, which takes something like the above and gives us the next states in a nice list.&lt;/p>
&lt;div class="highlight">&lt;pre tabindex="0" class="chroma">&lt;code class="language-python" data-lang="python">&lt;span class="line">&lt;span class="cl">&lt;span class="c1"># Parses the output of the proposal generator to extract the proposals&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="k">def&lt;/span> &lt;span class="nf">extract_proposals&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">text&lt;/span>&lt;span class="p">):&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">text&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">text&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">split&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="s2">&amp;#34;&lt;/span>&lt;span class="se">\n&lt;/span>&lt;span class="s2">&amp;#34;&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">text&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="p">[&lt;/span>&lt;span class="n">item&lt;/span> &lt;span class="k">for&lt;/span> &lt;span class="n">item&lt;/span> &lt;span class="ow">in&lt;/span> &lt;span class="n">text&lt;/span> &lt;span class="k">if&lt;/span> &lt;span class="s2">&amp;#34;Output&amp;#34;&lt;/span> &lt;span class="ow">in&lt;/span> &lt;span class="n">item&lt;/span>&lt;span class="p">]&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">proposals&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="p">[]&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">for&lt;/span> &lt;span class="n">x&lt;/span> &lt;span class="ow">in&lt;/span> &lt;span class="n">text&lt;/span>&lt;span class="p">:&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">x&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">x&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">split&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="s2">&amp;#34;left:&amp;#34;&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">if&lt;/span> &lt;span class="nb">len&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">x&lt;/span>&lt;span class="p">)&lt;/span> &lt;span class="o">==&lt;/span> &lt;span class="mi">2&lt;/span> &lt;span class="p">:&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">x&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">x&lt;/span>&lt;span class="p">[&lt;/span>&lt;span class="mi">1&lt;/span>&lt;span class="p">][:&lt;/span>&lt;span class="o">-&lt;/span>&lt;span class="mi">1&lt;/span>&lt;span class="p">]&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">proposals&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">append&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">x&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">return&lt;/span> &lt;span class="n">proposals&lt;/span>
&lt;/span>&lt;/span>&lt;/code>&lt;/pre>&lt;/div>&lt;p>Next, since this is a tree search, we’ll also need a heuristic to evaluate how good a particular state is. This is called the “&lt;strong>state evaluator”.&lt;/strong> It takes in a state and gives it a numeric score of how “good” it is.&lt;/p>
&lt;p>For the game of 24, we use the below prompt.&lt;/p>
&lt;div class="highlight">&lt;pre tabindex="0" class="chroma">&lt;code class="language-fallback" data-lang="fallback">&lt;span class="line">&lt;span class="cl">Evaluate if given numbers can reach 24 using basic arithmetic operations
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">(+, -, *, /).You can use each number only once,
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">but you can use the operations in any order and as many times as you want.
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Some examples are:
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Input: 10 14 -&amp;gt; 10 + 14 = 24. -&amp;gt; {Output: &amp;#34;sure&amp;#34;}
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Input: 4 9 10 13 -&amp;gt; (10- 4) * (13- 9) = 24. -&amp;gt; {Output: &amp;#34;sure&amp;#34;}
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Input: 20 10: Not possible -&amp;gt; {Output: &amp;#34;impossible&amp;#34;}
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">Can the numbers {INPUT} reach 24?
&lt;/span>&lt;/span>&lt;/code>&lt;/pre>&lt;/div>&lt;p>Since we are asking the model to output sure/impossible/likely, we need a parser to convert this text output to a decimal score&lt;/p>
&lt;div class="highlight">&lt;pre tabindex="0" class="chroma">&lt;code class="language-python" data-lang="python">&lt;span class="line">&lt;span class="cl">&lt;span class="c1"># Parses the output of the state evaluator to extract the score&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="k">def&lt;/span> &lt;span class="nf">extract_evaluation&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">text&lt;/span>&lt;span class="p">):&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">text&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">text&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">lower&lt;/span>&lt;span class="p">()&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">if&lt;/span> &lt;span class="s2">&amp;#34;impossible&amp;#34;&lt;/span> &lt;span class="ow">in&lt;/span> &lt;span class="n">text&lt;/span>&lt;span class="p">:&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">return&lt;/span> &lt;span class="mi">0&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">elif&lt;/span> &lt;span class="s2">&amp;#34;sure&amp;#34;&lt;/span> &lt;span class="ow">in&lt;/span> &lt;span class="n">text&lt;/span>&lt;span class="p">:&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">return&lt;/span> &lt;span class="mi">1&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">else&lt;/span>&lt;span class="p">:&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">return&lt;/span> &lt;span class="mf">0.5&lt;/span>
&lt;/span>&lt;/span>&lt;/code>&lt;/pre>&lt;/div>&lt;p>So now we have all the required components for our tree search. We’ll be doing a BFS search with the following steps&lt;/p>
&lt;ol>
&lt;li>Pass each state in the current set to the thought generator and generate proposals from it.&lt;/li>
&lt;li>Each run through the thought generator generates multiple proposals, and we pass each state multiple times through the proposal generator and store all the proposals.&lt;/li>
&lt;li>Evaluate all the proposals through our state evaluator, run each state through it multiple times, and average out the score.&lt;/li>
&lt;li>Sort the proposals by the score, and take only the top K.&lt;/li>
&lt;li>Repeat this until you run out of compute budget OR all your proposals have a crappy score.&lt;/li>
&lt;li>Output the best proposal&lt;/li>
&lt;/ol>
&lt;div class="highlight">&lt;pre tabindex="0" class="chroma">&lt;code class="language-python" data-lang="python">&lt;span class="line">&lt;span class="cl">&lt;span class="n">curr_states&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="p">[&lt;/span>&lt;span class="s2">&amp;#34;5 10 12 11&amp;#34;&lt;/span>&lt;span class="p">]&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="n">TREE_DEPTH&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="mi">3&lt;/span> &lt;span class="c1"># number of steps to take &lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="n">PROPOSAL_RUNS_PER_STATE&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="mi">2&lt;/span> &lt;span class="c1"># number of calls to proposal generator per state, we concatenate the proposals from these runs&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="n">EVAL_RUNS_PER_STATE&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="mi">3&lt;/span> &lt;span class="c1"># number of calls to state evaluator per state, we average the score from these runs&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="n">BRANCH_FACTOR&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="mi">3&lt;/span> &lt;span class="c1"># number of top - k states to keep at each step&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="k">for&lt;/span> &lt;span class="n">_&lt;/span> &lt;span class="ow">in&lt;/span> &lt;span class="nb">range&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">TREE_DEPTH&lt;/span>&lt;span class="p">):&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">proposal_and_score&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="p">[]&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="c1"># curr_states contains the best k states from the previous step OR the initial input state&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">for&lt;/span> &lt;span class="n">state&lt;/span> &lt;span class="ow">in&lt;/span> &lt;span class="n">curr_states&lt;/span>&lt;span class="p">:&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">proposal_prompt&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="sa">f&lt;/span>&lt;span class="s2">&amp;#34;Your goal is to use the given numbers and the basic arithmetic operations (+, -, *, /) to obtain the number 24. You can use each number only once, but you can use the operations in any order and as many times as you want. This task will take multiple steps. For the current step, you choose two numbers and perform an arithmetic operation on them. &lt;/span>&lt;span class="se">\n\n&lt;/span>&lt;span class="s2">Examples&lt;/span>&lt;span class="se">\n&lt;/span>&lt;span class="s2">Input: 4 9 10 13&lt;/span>&lt;span class="se">\n&lt;/span>&lt;span class="s2">Possible next steps:&lt;/span>&lt;span class="se">\n&lt;/span>&lt;span class="s2">Output1: 4 + 9 = 13 (left 10 13 13)&lt;/span>&lt;span class="se">\n&lt;/span>&lt;span class="s2">Output2: 10 - 4 = 6 (left: 6 9 13)&lt;/span>&lt;span class="se">\n&lt;/span>&lt;span class="s2">Output3: 10/9 = 1 (left: 4 1 13)&lt;/span>&lt;span class="se">\n&lt;/span>&lt;span class="s2">Output4: 9*13 = 117 (left: 4 10 117)&lt;/span>&lt;span class="se">\n\n&lt;/span>&lt;span class="s2">Input: 4 10 12 1&lt;/span>&lt;span class="se">\n&lt;/span>&lt;span class="s2">Possible next steps: &lt;/span>&lt;span class="se">\n&lt;/span>&lt;span class="s2">Output1: 12 - 10 = 2 (left: 4 2 1)&lt;/span>&lt;span class="se">\n&lt;/span>&lt;span class="s2">Output2: 4 * 10 = 40 (left: 40 12 1)&lt;/span>&lt;span class="se">\n&lt;/span>&lt;span class="s2">Output3: 12 + 1 = 13 (left: 4 10 13)&lt;/span>&lt;span class="se">\n&lt;/span>&lt;span class="s2">Output4 12/10 = 1 (left: 4 1 1)&lt;/span>&lt;span class="se">\n\n&lt;/span>&lt;span class="s2">Now for the below input&lt;/span>&lt;span class="se">\n&lt;/span>&lt;span class="s2">Input: &lt;/span>&lt;span class="si">{&lt;/span>&lt;span class="n">state&lt;/span>&lt;span class="si">}&lt;/span>&lt;span class="se">\n&lt;/span>&lt;span class="s2">Possible next steps:&amp;#34;&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="c1"># generate proposals for each state&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">proposals&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="p">[]&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">for&lt;/span> &lt;span class="n">_&lt;/span> &lt;span class="ow">in&lt;/span> &lt;span class="nb">range&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">PROPOSAL_RUNS_PER_STATE&lt;/span>&lt;span class="p">):&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">proposals&lt;/span> &lt;span class="o">+=&lt;/span> &lt;span class="n">extract_proposals&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">run_llm&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">proposal_prompt&lt;/span>&lt;span class="p">))&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="c1"># get the score for each proposal&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">for&lt;/span> &lt;span class="n">proposal&lt;/span> &lt;span class="ow">in&lt;/span> &lt;span class="n">proposals&lt;/span>&lt;span class="p">:&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">eval_prompt&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="s2">&amp;#34;Evaluate if given numbers can reach 24 using basic arithmetic operations (+, -, *, /).You can use each number only once, but you can use the operations in any order and as many times as you want.&lt;/span>&lt;span class="se">\n\n&lt;/span>&lt;span class="s2">Some examples are:&lt;/span>&lt;span class="se">\n&lt;/span>&lt;span class="s2">Input: 10 14 -&amp;gt; 10 + 14 = 24. -&amp;gt; {Output: &lt;/span>&lt;span class="se">\&amp;#34;&lt;/span>&lt;span class="s2">sure&lt;/span>&lt;span class="se">\&amp;#34;&lt;/span>&lt;span class="s2">}&lt;/span>&lt;span class="se">\n&lt;/span>&lt;span class="s2">Input: 4 9 10 13 -&amp;gt; (10- 4) * (13- 9) = 24. -&amp;gt; {Output: &lt;/span>&lt;span class="se">\&amp;#34;&lt;/span>&lt;span class="s2">sure&lt;/span>&lt;span class="se">\&amp;#34;&lt;/span>&lt;span class="s2">}&lt;/span>&lt;span class="se">\n&lt;/span>&lt;span class="s2">Input: 20 10: Not possible -&amp;gt; {Output: &lt;/span>&lt;span class="se">\&amp;#34;&lt;/span>&lt;span class="s2">impossible&lt;/span>&lt;span class="se">\&amp;#34;&lt;/span>&lt;span class="s2">}&lt;/span>&lt;span class="se">\n\n&lt;/span>&lt;span class="s2">Can the numbers&amp;#34;&lt;/span> &lt;span class="o">+&lt;/span> &lt;span class="n">proposal&lt;/span> &lt;span class="o">+&lt;/span> &lt;span class="s2">&amp;#34;reach 24?&amp;#34;&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">score&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="mi">0&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">for&lt;/span> &lt;span class="n">_&lt;/span> &lt;span class="ow">in&lt;/span> &lt;span class="nb">range&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">EVAL_RUNS_PER_STATE&lt;/span>&lt;span class="p">):&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">score&lt;/span> &lt;span class="o">+=&lt;/span> &lt;span class="n">extract_evaluation&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">run_llm&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">eval_prompt&lt;/span>&lt;span class="p">))&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">proposal_and_score&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">append&lt;/span>&lt;span class="p">((&lt;/span>&lt;span class="n">proposal&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">score&lt;/span>&lt;span class="o">/&lt;/span>&lt;span class="mi">3&lt;/span>&lt;span class="p">))&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="c1"># sort proposals by score&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">proposal_and_score&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">sort&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">key&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="k">lambda&lt;/span> &lt;span class="n">x&lt;/span>&lt;span class="p">:&lt;/span> &lt;span class="n">x&lt;/span>&lt;span class="p">[&lt;/span>&lt;span class="mi">1&lt;/span>&lt;span class="p">],&lt;/span> &lt;span class="n">reverse&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="kc">True&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="c1"># keep the top k proposals&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">curr_states&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="p">[&lt;/span>&lt;span class="n">item&lt;/span>&lt;span class="p">[&lt;/span>&lt;span class="mi">0&lt;/span>&lt;span class="p">]&lt;/span> &lt;span class="k">for&lt;/span> &lt;span class="n">item&lt;/span> &lt;span class="ow">in&lt;/span> &lt;span class="n">proposal_and_score&lt;/span>&lt;span class="p">[:&lt;/span>&lt;span class="n">BRANCH_FACTOR&lt;/span>&lt;span class="p">]]&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="c1"># print the best state&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="nb">print&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">curr_states&lt;/span>&lt;span class="p">[&lt;/span>&lt;span class="mi">0&lt;/span>&lt;span class="p">])&lt;/span>
&lt;/span>&lt;/span>&lt;/code>&lt;/pre>&lt;/div>&lt;p>That’s mostly ToT in a nutshell, although a really, really minimal and totally not working implementation of it. You can find the source code here- &lt;a href="https://github.com/ayushtues/tot_from_scratch/blob/main/replicate_tot.py" target="_blank" rel="noopener">ayushtues/tot_from_scratch (github.com)&lt;/a>.&lt;/p>
&lt;p>To enhance it, we can do the following.&lt;/p>
&lt;ol>
&lt;li>Prompt Engineering to improve the proposal and evaluation prompts by a LOT&lt;/li>
&lt;li>Multi-threading to run executions of different states in parallel&lt;/li>
&lt;li>Other search methods like DFS/A*&lt;/li>
&lt;li>Making the code more clean/extendable by following OOPS — A very good clean implementation, with a lot more features, is &lt;a href="https://github.com/kyegomez/tree-of-thoughts" target="_blank" rel="noopener">https://github.com/kyegomez/tree-of-thoughts&lt;/a>.&lt;/li>
&lt;/ol>
&lt;p>Anyway, this was just a small hack to learn more about LLMs, wasn’t really the aim to make it work perfectly, that&amp;rsquo;s it for the post as well!&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/a-hacky-minimal-tree-of-thoughts-implementation-47474934fa17" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Keeping up with ML</title><link>https://ayushtues.github.io/blog/keeping-up-with-ml/</link><pubDate>Wed, 25 Oct 2023 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/keeping-up-with-ml/</guid><description>&lt;figure>&lt;img src="https://ayushtues.github.io/blog/keeping-up-with-ml/image-1.jpeg"
alt="Generated by Bing Image Creator">&lt;figcaption>
&lt;p>Generated by Bing Image Creator&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Keeping up with ML is hard, period. Doing that, vis-a-vis a day job is harder. Add all the beautiful things that come with adulting ( YAAAAAAYYYY I AM AN ADULT NOW…..AM I?…WAIT, WTF, ME?) makes it seem like a seemingly labyrinthine task.&lt;/p>
&lt;p>So how do I do it? Well, I try, somehow, to do whatever I can to remain relevant in a field that seems to be reinventing itself every 12 months ( do I use GANs now or Diffusion models? LSTM or Transformers, which transformer? And don’t get me started on LLMs.)&lt;/p>
&lt;h3 id="reading-papers">Reading Papers&lt;/h3>
&lt;p>This one should be obvious: while I can’t do much research on my own while in a job that has nothing to do with it, I can still stand on the shoulders of giants by reading the latest papers. With the crazy amount of work published almost daily, finding relevant works is tricky, but the following sources work well for me.&lt;/p>
&lt;ul>
&lt;li>Twitter — Follow the right folks on Twitter, and your feed becomes the best source of staying updated&lt;/li>
&lt;li>&lt;a href="https://paperswithcode.com/" target="_blank" rel="noopener">PapersWithCode&lt;/a> — Has the list of the top papers floating around in social media/GitHub.&lt;/li>
&lt;li>Youtubers — &lt;a href="https://www.youtube.com/@YannicKilcher" target="_blank" rel="noopener">@YannicKilcher&lt;/a>, &lt;a href="https://www.youtube.com/@hu-po" target="_blank" rel="noopener">@hu-po&lt;/a>, &lt;a href="https://www.youtube.com/@gabrielmongaras" target="_blank" rel="noopener">@gabrielmongaras&lt;/a>, &lt;a href="https://www.youtube.com/@TheAIEpiphany" target="_blank" rel="noopener">@TheAIEpiphany&lt;/a>- make life super-easy by covering the top papers being released in an easy-to-understand fashion.&lt;/li>
&lt;/ul>
&lt;p>In the beginning, it used to take me a lot of time to understand if a paper was actually good, and I spent a lot of time reading papers which weren’t that good. Now, with years of practice ( I read my first ML paper four years back ), I can mostly glaze through a paper&amp;rsquo;s abstract/some figures, take a call on its quality and then decide if I want to spend time going deeper into it.&lt;/p>
&lt;p>If I really like a paper, or if I want to increase my understanding of a particular field, I follow the Feynman method and try to learn it by teaching it. This led to the birth of my own YouTube channel, &lt;a href="https://www.youtube.com/channel/UCGEWHZv8Gn10Lk56PzCeFMA" target="_blank" rel="noopener">RR with deku&lt;/a>, where I go deep into papers I find interesting, much inspired by &lt;a href="https://www.youtube.com/@YannicKilcher" target="_blank" rel="noopener">@YannicKilcher&lt;/a>.&lt;/p>
&lt;h3 id="blogsytbooks">Blogs/YT/Books&lt;/h3>
&lt;p>While information on the most cutting-edge research topics is mostly found in research papers, there is a lot to be said about the wonderful blogs written by experts and amateurs alike. If I have to learn about a new topic, I usually start by reading a few blogs about it on the internet since they explain the concepts in the most user-friendly manner.&lt;/p>
&lt;p>While the blogs I read are usually whatever surfaces in the Google results, here are some of the blogs I follow&lt;/p>
&lt;ul>
&lt;li>&lt;strong>Lilian Weng’s blog (&lt;/strong>&lt;a href="https://lilianweng.github.io/" target="_blank" rel="noopener">https://lilianweng.github.io/&lt;/a>) Has great overviews/survey sort of blogs on popular topics like Diffusion Models/LLMs/RL. Good to get a 100 feet view of the topics&lt;/li>
&lt;li>&lt;strong>Eugene Yan’s blog (&lt;/strong>&lt;a href="https://eugeneyan.com/" target="_blank" rel="noopener">https://eugeneyan.com/&lt;/a>) — Has good real-world ML/MLOps/Career/Engineering articles.&lt;/li>
&lt;li>&lt;strong>Chip Huyen’s blog (&lt;/strong>&lt;a href="https://huyenchip.com/blog/" target="_blank" rel="noopener">https://huyenchip.com/blog/&lt;/a>) — Similar to above&lt;/li>
&lt;/ul>
&lt;p>Aside from these, I also like to follow the engineering blogs of some of the major companies using ML to get an idea of &lt;em>how ML is actually used in products&lt;/em> instead of just focusing on the algorithms; some good ones are&lt;/p>
&lt;ul>
&lt;li>Netflix’s tech blog (&lt;a href="https://netflixtechblog.com/" target="_blank" rel="noopener">https://netflixtechblog.com/&lt;/a>)&lt;/li>
&lt;li>Uber’s Engineering blog (&lt;a href="https://www.uber.com/en-IN/blog/engineering/" target="_blank" rel="noopener">https://www.uber.com/en-IN/blog/engineering/&lt;/a>&lt;strong>)&lt;/strong>&lt;/li>
&lt;li>Spotify’s blog (&lt;a href="https://engineering.atspotify.com/" target="_blank" rel="noopener">https://engineering.atspotify.com/&lt;/a>, &lt;a href="https://research.atspotify.com/blog/" target="_blank" rel="noopener">https://research.atspotify.com/blog/&lt;/a>)&lt;/li>
&lt;li>Pinterest&amp;rsquo;s blog (&lt;a href="https://medium.com/pinterest-engineering" target="_blank" rel="noopener">https://medium.com/pinterest-engineering&lt;/a>)&lt;/li>
&lt;/ul>
&lt;p>Similarly, YouTube has played a big big role in my self-learning. It has videos on literally anything I want to learn more about. You can also watch a course from Stanford/MIT to gain structured knowledge about stuff OR watch the proceedings of a popular conference to get a whirlwind tour of all that’s happening in the field.&lt;/p>
&lt;p>Recently, I have been picking up a lot more code walkthroughs through YT, in which some wonderful people go through the code of the models that make DL happen, and I have found that to be very helpful; some of the good ones are&lt;/p>
&lt;ul>
&lt;li>&lt;a href="https://www.youtube.com/@TheAIEpiphany" target="_blank" rel="noopener">The AI Ephipany&lt;/a> — Aleksa has a lot of very detailed walkthroughs of the GitHub repos of all the popular Diffusion Models, which I found to be very helpful.&lt;/li>
&lt;li>&lt;a href="https://www.youtube.com/@MachineLearningwithPhil" target="_blank" rel="noopener">Machine Learning With Phil&lt;/a> — My go-to channel for anything related to implementing Reinforcement Learning&lt;/li>
&lt;li>&lt;a href="https://www.youtube.com/@umarjamilai" target="_blank" rel="noopener">Umar Jamil&lt;/a> — Has really good implementation videos for LLMs&lt;/li>
&lt;/ul>
&lt;p>While I don’t read a lot of technical books, I find them to be a good source of structured information. But I’ll still rate them lower than Blogs/YT. Some good technical books I read this year are&lt;/p>
&lt;ul>
&lt;li>&lt;a href="https://www.amazon.in/Designing-Machine-Learning-Systems-Production-Ready/dp/1098107969" target="_blank" rel="noopener">Designing Machine Learning Systems&lt;/a>&lt;/li>
&lt;li>&lt;a href="https://www.amazon.in/Machine-Learning-Design-Patterns-Preparation/dp/9385889214/ref=pd_lpo_sccl_1/261-7753377-1961202?pd_rd_w=tnhcJ&amp;amp;content-id=amzn1.sym.3e03ac8b-a005-4b9e-bd7f-ae35f0b8a1e6&amp;amp;pf_rd_p=3e03ac8b-a005-4b9e-bd7f-ae35f0b8a1e6&amp;amp;pf_rd_r=AG5VRZ9V2YVBVQZEFTYG&amp;amp;pd_rd_wg=BEqAP&amp;amp;pd_rd_r=c8c7afdc-135a-4d05-adcb-aad8182045bb&amp;amp;pd_rd_i=9385889214&amp;amp;psc=1" target="_blank" rel="noopener">Machine Learning Design Patterns&lt;/a>&lt;/li>
&lt;li>&lt;a href="https://www.amazon.com/System-Design-Interview-insiders-Second/dp/B08CMF2CQF" target="_blank" rel="noopener">System Design Interview&lt;/a> &lt;a href="https://www.amazon.com/System-Design-Interview-insiders-Second/dp/B08CMF2CQF" target="_blank" rel="noopener">— An insider’s guide&lt;/a>&lt;/li>
&lt;/ul>
&lt;h3 id="opensourceside-projects">OpenSource/Side-projects&lt;/h3>
&lt;p>The above was mostly dealing with the theoretical end of things. But theory would only take you so far, and I am someone who likes to hack things a bit and I love to code.&lt;/p>
&lt;p>This year, doing open-source contributions has emerged as a decent way for me to keep working on my machine-learning implementation skills. Open-source can be a hard nut to crack, and I have failed multiple times previously in the process. Thankfully, this year, I was able to contribute to Huggingface Diffusers (&lt;a href="https://github.com/huggingface/diffusers" target="_blank" rel="noopener">https://github.com/huggingface/diffusers&lt;/a>). Contributing three models to their core pipelines ( &lt;a href="https://huggingface.co/docs/diffusers/api/pipelines/kandinsky" target="_blank" rel="noopener">Kandinksy&lt;/a>, &lt;a href="https://huggingface.co/docs/diffusers/api/pipelines/consistency_models" target="_blank" rel="noopener">Consistency Models&lt;/a>, &lt;a href="https://huggingface.co/docs/diffusers/main/en/api/pipelines/blip_diffusion" target="_blank" rel="noopener">Blip Diffusion&lt;/a>). I also made a small fix (&lt;a href="https://github.com/openai/consistency_models/pull/37" target="_blank" rel="noopener">PR&lt;/a>) to the OpenAI Consistency Models repo. I think the best way to get started with open-source is to get familiar with any one project, dive deep into it and then just look for something you can add to it.&lt;/p>
&lt;p>Apart from open-source, I did a few side projects, like implementing a diffusion model from scratch in &lt;a href="https://github.com/ayushtues/diffusion_from_scratch" target="_blank" rel="noopener">Pytorch&lt;/a>, &lt;a href="https://github.com/ayushtues/diffusion_from_scratch_JAX" target="_blank" rel="noopener">Jax&lt;/a>. This was motivated by me wanting to learn exactly how a diffusion model works, and I thought the best way to do so would be to implement one from scratch. The same goes for learning JAX. Similarly, I wanted to play around with LLMs a bit, so I ended up creating an AMA bot for my personal website (&lt;a href="https://blog.gopenai.com/creating-an-ama-bot-for-my-personal-website-using-llms-5201a641deae" target="_blank" rel="noopener">blog&lt;/a>)!&lt;/p>
&lt;p>I think that is most of what I do to keep up with ML, along with my job. The most important thing driving me is probably curiosity and wanting to get better and better at my craft and remain relevant. I try, but it&amp;rsquo;s a long journey!&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/keeping-up-with-ml-3341560ed4d3" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Creating an AMA bot for my personal website using LLMs</title><link>https://ayushtues.github.io/blog/creating-an-ama-bot-for-my-personal-website-using-llms/</link><pubDate>Tue, 22 Aug 2023 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/creating-an-ama-bot-for-my-personal-website-using-llms/</guid><description>&lt;blockquote>
&lt;p>Yes, I am doing an LLM post after cringing about the gazillions of the already existing ones, but it&amp;rsquo;s my blog, so fuck it.&lt;/p>
&lt;/blockquote>
&lt;p>Having wanted to learn about LLMs for a while, what better way to do so than hacking away at something using them? And the ecosystem surrounding them has developed so rapidly that it was surprisingly easy to get something running &lt;em>very, very&lt;/em> quickly.&lt;/p>
&lt;p>I had a personal portfolio website hosted on Github, and while it is a good way to know more about me, it feels static ( cause well, Github only supports static pages), and I wanted to add something dynamic to it. Enter LLMs.&lt;/p>
&lt;p>Here’s a quick preview of what I made; try it out on my website: &lt;a href="https://ayushtues.github.io/" target="_blank" rel="noopener">Ayush Mangal | Academic (ayushtues.github.io)&lt;/a>.&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/creating-an-ama-bot-for-my-personal-website-using-llms/image-1.gif"
alt="Okay, this is a cherry-picked example; it will mostly give random answers lol.">&lt;figcaption>
&lt;p>Okay, this is a cherry-picked example; it will mostly give random answers lol.&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>So I created a QA bot over my personal information, which anyone can talk to, to know more about me. Why anyone would want to do that is a different question, which is beyond the scope of this article.&lt;/p>
&lt;p>Okay, so how did we do it? Simple&lt;/p>
&lt;ul>
&lt;li>Custom web scrapers in Python + Langchain to mine my personal information from various sources&lt;/li>
&lt;li>Langchain to create a QA bot over the data&lt;/li>
&lt;li>Replicate for the LLMs&lt;/li>
&lt;li>Huggingface for embeddings&lt;/li>
&lt;li>Streamlit for creating the webapp&lt;/li>
&lt;li>Huggingface spaces to host the web app for free&lt;/li>
&lt;li>Embedding the HF space iframe in my Github Website to make it dynamic!&lt;/li>
&lt;/ul>
&lt;p>Okay, wait, it doesn’t sound so simple now. Let’s break it down a bit, shall we?&lt;/p>
&lt;h3 id="the-data">The data&lt;/h3>
&lt;p>Any LLM trained on general web data doesn’t know anything about me; I mean, I am not that famous, &lt;em>yet. S&lt;/em>o we need to give the LLM information about me.&lt;/p>
&lt;p>Currently, I feed the LLM the following data&lt;/p>
&lt;h4 id="transcripts-from-my-ytchannel">&lt;strong>Transcripts from my YT channel&lt;/strong>&lt;/h4>
&lt;p>I have a YT channel where I discuss various research papers I am interested in, and I thought it would be cool if the LLM knew about it.&lt;/p>
&lt;p>Langchain has a nice YoutubeLoader tool to get all the transcripts from a YT video. So I only needed to get the video IDs for all the videos on my channel. There is an API from Google to do that, but I wrote a quick web scraper to get all the IDs from the YT page of my channel.&lt;/p>
&lt;div class="highlight">&lt;pre tabindex="0" class="chroma">&lt;code class="language-python" data-lang="python">&lt;span class="line">&lt;span class="cl">&lt;span class="kn">from&lt;/span> &lt;span class="nn">langchain.document_loaders&lt;/span> &lt;span class="kn">import&lt;/span> &lt;span class="n">YoutubeLoader&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="kn">import&lt;/span> &lt;span class="nn">requests&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="kn">import&lt;/span> &lt;span class="nn">re&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="c1"># some webscraping using requests and regex to get all IDs for my youtube videos&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="n">my_url&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="s2">&amp;#34;https://www.youtube.com/@rrwithdeku8677/videos&amp;#34;&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="n">r&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">requests&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">get&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">my_url&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="n">page&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="p">(&lt;/span>&lt;span class="n">r&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">text&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="n">pattern&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="sa">r&lt;/span>&lt;span class="s1">&amp;#39;watch\?v=([^&amp;#34;]+)&amp;#39;&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="n">matches&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">re&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">findall&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">pattern&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">page&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">re&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">IGNORECASE&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="n">ids&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="p">[&lt;/span>&lt;span class="n">x&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">split&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="s1">&amp;#39;=&amp;#39;&lt;/span>&lt;span class="p">)[&lt;/span>&lt;span class="o">-&lt;/span>&lt;span class="mi">1&lt;/span>&lt;span class="p">]&lt;/span> &lt;span class="k">for&lt;/span> &lt;span class="n">x&lt;/span> &lt;span class="ow">in&lt;/span> &lt;span class="n">matches&lt;/span>&lt;span class="p">]&lt;/span> &lt;span class="c1"># has all the IDs&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="n">base_url&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="s2">&amp;#34;https://www.youtube.com/watch?v=&amp;#34;&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="n">video_data&lt;/span>&lt;span class="o">=&lt;/span> &lt;span class="p">[]&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="c1"># Using YoutubeLoader from Langchain to get the transcripts of the videos&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="k">for&lt;/span> &lt;span class="nb">id&lt;/span> &lt;span class="ow">in&lt;/span> &lt;span class="n">ids&lt;/span>&lt;span class="p">:&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">loader&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">YoutubeLoader&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">from_youtube_url&lt;/span>&lt;span class="p">(&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">base_url&lt;/span> &lt;span class="o">+&lt;/span> &lt;span class="nb">id&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">add_video_info&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="kc">True&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="nb">print&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="s2">&amp;#34;got loader&amp;#34;&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">data&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">loader&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">load&lt;/span>&lt;span class="p">()&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">video_data&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">extend&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">data&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="c1"># video_data has all the YT transcripts now from my YT channel&lt;/span>
&lt;/span>&lt;/span>&lt;/code>&lt;/pre>&lt;/div>&lt;h4 id="medium-blog-posts--my-personalwebsite">Medium Blog Posts + My Personal Website&lt;/h4>
&lt;p>I thought it might be a good idea to add all my medium posts as well. Again Langchain has a cool WebBaseLoader, which can extract the text from any webpage. So I just needed to get the links to all my blogs, which I did by writing a custom web scraper and Beautiful Soup.&lt;/p>
&lt;p>I also thought since my personal website already has so much info about me, why not add that in the mix too?&lt;/p>
&lt;div class="highlight">&lt;pre tabindex="0" class="chroma">&lt;code class="language-python" data-lang="python">&lt;span class="line">&lt;span class="cl"> &lt;span class="kn">from&lt;/span> &lt;span class="nn">lxml&lt;/span> &lt;span class="kn">import&lt;/span> &lt;span class="n">etree&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="kn">from&lt;/span> &lt;span class="nn">langchain.document_loaders&lt;/span> &lt;span class="kn">import&lt;/span> &lt;span class="n">WebBaseLoader&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="kn">from&lt;/span> &lt;span class="nn">bs4&lt;/span> &lt;span class="kn">import&lt;/span> &lt;span class="n">BeautifulSoup&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">def&lt;/span> &lt;span class="nf">has_numbers&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">inputString&lt;/span>&lt;span class="p">):&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">return&lt;/span> &lt;span class="nb">any&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">char&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">isdigit&lt;/span>&lt;span class="p">()&lt;/span> &lt;span class="k">for&lt;/span> &lt;span class="n">char&lt;/span> &lt;span class="ow">in&lt;/span> &lt;span class="n">inputString&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="c1"># Get my profile page &lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">profile_url&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="s2">&amp;#34;https://ayushtues.medium.com&amp;#34;&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">response&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">requests&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">get&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">profile_url&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">soup&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">BeautifulSoup&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">response&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">content&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="s1">&amp;#39;html.parser&amp;#39;&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="c1"># Get the links to all my recent medium blogs from my profile page&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">links&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="p">[]&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">for&lt;/span> &lt;span class="n">link&lt;/span> &lt;span class="ow">in&lt;/span> &lt;span class="n">soup&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">findAll&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="s1">&amp;#39;a&amp;#39;&lt;/span>&lt;span class="p">):&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">x&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">link&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">get&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="s1">&amp;#39;href&amp;#39;&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">if&lt;/span> &lt;span class="n">x&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">startswith&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="s1">&amp;#39;/&amp;#39;&lt;/span>&lt;span class="p">)&lt;/span> &lt;span class="ow">and&lt;/span> &lt;span class="n">has_numbers&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">x&lt;/span>&lt;span class="p">)&lt;/span> &lt;span class="p">:&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">links&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">append&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">link&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">get&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="s1">&amp;#39;href&amp;#39;&lt;/span>&lt;span class="p">))&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">links&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="nb">list&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="nb">set&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">links&lt;/span>&lt;span class="p">))&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="c1"># has the links to all my Medium blogs&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">links&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="p">[&lt;/span>&lt;span class="n">profile_url&lt;/span>&lt;span class="o">+&lt;/span> &lt;span class="n">x&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">split&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="s1">&amp;#39;?source&amp;#39;&lt;/span>&lt;span class="p">)[&lt;/span>&lt;span class="mi">0&lt;/span>&lt;span class="p">]&lt;/span> &lt;span class="k">for&lt;/span> &lt;span class="n">x&lt;/span> &lt;span class="ow">in&lt;/span> &lt;span class="n">links&lt;/span>&lt;span class="p">]&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="c1"># Adding my personal website into the mix as well&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">links&lt;/span> &lt;span class="o">+=&lt;/span> &lt;span class="p">[&lt;/span>&lt;span class="s2">&amp;#34;https://ayushtues.github.io/&amp;#34;&lt;/span>&lt;span class="p">]&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="c1"># Use Langchain to get the data from the blog links&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">loader&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">WebBaseLoader&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">links&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">data&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">loader&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">load&lt;/span>&lt;span class="p">()&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">video_data&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">extend&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">data&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;/code>&lt;/pre>&lt;/div>&lt;h4 id="my-resume">My Resume&lt;/h4>
&lt;p>The document that contains the most updated professional knowledge about me! An essential ingredient in the mix. Again, Langchain has a good PyPDFLoader for this.&lt;/p>
&lt;div class="highlight">&lt;pre tabindex="0" class="chroma">&lt;code class="language-python" data-lang="python">&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="kn">from&lt;/span> &lt;span class="nn">langchain.document_loaders&lt;/span> &lt;span class="kn">import&lt;/span> &lt;span class="n">PyPDFLoader&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="c1"># download my resume from a public url&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">url&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="s1">&amp;#39;https://huggingface.co/spaces/ayushtues/personal-assistant/resolve/main/resume.pdf&amp;#39;&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">r&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">requests&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">get&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">url&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">stream&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="kc">True&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">with&lt;/span> &lt;span class="nb">open&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="s1">&amp;#39;resume.pdf&amp;#39;&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="s1">&amp;#39;wb&amp;#39;&lt;/span>&lt;span class="p">)&lt;/span> &lt;span class="k">as&lt;/span> &lt;span class="n">fd&lt;/span>&lt;span class="p">:&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">for&lt;/span> &lt;span class="n">chunk&lt;/span> &lt;span class="ow">in&lt;/span> &lt;span class="n">r&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">iter_content&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="mi">2000&lt;/span>&lt;span class="p">):&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">fd&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">write&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">chunk&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="c1"># Load using Langchain&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">loader&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">PyPDFLoader&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="s2">&amp;#34;resume.pdf&amp;#34;&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">pages&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">loader&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">load&lt;/span>&lt;span class="p">()&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">video_data&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">extend&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">pages&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;/code>&lt;/pre>&lt;/div>&lt;h3 id="creating-the-qabot">Creating the QA bot&lt;/h3>
&lt;p>There are already a thousand tutorials on creating a QA bot using Langchain. I found the documentation on their website to be the easiest: &lt;a href="https://python.langchain.com/docs/use_cases/question_answering/" target="_blank" rel="noopener">QA over Documents | 🦜️🔗 Langchain&lt;/a>.&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/creating-an-ama-bot-for-my-personal-website-using-llms/image-2.png"
alt="QA over Documents | 🦜️🔗 Langchain">&lt;figcaption>
&lt;p>&lt;a href="https://python.langchain.com/docs/use_cases/question_answering/" target="_blank" rel="noopener">QA over Documents | 🦜️🔗 Langchain&lt;/a>&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>I covered loading in the previous steps. After that, I used the RecursiveCharacterTextSplitter from Langchain to split the documents into splits of size 100. This is done because the documents are huge and can’t be fed into the LLM as a whole, and instead, we work with chunks.&lt;/p>
&lt;div class="highlight">&lt;pre tabindex="0" class="chroma">&lt;code class="language-python" data-lang="python">&lt;span class="line">&lt;span class="cl">&lt;span class="kn">from&lt;/span> &lt;span class="nn">langchain.text_splitter&lt;/span> &lt;span class="kn">import&lt;/span> &lt;span class="n">RecursiveCharacterTextSplitter&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="n">text_splitter&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">RecursiveCharacterTextSplitter&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">chunk_size&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="mi">100&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">chunk_overlap&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="mi">0&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="n">all_splits&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">text_splitter&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">split_documents&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">video_data&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;/code>&lt;/pre>&lt;/div>&lt;p>Now to actually search over these docs, we need to embed them. Most tutorials use OpenAI embeddings, but I had exhausted mine, and they seem to have a got protection against fake numbers. So instead, I turned to a free open alternative, HuggingFace Embeddings, which uses Sentence Transformer.&lt;/p>
&lt;div class="highlight">&lt;pre tabindex="0" class="chroma">&lt;code class="language-python" data-lang="python">&lt;span class="line">&lt;span class="cl"> &lt;span class="n">model_name&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="s2">&amp;#34;sentence-transformers/all-mpnet-base-v2&amp;#34;&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">model_kwargs&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="p">{&lt;/span>&lt;span class="s1">&amp;#39;device&amp;#39;&lt;/span>&lt;span class="p">:&lt;/span> &lt;span class="s1">&amp;#39;cpu&amp;#39;&lt;/span>&lt;span class="p">}&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">encode_kwargs&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="p">{&lt;/span>&lt;span class="s1">&amp;#39;normalize_embeddings&amp;#39;&lt;/span>&lt;span class="p">:&lt;/span> &lt;span class="kc">False&lt;/span>&lt;span class="p">}&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">hf&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">HuggingFaceEmbeddings&lt;/span>&lt;span class="p">(&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">model_name&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="n">model_name&lt;/span>&lt;span class="p">,&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">model_kwargs&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="n">model_kwargs&lt;/span>&lt;span class="p">,&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">encode_kwargs&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="n">encode_kwargs&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;/code>&lt;/pre>&lt;/div>&lt;p>Then we need to create a vector store to store these embeddings and efficiently search on them. I used FAISS for this.&lt;/p>
&lt;div class="highlight">&lt;pre tabindex="0" class="chroma">&lt;code class="language-python" data-lang="python">&lt;span class="line">&lt;span class="cl">&lt;span class="kn">from&lt;/span> &lt;span class="nn">langchain.vectorstores.base&lt;/span> &lt;span class="kn">import&lt;/span> &lt;span class="n">VectorStoreRetriever&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="kn">from&lt;/span> &lt;span class="nn">langchain.vectorstores&lt;/span> &lt;span class="kn">import&lt;/span> &lt;span class="n">FAISS&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="n">vectorstore&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">FAISS&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">from_documents&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">documents&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="n">all_splits&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">embedding&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="n">hf&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="n">retriever&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">VectorStoreRetriever&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">vectorstore&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="n">vectorstore&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;/code>&lt;/pre>&lt;/div>&lt;p>Finally, we need an LLM to actually perform the QA. With OpenAI blocked, I turned to Replicate (&lt;a href="https://replicate.com/" target="_blank" rel="noopener">https://replicate.com/&lt;/a>). I used the Llamav2 chat model.&lt;/p>
&lt;div class="highlight">&lt;pre tabindex="0" class="chroma">&lt;code class="language-python" data-lang="python">&lt;span class="line">&lt;span class="cl">&lt;span class="kn">from&lt;/span> &lt;span class="nn">langchain.llms&lt;/span> &lt;span class="kn">import&lt;/span> &lt;span class="n">Replicate&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="n">llm&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">Replicate&lt;/span>&lt;span class="p">(&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">model&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="s2">&amp;#34;a16z-infra/llama13b-v2-chat:df7690f1994d94e96ad9d568eac121aecf50684a0b0963b25a41cc40061269e5&amp;#34;&lt;/span>&lt;span class="p">,&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="nb">input&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="p">{&lt;/span>&lt;span class="s2">&amp;#34;temperature&amp;#34;&lt;/span>&lt;span class="p">:&lt;/span> &lt;span class="mf">0.75&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="s2">&amp;#34;max_length&amp;#34;&lt;/span>&lt;span class="p">:&lt;/span> &lt;span class="mi">500&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="s2">&amp;#34;top_p&amp;#34;&lt;/span>&lt;span class="p">:&lt;/span> &lt;span class="mi">1&lt;/span>&lt;span class="p">},&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;/code>&lt;/pre>&lt;/div>&lt;p>An LLM needs a prompt to work on. I used the following prompt.&lt;/p>
&lt;div class="highlight">&lt;pre tabindex="0" class="chroma">&lt;code class="language-python" data-lang="python">&lt;span class="line">&lt;span class="cl"> &lt;span class="kn">from&lt;/span> &lt;span class="nn">langchain.prompts&lt;/span> &lt;span class="kn">import&lt;/span> &lt;span class="n">PromptTemplate&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">template&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="s2">&amp;#34;&amp;#34;&amp;#34;Use the following pieces of context to answer the question at the end.
&lt;/span>&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="s2"> If you don&amp;#39;t know the answer, just say that you don&amp;#39;t know, don&amp;#39;t try to make up an answer.
&lt;/span>&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="s2"> Use three sentences maximum and keep the answer as concise as possible.
&lt;/span>&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="s2"> Always say &amp;#34;thanks for asking!&amp;#34; at the end of the answer.
&lt;/span>&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="s2"> &lt;/span>&lt;span class="si">{context}&lt;/span>&lt;span class="s2">
&lt;/span>&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="s2"> Question: &lt;/span>&lt;span class="si">{question}&lt;/span>&lt;span class="s2">
&lt;/span>&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="s2"> Helpful Answer:&amp;#34;&amp;#34;&amp;#34;&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">QA_CHAIN_PROMPT&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">PromptTemplate&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">from_template&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">template&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;/code>&lt;/pre>&lt;/div>&lt;p>With all the components of our bot ready, we create a RetrievalQA from Langchain.&lt;/p>
&lt;div class="highlight">&lt;pre tabindex="0" class="chroma">&lt;code class="language-python" data-lang="python">&lt;span class="line">&lt;span class="cl">&lt;span class="kn">from&lt;/span> &lt;span class="nn">langchain.chains&lt;/span> &lt;span class="kn">import&lt;/span> &lt;span class="n">RetrievalQA&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="n">qa_chain&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">RetrievalQA&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">from_chain_type&lt;/span>&lt;span class="p">(&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">llm&lt;/span>&lt;span class="p">,&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">retriever&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="n">retriever&lt;/span>&lt;span class="p">,&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">chain_type_kwargs&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="p">{&lt;/span>&lt;span class="s2">&amp;#34;prompt&amp;#34;&lt;/span>&lt;span class="p">:&lt;/span> &lt;span class="n">QA_CHAIN_PROMPT&lt;/span>&lt;span class="p">}&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;/code>&lt;/pre>&lt;/div>&lt;h3 id="creating-a-webappstreamlit">Creating a webapp — Streamlit&lt;/h3>
&lt;p>Now we have an LLM AMA bot ready. But we need a web app for someone actually to be able to use it. I used Streamlit for this, which has really great integration with Langchain to make it easy to create LLM web apps.&lt;/p>
&lt;div class="highlight">&lt;pre tabindex="0" class="chroma">&lt;code class="language-python" data-lang="python">&lt;span class="line">&lt;span class="cl">&lt;span class="kn">import&lt;/span> &lt;span class="nn">streamlit&lt;/span> &lt;span class="k">as&lt;/span> &lt;span class="nn">st&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="kn">import&lt;/span> &lt;span class="nn">os&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="n">st&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">set_page_config&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">page_title&lt;/span>&lt;span class="o">=&lt;/span>&lt;span class="s2">&amp;#34;🦜🔗 Ask a LLM to know more about me&amp;#34;&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="c1"># Set the title&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="n">st&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">title&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="s1">&amp;#39;🦜🔗 Ask a LLM to know more about me&amp;#39;&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="c1"># Get my Replicate API secret&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="n">os&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">environ&lt;/span>&lt;span class="p">[&lt;/span>&lt;span class="s2">&amp;#34;REPLICATE_API_TOKEN&amp;#34;&lt;/span>&lt;span class="p">]&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">st&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">secrets&lt;/span>&lt;span class="p">[&lt;/span>&lt;span class="s2">&amp;#34;REPLICATE_API_TOKEN&amp;#34;&lt;/span>&lt;span class="p">]&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="k">def&lt;/span> &lt;span class="nf">get_query_chain&lt;/span>&lt;span class="p">():&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="c1"># logic to get qa_chain&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">return&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="c1"># Initialize the QA chain&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="n">query_chain&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">get_query_chain&lt;/span>&lt;span class="p">()&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="k">def&lt;/span> &lt;span class="nf">generate_response&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">topic&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">query_chain&lt;/span>&lt;span class="p">):&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">result&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">query_chain&lt;/span>&lt;span class="p">({&lt;/span>&lt;span class="s2">&amp;#34;query&amp;#34;&lt;/span>&lt;span class="p">:&lt;/span> &lt;span class="n">topic&lt;/span>&lt;span class="p">})&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">return&lt;/span> &lt;span class="n">st&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">info&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">result&lt;/span>&lt;span class="p">[&lt;/span>&lt;span class="s1">&amp;#39;result&amp;#39;&lt;/span>&lt;span class="p">])&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="c1"># Create a form for the user to enter the question and submit it to the LLM &lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="c1"># when clicked on submit&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="k">with&lt;/span> &lt;span class="n">st&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">form&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="s1">&amp;#39;myform&amp;#39;&lt;/span>&lt;span class="p">):&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">topic_text&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">st&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">text_input&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="s1">&amp;#39;Enter keyword:&amp;#39;&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="s1">&amp;#39;&amp;#39;&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">submitted&lt;/span> &lt;span class="o">=&lt;/span> &lt;span class="n">st&lt;/span>&lt;span class="o">.&lt;/span>&lt;span class="n">form_submit_button&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="s1">&amp;#39;Submit&amp;#39;&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="c1"># generate the response from the LLM and display it&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="k">if&lt;/span> &lt;span class="n">submitted&lt;/span> &lt;span class="p">:&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="n">generate_response&lt;/span>&lt;span class="p">(&lt;/span>&lt;span class="n">topic_text&lt;/span>&lt;span class="p">,&lt;/span> &lt;span class="n">query_chain&lt;/span>&lt;span class="p">)&lt;/span>
&lt;/span>&lt;/span>&lt;/code>&lt;/pre>&lt;/div>&lt;p>This creates a web app which you can run on your localhost. But well, only you can access it. Not everyone on the internet.&lt;/p>
&lt;h3 id="hosting-the-webapp-using-huggingface-spaces">Hosting the webapp using Huggingface spaces&lt;/h3>
&lt;p>Huggingface, being the wonderful company it is, allows you to host your custom Streamlit apps on their website for FREE. You can read how to create a streamlit space here: &lt;a href="https://huggingface.co/docs/hub/spaces-sdks-streamlit" target="_blank" rel="noopener">Streamlit Spaces (huggingface.co)&lt;/a>. But it’s extremely easy.&lt;/p>
&lt;p>It creates a &lt;code>app.py&lt;/code> where you need to put all your web app code.&lt;/p>
&lt;p>Note that you need to specify all required libraries in a &lt;code>requirements.txt&lt;/code> file so that the space knows what must be preinstalled.&lt;/p>
&lt;p>You can find all code for my space here : &lt;a href="https://huggingface.co/spaces/ayushtues/personal-assistant/tree/main" target="_blank" rel="noopener">ayushtues/personal-assistant at main (huggingface.co)&lt;/a>.&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/creating-an-ama-bot-for-my-personal-website-using-llms/image-3_hu5457321535218710179.webp 400w,
/blog/creating-an-ama-bot-for-my-personal-website-using-llms/image-3_hu4070622353316344436.webp 760w,
/blog/creating-an-ama-bot-for-my-personal-website-using-llms/image-3_hu3032997670061240016.webp 1200w"
src="https://ayushtues.github.io/blog/creating-an-ama-bot-for-my-personal-website-using-llms/image-3_hu5457321535218710179.webp"
width="760"
height="448"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>To embed this app into my personal website, HF provides an iframe snippet, which I can directly put into the HTML of my website to embed the above space into my website!&lt;/p>
&lt;div class="highlight">&lt;pre tabindex="0" class="chroma">&lt;code class="language-xml" data-lang="xml">&lt;span class="line">&lt;span class="cl">&lt;span class="nt">&amp;lt;iframe&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="na">src=&lt;/span>&lt;span class="s">&amp;#34;https://ayushtues-personal-assistant.hf.space&amp;#34;&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="na">frameborder=&lt;/span>&lt;span class="s">&amp;#34;0&amp;#34;&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="na">width=&lt;/span>&lt;span class="s">&amp;#34;850&amp;#34;&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl"> &lt;span class="na">height=&lt;/span>&lt;span class="s">&amp;#34;450&amp;#34;&lt;/span>
&lt;/span>&lt;/span>&lt;span class="line">&lt;span class="cl">&lt;span class="nt">&amp;gt;&amp;lt;/iframe&amp;gt;&lt;/span>
&lt;/span>&lt;/span>&lt;/code>&lt;/pre>&lt;/div>&lt;p>So that is how I created an AMA bot and embedded into my static website. There are a lot of improvements to be done, some of them being&lt;/p>
&lt;ul>
&lt;li>Using a local LLM like GPT4ALL/PrivateGPT or something to remove the dependency on Replicate, since I will run out of quota there soon.&lt;/li>
&lt;li>Making the retrieval better, probably by making it hierarchical, based on the source&lt;/li>
&lt;li>Allowing for conversation using Chat features&lt;/li>
&lt;li>Reducing the latency&lt;/li>
&lt;li>Adding more data sources, like my Github, Spotify etc&lt;/li>
&lt;li>Better code structuring&lt;/li>
&lt;/ul>
&lt;p>Anyways, this was a fun learning hack to get a taste of making LLM-powered apps. Let’s see what all I can do with it as time goes!&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/creating-an-ama-bot-for-my-personal-website-using-llms-5201a641deae" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Leap of Faith</title><link>https://ayushtues.github.io/blog/leap-of-faith/</link><pubDate>Tue, 04 Jul 2023 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/leap-of-faith/</guid><description>&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/leap-of-faith/image-1_hu12718850413769067572.webp 400w,
/blog/leap-of-faith/image-1_hu1307002353015454589.webp 760w,
/blog/leap-of-faith/image-1_hu8210715738965793317.webp 1200w"
src="https://ayushtues.github.io/blog/leap-of-faith/image-1_hu12718850413769067572.webp"
width="760"
height="570"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>At this point, I feel like since no one actually reads this blog ( correction: no one I know reads this blog, each blog does actually get around 100 views usually on the internet ), I have started treating this more like a journal than anything. But if I do want to journal, then I should instead actually keep a private journal cause I can’t really say things openly here since it&amp;rsquo;s public, but I am too lazy to actually start a journal; makes sense? Nopes, doesn’t to me, either.&lt;/p>
&lt;p>Anyway, I just got back from my first trek (Kodachadri, Western Ghats) yesterday, well, not actually my first; I did go to Nishani Hills in Coorg last December, but compared to this trek, that was just a morning walk (albeit a really, really pretty morning walk lol). While the scenic beauty of the place was refreshing, it was the adventure and the oh-fuck-I-can-actually-die-if-I-am-not-careful moments in the trek which were the highlight for me.&lt;/p>
&lt;p>While I don’t claim to be an adventure sports freak, I have been slowly trying to put myself out there to give them a shot whenever I get the chance, having attempted paragliding (Nainital), water rafting (Rishikesh), and recently even Bungee Jumping (Rishikesh) from ~100m. This is something which is new for me; I didn’t indulge in adventure sports before college cause a) I was totally dependent on my parents, who would never let me do something like this, and b) I wasn’t exactly the adventurous kind anyway growing up.&lt;/p>
&lt;p>However, after trying some of these activities, I am starting to realize that I love adventure. I love the adrenaline rush you get when you know that you can actually die doing this. Every time I do something like this, at least for some time, I develop this hyper-realization of how fickle it all is and a sense that okay, whatever I have from now on is sort of a gift, which could have just been taken away from me, maybe that is what becoming grateful is? ( Although a hyper-privileged brat like me is &lt;em>so-sooooooo&lt;/em> not actually grateful for anything in his life, but well :P ).&lt;/p>
&lt;p>There is also a sense of giving up control while doing these activities, like Bungee Jumping, where you have to trust the operators with basically your life and just take a leap of faith, hoping that physics works and the rope doesn’t break or doesn’t break your spine or you don’t have a heart attack or whatever else can go wrong. It makes you realize that sometimes you just need to trust that things are going to work out and just go ahead with stuff ( or anyway, you’ll just die, and shit won’t matter to you; binary outputs are fun! ).&lt;/p>
&lt;p>It also involves a certain amount of apathy towards your life to do stuff like this, which I feel I am totally capable of, being ready to die just for the sake of doing something fun, idk; it probably means that I am a repressed psychopathic adrenaline junkie who is trapped in a soul-sucking corporate job and sometimes needs a refresh from life, who knows?&lt;/p>
&lt;p>Anyway, this was a shortie, that&amp;rsquo;s all for the blog!&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/leap-of-faith-6f8bb21fb10f" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>On being alone</title><link>https://ayushtues.github.io/blog/on-being-alone/</link><pubDate>Sat, 03 Jun 2023 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/on-being-alone/</guid><description>&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/on-being-alone/image-1_hu13907754153748645032.webp 400w,
/blog/on-being-alone/image-1_hu14243426232122200327.webp 760w,
/blog/on-being-alone/image-1_hu15377118878827883580.webp 1200w"
src="https://ayushtues.github.io/blog/on-being-alone/image-1_hu13907754153748645032.webp"
width="700"
height="375"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>No, this is not a post about introverts or how introverts are some sort of a special species with zen-like superpowers to somehow led the world towards utopia.&lt;/p>
&lt;p>This is me, on a trip to Kerela (Munnar-Thekkady-Alleypey, currently in Thekkady), with my family, sitting alone in the hotel room balcony, listening to songs, wanting to do a brain dump (read: rant) of random thoughts, while my family is out in the market, shopping (I for the life of me, never do random shopping lol).&lt;/p>
&lt;p>I am someone, who likes to be alone, by myself. I purposely live alone, instead of sharing a flat with someone, even if it is more expensive, and sometimes, honestly a lot worse, logistically. I don’t have any social media, except Twitter and Linkedin (that too only for professional reasons). I barely message or call anyone, even my family (much to their dismay, I seriously believe they’ll someday disown me), usually only interacting with the people physically around me, and finding it a drag to talk to anyone else purposely. I like being left to my own devices, doing my own stuff, lost in my own world.&lt;/p>
&lt;p>Now now, before you start calling me a sociopath, or worse, an introvert, I should clarify that I am fairly good at talking to people, infact I love talking to people, face to face, I can handle large groups, speak publicly, and have a fairly active social life, with a few friends I can rely upon to hang out with whenever needed. Also I don’t hate people, I love people. I am surrounded by a hell lot of super super good people, to the point I feel suprised why I don’t know &lt;em>any&lt;/em> bad people (seriously, where are all the bad people?).&lt;/p>
&lt;p>I just feel like being left alone, most of the time.&lt;/p>
&lt;p>There is something of a stigma surrounding wanting to be alone, you are expected to be hanging out with people, talking with everyone, being interested in everyone, being social, being amiable, knowing what’s going on in the life of the people around you, and also letting them know what’s going on in your life. You are supposed to say yes to plans, even though they might drain you, and distract you from doing the things that actually bring meaning to you.&lt;/p>
&lt;p>I like being alone. Being able to spend my time how I want, without having to think about anyone else. Doing things just for the heck of it, without a care for the world. And there are definitely a lot of things which require a group of people to execute, a single person can’t build a hugely complex thing, but I sometimes feel that in groups, you tend to get to look into things on a very narrow and high level, and sometimes I just want to dive deep into things and do them myself. Obviously this doesn’t scale well, and I also have to eventually participate in groups, but gosh, it feels soooooo good to say F you, and sit in your room alone, doing stuff.&lt;/p>
&lt;p>Being alone also lets me understand myself more, what I want, what actually matters to me, and what I actually want to do. A bit of soul searching goes a long way in my experience, and infact having a clear picture of self, I feel also makes you more geniuine in the interactions you have with other people, cause you have a better idea of how to behave in a way that is consistent with your values, instead of just going with the flow and picking up behaviour patterns in the moment from others and then later being like, did I really do that? I don’t do that usually….well yes dumbass, you don’t, you were just influenced by the behaviour of the people around you.&lt;/p>
&lt;p>Having alone time is also very good to skill up, and pick up new hobbies. I like being able to work on myself and skilling up slowly in things that make me happy. I am honestly very unskilled at almost everything right now, so it kinda feels like a blank canvas, which can be filled with anything at the moment, I can put efforts and learn whatever I want, being alone, not caring much about anyone, or what others are doing.&lt;/p>
&lt;p>Needless to say, it can be hard being alone, at times. Especially when shit hits the fan. I have terribly felt the need to be around people at times, like when I was ill, or going through a particularly tough patch. And hey, I do have people I can count on for those times, its just that, I don’t want to be in the face of those people all the time lol. Sometimes I don’t have people to help me out, and that’s okay, I have become used to handling problems alone, and living a solitary life. With the other option being to think about other people everytime I do something, I’ll rather prefer solace.&lt;/p>
&lt;p>Until the next one, Goodbye!&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/on-being-alone-fe6266d25f42" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>RL with diffusion — I</title><link>https://ayushtues.github.io/blog/rl-with-diffusion-i/</link><pubDate>Mon, 29 May 2023 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/rl-with-diffusion-i/</guid><description>&lt;h3 id="introduction">Introduction&lt;/h3>
&lt;p>There has been a lot of buzz surrounding diffusion models, which are used for a wide range of tasks ranging from text to image generation, video generation, and even protein synthesis.&lt;/p>
&lt;p>Given that they are primarily used in the image and text domains, or usually for generative modelling, it’s not trivial to see how they can be extended to be used in Reinforcement learning or why doing this would even make sense.&lt;/p>
&lt;p>Unless you start seeing that in RL, we are just generating a trajectory of states, which we want to optimize to have the highest reward. Sounds a bit like conditional generative modelling. We want to generate something conditioned on some property we want it to have ( match a text prompt, have a particular label &lt;em>or&lt;/em> have a high reward!).&lt;/p>
&lt;p>In RL, this sequence is usually generated autoregressively, which is a fancy way of saying that we observe a state and then predict the best action to go to the best next state. But what if we don’t do this autoregressively? What if we generate the entire sequence simultaneously &lt;em>in a single shot&lt;/em>?&lt;/p>
&lt;p>So now, we are generating an entire sequence of states connected in time together. Now, hold up, aren’t images also just a sequence of pixels connected in space? If we can use diffusion models to generate a series of pixels, why not use them to generate a series of states as well?&lt;/p>
&lt;p>Turns out, doing this has a lot of interesting properties, which were explored in the work we’ll be discussing today: &lt;strong>Planning with Diffusion for Flexible Behavior Synthesis (&lt;/strong>&lt;a href="https://diffusion-planning.github.io/" target="_blank" rel="noopener">Diffuser: Reinforcement Learning with Diffusion Models (diffusion-planning.github.io)&lt;/a>&lt;strong>).&lt;/strong>&lt;/p>
&lt;p>But I am getting ahead of myself; what the hell are even Diffusion Models, or what exactly does RL deal with?&lt;/p>
&lt;h4 id="reinforcement-learning">Reinforcement Learning&lt;/h4>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/rl-with-diffusion-i/image-1.jpg"
alt="A diagram depicting the RL setting">&lt;figcaption>
&lt;p>A diagram depicting the RL setting&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Going into only a few technical details about RL, on a high level, it consists of an agent acting in an environment. Say a robot (agent) walking in the world (environment).&lt;/p>
&lt;p>On each timestep t, an agent observes the environment through a State vector S_t, takes an action A_t, finds itself in a new state S_(t+1), and gets rewarded a reward R_(t+1). The task in RL is to learn a policy π(s), which tells the agent what action to take, given a state s, which will lead to the maximum reward.&lt;/p>
&lt;p>Now it gets tricky because we want to take actions that will lead to an &lt;em>overall&lt;/em> highreward, not just a high one in the current state. Imagine training an agent to drive a car from your home to the office and rewarding it for high speed. While it might make sense to hit the throttle and accelerate locally, it might lead to an accident, rendering you to gain higher rewards in the future and not reaching your goal.&lt;/p>
&lt;p>An alternative way to look at this is not considering a state one by one but looking at the &lt;em>entire&lt;/em> trajectory of states at once and optimizing the trajectory to have a high cumulative reward; this falls under the category of Trajectory Optimization. From the paper itself:&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/rl-with-diffusion-i/image-2_hu7593655255552374759.webp 400w,
/blog/rl-with-diffusion-i/image-2_hu6694711044663263586.webp 760w,
/blog/rl-with-diffusion-i/image-2_hu10414039933904461314.webp 1200w"
src="https://ayushtues.github.io/blog/rl-with-diffusion-i/image-2_hu7593655255552374759.webp"
width="760"
height="546"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;h4 id="model-based-reinforcement-learning">Model-Based Reinforcement Learning&lt;/h4>
&lt;p>Imagine you have to train an agent to drive a self-driving car. Here the environment is well, the real world roads. But can you really train your agent directly on real-world roads? It has two problems.&lt;/p>
&lt;ul>
&lt;li>&lt;strong>Safety&lt;/strong>: You don’t want to crash land your super expensive self-driving car while it’s still learning, and you definitely, definitely don’t want to end up hurting someone in the process. So directly training an agent in an environment might not be safe.&lt;/li>
&lt;li>Training in the real world is also very &lt;strong>expensive.&lt;/strong> You can’t really execute a hyperparameter sweep while training in the real world, that is, unless you have a shit ton of cars running like Tesla lol.Directly training in the real world can be prohibitively expensive or straight-up impossible.&lt;/li>
&lt;/ul>
&lt;p>Enter Model-based RL; what if you had a model of the real world, a sort of simulation, which can act as a proxy for the real world, in which you can train your agent as much as you want. Usually, these models are fast, quick to interact with, and can enable rapid training. NFL does kick in, though, since a model is always, well, a model. It’s not a substitute for the real world, and an agent performing well on a model of the environment, is not guaranteed to perform well in the real environment.&lt;/p>
&lt;p>So how do you learn a model? Wait, what is a model anyway? How is it represented? Mostly, you learn a dynamics model, which predicts what the next state (St+1) would be, based on the current state (St) and action (At)&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/rl-with-diffusion-i/image-3.png"
alt="Here f represents the model, which takes in the current state St and Action At, to give the next state St&amp;#43;1">&lt;figcaption>
&lt;p>Here f represents the model, which takes in the current state St and Action At, to give the next state St+1&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>The function is usually represented by a neural network and is trained by real-world interaction data.&lt;/p>
&lt;p>While you can learn a policy using a model, and there are approaches which do precisely that, there is another thing you can do. Think about chess; if you know the model or the rules by which chess work, you can think about what action to take right now, execute it in your imagination, see what states you’ll land into, and then plan again based on it. This is called model-based planning for trajectory optimization.&lt;/p>
&lt;h4 id="the-problem-with-model-based-rl">&lt;strong>The problem with Model-Based RL&lt;/strong>&lt;/h4>
&lt;p>Current approaches mostly use learning only for the dynamics model and leave the planning to classical trajectory optimizers with no learning involved.&lt;/p>
&lt;p>Now this poses a problem, if you have a classical planning algo, which was designed to work with the real environment, and assume the model they are working with is exactly accurate (ground truth), they’ll try to find a way to exploit the model they are given and find the best way to crack it open.&lt;/p>
&lt;p>However, in model-based RL, we are not dealing with the actual environment but a model of it. And a model, however good, is still just a model. It will have its errors in approximating the world. So, if a planning algo finds a way to exploit the model, it doesn’t directly mean that the hack would also transfer to the real world. This problem becomes even worse when we use Neural Networks to learn the model since NNs are prone to adversarial examples, which are examples which look like random garbage but can fool the model into believing that it’s just found the holy grail lol.&lt;/p>
&lt;p>So clearly, we need to address this disconnect between the learning process of the model and the planning done over it to avoid such exploitation.&lt;/p>
&lt;h4 id="the-proposed-solutiondiffusion">&lt;strong>The proposed solution — Diffusion!&lt;/strong>&lt;/h4>
&lt;p>But WTF are Diffusion models? Now this will require a blog of its own, which I might write someday; you can read more about them &lt;a href="https://lilianweng.github.io/posts/2021-07-11-diffusion-models/" target="_blank" rel="noopener">here&lt;/a> and also via a simple Google search lol.&lt;/p>
&lt;p>&lt;strong>TLDR:&lt;/strong> Imagine you have to learn a model which can generate the image of a cat from random noise. Now asking a model to learn how to directly convert random noise to a legible image seems overkill. However, there are models which do that, ie. GANs, VAEs and whatnot ( even something resembling diffusion models called Consistency Models are one-shot ).&lt;/p>
&lt;p>But what if I give you a slightly noised image of a cat and ask you to learn a model to denoise it back to the cat? That seems much more manageable. That is what diffusion models do; they learn to slightly denoise an image at each step. And then, they are applied iteratively on random noise to remove a little bit of noise at every step and lead to a clear image.&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/rl-with-diffusion-i/image-4_hu16467882459230760501.webp 400w,
/blog/rl-with-diffusion-i/image-4_hu1191123921649750709.webp 760w,
/blog/rl-with-diffusion-i/image-4_hu11607527051132983939.webp 1200w"
src="https://ayushtues.github.io/blog/rl-with-diffusion-i/image-4_hu16467882459230760501.webp"
width="727"
height="760"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>That was the intuition, I’ll just post a bit from the paper for this, to integrate the notation they use&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/rl-with-diffusion-i/image-5_hu12015727494753939259.webp 400w,
/blog/rl-with-diffusion-i/image-5_hu12239143929200889561.webp 760w,
/blog/rl-with-diffusion-i/image-5_hu306297914788056850.webp 1200w"
src="https://ayushtues.github.io/blog/rl-with-diffusion-i/image-5_hu12015727494753939259.webp"
width="760"
height="379"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;h3 id="rl-with-diffusion-diffusers">RL with Diffusion: Diffusers&lt;/h3>
&lt;p>No, not the hugging face library, lol, although they do have this work integrated (&lt;a href="https://huggingface.co/docs/diffusers/using-diffusers/rl" target="_blank" rel="noopener">here&lt;/a>).&lt;/p>
&lt;p>So now we want to integrate the model learning and the trajectory optimization as closely as possible; how do we do that?&lt;/p>
&lt;p>First, let’s think about what trajectory optimization means in the context of using diffusion models to generate trajectories. We now have a diffusion model, which can generate trajectories following a distribution pθ(τ ). This doesn’t really have any sense of what a good or bad trajectory is, aside from how likely the trajectory is in the probability distribution pθ(τ ).&lt;/p>
&lt;p>Trajectory optimization, however, wants trajectories that satisfy a particular condition; they should either have a high reward, end in a specific goal state, or some other constraint. So we want to make trajectories which satisfy these conditions more likely.&lt;/p>
&lt;p>Now let’s go back to generating images from diffusion models; let’s say we’re generating images of animals. Suppose we have no conditions or unconditional generative modelling. In that case, we’ll be fine if the output is a cat, dog, bird, or whatever. Still, if we want to get only images of, say, dogs, then we will have to &lt;em>condition&lt;/em> the model in some way to generate only images of dogs; this is called conditional generative modelling. In the case of images, this can be done by providing the image’s label to the model or a text prompt we want it to follow.&lt;/p>
&lt;p>Probabilistically, this means that we now want to sample from a new probability distribution :&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/rl-with-diffusion-i/image-6_hu11697297772742478306.webp 400w,
/blog/rl-with-diffusion-i/image-6_hu8684268880029995028.webp 760w,
/blog/rl-with-diffusion-i/image-6_hu2389184740166483770.webp 1200w"
src="https://ayushtues.github.io/blog/rl-with-diffusion-i/image-6_hu11697297772742478306.webp"
width="760"
height="431"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>So you convert your probability distribution by multiplying it by h(τ ), representing whatever condition you want your trajectories to follow.&lt;br>
Then you just sample from this distribution to get trajectories, which are both physically realistic under pθ(τ ) and high-reward (or constraint-satisfying) under h(τ )&lt;/p>
&lt;p>A neat this about this factoring is that all task-related information is contained in h(τ ), and is separated from the model pθ(τ ), so if you want to adapt to a new task, you can change h(τ ) according to the task, leaving the model as it is.&lt;/p>
&lt;p>So how do we actually integrate this conditional sampling? It is very similar to classifier (or classifier-free) guidance commonly used in image diffusion models, in which a classifier looks at the model’s output and slightly nudges the output based on how it classifies the output to belong to the class required. It’s called a guide, represented by Jφ in this work instead of a classifier.&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/rl-with-diffusion-i/image-7_hu13070462399777812392.webp 400w,
/blog/rl-with-diffusion-i/image-7_hu4604684886329540625.webp 760w,
/blog/rl-with-diffusion-i/image-7_hu2935233070844600598.webp 1200w"
src="https://ayushtues.github.io/blog/rl-with-diffusion-i/image-7_hu13070462399777812392.webp"
width="760"
height="339"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>So the steps are :&lt;/p>
&lt;ul>
&lt;li>We first train a diffusion model pθ(τ ) on the states and actions of all available trajectory data.&lt;/li>
&lt;li>We then train a separate model Jφ to predict the cumulative rewards of trajectory samples. The gradients of Jφ are used to guide the trajectory sampling procedure by modifying the means µ of the reverse process.&lt;/li>
&lt;li>On observing a state, we run the guided diffusion planning, get a trajectory starting from this state, execute the action it tells us to do, get into a new state, and then repeat the process (yeah, we replan in each and every state, kinda seems like an overkill doesn’t it, with all the local coherence and stuff)&lt;/li>
&lt;/ul>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/rl-with-diffusion-i/image-8_hu18435516623245375818.webp 400w,
/blog/rl-with-diffusion-i/image-8_hu13363680524069212526.webp 760w,
/blog/rl-with-diffusion-i/image-8_hu14957889250317372218.webp 1200w"
src="https://ayushtues.github.io/blog/rl-with-diffusion-i/image-8_hu18435516623245375818.webp"
width="760"
height="484"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;h3 id="designing-themodel">Designing the model&lt;/h3>
&lt;p>With the overarching concept of modelling RL as a diffusion process clear, let&amp;rsquo;s look at a few of the design choices that were made to implement this —&lt;/p>
&lt;h4 id="trajectory-representation">Trajectory Representation&lt;/h4>
&lt;p>Before even deciding what architecture to use, we need to figure out how to represent the input and the output. Although in the case of diffusion models, the input and output reside in the same space.&lt;/p>
&lt;p>Earlier, the dynamics model only predicted the state, s(t+1) from the current state and action; this was because we were not planning and didn’t care about what action we will be taking; we just wanted to know how the dynamics of the world work.&lt;/p>
&lt;p>But now, since we are integrating planning into the modelling itself, we will also need to predict actions, so our input will contain &lt;em>both&lt;/em> states and actions of the trajectory.&lt;/p>
&lt;p>And since we are predicting the entire trajectory, our input will be the series of all the states and actions in the trajectory. So we represent the inputs (and outputs) of Diffuser as a two-dimensional array:&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/rl-with-diffusion-i/image-9_hu13214895570665417014.webp 400w,
/blog/rl-with-diffusion-i/image-9_hu3347804589538883237.webp 760w,
/blog/rl-with-diffusion-i/image-9_hu8920817330082191669.webp 1200w"
src="https://ayushtues.github.io/blog/rl-with-diffusion-i/image-9_hu13214895570665417014.webp"
width="760"
height="240"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;h4 id="temporal-locality">Temporal Locality&lt;/h4>
&lt;p>While we want our trajectory to make sense as a whole globally, we also want to make it feasible locally because the states are connected in time. You can only reach a state in a trajectory from a previous state, so it should, in fact, make sense locally nearby that state that such a transition is possible.&lt;/p>
&lt;p>In typical RL, this is handled by the Markovian nature of things, where we predict the next state based on the current state and predict states autoregressively, i.e. one by one.&lt;/p>
&lt;p>However, in Diffusers, we predict the entire trajectory simultaneously, meaning we must enforce temporal locality more explicitly. This is done by using temporal convolutions, as seen in the below figure; the receptive field of a given prediction only consists of nearby timesteps, both in the past and the future. As a result, each step of the denoising process can only make predictions based on the local consistency of the trajectory.&lt;/p>
&lt;p>However, when we compose many of these denoising steps together, the local consistency can drive global coherence; think of it as how the effective receptive field in image models increases by applying more and more convolutional layers.&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/rl-with-diffusion-i/image-10_hu572193440201362132.webp 400w,
/blog/rl-with-diffusion-i/image-10_hu2403527856257727686.webp 760w,
/blog/rl-with-diffusion-i/image-10_hu8381922715297207346.webp 1200w"
src="https://ayushtues.github.io/blog/rl-with-diffusion-i/image-10_hu572193440201362132.webp"
width="737"
height="760"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;h4 id="architecture">Architecture&lt;/h4>
&lt;p>They use a U-Net, which is typical in image-based diffusion models, but with two-dimensional spatial convolutions replaced by one-dimensional temporal convolutions, since while in images, the pixels are connected through space, here the states are connected across time.&lt;/p>
&lt;p>Because the model is fully convolutional, the horizon of the predictions is determined not by the model architecture but by the input dimensionality; it can change dynamically during planning if desired. This is the same as fully convolutional image models being able to handle images of any resolution; the spatial resolution of images corresponds to the temporal length of the trajectory in this case.&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt=""
src="https://ayushtues.github.io/blog/rl-with-diffusion-i/image-11.gif"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;h4 id="goal-conditioned-rl-as-inpainting">&lt;strong>Goal Conditioned RL as Inpainting&lt;/strong>&lt;/h4>
&lt;p>One of the most incredible things I found with this work is how smoothly it integrates goal-conditioned RL. They treat it as an inpainting problem, which is already heavily explored in image models, where you mask out a part of the image, and ask the model to fill it based on some conditioning.&lt;/p>
&lt;p>Here, say if we want to reach a goal from a particular state, we treat it as an inpainting problem, with the first and last state specified and the other states to be inpainted by the diffusion model!&lt;/p>
&lt;p>What’s even cooler is that this doesn’t restrict to just a final goal state; if you want to constrain the trajectory to pass through certain intermediate states, specify them in the inpainting input, as simple as that!&lt;/p>
&lt;p>We can also see this as h(τ ) consisting of just 1’s at the first and last state and 0 at all other states, as shown in the below figure.&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/rl-with-diffusion-i/image-12_hu11612095743794256982.webp 400w,
/blog/rl-with-diffusion-i/image-12_hu4881048265689605113.webp 760w,
/blog/rl-with-diffusion-i/image-12_hu1396782395152524468.webp 1200w"
src="https://ayushtues.github.io/blog/rl-with-diffusion-i/image-12_hu11612095743794256982.webp"
width="760"
height="274"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt=""
src="https://ayushtues.github.io/blog/rl-with-diffusion-i/image-13.gif"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;h3 id="properties-of-diffusion-planners">&lt;strong>Properties of Diffusion Planners&lt;/strong>&lt;/h3>
&lt;p>&lt;strong>Learned long-horizon planning&lt;/strong> — The authors claim that since in diffusers, the planning is integrated into the dynamics modelling process, if the learnt model is good at predicting long-range trajectories, their process will also be good at long-horizon planning, since planning is the same as modelling, just with a conditioned probability distribution. They show their approach can avoid myopic failures and solve sparse reward settings where other methods struggle.&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/rl-with-diffusion-i/image-14_hu9631707702931266515.webp 400w,
/blog/rl-with-diffusion-i/image-14_hu16481327884469193583.webp 760w,
/blog/rl-with-diffusion-i/image-14_hu6761495828472533529.webp 1200w"
src="https://ayushtues.github.io/blog/rl-with-diffusion-i/image-14_hu9631707702931266515.webp"
width="760"
height="475"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>**Compositionality —**Markovian models, which are very local, and take decisions based on a particular state, can stitch together different parts of trajectories during training to generalize to new ones. Since the decision only depends on the nearby states, if they observe the grey trajectory, which is always going upward and the blue one, which is always going downward from the centre, it has learnt both motions around the centre state, so it can learn how to take a turn since it only depends on the local context, and not if the rest of the trajectory is going up or down.&lt;/p>
&lt;p>The authors claim that due to the local receptive field of the temporal convolutions, diffusers also adapt this property and can stick together trajectories to generalize to new ones.&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/rl-with-diffusion-i/image-15_hu10515732365147420769.webp 400w,
/blog/rl-with-diffusion-i/image-15_hu1733451176841570368.webp 760w,
/blog/rl-with-diffusion-i/image-15_hu4213044198383162742.webp 1200w"
src="https://ayushtues.github.io/blog/rl-with-diffusion-i/image-15_hu10515732365147420769.webp"
width="760"
height="656"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>&lt;strong>Variable-length plans:&lt;/strong> Because the model is fully convolutional in the time dimension, it can adapt to variable-length plans.&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/rl-with-diffusion-i/image-16_hu15307246304241322292.webp 400w,
/blog/rl-with-diffusion-i/image-16_hu5978926182612628956.webp 760w,
/blog/rl-with-diffusion-i/image-16_hu9584201089632001511.webp 1200w"
src="https://ayushtues.github.io/blog/rl-with-diffusion-i/image-16_hu15307246304241322292.webp"
width="760"
height="318"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>&lt;strong>Task compositionality&lt;/strong> — Since all task-related information is contained in h(τ ), and is separated from the model pθ(τ ), so if you want to adapt to a new task, you can just change h(τ ) according to the task, leaving the model as it is.&lt;/p>
&lt;p>You can get some crazy stuff from just training a single model pθ(τ ) and changing the constraint h(τ ); the model is able to adapt to a variety of stacking tasks&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/rl-with-diffusion-i/image-17_hu10527500064154267622.webp 400w,
/blog/rl-with-diffusion-i/image-17_hu7972788213772956114.webp 760w,
/blog/rl-with-diffusion-i/image-17_hu1772572109768461849.webp 1200w"
src="https://ayushtues.github.io/blog/rl-with-diffusion-i/image-17_hu10527500064154267622.webp"
width="760"
height="407"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;h4 id="warm-starting-diffusion-for-fasterplanning">&lt;strong>Warm-Starting Diffusion for Faster Planning&lt;/strong>&lt;/h4>
&lt;p>So all this is good, and it seems like this model has no flaws.&lt;/p>
&lt;p>Well, not really; one of the main limitations of the paper is that the approach could be faster. And this is a common problem with diffusion models in general; since they involve iteratively applying the model again and again t denoise, they are very slow.&lt;/p>
&lt;p>Add that to the fact that in planning, we run the entire process at every state; it gets even slower.&lt;/p>
&lt;p>To make the approach a bit faster, they say that we don’t need to entirely discard the plans generated at every state, since they are also helpful for nearby states.&lt;/p>
&lt;p>So you can take the plan from the previous state and use it as a warm start for the planning in this state. You do this by noising the previous plan a bit and then denoising it using your diffusion model. Still, since it’s mostly a well-structured plan with a little bit of noise added to it, you just need to run the diffusion for a few numbers of steps to tweak the plan based on your current state. This is shown to increase the speed a lot while hurting the performance slightly, as seen in the figure below.&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/rl-with-diffusion-i/image-18_hu2556884957591279424.webp 400w,
/blog/rl-with-diffusion-i/image-18_hu18040445422977615751.webp 760w,
/blog/rl-with-diffusion-i/image-18_hu7125908287036469741.webp 1200w"
src="https://ayushtues.github.io/blog/rl-with-diffusion-i/image-18_hu2556884957591279424.webp"
width="760"
height="719"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;h3 id="conclusion">Conclusion&lt;/h3>
&lt;p>So that concludes one of the ways in which Diffusion can be used for RL, there has been a cool follow-up which uses Diffusion for Offline RL : &lt;a href="https://arxiv.org/abs/2211.15657#:~:text=Our%20results%20illustrate%20that%20conditional%20generative,modeling%20is%20a%20powerful%20tool%20for%20decision-making." target="_blank" rel="noopener">[2211.15657] Is Conditional Generative Modeling all you need for Decision-Making? (arxiv.org)&lt;/a>, you can also find a list of other RL + diffusion works here: &lt;a href="https://github.com/opendilab/awesome-diffusion-model-in-rl" target="_blank" rel="noopener">opendilab/awesome-diffusion-model-in-rl: A curated list of Diffusion Model in RL resources (continually updated) (github.com)&lt;/a>.&lt;/p>
&lt;p>Overall I feel that this paper was quite novel in how it seamlessly integrated Diffusion with RL and showed that it enables a bunch of cool properties that might be desirable in RL. They also integrate with Diffusers, which is always nice for diffusion models.&lt;/p>
&lt;p>I might cover some other RL + Diffusion papers soon, or update this blog with some demos from the colab, so stay tuned!&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/rl-with-diffusion-i-d64c6e96d5ed" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Consuming Vs Creating</title><link>https://ayushtues.github.io/blog/consuming-vs-creating/</link><pubDate>Thu, 19 Jan 2023 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/consuming-vs-creating/</guid><description>&lt;blockquote>
&lt;p>A blank piece of paper is God’s way of telling us how hard it is to be God.&lt;br>
 — Sidney Sheldon&lt;/p>
&lt;/blockquote>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/consuming-vs-creating/image-1.png"
alt="Source : Create More Than You Consume if You Want to Worry Less and Feel More Fulfilled — OMAR ITANI">&lt;figcaption>
&lt;p>Source : &lt;a href="https://www.omaritani.com/blog/create-more-consume-less" target="_blank" rel="noopener">Create More Than You Consume if You Want to Worry Less and Feel More Fulfilled — OMAR ITANI&lt;/a>&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Okay so it’s the weekend, and you have lots of free time on your hands (on a side, it’s funny how many of my blogs are inspired by having shit loads of free time lol), and you’re thinking what to do with all this time. You want to watch that new episode of &amp;lt;insert the anime you’re binging currently&amp;gt;(one piece) or &amp;lt;insert latest episode of the latest podcast you’re following&amp;gt;(How I built this) or read the &amp;lt;insert latest book you’re reading&amp;gt;(A promised Land) or read the &lt;insert latest research paper you are interested in>(None lol), you get this gist, just consume, consume, consume! new and interesting stuff.&lt;/p>
&lt;p>But wait, you can also spend your time writing a blog, or sketching, or learning guitar, or brainstorming a solution to some coding problem, creating a video (for your almost dead YT channel :P), basically creating something new into existence (or atleast recreating your version of a previously existing thing).&lt;/p>
&lt;p>Well all this sounds good, and in an ideal world where we had infinite time, I could have consumed all I wanted, and created whatever I felt like, but alas we have limited time in this world and thus whenever I have any amount of free time, I’m constantly put in this dilemma of whether I should use it to consume or create. And yes, this sounds totally, totally like a first world problem, but okay, I don’t really have third world problems, and we need problems in our life to give meaning to it (all hail Victor Frankl for explaining the meaning of life to me), so why not create some first world problems for ourselves, shall we?&lt;/p>
&lt;blockquote>
&lt;p>Creativity is allowing oneself to make mistakes. Art is knowing which ones to keep.&lt;br>
 — Scott Adams, The Dilbert Principle&lt;/p>
&lt;/blockquote>
&lt;p>Now while at first thought, it might feel like I should spend most of my time on creating stuff, since it’s me actually producing something new in the world, using my own skills, and not just consuming stuff created by someone else, but let’s keep hold of the judgement cat inside us and give consuming a second thought.&lt;/p>
&lt;p>I feel that just like eating healthy and nutritious stuff is good for our body, and stuffing pizzas into our system is detrimental to our health, so is consuming good content a sort of mental food for our brain, to get new material for thinking, and for expanding our imagination. Whereas scrolling through Instagram reels, and mindlessly consuming endless memes, well, makes us mentally obese.&lt;/p>
&lt;blockquote>
&lt;p>***A reader lives a thousand lives before he dies . . . The man who never reads lives only one.&lt;br>
 —***George R.R. Martin&lt;/p>
&lt;/blockquote>
&lt;p>Reading books for example is a great way to increase our knowledge of the world via non-fiction, live multiple lives in one by reading through memoirs and biographies, and broadening our imagination of what is possible, and what might be possible through fiction (Fiction is soooooo underrated in its importance, that I’m planning to write an entire article about it someday, soon enough).&lt;/p>
&lt;p>Whereas listening to podcasts, not only helps you gain knowledge about the world (or about serial killers, if you’re into that sort of stuff lol), but the long form of conversation also reveals the human part of the guests, and gives you an insight into how the greatest people on the planet are in reality, how they react to questions, how they think, respond, and most importantly how they are also on this never ending process of sharing and consuming knowledge (okay, I’m talking more about interview styled podcasts, like my favourite The Lex Friedman show, How I built this etc, and I know podcasts don’t necessarily have to be like this, and can only be hosted by the host, but this is my blog, so fuck it).&lt;/p>
&lt;p>Also, as a technical person, I need to spend a fair amount of time, consuming technical articles, lectures, books and stuff either to learn something entirely new, or even stay updated with the latest happening in the familiar, but at the same time never boring, AI world. Now there’s an argument to be made here, that just consuming technical content doesn’t make you an expert or really improve your skills, and while I know there’s a lot of truth to that, personally, I enjoy reading about technical stuff that is no where even related to my day to day work, and it introduces me to a lot of new thinking patterns about how people approach problems in a different field, and also I feel it improves my meta learning, or my ability to learn new things, faster. (Huh, this also feels like something I should blog about someday, lets pin that to my notion list consisting of blog ideas that may or may not see the light of day ever lol).&lt;/p>
&lt;blockquote>
&lt;p>We write to taste life twice, in the moment and in retrospect.&lt;br>
 — Anaïs Nin&lt;/p>
&lt;/blockquote>
&lt;p>But okay, we have limited time, so why not spend it in creating new stuff, like writing this blog, which helps me do a brain dump, clear my head and also structure my thoughts a bit. Or maybe sketching, which well, just lets me focus for some time without any interruption, just with my sketchbook and my pencils (okay, also with the tutorial playing on my tablet 🙈). I think this is a think which is definitely better about creating and missing from consuming is the ability to focus on a task, like just sit down and use your brain to do something and create something and get that magical feeling of getting into the zone.&lt;/p>
&lt;p>Anyway, while creating does feel good, I don’t think there’s a problem with consuming good stuff, its just that we need to maintain a balance between the two, the time-management noob inside me, tried dividing the activities I do into consuming and creating, and scheduling time accordingly, failing speactacularly in the process, so I guess we just need to make sure we are just aware of this, and see how time unfolds.&lt;/p>
&lt;p>Until the next one, so long!&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/consuming-vs-creating-6b08d9d015e5" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Trusting the process OR asking the control freak inside you to STFU</title><link>https://ayushtues.github.io/blog/trusting-the-process-or-asking-the-control-freak-inside-you-to-stfu/</link><pubDate>Fri, 16 Dec 2022 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/trusting-the-process-or-asking-the-control-freak-inside-you-to-stfu/</guid><description>&lt;blockquote>
&lt;p>I wanted a perfect ending. Now I’ve learned, the hard way, that some poems don’t rhyme, and some stories don’t have a clear beginning, middle, and end. Life is about not knowing, having to change, taking the moment and making the best of it, without knowing what’s going to happen next. Delicious Ambiguity.&lt;/p>
&lt;/blockquote>
&lt;blockquote>
&lt;p>— Gilda Radner&lt;/p>
&lt;/blockquote>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/trusting-the-process-or-asking-the-control-freak-inside-you-to-stfu/image-1.jpg"
alt="Ahem ahem xD">&lt;figcaption>
&lt;p>Ahem ahem xD&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>With yet another weekend starting, and me not having any clue about what to do with all the time, and being shit tired to do anything productive, here we are back again, brain dumping the mess of thoughts inside my head, for well, no one to read really, other than maybe some random lost stranger, stumbling upon this nook of the internet guided by the all-powerful recommendation algorithms.&lt;/p>
&lt;p>Over the past few years, maybe because of the pandemic-induced lockdown or just cause I have so many changes going on in my life, it&amp;rsquo;s easy for me to feel lost and for life to feel out of control. But I am also starting to realise that it’s okay to give up control and just trust the process. But okay, why all this heavy talk? Let’s talk about something else, like maybe sketching!&lt;/p>
&lt;p>Now I was a very, very veerrrrrryyyyy nerdy kid in school, and the only two things going on in my life were my marks ( not even my subjects, I don’t think I actually cared about learning ) and computer games. I didn’t really have any hobby of sorts ( something which I am regretting a lot now and ever so slowly trying to fix ), but I did have one little hobby which used to sometimes pop out to remind me of its existence. I loved to draw!&lt;/p>
&lt;p>Wait wait, before you get any ideas, I am not a good painter, and there’s no Tare Zameen par going on here ( I am not even dyslexic, I may be ADHD or bipolar though, lol, IDK). Sometimes I feel my skills points for imagination and creativity got traded for my analytical and logical skills. If you ask me to imagine something or create something, I can’t, and the only thing I see when I close my eyes is pitch-black darkness. So I can’t really draw anything new, but wait, even if I can’t create something new, I can at least copy someone. So that’s what I do, I can’t imagine my own shit, but I like to copy other people’s work.&lt;/p>
&lt;blockquote>
&lt;p>I arise in the morning torn between a desire to improve (or save) the world and a desire to enjoy (or savor) the world. This makes it hard to plan the day.&lt;/p>
&lt;/blockquote>
&lt;blockquote>
&lt;ul>
&lt;li>E.B. White.&lt;/li>
&lt;/ul>
&lt;/blockquote>
&lt;p>We had this thing in our school where we were encouraged to draw something related to the pictures in our English/Hindi books in our notebooks, and I just used to love copying those paintings in my notebook. We also used to have a display board in the class for drawings, and there used to be a bunch of my drawings there as well. Again I was not very skilled, but I had this process where I used to draw rectangular grids around the original picture, and I used to draw scaled-out versions of the grid in my notebook and then just draw the part in each rectangle one by one, and just trusting that it will all fit in together, in the end and look something coherent and good, and it usually did. This also allowed me to convert like these very small drawings from my books to huge 2-page drawings in my notebook :P&lt;/p>
&lt;p>I stopped sketching for a long time since my JEE prep started, and I did one or two odd sketches during my 4 years of college, but recently I have been starting to pick this up again actively and find myself lost in sketching stuff after work.&lt;/p>
&lt;p>But wait, why all this talk about sketching? Isn’t the blog about trusting the process? Well…I feel through sketching I am actually realising the importance of just trusting in the process and not thinking too much about how things will end,, and just being optimistic about things. When I sketch something, in the beginning, it all looks pretty funny, just a bunch of odd circles and lines, and it doesn’t really feel like it will add up to anything much, or even when I am beginning to shade, it always feels like I’ll mess it up, and it will look flat and boring. But that’s where the magic of the process lies, I feel. You just have to trust that it will turn out well, or at least trust that the artist you are copying knows their shit and will somehow make things right at the end of the tutorial lol.&lt;/p>
&lt;blockquote>
&lt;p>Creativity cannot flourish and reach its deepest potential without the participation of its demons as well as its angels.&lt;br>
― Shaun McNiff&lt;/p>
&lt;/blockquote>
&lt;p>And I think this is true for soooooooo many things in my life; there are so many things where you just have to trust the process, making small efforts every day, investing a little bit of yourself into the process, and just hoping that it will all be worth it in the end. The intermediate state of things can sometimes look messy, and alright scary and hopeless, but as Jobs said, you can’t connect the dots while going forward, they’ll only make sense when you look backwards; you need to connect the dots backwards, and then everything will make sense, or maybe it won’t, maybe it will all just be like abstract art in the end, which won’t make any sense at all in itself, but will depend on how you interpret it.&lt;/p>
&lt;p>Even while writing this blog, I have to just trust the process and keep on writing, even though I don’t have anything much planned out, just hoping that the words will come to me and somehow I’ll be able to write something enough for an article, or well at least a long rant. ANYWAY, that’s a lot of brain-dump for an article; whenever I write one of these, I feel like I should do this more since it’s so relaxing; maybe I’ll, who knows.&lt;/p>
&lt;p>Until the next one — So long!&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/trusting-the-process-or-asking-the-control-freak-inside-you-to-stfu-c3abd6c899e9" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Life’s a Blur</title><link>https://ayushtues.github.io/blog/lifes-a-blur/</link><pubDate>Sun, 04 Sep 2022 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/lifes-a-blur/</guid><description>&lt;blockquote>
&lt;p>Ahem, this is what a quarter life crisis feels like I guess :P&lt;/p>
&lt;/blockquote>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/lifes-a-blur/image-1_hu13877147193073941880.webp 400w,
/blog/lifes-a-blur/image-1_hu2431967301938306886.webp 760w,
/blog/lifes-a-blur/image-1_hu7042391939318215148.webp 1200w"
src="https://ayushtues.github.io/blog/lifes-a-blur/image-1_hu13877147193073941880.webp"
width="760"
height="507"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>Aaaaaaaand I’m finally back in the mood to write again after so long, let’s see how this turns out, I don’t have a plan, I just feel like I should write something today, something, anything, just etch down what’s happening into something that’s permanently on the internet, a memory I can always come back to, with everything else in life passing away in a blur lol&lt;/p>
&lt;p>A lot has happened in the last few months, a change of the magnitude of the initial lockdown days, or coming to college for the first time. I’ve graduated college, and moved into Hyderabad, a fucking huge city, completely uncharted and unknown to me (well sort of, I know it somewhat now lol). The last few months have been completely crazy, first the last 2 months in college were the most fun ones in my entire span of college life ( and probably like my life also lol ). I spent the entire time hanging around with my friends, or making new ones, walking around the campus till the night faded away, or going on random trips, or giving random chapos to juniors, it all seems like a blur now, I barely remember anything, and I also remember so much at the same time.&lt;/p>
&lt;p>Then came moving into Hyderabad, while it has been 2 months since I came here, it feels like it’s been sooooooo long, and yet so less. Life’s become a routine, passing away in a blur, with going to work on the weekdays, and partying like hell, exploring the city on weekends, rinse and repeat. I’m meeting so many new and amazing people, holding on to the ones from college with all I have and also rekindling some relationships with a few older friendships. I’m trying out so many things, I went bowling, go-karting, played poker, monopoly, VR games which make you feel like a fucking Jedi for the first time, brushed up on my pool and foosball skills (finally get a bit better lol), picked up the guitar (again :P), its crazy how much life’s changed in such a short time really.&lt;/p>
&lt;p>But with all the thrill of a new life, there’s also something that feels wrong at times. With things changing so fast, you can’t help but wonder what’s actually stable in your life, what’s constant, what’s something you know you can hold on to? The very definition of who you are changes so fast, you don’t know what you stand for, what things you like, what you don’t, who are the people you can call yours, who are just temporary visitors, here to fade away into a distant memory. You stop having control on your own time, not knowing what to do with it anymore, just doing whatever seems interesting, whatever makes you feel alive, passing away life in a blur. There’s something soulless about it, about going anywhere and everywhere without a goal, just hoping to reach somewhere, oh which reminds me of this nice quote from Alice in Wonderland&lt;/p>
&lt;blockquote>
&lt;p>&lt;strong>Alice:&lt;/strong> Would you tell me, please, which way I ought to go from here?&lt;br>
&lt;strong>The Cheshire Cat:&lt;/strong> That depends a good deal on where you want to get to.&lt;br>
&lt;strong>Alice&lt;/strong>: I don’t much care where.&lt;br>
&lt;strong>The Cheshire Cat:&lt;/strong> Then it doesn’t much matter which way you go.&lt;br>
&lt;strong>Alice:&lt;/strong> …So long as I get somewhere.&lt;br>
&lt;strong>The Cheshire Cat:&lt;/strong> Oh, you’re sure to do that, if only you walk long enough.&lt;/p>
&lt;/blockquote>
&lt;p>With so much instability in life, I feel it’s important to have some constants in life, to give some meaning to all the madness, something you can always return to, as something &lt;em>you&lt;/em> do when you’re free, someone you know will be there for you, something that will always make you happy. For me I guess its just some people in my life, and some of my interests like music, reading, gaming or learning about random pointless stuff which has no practical use in my life lol. I guess these are the things that keep me sane, and give me some form of identity, making me, me. Oh this reminds me of another quote about how slowly things become a part of your life, from A Man Called Ove&lt;/p>
&lt;blockquote>
&lt;p>Loving someone is like moving into a house,” Sonja used to say. “At first you fall in love with all the new things, amazed every morning that all this belongs to you, as if fearing that someone would suddenly come rushing in through the door to explain that a terrible mistake had been made, you weren’t actually supposed to live in a wonderful place like this. Then over the years the walls become weathered, the wood splinters here and there, and you start to love that house not so much because of all its perfection, but rather for its imperfections. You get to know all the nooks and crannies. How to avoid getting the key caught in the lock when it’s cold outside. Which of the floorboards flex slightly when one steps on them or exactly how to open the wardrobe doors without them creaking. These are the little secrets that make it your home&lt;/p>
&lt;/blockquote>
&lt;p>I guess all this is normal, and expected when life’s going through a big change, and I shouldn’t think much about it, but well, what else are we supposed to do on a Sunday night, after the weekend passing by in a blur, partying lol. Let’s end it with a nice quote from the internet , and hope I keep on writing blogs regularly now lol&lt;/p>
&lt;blockquote>
&lt;p>The days blur into one, and the backs of my eyes hum with things I’ve never done.&lt;/p>
&lt;ul>
&lt;li>Radical Face&lt;/li>
&lt;/ul>
&lt;/blockquote>
&lt;p>That’s all folks, Byeeeeeee!&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/lifes-a-blur-ebf037874f37" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Context switching like my brain is a fucking CPU</title><link>https://ayushtues.github.io/blog/context-switching-like-my-brain-is-a-fucking-cpu/</link><pubDate>Wed, 15 Dec 2021 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/context-switching-like-my-brain-is-a-fucking-cpu/</guid><description>&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/context-switching-like-my-brain-is-a-fucking-cpu/image-1_hu9123580443307519225.webp 400w,
/blog/context-switching-like-my-brain-is-a-fucking-cpu/image-1_hu13666029645416820587.webp 760w,
/blog/context-switching-like-my-brain-is-a-fucking-cpu/image-1_hu18054845232646842131.webp 1200w"
src="https://ayushtues.github.io/blog/context-switching-like-my-brain-is-a-fucking-cpu/image-1_hu9123580443307519225.webp"
width="704"
height="400"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>They say our brain is sort of like a CPU, well agreed. But it’s not a multiprocessor which can manage multiple tasks in parallel, its a single core CPU which only supports a single thread at a time, and any attempts to balance multiple tasks at the same time, leads to so much context switching, that my brain usually ends up hanging like a CPU. A context switch is when a CPU switches from one task to another, and has to store some info about the old task ( to resume it later ) and also load some info about the new task to start working on it, this saving and loading leads to inefficiencies and multiple context switches leads to poor CPU utilisation. And that is exactly what my brain keeps on doing all the time these days it seems. I CAN’T FUCKING FOCUS!!!&lt;/p>
&lt;p>Okay, I should seriously do something about finding CS analogies about everything in my life, idk, maybe I like CS so much that I can’t help relate everything with CS (much to the annoyance of some of my friends though 😛), a few days back I was explaining a friend how love can be expressed as a constructive superposition of reward functions of two people ( which BTW I heard from Wojciech Zaremba, co-founder OpenAI on Lex Freidman’s Podcast ), and then I was complaining about how the market making ability of the dating market is so bad that it sounds like such an inefficient market, and then we were contemplating what would happen if I say such things on an actual date, well who knows, I don’t have one to do that experiment lol.&lt;/p>
&lt;p>As you can clearly see from the above paragraph, I find it very hard to focus on things, my mind wanders from one topic to the other very fast, which is also pretty evident when talking to me, cause I randomly switch to unrelated topics in the middle of conversations. I don’t know why that is so really, maybe it’s due to the generalist inside me, who likes to know about anything that seems interesting, which is in stark contrast with what I actually do, ie. research, which requires you to focus deeply on a topic, and carefully think about the problem, and its solution, maybe that’s why I don’t like doing research anymore.&lt;/p>
&lt;p>This blog is also a result of a lack of my focus, I was technically supposed to be thinking about my ongoing research projects, but I am too distracted to think about it, so here we are ranting in our blog to calm our mind a bit so that I can at least sleep without overthinking.&lt;/p>
&lt;p>The most valuable thing in our current generation is our attention, our focus. There are so many options, so many things that we can do that, that everyone is fighting for getting as much attention from others as possible, which is a bit weird if you think about it, since attention isn’t actually all that difficult to get &lt;em>if you tune out all the noise,&lt;/em> and puts anyone who can do so at such an unfair advantage in this distracted world. The products around us are made to compete for our attention, and in a free internet, our attention is itself the product which companies are selling to the highest bidders ( through ads, which infact makes up the majority of the revenue of Big-Tech giants like Google, Meta ).&lt;/p>
&lt;p>Sometimes when I am too distracted by all the context switching, I just put up some music and roam around in my room, it sort of helps with bringing myself back to focus, switching of my mobile helps as well, I think our mobiles are to blame a lot for this problem ( and we are to blame for using them in that way ). I can’t help but imagine how fucking awesome would it be to be able to have laser sharp focus on the task I am currently doing &lt;em>and only that task&lt;/em>, not caring about whatever the hell is happening in the world, maybe that’s where brilliance lies, maybe that’s how you do something unique — &lt;em>by cutting out the noise.&lt;/em>&lt;/p>
&lt;p>Well, with my brain back to reasonable amounts of clarity, that will be all for the post, if you found this post relatable, I can only say that it’s a real problem, and something that doesn’t get said often in today’s hustle culture obsessed with multi-tasking, so hey, just my thoughts but if you’re feeling distracted too, maybe write a bit about it :P&lt;/p>
&lt;p>Anyway that’s all for the post, so long!&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/context-switching-like-my-brain-is-a-fucking-cpu/image-2_hu13082928053486643770.webp 400w,
/blog/context-switching-like-my-brain-is-a-fucking-cpu/image-2_hu7131966780386037929.webp 760w,
/blog/context-switching-like-my-brain-is-a-fucking-cpu/image-2_hu13558755836345130327.webp 1200w"
src="https://ayushtues.github.io/blog/context-switching-like-my-brain-is-a-fucking-cpu/image-2_hu13082928053486643770.webp"
width="760"
height="507"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/context-switching-like-my-brain-is-a-fucking-cpu-133efa606817" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Living on Curiosity</title><link>https://ayushtues.github.io/blog/living-on-curiosity/</link><pubDate>Fri, 05 Nov 2021 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/living-on-curiosity/</guid><description>&lt;p>Curiosity killed the cat, but for some of us, it has been the only way to keep rolling through life in the last 2 years.&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/living-on-curiosity/image-1.jpg"
alt="I don’t know what memes to put, so I’ll just spam the entire blog with curiosity killed the cat memes 😛">&lt;figcaption>
&lt;p>I don’t know what memes to put, so I’ll just spam the entire blog with curiosity killed the cat memes 😛&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>I was not a curious kid, and mostly spent my time playing video games all day (which you can read more about &lt;a href="https://ayushtues.medium.com/ah-shit-here-we-go-again-35dae2acf8c4" target="_blank" rel="noopener">here&lt;/a> ), and rote-learning the answers teachers told us to memorise or practising the same questions enough time, to just learn the methods by heart. Well okay, my curiosity did start appearing through the cracks, while reading stuff about comic characters or the characters of the video-games on Wikipedia for hours, or while binge-watching stuff on History TV/Nat-Geo/Discovery all day long. ( Yep, I am a nerd, no point hiding it lol ).&lt;/p>
&lt;p>And now, I basically live on curiosity, some days its the only thing that keeps me going on ( ah I sound so depressed lol, which is totally not the case btw). With the world locked in a pandemic, and having practically nothing to do for months at times, I could only depend on my curiosity to find any reason to get through the day. Reading books, watching podcasts, learning about cool new tech stuff, or useless philosophy, became the way to pass my time and my curiosity my only constant companion.&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/living-on-curiosity/image-2_hu6195842716705802303.webp 400w,
/blog/living-on-curiosity/image-2_hu8605889564012048702.webp 760w,
/blog/living-on-curiosity/image-2_hu7760580272588607078.webp 1200w"
src="https://ayushtues.github.io/blog/living-on-curiosity/image-2_hu6195842716705802303.webp"
width="500"
height="500"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>I am not sure what came first, whether my curiosity attracted me to research, or doing research made me more curious, but there is something deeply satisfying about losing yourself in the process of learning more about a topic you’re passionate about and finding yourself even more and more curious about it ( I am sooooooooo curious about the whole Crypto-Blockchain space right now, that I am diving deeper and deeper into the rabbit hole lol ).&lt;/p>
&lt;p>And there’s this thing about curiosity, you don’t need any external incentive to pursue it, it’s just intrinsically rewarding to you. You will happily spend countless hours, satisfying it, while at the same time you’ll dream for a heft payment, for doing your job. In fact this intrinsic reward is such a powerful idea, that AI researchers have used “curiosity”, to train agents to learn complex behaviours without even having an external reward. The agent is just told to explore and do stuff which satisfies its curiosity, and in the process it learns these highly complex skills, which if you think about it, is true for humans too, when we allow ourselves to follow our curiosity, we can end up in places where we never imagined we’ll be capable of reaching 😛.&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/living-on-curiosity/image-3_hu4452662750438760190.webp 400w,
/blog/living-on-curiosity/image-3_hu7544053114955075761.webp 760w,
/blog/living-on-curiosity/image-3_hu15265621398115775053.webp 1200w"
src="https://ayushtues.github.io/blog/living-on-curiosity/image-3_hu4452662750438760190.webp"
width="523"
height="760"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>And now I feel there are just so many amazing things in the world, that I can’t imagine not being curious about it all the time. We’re living in an era where you can access any information you want with a simple google search, how can you not want to know more about stuff? And how the hell do you deal with the realisation that this small life of ours is not enough to satisfy our curiosity about really anything much, and how do you not want to try to learn anything and everything you can, while you still have the time? Lol for me this manifests in a weird way, where I sometimes ( read:often) put all my current responsibilities aside, and just start doing some shit reading for fun. Maybe that’s just a form of escapism, to avoid the work at hand, and instead get some instant intrinsic reward.&lt;/p>
&lt;p>One place where curiosity really helps I think is in your relationships with others. I feel if you just try to be curious about the other person, as a wonderful human, full of interesting perspectives, worth knowing, you can never run out of things to talk about. If you’re truly curious about the other person, you will eventually start connecting them on a much deeper level, than you’d if you just do some small talk, cause like, who doesn’t like talking about themselves ( yep, we are all narcissists ). This is something I have personally benefited, well with nothing to do, and not being an interesting person myself, I have tried to be curious about other people ( trying to put it here in the most uncreepy way I can xD), and trying to know their views on things, which has been blissful.&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/living-on-curiosity/image-4_hu12647264385377223127.webp 400w,
/blog/living-on-curiosity/image-4_hu1528808254733031117.webp 760w,
/blog/living-on-curiosity/image-4_hu4625945153404622053.webp 1200w"
src="https://ayushtues.github.io/blog/living-on-curiosity/image-4_hu12647264385377223127.webp"
width="544"
height="760"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>I would have probably not survived the year without my curiosity, what else are you supposed to hold on to when you’re all alone, cut-off from society? It’s like music, its a constant in my life. Even if there is nothing else to rely upon, I can always rely on good music, or my curiosity to keep myself busy.&lt;/p>
&lt;p>If curiosity is such an enriching thing for us, why do we try to suppress it? Why are we not encouraged to ask more questions? Why are people who ask more questions shunned off as nerds? Why do school teachers, who are supposed to educate us, lash off if we ask something out of syllabus? Hell, why are scientists shown as those crazy weird dudes and stereotyped? Most inventions we take for granted today, were the result of someone satisfying their curiosity, about what could be the limits of our society. I mean sure, curiosity did kill the cat, there are certain things were you just cant satisfy your curiosity, to bide by moral, ethical or societal rules, you can’t kill someone just cause you’re curious how it feels like ( no no, I am not a sociopath ), but apart from that, why are we forced to bottle in our curiosity?&lt;/p>
&lt;p>Well whatever might be the reason for that, I for one can’t stop my curiosity, and I don’t want to really. It’s been a driving force for me, both professionally and personally, and has mostly lead to positive outcomes. If nothing, it just ends up giving you that extra motivational boost, you sometimes desperately need lol&lt;/p>
&lt;p>Anyway, that was it for this post, with things coming back to normal, I am curious about how my life would be offline after so much time 😛&lt;/p>
&lt;p>Until the next one, Goodbye!&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/living-on-curiosity/image-5_hu1355862320167609000.webp 400w,
/blog/living-on-curiosity/image-5_hu17999805936068852295.webp 760w,
/blog/living-on-curiosity/image-5_hu6887797385489147431.webp 1200w"
src="https://ayushtues.github.io/blog/living-on-curiosity/image-5_hu1355862320167609000.webp"
width="760"
height="501"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/living-on-curiosity-f54ff980a947" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Why I don’t want to let go of being “a child”</title><link>https://ayushtues.github.io/blog/why-i-dont-want-to-let-go-of-being-a-child/</link><pubDate>Thu, 07 Oct 2021 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/why-i-dont-want-to-let-go-of-being-a-child/</guid><description>&lt;p>I have always been the “child” at my home, simply cause I have been the youngest one in the family for around 20 years, until the moment my nephew was born this year, taking the spot away from me.&lt;/p>
&lt;p>I think being the child of the home has affected me in ways that play out in interesting quirks in the real world, and I now kind of like playing the child in my other social groups in my college as well, although it is becoming harder and harder to maintain that image as I grow over the years, and people start expecting me to mature a bit, and take responsibility about my decisions, and even manage people lol. But honestly speaking, I feel that I don’t want to change the child-like personality into a mature one and thinking about that led into the birth of this post.&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/why-i-dont-want-to-let-go-of-being-a-child/image-1.jpg"
alt="Yeah, you would think xD">&lt;figcaption>
&lt;p>Yeah, you would think xD&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Why do we stop smiling as we grow up? Smiling about random incidents cooked in our mind, smiling to random strangers, to ourselves? Why do people expect us to let go of that wonderful naivety of the child, looking at the world through his rose colored glasses? I can’t stop smiling all the time, and really I smile for no reason at times, and find it super hard to control, even in serious situations ( which sometimes leads to tricky situations, like in a classroom….you can imagine 😛). Having a smile can diffuse the most tense of situations, and can in general do wonders in your life, so smile people, smile like a child ! I don’t want to ever grow up enough to stop smiling all the time, just the thought of such a future feels inexplicably dark and mundane.&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/why-i-dont-want-to-let-go-of-being-a-child/image-2.jpg"
alt="This is true on so many levels, with one being me not having a life">&lt;figcaption>
&lt;p>This is true on so many levels, with one being me not having a life&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Why do we have to stop trusting people naively? Call me a child, but I believe people are generally good and worthy of trust and good people are the rule, not the exception ( I am reading some Jordan Peterson on the side right now, and he’ll disagree a lot with this lol, but fuck it, I am naive ). I don’t want to lose the childish innocence of trusting people, because it has lead me to make blind leaps of faith in other people, and mostly people don’t disappoint really, they return your trust, sometimes, amplified. Now I do know that this is not always the case, and I shouldn’t trust random strangers all too easily ( like random people you meet on a trip? ), you &lt;em>could&lt;/em> end up in serious trouble someday, but do I really want to stop trusting people because someone someday &lt;em>might&lt;/em> break it? Doesn’t that seem a bit too harsh on people? I don’t want to grow up enough to not be able to trust people anymore.&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/why-i-dont-want-to-let-go-of-being-a-child/image-3.jpg"
alt="What the fuck happened to trusting people?">&lt;figcaption>
&lt;p>What the fuck happened to trusting people?&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Why do I have to grow up enough to be self-reliant, and not depend on others for help? Like really, is the world sooooooo easy that we can pretend that its possible to survive it on our own? Now I am not saying that we shouldn’t aim to grow, we should always try to become better than our past, what I am getting at is that I don’t want to grow enough that I stop asking people for help. Children don’t hesitate at all to ask for help from their peers and adults alike, it’s the only way they could survive really, since they don’t know much about really anything. But I don’t want to grow up enough to start believing that I know enough to be on my own from now on….I will always require people to help me figure my way through life out, and isn’t that how we make friends as well, by helping each other out?&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/why-i-dont-want-to-let-go-of-being-a-child/image-4.jpg"
alt="ALWAYS!">&lt;figcaption>
&lt;p>ALWAYS!&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>I also don’t want to grow up enough to stop being curious about the world, and stop seeing the wonder there is in all the things. Children see the world with a view of wonder, and are infinitely curious about things, because they don’t have a preimage of how the world should be, why do we have to stop being that way? Why do we have to grow into adults who have just accepted the fact that things are the way they are just because they are supposed to be that way? In the world of children, anything is possible, and maybe a child-like fresh view of the world might exactly be the thing we need to solve the most challenging problems that lie ahead of humanity, but maybe we are too busy growing up to see that.&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/why-i-dont-want-to-let-go-of-being-a-child/image-5.jpg"
alt="Keep calm, and stay curious!">&lt;figcaption>
&lt;p>Keep calm, and stay curious!&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Why do we suddenly need to have expectations from ourselves as we grow up? No one really expects anything from a child, and that is liberating in a sense, isn’t it? You’re free to do whatever you want and you do things cause you want to do them, not cause someone else expects you to do it. I don’t want to grow up enough to succumb into other people’s expectation of me, instead I want to remain a child, and just do whatever the hell I want to do.&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/why-i-dont-want-to-let-go-of-being-a-child/image-6.png"
alt="Expectations are only good in mathematics">&lt;figcaption>
&lt;p>Expectations are only good in mathematics&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>As time passes by, I will have to grow up, eventually. I will need to have a way to sustain myself, alone. I will have to take up on responsibilities, and be responsible for other people for as well. But that’s okay, that’s the good parts about growing up, it makes you stronger, what I don’t want is to lose some of the good parts about being a child while I grow up, I always want to remain a child in some ways ( lol, totally sounds like a quarter-life crisis doesn’t it? )&lt;/p>
&lt;p>Anyway, that’s it for the post, it does seem like things are returning back to normal slowly, and maybe this time they will stay normal lol, I hope the person reading this finds the courage to pass the time until things return back to normal, and hey, if you want to reach out to anyone, feel free to reach out!&lt;/p>
&lt;p>Until the next one, Stay Safe!&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/why-i-dont-want-to-let-go-of-being-a-child/image-7_hu13082928053486643770.webp 400w,
/blog/why-i-dont-want-to-let-go-of-being-a-child/image-7_hu7131966780386037929.webp 760w,
/blog/why-i-dont-want-to-let-go-of-being-a-child/image-7_hu13558755836345130327.webp 1200w"
src="https://ayushtues.github.io/blog/why-i-dont-want-to-let-go-of-being-a-child/image-7_hu13082928053486643770.webp"
width="760"
height="507"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/why-i-dont-want-to-let-go-of-being-a-child-bd98d71853ab" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>WORK, WORk, WOrk, Work, work…..why?</title><link>https://ayushtues.github.io/blog/work-work-work-work-work..why/</link><pubDate>Fri, 27 Aug 2021 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/work-work-work-work-work..why/</guid><description>&lt;p>Na na before you get any funny ideas, this wasn’t inspired by Rihanna or Fifth Harmony. Instead this is the result of having zero motivation to do any work ( and not having any work to either lol ), and the gradually increasing guilt that comes from not having written anything for a while.&lt;/p>
&lt;p>Anyway, while watching random videos on YouTube ( &lt;a href="https://www.youtube.com/watch?v=RPjL909DiVI" target="_blank" rel="noopener">link1&lt;/a>, &lt;a href="https://www.youtube.com/watch?v=cZnBNuqqz5g" target="_blank" rel="noopener">link2&lt;/a> if you’re interested in what I was watching)and diving deeper and deeper to the rabbit hole it is, I started thinking about a question : &lt;em>&lt;strong>Why do we work?&lt;/strong>&lt;/em> Don’t ask me why I was thinking about philosophical questions while watching shit content on youtube, the internet is an abyss, stare deep into it, and it stares back at you and messes with your head.&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/work-work-work-work-work..why/image-1.jpg"
alt="I honestly have no idea">&lt;figcaption>
&lt;p>I honestly have no idea&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Well the answer to “why &lt;em>we&lt;/em> work” is not for me to answer, that’s something for the world and all its people to decide, so I’ll try a different version of the question, sort of like a single sample estimate of the problem : why do &lt;em>I&lt;/em> work?&lt;/p>
&lt;p>Since I am in my 4th year now, and will start “working” in whatever sense the world defines it in the near future, this question was a bit hard for me to think about. But the question is more general than just the work I’ll do for a company, work encompasses a large part of my current life as well, as a self-proclaimed workaholic, the work I do is what defines me for me. If there is no meaning to why I do work, then is there even a meaning to me? Well about that, there is no meaning to my or anyone’s existence lol, but let’s not go there today.&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/work-work-work-work-work..why/image-2_hu9400537340590594860.webp 400w,
/blog/work-work-work-work-work..why/image-2_hu6931451749401026690.webp 760w,
/blog/work-work-work-work-work..why/image-2_hu9264097348186518319.webp 1200w"
src="https://ayushtues.github.io/blog/work-work-work-work-work..why/image-2_hu9400537340590594860.webp"
width="640"
height="480"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>Why do I work for? The first obvious thing that comes to my mind is money. I work so that I can potentially financially support myself in life. And this is probably why most people work, to support for themselves and their family, and I don’t want to demean this, it’s a noble task to work to support for those you care about and also for yourself, to live whatever standard of lifestyle you wish for. But is money all that I work for? Not really.&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/work-work-work-work-work..why/image-3_hu16549087837115049846.webp 400w,
/blog/work-work-work-work-work..why/image-3_hu14920456084338574205.webp 760w,
/blog/work-work-work-work-work..why/image-3_hu16184580354305656899.webp 1200w"
src="https://ayushtues.github.io/blog/work-work-work-work-work..why/image-3_hu16549087837115049846.webp"
width="760"
height="733"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>If money was all I worked for, it wouldn’t explain the hours I have spend reading about stuff which has no connection with my job, or for the times when I worked for free, working just for the sake of it, working purely out of intellectual curiosity, and atleast till now, I have worked harder here than when I worked when anyone paid me for the task. So maybe I work, just because I like working? But I like a lot of other things as well, then why work if I can do those other things, if I am not getting any special reward to do the work, then it’s just another one of the things I can possibly do, like doing nothing, I love that as well xD. So maybe that’s why we give people money to work, so that they choose it over other things they equally enjoy?&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/work-work-work-work-work..why/image-4_hu4363859289706378111.webp 400w,
/blog/work-work-work-work-work..why/image-4_hu9280061592399331680.webp 760w,
/blog/work-work-work-work-work..why/image-4_hu12049229125435108786.webp 1200w"
src="https://ayushtues.github.io/blog/work-work-work-work-work..why/image-4_hu4363859289706378111.webp"
width="500"
height="300"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>There’s also the concept of a 5day, (or a 4day) workweek, which I don’t seem to get. Does that mean that work is not a necessity in your life, but just a part-time affair? Why do we need a break from work, if work is such an integral part of our life, why does even the bible say we need a day off?&lt;/p>
&lt;blockquote>
&lt;p>Exodus 34:21: Six days thou shalt work, but on the seventh day thou shalt rest: in earing time and in harvest thou shalt rest&lt;/p>
&lt;/blockquote>
&lt;p>Would we not need a day off, if we were not working for money? Think about all the extra work we’ll get done off, Would that extra day of work be a good thing for society, or would that be a bane?&lt;/p>
&lt;p>Or do we work just to give meaning to our life? Maybe we work just to give an excuse for why we are alive on this earth, cause hey, I am working, that means I am doing something meaningful and thus I deserve to live? Or maybe that’s just the workaholic inside me trying to justify its existence lol, Or maybe we work just to distract ourself from the problems in our life? If we can appear to be busy in our “work”, we don’t need to focus on other things in our life which we don’t want to encounter head on.&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/work-work-work-work-work..why/image-5_hu7015081519550014221.webp 400w,
/blog/work-work-work-work-work..why/image-5_hu14497836329529217811.webp 760w,
/blog/work-work-work-work-work..why/image-5_hu11762129238754396517.webp 1200w"
src="https://ayushtues.github.io/blog/work-work-work-work-work..why/image-5_hu7015081519550014221.webp"
width="600"
height="450"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>Alas, I don’t know the answer to the question of Why do we work? and it seems like I posted a lot more questions than answering anything lol, but hey, that’s what the blog is for, to dump in whatever shit’s going in my brain and get some clarity ( or get even more confused ). Anyway, the reality is that I will have to work in one form or another for the rest of my life, even if there is no meaning to it, cause who said even life has any meaning at all?&lt;/p>
&lt;p>Well anyway, that’s it for this blog, it’s funny how I can waste time thinking about stuff to write about, and then get an idea while watching random videos on YouTube….maybe there’s something seriously wrong with me, but okay, until the next time!&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/work-work-work-work-work..why/image-6_hu15314339617622266958.webp 400w,
/blog/work-work-work-work-work..why/image-6_hu1559227018599105879.webp 760w,
/blog/work-work-work-work-work..why/image-6_hu6519156482396565333.webp 1200w"
src="https://ayushtues.github.io/blog/work-work-work-work-work..why/image-6_hu15314339617622266958.webp"
width="760"
height="517"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/work-work-work-work-work-why-21e458df68bc" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>So….what are your hobbies?</title><link>https://ayushtues.github.io/blog/so.what-are-your-hobbies/</link><pubDate>Wed, 21 Jul 2021 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/so.what-are-your-hobbies/</guid><description>&lt;blockquote>
&lt;p>TLDR : Got anxious, mind started wandering, started overthinking about useless stuff, and finally ended up in an existential crisis as usual, wrote a blog in the process, nothing new, seriously…….. skip this, and go read something better &lt;em>😛&lt;/em>&lt;/p>
&lt;/blockquote>
&lt;p>Imagine this, you just met a new person , or got into a new social group of some kind or maybe you are just interviewing for a potential career opportunity. You are brimming with excitement to introduce yourself, to put yourself out there, and let people see how fucking awesome you are, and it goes on good for a while, that smug confidence you take pride in, is working its charms and people seem interested in knowing your more, but then you sense something happening…..ah fuck, no please no, don’t, don’t say it, not again…….you try to quickly maneuver the conversation, hoping to avoid the inevitable, and then bam, it happens again, and you crash-land on your face…. someone asks, &lt;em>So…..what are your hobbies?&lt;/em>&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/so.what-are-your-hobbies/image-1_hu6038185465017993957.webp 400w,
/blog/so.what-are-your-hobbies/image-1_hu11609960814824555240.webp 760w,
/blog/so.what-are-your-hobbies/image-1_hu14968227206652065452.webp 1200w"
src="https://ayushtues.github.io/blog/so.what-are-your-hobbies/image-1_hu6038185465017993957.webp"
width="600"
height="347"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>There are many questions that put me into an existential crisis, making me question my entire existence, like &lt;em>what’s the point of life&lt;/em> ( there’s no point )&lt;em>, what’s your future plan&lt;/em> ( I have none )&lt;em>, what am I doing with my life&lt;/em> (wasting it ), ormy mom’s &lt;em>what do you want to eat for dinner?&lt;/em> ( she ends up deciding it herself anyway). I can probably write a blog about each of them, but these are mostly questions that I have to face from people relatively closer to me and are relatively infrequent ( other than the dinner bit ), but today I wanted to talk about a question I have to answer again and again , all the time, in literally &lt;em>any&lt;/em> situation I have to introduce myself to others, like seriously, who the fuck invented this torturous ritual of asking people, &lt;em>So…..what are your hobbies?&lt;/em>&lt;/p>
&lt;p>Whenever I am asked this question, I get very confused, and anxious, since I don’t really feel like I have any hobby, atleast like other normal people seem to have. Sure I listen to music, keep on reading/watching random stuff on the internet, but they don’t seem to fall in the normal “cool” hobbies like painting, or dancing etc that people usually expect you to answer. This leads to me questioning my existence, and I often ( reluctantly ) end up answering something like “I have no hobbies” or “ah….watching animes?” And it doesn’t take a mentalist to guess what the other person’ feels about me after this xD&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/so.what-are-your-hobbies/image-2.gif"
alt="GIF">&lt;figcaption>
&lt;p>When you tell someone you don’t have a hobby :P&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Anyway, let’s first define what a hobby is, we’ll use our usual trusty go-to source of wisdom, Wikipedia ( as a side note, I recently came to know that even in industry settings, information on Wikipedia is considered to be kind of a gold standard in quality, which seems a bit weird, since I literally go edit any page on the website right now lol)&lt;/p>
&lt;blockquote>
&lt;p>A hobby is considered to be a regular activity that is done for enjoyment, typically during one’s leisure time. Hobbies include collecting themed items and objects, engaging in creative and artistic pursuits, playing sports, or pursuing other amusements. Participation in hobbies encourages acquiring substantial skills and knowledge in that area&lt;/p>
&lt;/blockquote>
&lt;p>Now it is definitions like these, that make answering this question such a hard job. Firstly it says a hobby needs to be something you do regularly, well you know what, I can’t keep on doing &lt;em>anything&lt;/em> regularly, I can’t schedule a time everyday to journal ( failed attempt count to keep a journal tending to infinity ), or to practice an instrument ( same story here as well ), something like this just doesn’t work for me. Hell, I can’t even watch animes regularly, instead it usually happens in like short periods of binging, where I end up watching an entire series in a single sitting. So hell no, if a hobby is something you have to do regularly, I’ll remain hobby-less forever.&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/so.what-are-your-hobbies/image-3.gif"
alt="GIF">&lt;figcaption>
&lt;p>What’s your hobby?…..uhh….the internet?&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Next, hobby is something done for enjoyment, typically during one’s leisure time. Ah well, lol, HELL NO. Not unless enjoyment means escaping from your boring normal life, which seems so mundane, that you can’t stand another minute of it, so you hop on to that random youtube video, or that new cool anime, or even start writing your blog, &lt;em>even&lt;/em> during your work hours ( usually in your work hours lol ), since you anyway can’t concentrate on the boring work you are doing. The fact that you classify some stuff you are doing as “fun/hobby” and other as work, is a signal that something’s wrong somewhere according to me.&lt;/p>
&lt;p>&lt;em>Hobbies include collecting themed items and objects, engaging in creative and artistic pursuits, playing sports, or pursuing other amusements. Participation in hobbies encourages acquiring substantial skills and knowledge in that are.a&lt;/em> Well my “hobbies” consists of binging animes, watching random youtube videos, just listening to songs for hours, and overthinking about stuff, so does that mean I haven’t got a hobby? If a hobby is meant to be something you do for fun, why do we need to associate it with a few cool stuff like painting or reading or playing a sport….why isn’t binging random stuff on the internet considered a hobby? Also well I don’t know about you, but I don’t consider myself having substantial skill and knowledge about anything, well sure I know my fair-share of anime trivia, like what’s happening in the latest chapter of One Piece or My Hero Academia, but I don’t think that something that’s meant to be done for fun, should necessarily be related to acquiring skills, can’t we just do things for fun?&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/so.what-are-your-hobbies/image-4.gif"
alt="GIF">&lt;figcaption>
&lt;p>Sabke paas nahi hoti laxman&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Okay, I guess that’s enough rant for a blog, but I do feel if hobbies are meant to be something you do for fun, you don’t need someone else’s pre-existing definition of what a hobby is to bind you, and make you feel bad about not having a “cool” hobby. So maybe the next time when someone asks you what are your hobbies, just try being honest about what you enjoy doing in your free time, it could be something as weird as watching people reacting to people reacting to people reacting to people xD&lt;/p>
&lt;p>Anyway, that’s it for this blog, as usual feel free to hit me up, and let’s have a good chat!&lt;/p>
&lt;p>Until then — Stay Safe!&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="GIF"
src="https://ayushtues.github.io/blog/so.what-are-your-hobbies/image-5.gif"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/so-what-are-your-hobbies-b438626a764f" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>The Man with(out) a Plan</title><link>https://ayushtues.github.io/blog/the-man-without-a-plan/</link><pubDate>Sun, 20 Jun 2021 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/the-man-without-a-plan/</guid><description>&lt;p>What’s your plan? You’re entering your final year, what are you planning to do after college, you must have thought about it by now, right?&lt;/p>
&lt;p>Well….&lt;em>HEEEELLLLLLL no&lt;/em> 🙈&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/the-man-without-a-plan/image-1.jpg"
alt="Like seriously, have you ever seen me rant 🙈">&lt;figcaption>
&lt;p>Like seriously, have you ever seen me rant 🙈&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>I don’t know why, but people always seem to assume that I have always got my stuff figured out. This is something that has stuck with me since my school days, and it probably seems that way from the outside, when you &lt;em>just&lt;/em> see the outputs and the achievements and stuff while the reality can’t be farther from the truth, and the peeps who really know me, know just how childish fickle-minded, volatile, gut-feely guy I am, taking most of my major decisions just because my instincts tell me so, and also how I am &lt;em>totally&lt;/em> not smart enough, or focused enough to make any long terms plan anyway.&lt;/p>
&lt;blockquote>
&lt;p>Tera to sab set hai, tujhe kya hi chinta hai”&lt;/p>
&lt;/blockquote>
&lt;blockquote>
&lt;p>&lt;em>- I HATE these words&lt;/em>&lt;/p>
&lt;/blockquote>
&lt;p>Anyway, enough with the rant, let’s talk come back to the blog. Well, I don’t really have a plan about what to write in the blog, I have observed that most of my good blogs happen without prior planning. It’s like I can spend a day thinking about what to write in the next blog, and it’s all blank, but then bam, days later, when I am not even thinking about it, the idea for the next one strikes, and I usually end up writing them within an hour or two. Incidentally, this also happens a lot with the books I read, I spend a lot of time searching for the next book to read, try out reading some pages of a few of them, get bored, tired or both, and end the search for the day, and then miraculously when I am not looking for a book, I get a book suggestion somewhere, and it turns out to be so lovely that I end up binging that book in the next day or two.&lt;/p>
&lt;p>Why does everyone expect us to have a plan? Yeah, I am sure all of your plans for 2020 turned out well 😛 . When the world is soooooooooo random, what’s the use of planning anything anyway. Also, why do we suddenly need to have a plan when we grow up, as a child I could literally spend the entire day doing &lt;em>nothing&lt;/em>, having to plan out &lt;em>nothing&lt;/em>, and expect &lt;em>nothing&lt;/em> out of life. Hell, I am always amazed how you suddenly need to have a plan, when you move from 9th to 10th, or enter your final year…..and nobody expects anything from you in your 9th grade or junior year, but no, the transition happens, you advance a grade/year, and suddenly everyone expects you to have a plan, even though you are literally the same person you were a few months back, still as clueless, or maybe even more 🙈&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/the-man-without-a-plan/image-2.jpg"
alt="Imagine there’s no heaven, It’s easy if you try, No hell below us, Above us, only sky, Imagine all the people Livin’ for today">&lt;figcaption>
&lt;p>Imagine there’s no heaven, It’s easy if you try, No hell below us, Above us, only sky, Imagine all the people Livin’ for today&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>I have entered my final year at college, and I have no clue what to do with my life. Somedays I want to be an AI researcher, others an entrepreneur, somedays I want to maybe do an MBA and yet another day I just want a comfy job, some days back I was interested in robotics, computer graphics, finance, psychology, graphic designing and there was a time when I was absolutely sure I wanted to work on making a self-driving car, and oh currently I have got this newfound fascination with the whole blockchain-crypto finance world which absolutely doesn’t fit into &lt;em>any&lt;/em> of my possible future plans, but I still ended up spending a lot of time the last month digging into it. TLDR, I DONT HAVE A PLAN&lt;/p>
&lt;p>And here’s the catch, I don&amp;rsquo;t think I &lt;em>need&lt;/em> a plan. I might just want to let things roll out on their own, and just try to have fun along the way, doing whatever the hell I want to do, and accepting wherever that leads me to. Cause I am just a kid right now, I just can’t dictate my life following some kind of “plan” to achieve some kind of stability that somehow society wants you to achieve after college. Sure some of my friends have a plan and are very focused about it, make dedicated efforts to achieve that, and I am sure having a plan would help them a lot to achieve whatever they want, and I often get anxious seeing them having it all figured out, ahem, maybe they also don’t have shit figured out, and it just appears so, but at least they say they are sure they want to do a PhD or something, would give anything for even that little bit of clarity 🙈. But okay, I have turned out just fine till now, allowing fate to take the reins for me, and I’ll probably end up fine now as well….&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/the-man-without-a-plan/image-3_hu847666928338101790.webp 400w,
/blog/the-man-without-a-plan/image-3_hu12845971699413036702.webp 760w,
/blog/the-man-without-a-plan/image-3_hu13064370251363461792.webp 1200w"
src="https://ayushtues.github.io/blog/the-man-without-a-plan/image-3_hu847666928338101790.webp"
width="537"
height="760"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>Anyway, this turned out more to be a therapeutical rant kind of thingy to calm down my final year entering self, rather than some kind of blog, which is totally fine, cause after all, what’s wrong about being “the man without a plan”.&lt;/p>
&lt;p>That’s all for this piece I guess, if someone even reads this stuff anymore, feel free to ping me up, let&amp;rsquo;s catch up, surely we’ll both have a lot to talk about in this mad year 😛&lt;/p>
&lt;p>Until the next one, Stay Safe!&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/the-man-without-a-plan/image-4_hu13082928053486643770.webp 400w,
/blog/the-man-without-a-plan/image-4_hu7131966780386037929.webp 760w,
/blog/the-man-without-a-plan/image-4_hu13558755836345130327.webp 1200w"
src="https://ayushtues.github.io/blog/the-man-without-a-plan/image-4_hu13082928053486643770.webp"
width="760"
height="507"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/the-man-with-out-a-plan-1791e03be39a" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>How the law of diminishing marginal utility messes with our happiness</title><link>https://ayushtues.github.io/blog/how-the-law-of-diminishing-marginal-utility-messes-with-our-happiness/</link><pubDate>Mon, 17 May 2021 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/how-the-law-of-diminishing-marginal-utility-messes-with-our-happiness/</guid><description>&lt;p>No, no, this is &lt;em>totally&lt;/em> not a post I thought about in my Econ 101 class, although I should really study the basics of Economics again; hmm, maybe I will, maybe after I finish writing this piece up, that’s what I’ll do.&lt;/p>
&lt;p>Now while I am actually a CS student, at least that’s what my degree would say, or my obituary ( in 2021….you never know ), people who interact with me regularly know that I love throwing financial or economic terms here and there in my conversations cause idk some of the theories in these subjects are really widely applicable in our day to day lives, and also because it makes me sound more intelligent and also partially because I have &lt;em>some&lt;/em> research background in finance. Oooooo btw talking about research, check out this cool work at the intersection of AI and economics by Salesforce in this awesome blog post &lt;a href="https://blog.einstein.ai/the-ai-economist/" target="_blank" rel="noopener">here&lt;/a> ( or if you are feeling up to it, then the paper &lt;a href="https://arxiv.org/abs/2004.13332" target="_blank" rel="noopener">here&lt;/a> ).&lt;/p>
&lt;p>Anyway, back to the blog post, so let’s first talk about how the blog post came out to be, cause that itself includes a good example of the Law of Diminishing Marginal Utility.&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/how-the-law-of-diminishing-marginal-utility-messes-with-our-happiness/image-1.png"
alt="When you reach the point of zero marginal utility, but you are still happy :P">&lt;figcaption>
&lt;p>When you reach the point of zero marginal utility, but you are still happy :P&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>So I got my &lt;em>third&lt;/em> research paper accepted a few days back, again in one of the top conferences in the world, and well, honestly, I couldn’t care less about it. I even thought a lot before bothering to tell anyone about it cause it just didn’t seem much of a big deal. Contrast this to my first paper a year back, where I was just jumping around with excitement and couldn’t stop myself from telling the world about it,. I even publicised it on social platforms and stuff ( contrast to now, when I don&amp;rsquo;t even have a social platform anymore :P ). And this got me thinking, why do I not feel the same excitement about this anymore? I mean, it is a big deal; I put a lot of effort into it, I should allow myself to feel happy about it, but instead, I didn’t really feel anything, and I forgot about it in like an hour or something and just moved on to the next thing. This reduction in excitement or gratification from getting more of something, after you have already had some of it before, is what is, in essence, the Law of diminishing marginal utility, and it seriously fucks up your happy moments more than you realise.&lt;/p>
&lt;p>Okay, that might not have been the best example, but well it’s my blog, I’ll rant about whatever I want. But yeah, let’s define the stuff formally know, so kids, welcome to Econ 101 with Deku sensei&lt;/p>
&lt;p>Straight off from &lt;a href="http://www.investopedia.com" target="_blank" rel="noopener">investopedia&lt;/a> ( my go to guide for anything related to finance ) the law is defined as :&lt;/p>
&lt;blockquote>
&lt;p>The law of diminishing marginal utility explains that as a person consumes an item or a product, the satisfaction or utility that they derive from the product wanes as they consume more and more of that product.&lt;/p>
&lt;/blockquote>
&lt;p>In other words, &lt;strong>less is more.&lt;/strong> The more of something you get, the more you get used to it, and the less satisfaction you derive from each additional unit of that thing.&lt;/p>
&lt;p>A few technical terms that might get you confused :&lt;/p>
&lt;ol>
&lt;li>&lt;strong>Utility&lt;/strong> — Just a fancy name for the amount of satisfaction you get from something&lt;/li>
&lt;li>&lt;strong>Marginal utility&lt;/strong> — The amount of utility you get from one additional unit of something&lt;/li>
&lt;/ol>
&lt;p>A good example of this ( that I again found online ) is this : Have you ever gone to Domino’s and ordered lot’s of pizzas, probably more than you can eat. When you eat the first few slices, you get a lot of satisfaction, when you eat the next few slices, you get a little bit full, and the satisfaction you had from each slice decreases. Then when after a few slices, you get completely full, it doesn’t matter how many more slices you have remaining, you simply don’t want them, but well, you did pay for them, and then they go to waste, and that results in a negative utility.&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/how-the-law-of-diminishing-marginal-utility-messes-with-our-happiness/image-2.gif"
alt="A nice little example of the law ( and it also hints at why it makes things worse for us due to the anticipation )">&lt;figcaption>
&lt;p>A nice little example of the law ( and it also hints at why it makes things worse for us due to the anticipation )&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Okay you might be saying, I get what the law is, but just like the laws of physics, why do I care, it’s just another law ( hmm, this is interesting, just think about it, how important are the laws of physics, and how little do you care about them, or maybe there are just a too many of laws, for you to care about ?).&lt;/p>
&lt;p>Well, you might care about this if you want to remain being capable of feeling happiness ( or you might end up being numb to anything like someone like me, and then well, nothing matters ). Consider this, if the amount of satisfaction you get from each individual unit decreases, but you still want to maintain the amount of happiness you had in the past, what can you mathematically do? You will simply have to ramp up the amount of stuff you consume, and that is the essence 21&amp;rsquo;st century consumerism.&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/how-the-law-of-diminishing-marginal-utility-messes-with-our-happiness/image-3.jpg"
alt="This should be named the marginal utility meme, ooooo also just consider how marginal utility fuels your obsession with new memes.">&lt;figcaption>
&lt;p>This should be named the marginal utility meme, ooooo also just consider how marginal utility fuels your obsession with new memes.&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Everyone just throws stuff at you online or gives you these small happiness hits by giving you instant gratification, and then the marginal utility of stuff decreases for you, no wonder you feel that nothing interests you anymore, your marginal utility has decreased to the point that &lt;strong>the only way you can maintain the previous level of happiness is by consuming more and more of the same stuff, &lt;em>or&lt;/em> byfinding new products to consume&lt;/strong> ( which still have high marginal utility ), no wonder we change mobiles, items, or even relationships as frequently as we do, it’s all about the marginal utility you derive from these things. It is what drives the supply-demand dynamics of the financial markets, it’s just everywhere.&lt;/p>
&lt;p>So okay, I think I made my point clear why it’s fucking you up. But what to do about it. Well, honestly I have no idea. Personally the fact that I don’t get satisfaction from pumping out research papers anymore, is kind of a first-world problem, which is not something I should really care about ( or write a blog on ), cause the fact remains that the reason you are experiencing low marginal utility, is that you are already at a very high level of utility, you are already killing it. So maybe you can stop complaining about not being able to feel happy about stuff anymore.&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/how-the-law-of-diminishing-marginal-utility-messes-with-our-happiness/image-4.jpg"
alt="Huh, &amp;lt; = &amp;gt; , that’s a pretty cool way of writing it isn’t it ?">&lt;figcaption>
&lt;p>Huh, &amp;lt; = &amp;gt; , that’s a pretty cool way of writing it isn’t it ?&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>But maybe, you can also try to consume less? Like maybe focus on quality over quantity, like maybe I should focus on working super hard on one extra-ordinary piece of research work, instead of producing a lot of not so ground breaking papers. Yeah maybe that should help, I don’t know. One thing that can totally be done is reducing your online content consumption, although that’s really hard in this pandemic ( believe me, I have tried….a lot). Idk, they say you should appreciate the small things in your life, maybe that will also help, maybe try looking more into finding happiness in the everyday occurrences of life, I am not sure, I struggle with the problem a lot myself. Maybe write about it ? I’ll have to see in the coming days if that helps xD&lt;/p>
&lt;p>I think that’s all I want to write in this one, it turned out to be more of a rant than a blog post, but well whatever, it’s my blog. As always, to anyone reading this, if you are someone I know, and we haven’t talked for a while, feel free to ping me up, I am dying from boredom, and I am sure we all can do more of conversations with friends in these times ( until, their marginal utility goes down as well, although I hope it doesn’t ). Anyway, until the next one&lt;/p>
&lt;p>Stay Safe!&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/how-the-law-of-diminishing-marginal-utility-messes-with-our-happiness-cf0ec24f8f31" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>The DFS vs BFS way of living your life</title><link>https://ayushtues.github.io/blog/the-dfs-vs-bfs-way-of-living-your-life/</link><pubDate>Sun, 18 Apr 2021 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/the-dfs-vs-bfs-way-of-living-your-life/</guid><description>&lt;p>Framing the problem in terms of search algorithms is probably not the best way to put it, and totally makes me sound like a computer geek ( which I proudly am ), but it’s the way I discuss this with one of my friends, and so that’s that. You may have also heard about it as the specialist vs generalist problem. You can explore life in a DFS fashion, diving deep into a single/few fields and becoming a specialist or follow BFS, and explore the breadth of life, and become a generalist who has a broad skill set, although not being extremely talented in any.&lt;/p>
&lt;blockquote>
&lt;p>I fear not the man who has practiced 10,000 kicks once, but I fear the man who &lt;br>
has practiced one kick 10,000 times.&lt;br>
 — Bruce Lee ( ah and what if life puts you in a boxing match, will you kick your way out of it?)&lt;/p>
&lt;/blockquote>
&lt;p>You might have heard the above quote, and well it kind of represents the 21st centuries’ infatuation with specialists. We are moulded into being specialists from our childhood. Don’t take risks, don’t explore much, do what you are good at, become better at it, become a specialist in it, nobody will value you if you’re mediocre etc. And don’t get me wrong there’s no problem in being a specialist, in fact, I am probably one as well ( though I am working to change that bit by bit, still, I have a very limited skill set as of now). We put people in a specialised major in colleges, basically telling them to restrict their curiosity to a particular field, and then again we stick to a particular subfield based on our interests and well the job market.&lt;/p>
&lt;p>However, I don’t know why, but I am fascinated by generalists, I find people who are able to do a wide variety of things extremely awesome, and honestly, I would judge someone’s “intelligence” based on the breadth of things they know, rather than the depth. And well it’s not really a fair thing to do, but well whatever, we all have our biases. Given how complex reality is, and how often it pushes you into unprecedented circumstances *cough* COVID *cough*, I don’t really see the sense in becoming a specialist and making yourself fragile to the changing whims of reality. But okay, being a generalist is hard, it comes with an amount of uncertainty which not everyone can be comfortable with. We are inherently not well suited for multi-tasking, and can only focus well on a single task at a time, and so handling too many diverse stuff together might be a recipe for a disaster ( leading to idioms like, don’t chase two rabbits at once ).&lt;/p>
&lt;p>Now I shouldn’t exactly be writing this stuff, since I am closely associated with some academics, but my problem with specialists is also what makes me not like academia much now. I just feel that spending all your time working in a small niche field and becoming a specialist in it is kind of wasteful since life has so many interesting problems to offer. Just the idea of working on a single thing makes me suffocate, and I have no idea how people pull it off. I have a track record of leaving things the moment I get bored of them, so maybe it’s just the way I am. And well I get it, there is a lot of stuff that needs a specialist, complex stuff where you need someone who knows the ins and outs of everything required to get shit done, you need researchers to move any field forward, and it’s not an easy or trivial task, but well okay, it just doesn’t feel like the best way to live your life to me.&lt;/p>
&lt;p>But well, just like there is no one single optimal search algorithm ( No, A* isn’t optimal, well, it is optimal in a sense, but the better question to ask is &lt;em>what&lt;/em> is it optimal on, a good discussion on this can be found &lt;a href="https://www.youtube.com/watch?v=adrVlZegiR0" target="_blank" rel="noopener">here&lt;/a>), there’s no one optimal way to search for your purpose in life. Maybe you prefer DFS, or maybe you want to do a BFS, or maybe you want to do some crazy hybrid shit, it’s up to you, cause hey, even researcher’s are trying to figure out what’s the best way to search, and this is life you are talking about, there are no experts here :P&lt;/p>
&lt;p>Anyway, that’s it for this post, I don’t know why, but the 2nd wave of COVID scares me waaaaaaay more than the first one, so dear reader, until the next one&lt;/p>
&lt;p>— Stay Safe!&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/the-dfs-vs-bfs-way-of-living-your-life-a57dad923525" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>The one due to my insomnia</title><link>https://ayushtues.github.io/blog/the-one-due-to-my-insomnia/</link><pubDate>Wed, 31 Mar 2021 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/the-one-due-to-my-insomnia/</guid><description>&lt;figure>&lt;img src="https://ayushtues.github.io/blog/the-one-due-to-my-insomnia/image-1.jpg"
alt="It’s really all white noise inside my head right now lol ( Source: https://www.reddit.com/r/Art/comments/7drqcz/insomnia_digital_4000x4000/)">&lt;figcaption>
&lt;p>It’s really all white noise inside my head right now lol ( Source: &lt;a href="https://www.reddit.com/r/Art/comments/7drqcz/insomnia_digital_4000x4000/" target="_blank" rel="noopener">https://www.reddit.com/r/Art/comments/7drqcz/insomnia_digital_4000x4000/&lt;/a>)&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>So here I am back again on my laptop, with a foggy brain after a futile attempt at sleeping, maybe writing shit would make it easier for me to calm down and sleep.&lt;/p>
&lt;p>Sleeping has always been a hard thing for me, due to different reasons over time though. During my childhood, I was very scared of the dark, ( I wasn’t exactly a scaredy-cat generally, playing downright gory video-games all day ), but there was something about the silent darkness that comes with the night that used to scare the shit out of me. And however embarrassing it might be to say, I used to be shit scared of ghosts during the night 🙈 I used to have a lot of hallucinations I guess during the night, seeing weird figures and stuff, which were obviously not present, and usually were the result of my overactive imagination finding patterns in the shadows. Although I did use to see things moving sometimes ( I distinctly remember a photo frame in my room which I used to find shaking all the time ), I am pretty sure that was just my brain at work again. It’s probably the reason why I don’t watch horror movies to date, and there were these shows like “Aahat” and stuff popular during those days, which played quite an important role in my phobia lol. Anyway, I used to be so scared that I couldn’t even cross the rooftop during the night, but okay, that’s enough embarrassing details for one blog.&lt;/p>
&lt;p>So yeah the reason why I couldn’t sleep in childhood was me being scared of the dark, it would often want me to wish that somehow the night would just pass by and I would sometimes stay awake all night, hoping for the day to come by ( also we used to wake up at 5:30 for school those days, so well, that’s that ). Anyway as I grew a bit older, and my phobia of the dark vanished completely and instead got converted into a love of the dark and silence. I love dark rooms now and often used to roam around the college campus late at night with my friends to enjoy the calm environment of the night ( which was usually spoiled with all the couples roaming around during the night, but ah well ).&lt;/p>
&lt;p>But it came with its own problems. Now it turns out that I am a bit too sensitive to the dark and silence. I can only sleep when there’s absolutely little to no light in the room and complete silence. This also makes sleeping during the day almost an impossible task for me, which used to present a lot of problems during JEE, where it would be necessary to sleep after being tired from attending coaching for the entire day, but I couldn’t really sleep and had to continue studying half tired ( I still can’t sleep during the day, and have to work half tired if I get tired during the day ). It is probably also the reason why I sleep late at night, and also the reason why I couldn&amp;rsquo;t sleep today because there was way too much light in my room.&lt;/p>
&lt;p>Anyway, it’s weird to think how some people can struggle so much with something as simple as sleeping, while it’s the most natural thing to do for other people. It kind of makes you realize how much we take for granted in life. I am so jealous of people who can sleep anytime anywhere. Maybe we should all be a little more aware of the subtle privileges we all enjoy, especially in this weird year, when so many people lost so many things, that pity problems ( like not being able to sleep 🙈 ) are really not that important, and we should just suck it up lol.&lt;/p>
&lt;p>Also, it’s interesting to think how our relations with things change as we grow up, when I was a child, I used to hate to sleep, but now, I would not give up my sleep for anything. It’s THE best part of the day, and I can’t get enough of it ( literally ). Our priorities change so fast, that we end up spending a lot of time worrying in the present about things, that might not be at all important to us in the future, and might ignore things, that may become very important to us in the future ( like developing a healthy sleeping routine ). It’s like the fact that we always wanted to grow up quickly as a child, but now growing up is the last thing I want to happen lol.&lt;/p>
&lt;p>Anyway, I think that’s enough random late-night bickering for the day, it would be 2 soon, and it’s definitely not a good choice to keep on writing until then, cause as someone said “Nothing good happens after 2 AM”. So well, I should probably give sleeping another shot, although I don’t really have high hopes for it, but well anyway, until the next one —&lt;/p>
&lt;p>Stay safe!&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/the-one-due-to-my-insomnia-41339accbf2b" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>How often do we try to actually understand stuff?</title><link>https://ayushtues.github.io/blog/how-often-do-we-try-to-actually-understand-stuff/</link><pubDate>Tue, 30 Mar 2021 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/how-often-do-we-try-to-actually-understand-stuff/</guid><description>&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/how-often-do-we-try-to-actually-understand-stuff/image-1_hu13567721340546720858.webp 400w,
/blog/how-often-do-we-try-to-actually-understand-stuff/image-1_hu16920313422613684644.webp 760w,
/blog/how-often-do-we-try-to-actually-understand-stuff/image-1_hu33606692558673108.webp 1200w"
src="https://ayushtues.github.io/blog/how-often-do-we-try-to-actually-understand-stuff/image-1_hu13567721340546720858.webp"
width="760"
height="760"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>So a few days back I finished reading “Surely you are joking, Mr Feynman!”, the biography of Nobel prize-winning physicist Richard Feynman, best known for his contribution to theoretical quantum physics ( and less known for a whole lot of other stuff, like helping to create the atom bomb, having the patent for a lot of nuclear stuff, being an awesome painter, musician, an expert safe-lock picker and totally a ladies man xD ) and the author of what can be called the most popular physics book: The Feynman Lectures on Physics.&lt;/p>
&lt;p>Now I have always been a lover of physics, or so I would like to say at least. It was my favourite subject in school, JEE, and it is probably because I used to spend a lot of time trying to go deep and understand stuff in physics during JEE, that I ended up becoming a research enthusiast in college ( although, not in physics though lol ). Anyway, since I had not really looked much into physics since college started, I thought it would be a good idea to read the Feynman book. I began reading the book with the expectation that it would have a lot of cool physics stuff, much like Hawking’s “A brief history of time” ( another book pretty high on my reading list ). But well, I couldn’t have been more wrong lol, there was very little physics in the book, with a few real-life examples here and there, but most of the book consisted of Feynman describing his day to day experiences ( actually the book I should have read was the Lectures book, but whatever lol). And I’ll be honest, I don’t know many genius people in history, but reading the book, I legit felt that Feynman was one of the smartest people I have ever read about ( kind of a similar feeling to reading Elon Musks’s bio ).&lt;/p>
&lt;p>But I am not going to write the post about how I found him extremely intelligent ( that probably deserves a separate post ), but reading the book made me realise a thing I had forgotten about life.&lt;/p>
&lt;p>Like any good physicist, Feynman was a follower of looking at things from “First Principles”. He always tried to &lt;em>actually&lt;/em> understand stuff, way down from the basics. He didn’t try to skip any level of complexity as being “obvious”, and didn’t try to hide stuff in the name of using complex terms. All the examples he mentioned in the book were very very easy to follow, and you could actually understand what he was talking about, which is not a simple feat since he’s a guy of the likes of Newton and Einstein and easily one of the top-10 physicists of all time.&lt;/p>
&lt;p>And it got me thinking, ever since entering into college, or even during my childhood, how often have I tried to &lt;em>actually&lt;/em> understand things? The “getting shit done” attitude, often trades understanding for efficiency. And just like this tradeoff leads to the popular &lt;em>technical debt&lt;/em> in tech, I think this habit has also lead to an &lt;em>understanding debt&lt;/em> in my overall knowledge about stuff. Like I very often just understand stuff from a very bird’s eye view, and make myself believe that I know it, while in reality, my knowledge about it is full of holes.&lt;/p>
&lt;p>And society promotes this kind of behaviour, from childhood we are conditioned to not ask too many questions. You are made to feel dumb for asking doubts, and our textbooks hardly explain stuff in detail, giving incomplete analogies about stuff( the book had a section about Feynman lamenting about the time when he was made to review the school textbooks of US ). A good example of this in the book was “energy”.&lt;/p>
&lt;p>The textbooks introduced energy as — Energy keeps the wound-up toy moving, energy keeps the fan moving, energy allows you to pick up a stone. But isn’t this dumb? It’s just a definition of a word, I can replace Energy by X here and it doesn’t matter. I can say “LoremIpusum keeps the wound-up toy moving, LoremIpsum keeps the fan moving”. How is this actually supposed to teach you anything about energy?&lt;/p>
&lt;p>More so in the tech world, there goes the saying that if stuff works, then it works. Don’t try to change it. People don’t understand why their own code is working half of the time. Nobody really understands how AI works anymore, not even the guys creating it. Is society really heading towards one where we use stuff &lt;em>just because it works?&lt;/em> In college we are supposed to cover so many topics at once, that we hardly have time to actually understand any topic in detail, no wonder the research culture in India sucks.&lt;/p>
&lt;p>Why are we so hell-bent on productivity? Why can’t we slow down a bit and actually try to understand stuff. Is there a reason why most of us feel like imposters? Is there a reason why we are always scared that people might find out that we don’t understand shit about what we are talking about. Is it okay for society to tell people that it’s OK that they don’t know what they are talking about? Is it ok to “move fast and break stuff”, instead of settling down and making stuff from small building blocks? Why do I have to feel guilty about trying to spend my time to understand the world in deep, and not spend my time “productively” on building flashy apps and products?&lt;/p>
&lt;p>Well, that’s a lot of why’s, and a lot to understand about. But well, will you actually try to understand this, or just skip over? I am not sure about even myself, let anyone else, but well whatever, until the next one —&lt;/p>
&lt;p>Stay Safe!&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/how-often-do-we-try-to-actually-understand-stuff/image-2_hu1077102406100028474.webp 400w,
/blog/how-often-do-we-try-to-actually-understand-stuff/image-2_hu13591596462033225112.webp 760w,
/blog/how-often-do-we-try-to-actually-understand-stuff/image-2_hu1292749330699234213.webp 1200w"
src="https://ayushtues.github.io/blog/how-often-do-we-try-to-actually-understand-stuff/image-2_hu1077102406100028474.webp"
width="760"
height="760"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/how-often-do-we-try-to-actually-understand-stuff-170733c7436a" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>An honest review of (metaphorically) living under a rock</title><link>https://ayushtues.github.io/blog/an-honest-review-of-metaphorically-living-under-a-rock/</link><pubDate>Sun, 28 Mar 2021 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/an-honest-review-of-metaphorically-living-under-a-rock/</guid><description>&lt;blockquote>
&lt;p>What does living under a rock feels like in today’s hyperconnected age?&lt;/p>
&lt;/blockquote>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/an-honest-review-of-metaphorically-living-under-a-rock/image-1.jpg"
alt="Yes, Yes I do! ( Source: https://www.deviantart.com/aelthwyn/art/Yes-I-Live-Under-a-Rock-549899331)">&lt;figcaption>
&lt;p>Yes, Yes I do! ( Source: &lt;a href="https://www.deviantart.com/aelthwyn/art/Yes-I-Live-Under-a-Rock-549899331" target="_blank" rel="noopener">https://www.deviantart.com/aelthwyn/art/Yes-I-Live-Under-a-Rock-549899331&lt;/a>)&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>For the better or the worse it seems like social media has become an integral part of our lives, more so than ever since the pandemic began, and it became kind of the only way we could interact with society. It has become such a central part of our day to day interaction with people, that it is frightening to think about living without it, isn’t it? Well, maybe it isn’t lol&lt;/p>
&lt;p>Now while I can’t claim to be completely free of using social media myself, I don’t really use &lt;em>much&lt;/em> of it. I only use WhatsApp and Slack ( does Medium also count? ), with WhatsApp being the place for personal conversations with close friends and family and people that I &lt;em>actually&lt;/em> know, and Slack for professional or semi-professional conversations with some groups I am associated with. Given the digital era we live in, and the pandemic, I find it very difficult to get away with these two. I have been slowly reducing my social media consumption since 2019, and have been out of mainstream apps like Facebook, Insta, Twitter, Reddit, Snapchat etc since near the end of 2019, with the most recent reduction being quitting LinkedIn.&lt;/p>
&lt;p>Anyway, given all this, I thought it might be interesting to comment on how is my experience “living under the rock” since people always give funny reactions when I mention that I don’t use social media, and how most people think it’s impossible to live without it. I’ll be completely honest, and it’s not my intention to convince anybody to quit social media or stuff, it’s a personal choice ( although, I would like to remind people, that it’s still a &lt;em>choice&lt;/em> and not a &lt;em>necessity&lt;/em> )&lt;/p>
&lt;p>One of the first responses I usually get is about how I make connections without social media, how do I make friends and get to meet new people? Well honestly speaking, it’s hard alright. I have very few connections compared to “normal” standards, and yes not being on social media is one of the major causes of that. I am a college student, in my 3rd year, and I hardly know any of my juniors or seniors or hardly have any friends outside a small circle, which does make me a bit sad from time to time and is usually the only reason that makes me sometimes reconsider my choice, since I am conversationalist, who loves talking to people more than anything.&lt;/p>
&lt;p>That being said, however, I feel that the few connections I do have are usually sufficient for me. In the current hyperconnected world, we are used to thinking that the more connections you have, the better it is, and maybe it is, but I don’t really have enough mental energy to maintain a large number of connections and so I like the idea of having a few good connections. I have a very very weak memory and with very few connections, I find it easy to remember stuff about people and have meaningful conversations with them, and just really &lt;em>remember them.&lt;/em> I might have to trade off being &lt;em>known&lt;/em> by many people, but I am fine with that and instead try to be &lt;em>loved&lt;/em> by a few….&lt;/p>
&lt;p>Now when I say the above to people, they come up with another counterpoint, what do I do about professional networking? Don’t I miss professional opportunities from being absent on social media? Well this one, I am sure I don’t mind at all. Now it is true that platforms like LinkedIn do make it easier for you to keep updated about the professional world, and build up “connections”, I have used these platforms for a while, and honestly, I never found any actually good opportunity there. Most people on these platforms are too busy pushing themselves forward, to be of any help to you. And since it’s open for all, it’s very hard for you to be notices amongst all the other people anyway. Also, there’s the fact that if you are &lt;em>really&lt;/em> good at something, people will automatically find you, and you’ll end up making a network anyway.&lt;/p>
&lt;p>Now another problem that occurs is that of well “living under the rock”. Ever since I have quit social media, I am no longer able to keep updated with the latest memes and social media fads, and this often makes for some very weird moments in conversations, when I end up having no idea about what the other person is referring to, and then the painful process of explaining to them why I don’t know it, followed by some low-key judging and then them having to explain the fad to me.&lt;/p>
&lt;p>Honestly speaking I consider this a positive point really, the internet is full of so much shit content, that the less you trash you feed to your mind, the better it is. Most people are so trapped by the constant stream of content being fed to them, that they don’t even realize that they are trapped. It is when you try to reduce the consumption, that you realize how addicted you are to content. Your brain gets used to abnormally high dopamine levels, gained from those byte-sized awesome content, and when you remove that input, your brain still wants to maintain those high dopamine levels and you feel abnormally bored from normal day to day stuff, and rush to your phone, the first moment you can get your hand on it. I have started to become very conscious of what and how much content I consume, ever since I reduced my social media consumption, and actually, you know what is better than consuming content, &lt;em>creating content.&lt;/em>&lt;/p>
&lt;p>Talking about happiness levels, I strongly believe that social media causes depression, and your mental wellbeing and happiness levels are way better without it. 5 mins of success stories on LinkedIn are enough to depress me for the entire day, a recent enough example of this is that since our college campuses have reopened, and some lucky students are able to get back, social media feeds are full of pictures by these guys, flaunting their return, and making guys like me who are still at their house feel really really sad. It is good that I am not present on social media, I wasn’t even aware of all this until some of my friends told me about it, and I was glad that I didn’t know about it lol.&lt;/p>
&lt;p>Quitting social media allows you to escape the constant surveillance and judgement of society, and gives you the power to do whatever the hell you want. You can do anything, and no one would judge you, because people are busy judging people online lol. You can work hard in silence, in peace, and hone your skills. Social media really takes up a lot of our free time, and you won’t realize how much time you spend on it until you actually cut yourself off. I have so much time on my hand without social media that I don’t what to do with it lol. I am sure I wouldn’t have been able to spend time learning so many new things, and reading so many books, if I had kept on scrolling through my social media feed, watching memes, and short videos, battling hate speech and whatnot. And it’s not as if these things are wrong in moderation, but I do think they are wrong in excess, and well, with infinite scrolling, usually stuff is in excess ( nowadays I run away from an app whenever I see infinite scrolling in it 🙈)&lt;/p>
&lt;p>Personally quitting social media has made me more “social”, I know this is a bit contradictory, but I find myself being more responsive to people’s emotions ever since I quit social media. Social media makes it so easy to bitch about people, that people get used to trolling others, and there is so much hate on it, that it makes me sad. So I like to be away from all of it, and be kind to people around me, and not indulge in useless bitching and rumour spreading and stuff, cause honestly I don’t have enough mental energy for all that. Being aware of the fact that I have limited mental energy in a day, is one of the main reasons I avoid social media because it really drains you up pretty quick. Instead, I try to use that extra time to replenish that mental energy by doing stuff like taking a power nap ( I love power naps ) or reading a book or a research paper, or talking to a close friend, or even my family members ( I have started talking more to my family than ever since I reduced my social media consumption ).&lt;/p>
&lt;p>All in all, I do personally think that it&amp;rsquo;s possible to survive living under a rock even in 2021. Social media is a choice and not a rule, and it should really be your personal choice if you want to use it. My only aim was to provide a perspective about the fact that it is possible to live without it, and most of the doubts we have while quitting are really not all that important actually. Now I don’t aim to convince you to quit it right away, but maybe just try to be a bit more aware of how it affects you, and you’ll probably figure out the balance that works the best for you anyway.&lt;/p>
&lt;p>Anyway, I guess that’s it for the post, I realised that I always end up my posts saying that if we haven’t talked in a while, then feel free to hit me up, but as it so happens, I can be a bit hard to find, given that I am not on social media, but well, maybe that’s the way it should be, anyway I don’t think anyone reads this stuff anymore anyway lol.&lt;/p>
&lt;p>So well, that’s that, until the next one — Stay Safe!&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/an-honest-review-of-metaphorically-living-under-a-rock-825fa05a2bf5" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>How brittle is your happiness?</title><link>https://ayushtues.github.io/blog/how-brittle-is-your-happiness/</link><pubDate>Sat, 20 Mar 2021 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/how-brittle-is-your-happiness/</guid><description>&lt;blockquote>
&lt;p>What makes you happy? What motivates you to get out of bed every day? What do you associate your self worth to? What really matters at the end of the day?&lt;/p>
&lt;/blockquote>
&lt;p>I think these are important questions to ponder about, it saves a lot of time to introspect a bit and figure out what are you &lt;em>actually&lt;/em> doing stuff for. Cause if you think about it, sometimes you are doing things for no reason at all, sometimes you do shit just for the sake of doing shit, and that gets pretty exhausting man, and eventually leads to a burnout after which you start questioning, &lt;em>“Why am I doing all this?”.&lt;/em> Why not save yourself all this trouble, and focus on things that actually make you happy.&lt;/p>
&lt;p>And although it’s a bit embarrassing to admit, like many people, I am guilty of attaching my self-worth/happiness to achievements. Whether it be getting straight A’s in school, or getting that awesome internship or research publication in college, it seems like I was always looking for more and more achievements to make myself happy. I don’t really even know why do I chase behind stuff like this, it’s just something I think I mindlessly used to do, it felt like the right thing to do, cause hey, isn’t it success that’s supposed to make you happy? Shouldn’t everyone try to become better and better and hustle and climb up the social ladder?&lt;/p>
&lt;p>Well, I guess, you have to give one thing to the last year, however shitty it was to stay isolated in a room for so long, it did make me introspect more and figure out what things actually make me happy, and how brittle and fragile was attaching my self worth to achievements and success. I have started to realise that achievements don’t make you happy, at least not for long, they do feel a bit good for like 1–2 hrs, but then you get normal, and then you start looking for the next potential achievement to make you happy, it’s almost like an addiction really, you enjoy the “hit” that lasts for a while, and you are never satisfied and want more and more of it. It’s a vicious cycle, that knows no end, and gets worse day by day.&lt;/p>
&lt;p>Man, how dumb was I to achieve my self-worth to my school grades, when they are so random and subject to the whims of people and not reflective of anything really. It doesn&amp;rsquo;t take much for fate to roll a twisted set of events, leading you to get a fail grade, and then what, would you go into depression and shit? Does your self worth decrease since you failed a subject and you let down people, really? Is that all is to your being as a human? What about the fact that you are a kind person, what about all the times when you helped people or made people smile, doesn’t that count as something to be proud of? Why do we associate value with the wrong things and make life miserable for us? Don’t you feel infinitely happier when you complete that personal project, which you do for yourself, for your own curiosity and not for validation from anyone else. Don’t you feel happier when you have a meaningful conversation with a friend after a long time? Don’t you feel happier when you slowly let life pass by, and just live?&lt;/p>
&lt;p>I can’t express how good it feels when you shift your self worth from flimsy achievements to the thing that actually matters, YOU! You start making decisions, not for the end achievement, not for pleasing anyone else, but after considering how the &lt;em>process&lt;/em> would make you feel. You allow yourself to relax, and work on side-hustles ( like this blog ) that give meaning to your life, and even though this might lead you to a less “successful” life, in whatever terms society defines it, but you sure as hell end up living a more &lt;em>happy&lt;/em> life.&lt;/p>
&lt;p>Anyway, it’s not as if I have everything figured out yet, I still find myself chasing after the wrong things ( and the wrong people ), but I do think that things are getting a bit better, and maybe this is a part of “growing up” and finding yourself and all that catchy words lol. I guess this is it for the post, and if anyone still reads my posts ( I heavily doubt it ), feel free to ping me if you would like to talk, cause I know for sure, that I would LOVE to!&lt;/p>
&lt;p>Until then, Stay Safe!&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/how-brittle-is-your-happiness-3ce982def1a5" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Let’s talk about randomness</title><link>https://ayushtues.github.io/blog/lets-talk-about-randomness/</link><pubDate>Fri, 12 Mar 2021 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/lets-talk-about-randomness/</guid><description>&lt;blockquote>
&lt;p>This is another one of my “casual” posts, I usually write well-drafted blogs, but they kind of felt artificial, and something which I wrote for seeking validation from others in a sense, and not to express myself, so I thought I should drop that format and just go informal lol.&lt;/p>
&lt;/blockquote>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/lets-talk-about-randomness/image-1_hu17738448152940413148.webp 400w,
/blog/lets-talk-about-randomness/image-1_hu15231724800724526646.webp 760w,
/blog/lets-talk-about-randomness/image-1_hu6367277247262804214.webp 1200w"
src="https://ayushtues.github.io/blog/lets-talk-about-randomness/image-1_hu17738448152940413148.webp"
width="760"
height="397"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>Is anything random? Or is the world deterministic? Well, even einstein was amused by this question, but let’s leave the correct answer to theoretical physicists ( or whoever does such type of research lol). Having an hour left before bed, and being bored out of hell from editing stuff for my research papers, I am here to edit even more stuff in my blog. ( Sometimes I think I should take up a job as a writer or editor somewhere since I spend more time writing/editing/reviewing stuff than actually doing CS stuff. )&lt;/p>
&lt;p>Anyway, since the amount of “thoughts about life” in my mind start growing exponentially as bedtime approaches ( with a threshold at 2 AM after which it’s just hell ), I started wondering about the effect of serendipity ( a word I misspelt twice since I starting writing this article) in life. It kind of feels to me that all-important things in life happen randomly and there is actually not much we can do about the stuff that really has an impact on our life. So why do we try so hard to control things? Why not let the fates control our lives, is it really too lazy and non-virtuous to do so? ( you know the ancient greeks used to place the three fate sisters Clotho, Lachesis, Atropos above the all-powerful 12 Olympians like Zeus, Poseidon, Hades, with even the gods being subject to the dictates of fate )&lt;/p>
&lt;p>Almost all good things ( and the bad things as well )in the last year I can think of happened randomly, well the entire year was a black-swan event anyway. I got connected to people almost randomly and out of nowhere ( btw, NEVER tell anyone that you just randomly texted them, if you haven’t talked in a while, you’ll get a nice backlash lol, people want to be felt important), all big professional achievements happened almost randomly, when I was not even trying hard, in fact, I have repeatedly observed that good things usually come to us when we are not looking for them, but that doesn’t make any sense right? If that’s the case, should we all relax and let fate play it’s cards and just deal with whatever hand we get? Is the concept of working hard to make our own destiny fundamentally flawed owing to the non-determinism inherent in the world?&lt;/p>
&lt;p>Although the email ID handed to me by my college administration literally means “unlucky”, I personally believe that I have the devil’s luck at my disposal. I often take weird steps without thinking twice, leaving it to blind luck to make stuff work out for me, which it more than often does. And I don’t want to say this in a kind of bragging fashion, but my friends do now how much I rely on sheer dumb luck ( although I have kind of hidden this fact at college to hide my imposter syndrome, I was infamous in my school and coaching for my “tukkas” lol ). In fact, most good things happen when I let things go on autopilot. So does that make me an imposter? Or does that make me someone using statistics to his advantage? Well who knows, maybe when things go bad, it makes me an imposter relying on his luck, when things go good, it makes me lucky lol or maybe that’s just a self-serving contract.&lt;/p>
&lt;p>Or maybe the randomness has something to do with the subconscious part of our brain. I have read somewhere that the best way to solve a problem is 1.) Do a lot of research about it 2.) Stop thinking about it completely, and try to do something as unrelated as possible 3.) Just wait for a eureka moment for the solution to automatically come to you ( and believe that this would happen ), and I think you would agree ( more than you would like to ), that this is how you get your best ideas, so maybe the creative process itself is random, or at least out of our conscious control?&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/lets-talk-about-randomness/image-2.jpg"
alt="1–2–3–4–5–6, hey I can be random too?">&lt;figcaption>
&lt;p>1–2–3–4–5–6, hey I can be random too?&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Then is the life we are currently living just a result of random processes? And if so, does that reduce it’s worth, since it’s just one of the possible things that could have happened, and life could have gone any other way. Or does it increase it’s worth since we should be grateful that this is the one that came true and not the one where a meteorite or a nuclear war wiped everyone out? I personally find it a bit relaxing to have the burden of my future being slightly reduced from my shoulders and leaving somethings to fate, cause let’s face it, the world is a bit too complex for us to plan about. It’s a weird culmination of various complex dependencies between shit loads of random processes, and we are all but a tiny tiny part of it, with little to no actual control over it, for all the cosmos care, we are just random noise, that fades away when averaged out across people, planets, solar systems, galaxies and maybe even universes?&lt;/p>
&lt;p>Then comes the all-powerful butterfly effect, where seemingly minor changes in the present, when compounded over time, lead to drastic changes in the future, while the effect is only observable in the past, like when you start considering that how that one seemingly mundane conversation changed your whole life, or how a guy’s taste for bat soup changed the world xD. If you take this effect seriously, then you can either be hyperconscious of all your choices, worrying that they’ll have a huge impact on your future, or just accept your inability to comprehend the subtle but powerful effects of compounding, and leave stuff to fate, and just be present.&lt;/p>
&lt;p>Anyway, whether we like it or not, randomness rules our life more than we would like to accept, and hard work isn’t the only thing that matters, although relying on luck is itself not a good way to live life, so maybe we should try to find a good middle ground between them. But okay, I think I should sleep now, I’ll publish this tomorrow I guess, making sure I didn’t babble something weird at night ( which I do a lot ). Like always feel free to reach out to me if we haven’t talked in a while, cause you never know, maybe we are just one step away from a random conversation that changes our life?&lt;/p>
&lt;p>Until then, Stay Safe!&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/lets-talk-about-randomness-ec07993aa146" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Why do we ( I )find doing nothing, harder than doing something?</title><link>https://ayushtues.github.io/blog/why-do-we-i-find-doing-nothing-harder-than-doing-something/</link><pubDate>Sun, 28 Feb 2021 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/why-do-we-i-find-doing-nothing-harder-than-doing-something/</guid><description>&lt;blockquote>
&lt;p>This is probably going to be a very short one, and it’s kind of going to be raw and abstract, with a lot of errors, basically, for all practical purposes, it&amp;rsquo;s just a rant lol, so feel free to skip over to something more indulging and interesting.&lt;/p>
&lt;/blockquote>
&lt;p>So this post has a lot to do with the fact that I am writing this post lol, cause it’s a Sunday, a day meant for people to relax and take it off, but here I am, crazily running experiments for my next research paper deadline ( again ), and you would think that while the experiments are running, I must be chilling and blowing off some steam, but hell no, I can&amp;rsquo;t do anything so sacrilegious to the religion of hyper-productivity that I supposedly became a part of in order to adjust with the “hustle” culture college grads oh-so-religiously follow, and here I am writing a blog instead.&lt;/p>
&lt;p>I have observed it as a fact, that I am unable to take off 10 mins and just do nothing during it, my brain starts rummaging about unnecessary things and I start feeling restless to do something, &lt;em>anything.&lt;/em> So I either check my mobile for new notifs, or start reading some random book or article, or do something else. Even when I am completely free, I don&amp;rsquo;t like being free, its almost disturbing for me to be free or relaxed, I want that dopamine hit that comes when you feel you are being productive, maybe it’s an addiction, just like other drugs. And I have observed that although I do a lot of things in this meaningless fashion, making my self appear busy, I actually get nothing done lol, so maybe it&amp;rsquo;s just a facade for my inability to actually get shit done, that I keep on doing shit.&lt;/p>
&lt;p>Now there are a lot of reasons for this, one of them being the rigorous JEE prep I had to go through, where basically you were assumed to be studying all the time, and that leads to me being unable to feel at peace until I was studying something, it has made me treat myself as a machine, context switching between multiple tasks to optimize its CPU utilization ( see, even in blogs I am unable to stop referring to CS stuff, what the hell is wrong with me, why can&amp;rsquo;t I just tune off shit from my head ). Another reason is the constant conditioning by the various social media apps, with their endless stream of notifications, making me check them again and again, and reducing my attention span to maybe 30 secs. Or maybe it’s just the way I am lol. But I think its more of a cultural problem, rather than my individual problem. For people, especially in the IT world, everything becomes about optimizing metrics, whether it might be work-related metrics, like KPI or stuff or unfortunately “metrics” for their personal life ( whatever they might think are metrics here ).&lt;/p>
&lt;p>Now sometimes I think what’s bad about this, after all, it just means I am being productive, isn’t that good? Well yes, and no, I think that if such a routine is followed for a long time, it eventually leads to a massive burnout ( I had one last year and it took me around 1.5 months of doing nothing and just reading novels to get out of it ) and also it leads to a lack of creativity. It is said that boredom is important because it is when you are bored, that you figure out what you really want to do, what inspires you to get out of that boring state, and unleash your creative juices. Also, it is not good for your mental health, I suffer from a lot of headaches, and always feel tired, which is weird, since I am not at all sleep-deprived lol, compared to the pre-covid college days, where I was definitely sleep-deprived, but a hell lot more energetic. Also, I guess such routines make you less focused on any particular task since you have a lot of stuff going on your mind ( although they do help you to find correlations among unrelated stuff, that might be hard to find otherwise ).&lt;/p>
&lt;p>I have been thinking about this a lot for quite some time, and have developed my own solutions for it. One of them is well, blogging, some people prefer journalling, but both help you to get stuff out of your brain. I have also disconnected myself from most social media and sometimes, often on Sundays, I completely switch off the internet for the evening time and instead just read books or stuff ( something that’s not possible today, due to a deadline). In college I used to leave my mobile in my room sometimes, just to break the notification hook (things like do not disturb, etc which can be easily reversed with a tap, never worked for me ). Music helps a lot as well, listening to calming music relaxes my brain and helps me gain some clarity. But all things said and done, I still struggle from it, and maybe it’ll take a really serious burnout to finally put my mind to rest.&lt;/p>
&lt;p>Anyway, I guess that’s all I wanted to say, for now, I could go on and on and on, but idk if that would make things any better, so well, until then, so long!&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/why-do-we-i-find-doing-nothing-harder-than-doing-something-2e90a809acc3" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>How I would like to experience music in the future?</title><link>https://ayushtues.github.io/blog/how-i-would-like-to-experience-music-in-the-future/</link><pubDate>Tue, 23 Feb 2021 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/how-i-would-like-to-experience-music-in-the-future/</guid><description>&lt;figure>&lt;img src="https://ayushtues.github.io/blog/how-i-would-like-to-experience-music-in-the-future/image-1.png"
alt="Source: https://in.pinterest.com/pin/5348093295633753/">&lt;figcaption>
&lt;p>Source: &lt;a href="https://in.pinterest.com/pin/5348093295633753/" target="_blank" rel="noopener">https://in.pinterest.com/pin/5348093295633753/&lt;/a>&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Although I have not been a big music fan from my childhood ( honestly I didn’t even listen to songs on my own before 12th). But after becoming an undergrad, music has become an integral part of my life. It’s one of the very few things that I can say I truly enjoy and Spotify feels like my own personal therapist ( more so after I started paying a yearly premium for it lol ).&lt;/p>
&lt;p>I usually use Spotify for music, depending on it and people in my close circle for recommendations. I also love listening to different versions of the same song ( covers ) on Youtube, sometimes I spend hours just listening to music late at night, and I never regret “wasting” that time, however much of a productivity freak I might seem to be.&lt;/p>
&lt;p>However, an interesting thought crossed my mind a few days back, “&lt;strong>How would I like to experience music in the future?”&lt;/strong> Now I don&amp;rsquo;t know the answer to the question or the fact that why do such questions come into my mind when I have tons of other questions to find answers to, but I thought it might be interesting to think about, and well, that’s pretty much the blog.&lt;/p>
&lt;p>Let’s discuss some of the ways I think we may be able to experience music in the future, I am allowing myself to think as wild as possible, and not think about the technological constraints, cause hey, it’s the future, who knows what might be possible :P&lt;/p>
&lt;h3 id="connection">Connection&lt;/h3>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/how-i-would-like-to-experience-music-in-the-future/image-2.jpg"
alt="Source — https://christianartistslive.com/concert-promotion-tips/">&lt;figcaption>
&lt;p>Source — &lt;a href="https://christianartistslive.com/concert-promotion-tips/" target="_blank" rel="noopener">https://christianartistslive.com/concert-promotion-tips/&lt;/a>&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>With the increasing rise of Augmented reality and Virtual reality, I think it is quite possible that we might be able to attend music concerts in a virtual space. I always feel sad when I watch concerts on youtube and think that I cannot attend the concert of an artist from another country, because I cannot simply be physically present there, in another country. But I hope in the future, artists might be able to hold virtual concerts, liberating us of these physical boundaries and enhancing one of the major purposes of music — “connecting people”. I recently made a friend out of a stranger, solely based on our common interests in music, and I heartily believe in the power of music to connect people.&lt;/p>
&lt;p>However, I am not excited about just looking at my laptop with my earphones attached and roaming around in a VR avatar in a VR concert, with the help of my keyboard. I want to be able to actually experience the physical experience of a concert with other people, virtually. I want to feel the soundwaves shaking the air around, and the sheer madness of the crowd’s energy, not some digital shit. Lol sounds crazy I know, but hey again, it’s the future. With the rise of haptic tech and AR/VR who knows what the future holds. Maybe the day will come when we can experience the virtual world physically and surely that will change the way we experience music, amongst other things.&lt;/p>
&lt;h3 id="customisation">Customisation&lt;/h3>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/how-i-would-like-to-experience-music-in-the-future/image-3.jpg"
alt="Source: https://www.istockphoto.com/photos/music-box">&lt;figcaption>
&lt;p>Source: &lt;a href="https://www.istockphoto.com/photos/music-box" target="_blank" rel="noopener">https://www.istockphoto.com/photos/music-box&lt;/a>&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>We live in an era of customised content, all of the things we see on the internet are customised for us by a myriad of algorithms. I am sometimes amazed and downright scared by the fact that we all experience our world in such a personalised fashion to the extent that we practically live in separate realities ( through personalised content, people might be exposed to different interpretations of the same thing, thus leading to their different views and thus leading to increasing differences in the society )&lt;/p>
&lt;p>But I am not talking about some music recommendations, I am talking about being able to customise the experience of music itself. Imagine how cool it might be to be able to change the instruments in the background of a song, automatically converting it into a lo-fi version, or a jazzy version lol. How cool would it be to be able to change the singer of the song, and hear any song in anyone’s voice? And hell, how cool it would be to automatically translate a song into your own language. Maybe instead of attending a virtual concert, we can create our own virtual concerts, maybe I can experience Michael Jackson dancing to Billie Jean in a virtual world of my creation, and maybe I can change the location/set anyway I want. Maybe for Billie-Jean, I can have a jazz club environment, and for “They Don&amp;rsquo;t Care about us” maybe I can have a dystopian environment, kind of like a WW2 scene. How awesome it would be to able to experience these legends, customizable at our fingertips…&lt;/p>
&lt;h3 id="creation">Creation&lt;/h3>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/how-i-would-like-to-experience-music-in-the-future/image-4.png"
alt="Source — https://superpowered.com/low-latency-audio-android-ios">&lt;figcaption>
&lt;p>Source — &lt;a href="https://superpowered.com/low-latency-audio-android-ios" target="_blank" rel="noopener">https://superpowered.com/low-latency-audio-android-ios&lt;/a>&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Everyone has a particular taste in music. I know what kind of beats I like, what type of voice I find good and what words I want to hear according to my current mood, what response different sounds induce in me. But ah well, I am mostly unable to produce them for myself …..yet ( I will surely learn at least one instrument soon, already working on my guitar skills, but again, the experience will be quite limited by my own talent). But what if in the future, music creation is so greatly simplified, that it becomes possible for “normal” people like you and me to create music we like, ourselves, without being a musical maestro.&lt;/p>
&lt;p>Maybe I can just think about what type of music I want to listen to and the rest is handled by some software that creates exactly the type of song that I want to hear out of thin air. Maybe a machine can just interpret my thoughts and estimate my mood and create music that would be best suited for my current state, although this does sound a bit out of an authoritarian post-apocalyptic dystopia of the likes of George Orwell’s 1984, but well, maybe it&amp;rsquo;s not so bad if limited to music lol?&lt;/p>
&lt;h3 id="sensation">Sensation&lt;/h3>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/how-i-would-like-to-experience-music-in-the-future/image-5.jpg"
alt="Source — https://www.pinterest.de/pin/297167275387524552/">&lt;figcaption>
&lt;p>Source — &lt;a href="https://www.pinterest.de/pin/297167275387524552/" target="_blank" rel="noopener">https://www.pinterest.de/pin/297167275387524552/&lt;/a>&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Now I love listening to music, but there’s always been a tradeoff of sorts associated with listening to it. I can either a.) Wear a headphone/earphone and keep my music private to myself, but deal with the health consequences of wearing something around your head all the time/ having sound directly being fed into your ears or b.) Turn up the music on a speaker, losing both privacy and audio quality lol, so well I usually go with the headphone/earphone option, but they are leading me to lose my hearing ability a bit ( even if it’s not significant enough to be noticeable at the moment ).&lt;/p>
&lt;p>But what if I can directly get music transmitted to my brain? After all, music is in the end just brain-signals, created by our brain’s interpretation of variations in pressure. Why not directly create those signals somehow and be able to experience music, without listening to it? I always find background noise disturbing, so it might be great to somehow remove all background noise by temporary disabling your hearing and just being able to focus on the music. And with techs like NeuraLink and stuff on the horizon, maybe this idea is not too wild lol. Or maybe I can artificially enhance my sense of pitch, and give myself perfect pitch, and be able to experience music with amazing clarity.&lt;/p>
&lt;p>Now I know most of these are way too farfetched technologically speaking to be realised, at least while I am alive. And the way we experience music has more or less stayed the same for centuries ( although there has been a rise of new genres and listening to music has become more portable, and recommendations have given it a sense of personalization ) it has more or less been the same, in its experience and I hope there’s some awesome innovation in the way we experience music lying somewhere around the corner, which might make it more enjoyable, cause hell, I don’t know what I’ll do if I ever get bored of music.&lt;/p>
&lt;p>Anyway, that’s it for this one, as usual, feel free to contact me if you have any feedback about my blogs, or if we haven’t talked in a while, let’s talk! Maybe we can just share a few music recommendations lol, if not much :P&lt;/p>
&lt;p>Until then, Stay safe!&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/how-i-would-like-to-experience-music-in-the-future-d46569cc9c5e" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>What doing research as an undergrad can teach you</title><link>https://ayushtues.github.io/blog/what-doing-research-as-an-undergrad-can-teach-you/</link><pubDate>Sat, 23 Jan 2021 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/what-doing-research-as-an-undergrad-can-teach-you/</guid><description>&lt;figure>&lt;img src="https://ayushtues.github.io/blog/what-doing-research-as-an-undergrad-can-teach-you/image-1.png"
alt="Source — https://www.nature.com/articles/d41586-020-02847-8 ( yeah I thought I should mention the source of the images from now on, its a good practice )">&lt;figcaption>
&lt;p>Source — &lt;a href="https://www.nature.com/articles/d41586-020-02847-8" target="_blank" rel="noopener">https://www.nature.com/articles/d41586-020-02847-8&lt;/a> ( yeah I thought I should mention the source of the images from now on, its a good practice )&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Although my coaching teachers used to say now and then ki&lt;/p>
&lt;blockquote>
&lt;p>“itna deep me jaake kya karega topic me, research karni he kya tereko isme?”&lt;/p>
&lt;/blockquote>
&lt;p>If someone told me in my first few months at college, that I’ll be spending most of my time pursuing research as an undergrad, I would have probably laughed it off. And honestly, it was accidental for most of the part, a random conversation with a senior and some fateful string of events made me consider research as a direction that I would like to explore and then it was just one thing after the other, and it’s not as if I am considering research as a life-career, it’s more of a just for fun kind of activity.&lt;/p>
&lt;p>Also, I am not a full-fledged researcher, I am just an undergrad lol, with only a few small research experiences ( you can find more about it &lt;a href="https://ayushtues.github.io/" target="_blank" rel="noopener">here&lt;/a>). Most of them happened under the supervision of highly experienced people, and it was more of me doing stuff for their research and not a research of my own per se, and so I shouldn’t be writing this blog, but there are two reasons I am doing so :&lt;/p>
&lt;ul>
&lt;li>I probably won’t continue to pursue research as a career in future, so I thought I should write down the feeling, the lessons learnt while they are still fresh, just to come back to them after some years and maybe get inspired or stuff.&lt;/li>
&lt;li>It’s been way too long since my last blog post, and I wanted to write about something ( the real reason lol)&lt;/li>
&lt;/ul>
&lt;p>So, I strongly believe, that everyone should try to get some research experience in their undergrad, even if you don’t want to pursue academic research, &lt;strong>especially&lt;/strong> if you don’t want to pursue it as a career. But you might find this statement contradictory and ask, what would research teach me? If I don’t want to pursue it as a career, why is this for me?&lt;/p>
&lt;p>Well, that’s what I will talk about in this blog —&lt;/p>
&lt;h3 id="bringing-order-tochaos">Bringing Order to Chaos&lt;/h3>
&lt;blockquote>
&lt;p>What is research but a blind date with knowledge?&lt;/p>
&lt;/blockquote>
&lt;blockquote>
&lt;p>— Will Harvey&lt;/p>
&lt;/blockquote>
&lt;p>The most obvious reason for getting research experience that comes to my mind is that it teaches you how to approach an abstract, unstructured problem and gradually bring order to all the chaos, structure stuff up, and devise a solution to a seemingly unsolvable problem.&lt;/p>
&lt;p>Usually, research problem statements are poorly specified, and there is no right or wrong answer to them, no one knows how to solve them accurately, and you get to work on the frontiers of human knowledge, and although it’s doubtful that you’ll be able to make the next great discovery that will lead humankind to salvation, you might be able to take a small small part in the conversation of experts of the world and maybe help shift the conversation from one point to another ( or you can get effectively remain unknown and practically dead for the community, and well, there’s a graveyard for all such dead ideas — &lt;a href="https://arxiv.org/" target="_blank" rel="noopener">arxiv.org&lt;/a> 😛)&lt;/p>
&lt;p>This ability is precious in whatever career you want to pursue, problems in the real world are like research problems too, they don’t have one correct answer ( maybe they don’t have any answer at all ), but you still need to make hypotheses, test it, and then repeat, and that’s precisely what doing research teaches you. I personally feel more confident about tackling any problem in any field after doing some research work, because it has taught me how to approach a problem statement scientifically and break it down into smaller manageable subproblems.&lt;/p>
&lt;h3 id="standing-on-the-shoulder-ofgiants">Standing on the Shoulder of Giants&lt;/h3>
&lt;p>A researcher who believes that he can figure it out all by himself is probably not a very good one. Good researchers respect the opinions of people who came before them, dive deep into their work, what they did, what problems they discovered and how they tackled them. A lot of good research work has been done by taking inspiration from something done by someone lots of years back, which was forgotten over time, and redesigning it into the present. In fact, Issac Newton famously quoted —&lt;/p>
&lt;blockquote>
&lt;p>If I have seen further it is by standing on the shoulders of Giants&lt;/p>
&lt;/blockquote>
&lt;blockquote>
&lt;p>— Issac Newton&lt;/p>
&lt;/blockquote>
&lt;p>And well, I don’t think anyone is vain enough to believe they are smarter than this guy, so if the biggest daddy out there in research says that respect those who came before, you bet we do precisely that.&lt;/p>
&lt;p>Also, research is a highly collaborative environment. People openly discuss ideas ( albeit, there are some constraints due to the ideas being an intellectual property ), they welcome criticism of their works since they believe that the problem statement they are working on is yet unsolved, they are probably not lucky enough to have it all figured out correctly. And I think this is a bit more prevalent in the US and a bit less in India as of now, but hopefully, that would change in the future.&lt;/p>
&lt;p>Working in such an environment would pay off in alternate careers, because well, developing a collaborative spirit is quintessential to success. You get comfortable with putting your “ego” aside and look for ideas from everywhere. You become intellectually humble, and begin to realise that one cannot possibly know everything about anything and also begin to respect other people’s work. You stop discarding things as inherently easy because you know that there is probably some really complex research going on even for something seemingly simple. Personally doing research has made me more open to asking for help from others ( sometimes I think I ask for help a little too much lol, but that’s okay ) and also appreciate the inherent complexity of stuff around us.&lt;/p>
&lt;h3 id="the-art-of-storytelling">The art of storytelling&lt;/h3>
&lt;blockquote>
&lt;p>The human species thinks in metaphors and learns through stories&lt;/p>
&lt;/blockquote>
&lt;blockquote>
&lt;p>— Mary Catherine Bateson&lt;/p>
&lt;/blockquote>
&lt;p>A critical part of life as a researcher is making other people believe that your work is actually worth considering in the larger conversation going about that problem. And what better than storytelling to make a large group of people believe in something ( take religion, for example, it’s mostly storytelling used to unite a people under a common idealogy ).&lt;/p>
&lt;p>Even while writing a research paper, a lot of effort goes into making stuff sound coherent and in line with the overall “story” you are putting forward. You conduct experiments that further strengthen your story’s validity. You review other people’s version of the story and try to add some quirks of your storytelling into it. A PhD thesis is basically a story of your beliefs, and a PhD defence puts that story to the test, hoping that you can make others believe it.&lt;/p>
&lt;p>Again this art of storytelling is a vital skill in any career. Say you want to be an entrepreneur, I can’t imagine someone being successful as an entrepreneur without storytelling. You need to make people believe in your story for making them work with you, and you need customers to believe in your story to buy services/products from you. Political parties need people to believe in their story if they want to come to power. The human population has grown to such a level, that to unite these many people, what works the best is making them believe in a common story.&lt;/p>
&lt;p>Also, as an added bonus, as a researcher, you have to give a lot of presentations ( since you need to present your research work to multiple peoples, at various conferences, guest lectures, reading groups etc. etc. ). So you also get to learn the crucial skill of making great presentations quickly and also a lot of time you have to present complex research work to people who don’t necessarily have the same level of expertise in that area, so you also learn how to simplify complicated stuff and explain it in a simple manner, and that my dear reader, is genius.&lt;/p>
&lt;h3 id="do-or-do-not-there-is-notry">Do or Do not. There is no try.&lt;/h3>
&lt;blockquote>
&lt;p>Research teaches a man to admit he is wrong and to be proud of the fact that he does so&lt;/p>
&lt;/blockquote>
&lt;blockquote>
&lt;p>— H.E. Stocher&lt;/p>
&lt;/blockquote>
&lt;p>This one is probably the most critical thing research has perhaps taught me personally. Rejection becomes a daily thing for people doing research and failure your only true companion. You work your ass off on a research paper, spending countless hours, doing tedious experiments, and submit it for peer review, and then get it rejected by the whims of a moody reviewer. Also, especially as an undergraduate, research opportunities are sparsely available. You get rejected a lot while applying for research positions, hell most of the people don&amp;rsquo;t even bother replying to undergrads lol.&lt;/p>
&lt;p>Also doing research requires you to be quite persistent and intrinsically motivated. It’s quite common to feel lost in the process of finding the answer to your question, and help is not readily available, cause well, its a research problem, nobody really knows/cares about it other than you and probably a few more people. You need to stay motivated and keep trying and trying until you finally overcome it or decide to pivot around the issue ( knowing when to pivot and change your hypothesis when stuff isn’t working is also a crucial skill in itself).&lt;/p>
&lt;p>When failure becomes a daily thing for you, you start getting used to it; you become used to give it your all, knowing that you might fail. This ability is vital in any career, and it makes you fearless and bold, and more willing to take risks. I personally feel that after getting rejected so many times in my early applications in finding research, I have become quite shameless about failing and have become used to doing whatever it takes to get shit done.&lt;/p>
&lt;h3 id="some-otherstuff">Some other stuff…&lt;/h3>
&lt;p>Here are some other points that come to my mind, but I don’t want to drag the post too long, so I’ll cover them quickly.&lt;/p>
&lt;ul>
&lt;li>It makes you a faster reader and makes you able to gather knowledge pretty quickly, which is a bit obvious since you spend most of the time reading research papers and seeing other people’s work.&lt;/li>
&lt;li>It makes you more creative since you have to figure out novel solutions to problems day in and out. You can’t just repeat a previous idea, you’ve got to think something new, hell, you might be trying to solve something that no one else did before, you obviously have to be highly creative.&lt;/li>
&lt;li>It teaches you to respect the small details and how to follow rigorously meticulous experimental procedures. I think doing research has made me better at organising stuff and following the proper procedure of things even if it’s boring.&lt;/li>
&lt;li>You become better at expressing your ideas, concisely and clearly. Research papers usually have an upper bound on the length of pages, and thus you get used to explaining stuff in as little words as possible ( although you do tend to start using more jargons, but ah well there is no free lunch )&lt;/li>
&lt;li>You get used to breaking down a problem to its core and doing things from first principles ( like a physicist ), an approach used by highly successful people like Elon Musk to reimagine the world.&lt;/li>
&lt;/ul>
&lt;p>While I might not continue pursuing research as a career, I am grateful for whatever research experience I got in the last year. It fundamentally changed the way I approach problems. I hope this article might motivate more people to give it a shot even if they don’t want to become a scientist and remove the stigma surrounding research from their mind ( cause well, even I thought research was for weirdos with a grey beard when I started xD ).&lt;/p>
&lt;p>As usual, feel free to hit me up if you have any feedback for my writing, I would highly appreciate that! Also hey, let’s talk anyway! Its been a long year, and I am sure you have a lot of interesting stuff to share :P&lt;/p>
&lt;p>Until then….Stay safe!&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/what-doing-research-as-an-undergrad-can-teach-you-fa6ea1836b97" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>The Exploration Vs Exploitation Dilemma in RL</title><link>https://ayushtues.github.io/blog/the-exploration-vs-exploitation-dilemma-in-rl/</link><pubDate>Fri, 27 Nov 2020 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/the-exploration-vs-exploitation-dilemma-in-rl/</guid><description>&lt;p>Hi Guys, this is the first of my 3 part series on Reinforcement Learning:&lt;/p>
&lt;ul>
&lt;li>The exploration vs exploitation Dilemma&lt;/li>
&lt;li>Value and Policy Iteration&lt;/li>
&lt;li>How DeepMind beat the world Go Master&lt;/li>
&lt;/ul>
&lt;p>Reinforcement Learning is fucking cool, period. I have been reading about it for quite a time now and thought it would be a great idea to make a blog series, describing some basic RL algorithms and the problems associated with it. It would also serve the purpose of fulfilling my long-standing dream of writing my first technical blog.&lt;/p>
&lt;p>Okay, having successfully added enough text to make the preview appear like its a tutorial. For all the AI geeks out there, who thought this was a cool tutorial blog on Reinforcement Learning (RL), my sincerest apologies…..this is not what you are looking for (although seen from a different perspective, maybe it is). But hopefully, the title at least fooled the Medium recommendation algorithms a bit xD.&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/the-exploration-vs-exploitation-dilemma-in-rl/image-1.gif"
alt="GIF">&lt;figcaption>
&lt;p>If you thought this was just another RL tutorial, SURPRISE!&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Having established my identity as a guy lost in vanity, playing dirty tricks like clickbaity titles to increase a few views of his blog ( that almost no one reads anyway) and managing to hurt some sentiments. Let&amp;rsquo;s get started with the real stuff.&lt;/p>
&lt;p>So first off what’s the blog actually about, well, loving word-plays and being a self-pronounced master of pointless puns cringe enough to make people laugh in pity, the title was a word-play as well. The blog is not about RL, at least not the RL that’s Reinforcement Learning, it&amp;rsquo;s about something more important and complex and of practical utility than teaching some robot do stuff in a simulator ( which is also damn cool though). It’s about the exploitation vs exploration dilemma that’s omnipresent in &lt;strong>RL ie.&lt;/strong> &lt;strong>Real Life.&lt;/strong>&lt;/p>
&lt;h3 id="the-dilemma">The Dilemma&lt;/h3>
&lt;p>Surely you have faced this dilemma before, even though you might not know that this is a heavily researched problem in AI. You have two choices in almost anything in life: 1.) you can &lt;strong>exploit&lt;/strong> the trusted option that is sure ( or at least highly expected ) to give a good reward or 2.) you can take a leap of faith and &lt;strong>explore&lt;/strong> into unchartered territories, either hitting a jackpot and getting rewards beyond your wildest expectations or maybe, hitting rock bottom and failing hard or anything in between.&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/the-exploration-vs-exploitation-dilemma-in-rl/image-2_hu11881637642075134861.webp 400w,
/blog/the-exploration-vs-exploitation-dilemma-in-rl/image-2_hu4019147475250708125.webp 760w,
/blog/the-exploration-vs-exploitation-dilemma-in-rl/image-2_hu14999688424854163954.webp 1200w"
src="https://ayushtues.github.io/blog/the-exploration-vs-exploitation-dilemma-in-rl/image-2_hu11881637642075134861.webp"
width="760"
height="430"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;h3 id="some-examples-from-my-ownlife">Some examples from my own life&lt;/h3>
&lt;p>The dilemma is such a huge part of our life, that it felt worthwhile to write about it. And what do you know, even while writing this blog, I had the pleasure of facing it again. Which music to listen to while writing the blog? Whether to go with the trusted “Liked Songs” on Spotify ( Exploit ) or explore a new cool genre. Well, after some internal deliberation, exploration it was! Am writing this while listening to some cool French songs ( well, French is said to be the best language for music, and I also loved this youtube called &lt;a href="https://www.youtube.com/channel/UCSiPjfAJBgbFlIUsxOWpK0w" target="_blank" rel="noopener">Pomplamoose&lt;/a> which did cool French/English covers, so well, thought why not try it out lol).&lt;/p>
&lt;p>Also with today being the last day of the semester, and an entire month of holidays ahead, the dilemma of whether to use the time to explore new stuff and learn new skills, or to exploit my existing skills in AI research and do some more cool work in it strikes again. Thought maybe writing this blog would help me find the answer.&lt;/p>
&lt;p>Now with the AI community having discussed this problem in such great detail and being an AI practitioner myself, it felt almost sacrilegious to not derive from the various policies established there, for addressing the problem. So let’s see how those policies transfer to the real world :&lt;/p>
&lt;h3 id="policy-1always-exploiting">Policy #1 — Always Exploiting&lt;/h3>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="GIF"
src="https://ayushtues.github.io/blog/the-exploration-vs-exploitation-dilemma-in-rl/image-3.gif"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>Well, most of us are guilty of living inside that sweet comfy blanket that is our comfort zone. It just feels so safe to go with the option that we know has a reliably high chance of success, whether it&amp;rsquo;s about choosing to climb the corporate ladder or about going to your favourite restaurant for that monthly dinner with the family ( I pretty much order the same dishes every time as well xD ). Especially in India, this policy is deeply ingrained in the minds of children from the starting, with parents forcing their children to choose the tried and time-tested conventional career options.&lt;/p>
&lt;p>But as is the case with all greedy policies, the policy often leads to you being stuck at a kind of “local” maxima. What that means is that if you just exploit all the time, repeating the same thing again and again, you’ll never exploit the undiscovered surprises life has kept hidden for you. It’s the equivalent of being trapped in the proverbial “Rat Race” ( which reminds me, you should totally check out this awesome book called Rich Dad Poor Dad, you’ll never be the same about the importance of learning personal finance again )&lt;/p>
&lt;h3 id="policy-2-always-exploring">Policy #2— Always Exploring&lt;/h3>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/the-exploration-vs-exploitation-dilemma-in-rl/image-4.gif"
alt="GIF">&lt;figcaption>
&lt;p>Don’t be like this dumbass explorer&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Now rarely anyone follows this policy, but this extreme end of the spectrum is also worth contemplating about. What will happen if you always explore new things. Well as an initial thought, you might be tempted to say that life would be amazing, you’ll feel alive and stuff, because, in the recent world scenario, everyone tells you to explore new horizons, but might be wrong with that?&lt;/p>
&lt;p>Well first of all if you just keep on shifting from one thing to another, you might end up having knowledge about nothing. To get good at something, you have to put in the time and repeat it again and again ( exploit it ). Have literally spent months in this quarantine period, mindlessly exploring stuff and while it was super fun to learn all these new things, now that I think about it, didn’t really get a solid understanding in any of them to do anything meaningful about it. Although yeah, one thing is true as hell, you NEVER regret spending time exploring new stuff. Also well, exploring always might end up you hitting rock bottom, as it did quite literally for our dumbass Dora the explorer in the above gif xD&lt;/p>
&lt;h3 id="policy-3-epsilongreedy">Policy #3— Epsilon Greedy&lt;/h3>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/the-exploration-vs-exploitation-dilemma-in-rl/image-5.png"
alt="Its fun to explore once in a while, right?">&lt;figcaption>
&lt;p>Its fun to explore once in a while, right?&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>This is the defacto strategy in the AI community, which is simplistic but turns out to be surprisingly hard to beat. What it says is that you keep on sticking to your tried and tested options for most of the time, but, time and again you randomly explore some crazy stuff. This will make sure that you keep on leading a smooth life ( due to the exploitation ) but are also able to discover new aspects of yourself through the occasional exploration.&lt;/p>
&lt;p>I think this is the optimal policy for your life as well, keeping some kind of mechanisms, which you know would keep your life stable and then occasionally just exploring random shit to reinvent yourself. It promotes the idea of being a life long learner, who has some kind of a main job which supports for his/her family, but also has these tangential side interests which keep the exploring child inside him alive :P&lt;/p>
&lt;h3 id="policy-4-decreasing-epsilongreedy">Policy #4— Decreasing Epsilon Greedy&lt;/h3>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/the-exploration-vs-exploitation-dilemma-in-rl/image-6_hu5442794882879980757.webp 400w,
/blog/the-exploration-vs-exploitation-dilemma-in-rl/image-6_hu13748235346273299078.webp 760w,
/blog/the-exploration-vs-exploitation-dilemma-in-rl/image-6_hu16209468247040372143.webp 1200w"
src="https://ayushtues.github.io/blog/the-exploration-vs-exploitation-dilemma-in-rl/image-6_hu5442794882879980757.webp"
width="734"
height="429"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>While the epsilon greedy policy is the most optimal one in regards to life, this is the policy that most people actually end up following. In this policy, you gradually stop decreasing the amount of exploration you do over time and do more and more exploitation. This is similar to humans, who in their childhood are brave explorers, but over time, as adults become more and more conservative in nature.&lt;/p>
&lt;p>And while this policy was introduced in the AI community to allow for allowing the agent to gradually end up having a smooth transition into its end state ( which is death in the case of humans). I strongly feel that no matter how old we get, no matter if we are still in college or an industry veteran, we shouldn’t stop exploring, because you never know what new interesting things might have popped in this awesome wilderness called the Earth.&lt;/p>
&lt;h3 id="policy-5-contextual-epsilongreedy">Policy #5 —Contextual Epsilon Greedy&lt;/h3>
&lt;p>This is the fight or flight version of exploration strategies for your life. The rule is simple, when your life is going smoothly, explore more and when you are having a hard time, exploit more and get yourself back on track. If you have an exam tomorrow, study for it, don’t start on exploring and start learning that new song ( speaking from personal experience lol ). If its the start of the semester, and life’s going on smoothly, take risks, explore new stuff.&lt;/p>
&lt;p>Now usually it&amp;rsquo;s the opposite that happens, people explore less when their life is going on smoothly, they like living in the comfort zone and when life gives them a sudden shock, tragedy strikes, then they explore a wide variety of techniques to mitigate and get back on track. But just think about, what if doing the opposite makes more sense?&lt;/p>
&lt;p>Having mentioned all these policies, its important to remember that there is no one glove fit alls policy for life, and well life’s a pretty complex thing, even for something relatively easier like Reinforcement learning, the exploration vs exploitation dilemma is pretty much unsolved. Regardless it’s an interesting problem to confound upon and it also goes on to show that we can take inspiration from science to solve real-life problems ( and the other way around ). The dilemma is close to my heart, both because I love Reinforcement Learning and AI in general and also because I find myself in the same dilemma almost every day, especially in the COVID times due to the extra free time.&lt;/p>
&lt;p>Anyway, that’s all for the blog post folks, as always it would be a pleasure to get any kind of feedback for my blogs. And if we are acquaintances, and haven’t talked in a while, let&amp;rsquo;s each explore a bit and connect again! Until then, so long ….&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/the-exploration-vs-exploitation-dilemma-in-rl/image-7.gif"
alt="GIF">&lt;figcaption>
&lt;p>ByeeeeeeeeeeeeEEEEEEE!&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>&lt;strong>References:&lt;/strong>&lt;/p>
&lt;p>[1] &lt;a href="https://medium.com/@dennybritz/exploration-vs-exploitation-f46af4cf62fe" target="_blank" rel="noopener">https://medium.com/@dennybritz/exploration-vs-exploitation-f46af4cf62fe&lt;/a>&lt;/p>
&lt;p>[2] &lt;a href="https://medium.com/data-science-for-everyone/the-explore-exploit-dilemma-436cb1edff0d" target="_blank" rel="noopener">https://medium.com/data-science-for-everyone/the-explore-exploit-dilemma-436cb1edff0d&lt;/a>&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/the-exploration-vs-exploitation-dilemma-in-rl-fca0811142ec" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Cool products that I would love (someone) to make if they don’t exist yet :P</title><link>https://ayushtues.github.io/blog/cool-products-that-i-would-love-someone-to-make-if-they-dont-exist-yet-p/</link><pubDate>Tue, 13 Oct 2020 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/cool-products-that-i-would-love-someone-to-make-if-they-dont-exist-yet-p/</guid><description>&lt;figure>&lt;img src="https://ayushtues.github.io/blog/cool-products-that-i-would-love-someone-to-make-if-they-dont-exist-yet-p/image-1.png"
alt="Please, please, pleeeeeease?">&lt;figcaption>
&lt;p>Please, please, pleeeeeease?&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Have you ever had that moment… when you are facing a problem in your everyday life and you think “Why the hell hasn’t anyone made a product to deal with this”…..well, am pretty sure you had. Anyway, I most certainly have, and quite frequently too. But I never did anything about it and the thoughts used to die after the moment was gone and were sent to some weird graveyard of lost ideas and hopes ( &lt;em>**cough** no reference to Arxiv **cough**&lt;/em>).&lt;/p>
&lt;p>Anyway, having all the free time in the world due to COVID, last weekend I decided to start keeping a note of all such weird ideas that pop into my mind about cool stuff that should be made ( &lt;em>by someone else&lt;/em> ) for making life a bit easier, which resulted in this article. Now well, it’s a bit embarrassing to make this post, since being a CS undergrad, we are expected to make solutions to our own problems, but the truth is….I suck at coding stuff up…so that’s that. Can’t expect to make any of this on my own, but hey maybe you can do something about these? At least please please do let me know if solutions to these already exist :P&lt;/p>
&lt;p>Since this is something I started doing pretty recently ( &lt;em>like 3–4 days ago lol&lt;/em> ), I have come up with only 5 products that I would love to see someone to make and would happily pay for (&lt;em>well…maybe xD&lt;/em> ). Here goes :&lt;/p>
&lt;h3 id="why-cant-we-have-a-smartphone-for-minimalists">Why can’t we have a smartphone for minimalists?&lt;/h3>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/cool-products-that-i-would-love-someone-to-make-if-they-dont-exist-yet-p/image-2.png"
alt="The internet suggested me hundreds of choices for this, so much for minimalism :P">&lt;figcaption>
&lt;p>The internet suggested me hundreds of choices for this, so much for minimalism :P&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>I was recently intrigued by the concept of minimalism ( &lt;em>after watching this awesome documentary called “The Minimalists”&lt;/em> ), and well keeping the explanation minimal, what it means is that while in modern society we are following the “American Dream” of a previous age, where the more there was, the better. Maybe….just maybe…less is more? Anyway, I am not going to talk much about minimalism here( &lt;em>maybe next time ?&lt;/em> ).&lt;/p>
&lt;p>But hey, just think about it yourself, with computational capabilities following Moore’s law and rising exponentially, our mobiles are getting loaded with more and more stuff ( &lt;em>read&lt;/em>: crap ). We load them with one app after another, and even the mobile loads us with more notifs after notifs. And everything is super-customisable, you have like countless possible combinations to set-up your mobile in with all the themes, widgets, launchers, setting configurations etc etc.….and ah well, personally, for someone like me who goes to an ice-cream parlour, gets confused by all the options and ends up taking the plain vanilla scoop at the end, I’ll love a minimalistic mobile, and I guess others will also, right?&lt;/p>
&lt;p>Maybe it can have very few options for stuff, or maybe have an AI algorithm that automatically tunes everything. And it only has a single app for every purpose, just one pdf reader, just one calendar app, just one music app. And no social media please ( &lt;em>now this is just customising it a little bit for me lol&lt;/em> ) and minimal notifs. Now I know this will have a hard hit on the monetary side of things and would be a turn-off for most people, but for some weirdos with even weirder OCD’s (&lt;em>like me&lt;/em>), it would be heaven-sent :P&lt;/p>
&lt;h3 id="why-cant-i-get-that-magical-feel-of-touching-paper-while-readingebooks">Why can’t I get that magical feel of touching paper while reading ebooks?&lt;/h3>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/cool-products-that-i-would-love-someone-to-make-if-they-dont-exist-yet-p/image-3.png"
alt="Would be a dream come true">&lt;figcaption>
&lt;p>Would be a dream come true&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>A quick piece of info, I fucking reading love books, period. And most of the reading I do is in ebook format, which is all great and flexible and all, but there’s that one super-important element missing from them — The magical feel of the texture of the real paper. Having a personal library full of paperbacks is high on my bucket list ( &lt;em>yes am a nerd :P&lt;/em> ), but well, it&amp;rsquo;s hard to realise until I start earning some money. Until then, I would love someone to make ebook reader with the texture of real paper ( &lt;em>and maybe with the blissful smell of books as well?&lt;/em> ).&lt;/p>
&lt;p>I had a design in my mind, which may or may not be feasible. So basically it was inspired from that magnetic slate kind of thing we had when we were way too young ( &lt;em>you know, the one with the pen, and a slider below to like slide and erase stuff&lt;/em> ). Maybe we can have a controller on the front, which is basically a digital interface for like users to select the book to read and like change the pages etc. And then we had like 10 pages constituting the book, having texture exactly like real paper, displaying the current 10 pages, which worked on the same concept of using magnets to orient iron on the paper to display words, and maybe we can combine it with some tech from Braille for blind people to give the words some thickness, idk…sounds cool lol, would certainly buy one.&lt;/p>
&lt;h3 id="where-is-the-stuff-i-stored-somewhere-in-my-house-agesago">Where is the stuff I stored somewhere in my house ages ago?&lt;/h3>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/cool-products-that-i-would-love-someone-to-make-if-they-dont-exist-yet-p/image-4.png"
alt="Well, google kind of knows all about us, why not this as well?">&lt;figcaption>
&lt;p>Well, google kind of knows all about us, why not this as well?&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Well, it’s a sad truth of life, that not everyone is like Sherlock Holmes, having a “mental palace” in their mind, with a perfect replica of whatever they have seen. If you are anything like me, you would have limited space in your brain for stuff ( &lt;em>and if you are more like me, that’s already full of useless shit :P&lt;/em> ). And ever so often we forget about where we kept our stuff in our house ( &lt;em>which can be pretty large for some people and/or full of too much stuff&lt;/em> ) making finding something small like say a pendrive, a needle in a haystack problem.&lt;/p>
&lt;p>Now imagine, having a mental palace of your own just like Sherlock, through an app. Which has a great 3D model of your entire house, and in which you can keep track of what stuff goes where and easily search for where was your stuff stored. Also if it’s a 3D model, maybe it can help you in organising your house well, like optimising the space allocation of your furniture to give the optimal distribution.&lt;/p>
&lt;p>Well constructing a 3D model may be hard from scratch, but remember every house has a well-detailed drawing made by the architect, which is also used for the construction of the house. So maybe we can have an app which takes in that drawing and create a 3D model out of it.&lt;/p>
&lt;p>And well maybe, we don’t need such a fancy solution, maybe we just want a simple app to keep track of where we kept things in our house, who knows?&lt;/p>
&lt;h3 id="what-to-make-today-fordinner">What to make today for dinner?&lt;/h3>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/cool-products-that-i-would-love-someone-to-make-if-they-dont-exist-yet-p/image-5.png"
alt="Don&amp;rsquo;t tell this to my Mom xD">&lt;figcaption>
&lt;p>Don&amp;rsquo;t tell this to my Mom xD&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Okay, this is not my problem, but its a question that haunts my mother almost every day lol. What exactly should we make for lunch/dinner today, like really?&lt;/p>
&lt;p>Now given the limited amount of things that different people eat in our family and the inventory of stuff available in the fridge, and the recency of stuff in the past few days, there’s only a limited amount of stuff that can be made. And solving this kind of optimisation sounds like just the perfect thing for an app.&lt;/p>
&lt;p>Oh and we can get all fancy with it if we want, we can have all these nutrition assigned with all the dishes, and your mother can like, slide this bar from tasty — healthy and get an entire spectrum of combinations. Also, it can get real-time feedback from my dad who is otherwise in office, about what he would like to eat when he comes back home, there can also be some AI added, which learns our preferences and stuff….and hey, who said this is limited to my house, it could be scaled up to hotels and cafeterias and what not.&lt;/p>
&lt;h3 id="why-is-it-so-hard-to-find-research-opportunities-as-a-beginner">Why is it so hard to find research opportunities as a beginner?&lt;/h3>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/cool-products-that-i-would-love-someone-to-make-if-they-dont-exist-yet-p/image-6.png"
alt="Ah the terrors">&lt;figcaption>
&lt;p>Ah the terrors&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>If you think getting interns are hard, try getting a research intern as an undergrad, it&amp;rsquo;s hellishly hard. And for good reason as well, there’s not much of a place for students to connect with researchers for opportunities. You don’t have an AngelList/Internshala for research interns. And the only option you are left with is to go through the dreaded cold mailing process, which is more or less, just a random process.&lt;/p>
&lt;p>Now I know it sounds cliche, but it would be great to have some kind of social platforms for less experienced researchers to connect with experienced ones and get work under them. The process is already super common in the software engineering domain, then why not in research? Interns are cheap ( &lt;em>some like me, even work for free&lt;/em> ), they are super enthusiastic and can endlessly carry out mindless experimentation, and hey, professors do want people like that lol.&lt;/p>
&lt;p>Now there certainly are websites like Reserachgate etc which aim to do this, but maybe having something specific for research interns would help a lot.&lt;/p>
&lt;p>Anyway, I wanted this article to be a quick one, so it might have been a bit rough and unpolished, but please please do let me know if anyone of these ideas has been already implemented. And as usual, if you are someone with whom I have not connected in a while, do drop a message, would love to talk!&lt;/p>
&lt;p>So, until the next one :P&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="GIF"
src="https://ayushtues.github.io/blog/cool-products-that-i-would-love-someone-to-make-if-they-dont-exist-yet-p/image-7.gif"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/cool-products-that-i-would-love-someone-to-make-if-they-dont-exist-yet-p-4fd25b7eb03d" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Please don’t let them find out I am just an imposter!</title><link>https://ayushtues.github.io/blog/please-dont-let-them-find-out-i-am-just-an-imposter/</link><pubDate>Sun, 20 Sep 2020 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/please-dont-let-them-find-out-i-am-just-an-imposter/</guid><description>&lt;blockquote>
&lt;p>What’s talent but the ability to get away with something ?&lt;/p>
&lt;/blockquote>
&lt;blockquote>
&lt;p>— Tennessee Williams ( just some guy google suggested )&lt;/p>
&lt;/blockquote>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/please-dont-let-them-find-out-i-am-just-an-imposter/image-1.png"
alt="The eternal nightmare">&lt;figcaption>
&lt;p>The eternal nightmare&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>A couple of months ago, I was reading this awesome book by Nassim Nicholas Taleb called Fooled by Randomness which was supposed to be about some common finance-related misconceptions, how the human brain is painfully bad at dealing with probabilities, and how much does luck have to do with success….reading it….. kind of had a weird side effect, and a lot of introspection, which culminated in this post on something that’s on everyone’s mind and which not many people talk about.&lt;/p>
&lt;p>So, you are tagged as “successful” by your friends and family, everyone looks up to you, and believe that everything is going on great with you. And well…for most of it, its actually true, you are really on the fast-lane and things are looking pretty well, until…….&lt;/p>
&lt;p>Until something creeps out from a dark abyss inside your brain and whispers into your ear — “ Lol, but you do know you don’t &lt;em>actually&lt;/em> deserve all this, right? You’re just faking it, right? People will eventually find out, and that will be the end of you ” — Now apologies if that just put you into an existential crisis, but the whispers of this little asshole of a devil, is something that well, more or less all of us have to hear to.&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/please-dont-let-them-find-out-i-am-just-an-imposter/image-2_hu6699637831104031854.webp 400w,
/blog/please-dont-let-them-find-out-i-am-just-an-imposter/image-2_hu6709612508137477182.webp 760w,
/blog/please-dont-let-them-find-out-i-am-just-an-imposter/image-2_hu5579073171564767080.webp 1200w"
src="https://ayushtues.github.io/blog/please-dont-let-them-find-out-i-am-just-an-imposter/image-2_hu6699637831104031854.webp"
width="760"
height="374"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>It turns out that approximately 70% of people feel that they are just an “imposter”, that we are just lucky and have no real basis for our successes and someone would eventually find this out and that will the end of it. Well, 70% is just a figure from the net, hell….if you are not one of those narcissists or aren’t dumb enough to even know that there’s a scope of improvement in your current abilities, it&amp;rsquo;s highly likely you’ll experience this. And well, the feeling is completely normal, and usually goes away in a few hours/days, but its when it&amp;rsquo;s stuck, is where the problem begins.&lt;/p>
&lt;p>It&amp;rsquo;s when these thoughts are recurring; when they lower your confidence and self-esteem; when they make you overly self-critical and fearful of criticism is when the problem occurs. And well it’s actually so common, that it even has a name ie. “Imposter Syndrome”, and believe me, it&amp;rsquo;s not pretty.&lt;/p>
&lt;p>Since I have been myself dealing a lot with this syndrome during the last few months ( won’t go into the details here, but let&amp;rsquo;s just say that you have to be pretty careful to not bite more than you can chew :P ), I decided to write about it as well, maybe as a kind of therapy? Anyway….. let&amp;rsquo;s discuss some of the causes of this psychological phenomenon and some ways to get out of this mindset:&lt;/p>
&lt;h3 id="causes">Causes :&lt;/h3>
&lt;h3 id="perfectionism">Perfectionism&lt;/h3>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/please-dont-let-them-find-out-i-am-just-an-imposter/image-3.gif"
alt="GIF">&lt;figcaption>
&lt;p>It&amp;rsquo;s never good enough, is it?&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Sometimes it&amp;rsquo;s so hard to believe that the work you are doing is actually good. The perfectionist inside you says that “ Dude, you call this piece of trash, ok? You can’t fool people for long with this quality of work ”. Whether it be a piece of code or art or a blog, a perfectionist never feels that what they have done is good. And yeah, some level of perfectionism is great, after all the devil is in fact in the details. But sometimes you just need to let it go and publish your work, and more often than not, people are actually quite welcoming of your work and help you improve it with constructive feedback, rather than dissing you off as an imposter.&lt;/p>
&lt;h3 id="the-socialdilemma">The Social Dilemma&lt;/h3>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/please-dont-let-them-find-out-i-am-just-an-imposter/image-4.png"
alt="Awesome documentary!">&lt;figcaption>
&lt;p>Awesome documentary!&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>With yet another documentary about the detrimental effects of social media on the “self”, ironically going viral on social media ( The Social Dilemma ), it’s enough to say that social media has finally converted into the Skynet that movies predicted, just not in the form of a buffed up humanoid robot Arnold Schwarzenegger. And yes, sure enough, from inside the pandora&amp;rsquo;s box it has opened, also came out imposter syndrome.&lt;/p>
&lt;p>When you open up platforms like LinkedIn, Facebook, Twitter, Reddit etc. ( am losing count honestly ), the place is full of over-achievers. For each one of your achievement, there would be someone 10x times better posting about his/her achievements and hard work and success story etc. And it gets really frustrating when you are working your ass off, but still can’t seem to catch up with these people, which further fuels your imposter syndrome of being “just lucky”. I can’t count how many times have opened up LinkedIn and ended up stuck in an existential crisis.&lt;/p>
&lt;h3 id="friends">F.R.I.E.N.D.S.&lt;/h3>
&lt;p>Oh, and if you thought plugging out of the social media world would finally help you get some piece, lol, not so fast bro! You can’t escape your immediate social circle. If you are surrounded by over-achievers, braggers, then it&amp;rsquo;s bound to take a toll on your self-esteem. Now am not saying that having over-achievers in your social group is bad, it&amp;rsquo;s great! but just try not to compare yourself with them too much.&lt;/p>
&lt;p>Also, keep away from the people who always say that “Oh you are so awesome, Oh you have everything figured out and sorted”. An honest compliment is a great thing and a rare thing to find, but if someone repeatedly showers you with such half-hearted praises, day in and day out, you start connecting your self-worth with it and that’s when all the problems begin.&lt;/p>
&lt;h3 id="how-the-hell-to-deal-withit">How the hell to deal with it?&lt;/h3>
&lt;p>Now you might have realised that you, just like most of us are suffering from imposter syndrome, but hey, it&amp;rsquo;s alright! While its something which cannot be completely fixed ( and maybe should not be fixed ) there are some things you can do to deal with it positively :&lt;/p>
&lt;h3 id="actively-seek-feedback-and-criticism">Actively seek Feedback and Criticism&lt;/h3>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/please-dont-let-them-find-out-i-am-just-an-imposter/image-5.gif"
alt="GIF">&lt;figcaption>
&lt;p>Feedback for this blog is also highly appreciated!&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Instead of being scared of blowing up your farce, and being exposed as an imposter, it&amp;rsquo;s better to actively seek constructive criticism. Ask people that are close to you, like your mentors or close friends what shortcomings they find in your work and where you can improve. I personally feel that as a society, especially in India, we should see criticism in a more positive light and should create an environment where people are not scared to ask silly questions. This culture is already well present in countries like the US, Japan etc where people encourage others to ask even the silliest of questions ( which although can be quite intimidating as a beginner, like when you ask a super-dumb basic question to people who are expert researchers in the field, but that&amp;rsquo;s OK ). In the present scenario, it&amp;rsquo;s pretty hard to find people who give you honest feedback for your work, which also leads you to be confused and ultimately scared of all the potential loop-holes in your work.&lt;/p>
&lt;h3 id="improve-your-socialcircle">Improve your social circle&lt;/h3>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/please-dont-let-them-find-out-i-am-just-an-imposter/image-6.gif"
alt="GIF">&lt;figcaption>
&lt;p>Whenever you meet someone toxic — PIVOT! PIVOT! PIVOT!&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>They say you can’t choose your family, but you can sure as hell choose your friends. Try to stay away from people who are always talking about their achievements making you feel smaller and smaller (…..damn, even I belong to this category, sincere apologies to anyone reading this who had to deal with me bragging around 😅 ). Also, try to avoid people who always seem to measure your worth based on your achievements and shower you with half-assed praises. Instead, try to surround yourself with people, who actually enjoy being with you, as a person ( its easy to figure out who is actually this way, on your bad days :P ) and people who both offer constructive criticism and are also supportive of you.&lt;/p>
&lt;h3 id="remember-you-are-awesome">Remember, You are awesome 😃&lt;/h3>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/please-dont-let-them-find-out-i-am-just-an-imposter/image-7.gif"
alt="GIF">&lt;figcaption>
&lt;p>Even though you might not feel like it&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Remind yourself of all the great things you have achieved. Give yourself a pat on the back for all the seemingly small accomplishments you make and the big ones as well!. Maybe luck really has to do something with bringing you where you are, but hey! it&amp;rsquo;s OK. Luck can only get you so far, you must really have had the talent to be where you are, to have &lt;em>stayed&lt;/em> at that position. The world is a highly competitive place, and as the father of economics, Adam Smith says in his &lt;em>Wealth of Nations.&lt;/em> In such competitive and open environments like the current global world,there exists a sort an “Invisible hand” which automatically pushes people to best well-suited place according to their abilities ( actually the theory he proposed was for market prices….but well, it&amp;rsquo;s more generically applicable). And believe me, the entire modern financial theory is based on this fact, and you are not an outlier! You are where you are, because of who you are!&lt;/p>
&lt;h3 id="learn-to-embrace-your-impostersyndrome">Learn to embrace your imposter syndrome&lt;/h3>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/please-dont-let-them-find-out-i-am-just-an-imposter/image-8.png"
alt="NOT TODAY!">&lt;figcaption>
&lt;p>NOT TODAY!&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>All said and done, the imposter syndrome is not really such a bad thing if you look at it from a different perspective. Sure, if you crumble from its weight and go all emo and stop working, then it&amp;rsquo;s awful, but if you have a solid will-power and a lust for learning, you can turn it around for your advantage. Feeling like an imposter means that you are aware of where you lack in your abilities and that insight, dear reader is a godsend. Once you start working on filling those holes in your abilities, you’ll slowly start overcoming the imposter syndrome, start feeling more confident in your abilities, until well….the next wave strikes :P&lt;/p>
&lt;p>So well that was it for the post, the imposter syndrome is something that I struggle with regularly and hence the post was written as a kind of self-therapy as well and also with a slight hope that it may help some other lost soul as well, lol, I hope it did…… Anyway, as always, if you are someone I know, and we haven’t connected in a long while, feel free to drop by a message, would love to reconnect and also get some feedback on the blog! Until then, so long!&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="GIF"
src="https://ayushtues.github.io/blog/please-dont-let-them-find-out-i-am-just-an-imposter/image-9.gif"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/please-dont-let-them-find-out-i-am-just-an-imposter-19c27ae6302b" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Why the Hell do Companies even Hire an Intern?</title><link>https://ayushtues.github.io/blog/why-the-hell-do-companies-even-hire-an-intern/</link><pubDate>Thu, 30 Jul 2020 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/why-the-hell-do-companies-even-hire-an-intern/</guid><description>&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/why-the-hell-do-companies-even-hire-an-intern/image-1_hu12655424661883216311.webp 400w,
/blog/why-the-hell-do-companies-even-hire-an-intern/image-1_hu17878623435370666549.webp 760w,
/blog/why-the-hell-do-companies-even-hire-an-intern/image-1_hu18093667024987336697.webp 1200w"
src="https://ayushtues.github.io/blog/why-the-hell-do-companies-even-hire-an-intern/image-1_hu12655424661883216311.webp"
width="760"
height="331"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>Internships are a big deal, right? They are awesome opportunities to allow one to get a taste of the big-bad real world of adults that fascinates us since our childhood.&lt;/p>
&lt;p>And well, as it so happens, having just completed my sophomore year of engineering, it seems like the corporate companies have decided that it may be time to welcome me and my friends to their all starred glamorous world. But wait wait….they are not like Willy Wonka, it&amp;rsquo;s not as if they’ll just let us in by pure chance hiding gold tickets in chocolate bars ( That would be sad since I don’t eat chocolate XD ). Instead, to eliminate any kind of randomness, they came up with a great solution ie. standardised tests for screening applicants. Quite a brilliant move to eliminate randomness I must say. Now, this post isn’t a rant about standardised screening tests ( maybe I’ll do that some other time).&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="GIF"
src="https://ayushtues.github.io/blog/why-the-hell-do-companies-even-hire-an-intern/image-2.gif"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>Instead, in the process of preparing for similar standardised tests for my intern season, I asked the question that I guess most of you must have asked before an exam/test —&lt;/p>
&lt;blockquote>
&lt;p>Why do we even have these exams/tests?&lt;/p>
&lt;/blockquote>
&lt;p>Well, I couldn’t find a great answer to that question, but thinking about that question led to another, surprisingly interesting question ( Lol, I think that instead of thinking about such irrelevant things, it would be way more useful for me to think about studying Data Structures and Algorithms for my intern season….but ah well…..fuck it :P )&lt;/p>
&lt;blockquote>
&lt;p>Why do companies even hire interns at all ?”&lt;/p>
&lt;/blockquote>
&lt;p>On the face of it, it seems like its a stupid question to ask ….cause well…it&amp;rsquo;s such a common thing for companies to hire interns and we know how cost-effective these organisations are…..that it must be a useful thing and that’s why so many companies are doing it. It’s surely not something complex enough to write a blog about, is it? Well, that’s what I thought first, but then I thought a bit more about it……and it started to look weird to me.&lt;/p>
&lt;p>I couldn’t understand why such well-established companies would actually hire such inexperienced undergrads to do their work. Since according to my ( yet so naive) understanding of how businesses work, obviously, the work we’ll be assigned wouldn’t be a high priority/importance job for the organisation, since then they would get someone experienced to do it instead of giving it to a new guy. So they are essentially paying us to do some work that’s actually not that useful, which seemed really weird since companies like to cut their costs. I thought that maybe an internship is just an elaborate screening procedure, to filter out potential employees…..but that seemed way too wasteful.&lt;/p>
&lt;p>Amidst the confusion, I decided to do what everyone does when in doubt…..Google. I read a few nice articles explaining why it&amp;rsquo;s beneficial for companies to hire interns, and I found some interesting insights, which will be useful for someone who is going to be hired as an intern in understanding the reason behind his/her hiring…so as to get an idea of the bigger picture. And well, being thoroughly bored due to COVID and from the internship preparation process itself ( which sucks). I decided to write an article about it myself….. if not for anything, maybe just as something to laugh on during my internships?&lt;/p>
&lt;p>So well, let&amp;rsquo;s discuss some of the reasons why it makes sense for companies to hire interns:&lt;/p>
&lt;h3 id="lets-take-a-test-drive">Let’s Take a Test Drive 🚗&lt;/h3>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="GIF"
src="https://ayushtues.github.io/blog/why-the-hell-do-companies-even-hire-an-intern/image-3.gif"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>This is the most obvious reason why companies hire interns. Roughly, about 65% of paid internships become full-time job offers, while 39% of unpaid internships become full-time employment. Internships allow organisations to gauge the skills of the interns and check how they fit in the workforce. Kind of like a test drive before actually buying a car for the next 5–10 years. It’s also the case statistically that interns tend to be more likely to be long term employees and have a higher retention rate which is pretty profitable for organisations since hiring is actually quite an expensive process for the company.&lt;/p>
&lt;h3 id="a-fresh-set-of-eyes-doesnthurt">A Fresh Set of Eyes Doesn’t Hurt?&lt;/h3>
&lt;p>Being dumb (: read inexperienced) can actually be quite useful in highly specialised environments. People who are way too familiar with their work sometimes tend to miss out some obvious things, which can be easily spotted through a change of perspective. Interns, being usually completely new to the work think different and creative ways to tackle problems…..they think out of the box since they don&amp;rsquo;t even know what the box is :P. This often helps companies to engineer new and creative solutions which are highly profitable.&lt;/p>
&lt;h3 id="the-senseifeeling">The Sensei Feeling&lt;/h3>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/why-the-hell-do-companies-even-hire-an-intern/image-4.gif"
alt="GIF">&lt;figcaption>
&lt;p>That moment when someone does what you teach them&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>If you have ever had a junior or ever taught someone, I am sure you will relate. It’s a great feeling to mentor someone and watch them grow, and also it helps you grow as well. Similarly, handling and mentoring interns make an excellent opportunity for the newly appointed managerial staff and junior employees to develop their own leadership and managing skills. Since managing employees of a similar level as they are might be a bit hard for them, mentoring a total newbie like an intern helps them practice these skills in a less stressful manner. After all, who’s better than an intern for being the scapegoat.&lt;/p>
&lt;h3 id="dirty-deeds-done-dirt-cheap">Dirty Deeds Done Dirt Cheap :&lt;/h3>
&lt;p>Another obvious point — Interns are dirt cheap, period. While the stipends provided to interns are increasing by the day, they are still a far cry to the expenses of an actual employee. On the flipside, interns being young are literal powerhouses who want to prove themselves to the world. They are highly motivated and are more willing to work upon challenging problems than a normal employee. So you can get a lot of work done, with minimal expenses.&lt;/p>
&lt;h3 id="whats-the-newfad">What’s the New Fad?&lt;/h3>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/why-the-hell-do-companies-even-hire-an-intern/image-5.gif"
alt="GIF">&lt;figcaption>
&lt;p>You just can’t underestimate social branding&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Being a tech undergrad, looking at cutting edge technology companies it becomes easy to forget how tech-savvy our generation is compared to the previous generation. Interns can help bring in the latest technology to improve performance, simply because they have grown up using technology to solve all their problems. Also, being young they are aware of all the latest stuff going around in social media and the current trends. So they can be highly useful in developing the social brand of the company since they are more adept in understanding what&amp;rsquo;s popular nowadays than the older generation.&lt;/p>
&lt;h3 id="once-a-ranger-always-a-rangerp">Once a Ranger, Always a Ranger :P&lt;/h3>
&lt;p>Internships are a huge huge huge deal for the interns, so it&amp;rsquo;s obvious they talk a lot about it in their social circle and well…word of mouth is a powerful force. I have seen it myself, I usually talk to my seniors about which companies are good, which they usually describe from their own experiences. So give an intern a positive internship experience and you have a walking brand advocate amongst the great pool of talent that exists on college campuses who will make sure everyone knows about you. Internships often open up unseen collaborations that benefit both the intern and the company.&lt;/p>
&lt;h3 id="lets-light-this-place-up--no-noi-am-not-an-arsonist">Let’s light this place up ( No no…..I am not an arsonist)&lt;/h3>
&lt;p>Having interns creates a positive environment for the entire workforce since they help reduce the workload on employees by taking on some of their work. But also because interns are filled with energy, that energy quickly catches its way to the other employees and improves the overall mood of the entire place. Think of it as similar to how kids seemingly light up any place they are, and everyone enjoys interacting with them ( Hell…..I am constantly treated as a kid on all of my interns )&lt;/p>
&lt;h3 id="the-subtle-art-of-giving-afuck">The Subtle Art of Giving a Fuck&lt;/h3>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/why-the-hell-do-companies-even-hire-an-intern/image-6.gif"
alt="GIF">&lt;figcaption>
&lt;p>Well, not everyone is wicked :P&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>We may view corporates as some evil organisations, that are solely concerned about their profits….but well let&amp;rsquo;s be a bit more optimistic shall we. They do give a fuck about others. It must feel great for companies to see young people interested in their work, and they might feel like it&amp;rsquo;s their responsibility to give back to the community by helping the young guns in boosting their career. In return, it also helps them develop their public image as an organisation that’s more thoughtful about society and well avoid all that monopoly crap that ruined Microsoft once.&lt;/p>
&lt;p>Well, whatever the reason might be, I am glad that companies hire interns. Although the general hiring process for college students in India is a tad bit stupid, its still worth the pain. Having had previous internship experiences myself, I can say that its a really great learning experience and it changes you in ways other than just professionally. This post was just a way for me to take a break from the weird almost psychotic havoc that’s created around the internship season at my college. I hope you liked it and it gave you a bit of an understanding of what your role really is as an intern, which may help you improve your internship experience. I am by no means qualified to write about this topic (heck I don’t know shit about it) so take this with a grain of salt and for all the peeps preparing for their next internship……Hope you rock it!&lt;/p>
&lt;p>So long, until the next post :P&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="GIF"
src="https://ayushtues.github.io/blog/why-the-hell-do-companies-even-hire-an-intern/image-7.gif"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/why-the-hell-do-companies-even-hire-an-intern-525d9b0d99f9" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>The Side Effects of Being Studious</title><link>https://ayushtues.github.io/blog/the-side-effects-of-being-studious/</link><pubDate>Sat, 04 Jul 2020 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/the-side-effects-of-being-studious/</guid><description>&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/the-side-effects-of-being-studious/image-1_hu8870727023551566781.webp 400w,
/blog/the-side-effects-of-being-studious/image-1_hu13073960410731849325.webp 760w,
/blog/the-side-effects-of-being-studious/image-1_hu11894983087480305508.webp 1200w"
src="https://ayushtues.github.io/blog/the-side-effects-of-being-studious/image-1_hu8870727023551566781.webp"
width="760"
height="352"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>Let&amp;rsquo;s take a trip down memory lane. Not really that deep though, not till the beginning of the universe or the big bang, just a bit deep. Back to the good old rose coloured school days (ahem, actually I am still a student so it’s not much of a trip for me :P). It was super fun, right? Each day being a new adventure, without any worry in the world. Also for most of the students, the school wasn’t really a place to study, it was more of a hangout place, an elaborate cafe so as to say. Heck, you can almost say that it was a library-themed cafe, where you were surrounded by books and had to socialize in hushed voices due to the metaphorical librarian, ie. the teacher and pay a hefty price for the experience like you know, cafes (why the hell are cafes so overpriced).&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/the-side-effects-of-being-studious/image-2.gif"
alt="GIF">&lt;figcaption>
&lt;p>Ah well, Recess at school was fun :P&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>But even then, there was always that one guy, that poor soul, who took things waaaaaaaay too seriously. His entire world revolved around books and grades and assignments and the only thing he dreaded more than death was getting a fail grade ( and well….maybe public speaking?). Also, however, embarrassing that might be to accept…..I was one of those guys and some would argue that I still am.&lt;/p>
&lt;p>But wait! this post is not about me. Well, it’s partially based on my own experience, but that’s just a way to represent people like me who are extremely studious. Take it as an estimation of the entire class of studious people by one point sample viz-a-viz me, although since it’s a single sample estimate, it’ll be a bit biased ( I kind of love probability and statistics, so umm, apologies for the lame jargons and analogies).&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/the-side-effects-of-being-studious/image-3.gif"
alt="GIF">&lt;figcaption>
&lt;p>I’ll admit to being guilty of similar behaviour😅&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Honestly speaking, being studios has its perks, I’ll have to admit that. But now, in my sophomore-junior-ish year at college (it&amp;rsquo;s kind of in an indeterminate position currently, due to COVID), a few years out of school, looking back, I kind of regret being studious. It lead to me thinking about the ways it has affected my life ( and other things in general) and while it has lead to many, many good things, mostly in my professional career, but it did have a lot of side effects (There really is no free lunch, is there?). This lead to me, writing this blog post, reflecting on what I think, are the side-effects of being studious. Let’s begin:&lt;/p>
&lt;h4 id="you-tend-to-become-a-perfectionist">You tend to become a perfectionist&lt;/h4>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/the-side-effects-of-being-studious/image-4.gif"
alt="GIF">&lt;figcaption>
&lt;p>Well, maybe I don&amp;rsquo;t :P&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>When you spend your entire childhood in getting “perfect A’s”, you can’t help but develop into a perfectionist. It’s just never enough for you. You always try to find out loopholes in your existing work, be it your preparation for the next exam (or a piece of blog you are working upon :P) and while perfectionism is not necessarily a bad thing, it can lead to a very subtle kind of procrastination where you keep on delaying things, on the prospect on being “not yet just right”. It becomes very hard for people like me, to realise it, let alone fix such procrastination since it gives a false sense of superiority.&lt;/p>
&lt;h4 id="you-tend-to-leave-school-with-very-fewmemories">You tend to leave school with very few memories&lt;/h4>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/the-side-effects-of-being-studious/image-5.gif"
alt="GIF">&lt;figcaption>
&lt;p>Whenever someone reminds me of school&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>This is probably the single biggest regret of my school life. After 4 years, I barely have any good memories of my school life. It’s almost like I never even went there which apart from being a bit spooky is really really sad ( I should get my memory checked up, maybe it’s an artificially induced amnesia to delete the memories of some super-secret alien species I discovered in the school. I always suspected some of my teachers are aliens :P). People like me tend to be so focused on their academics, that we forget to enjoy the simple things in life. I can probably give an almost exact description of how my daily day at school went, and that would describe each day of my school life, owing to the monotonicity. We also make very little friends, due to being framed as the “nerds” which reduces the chances to develop meaningful memories with friends even further.&lt;/p>
&lt;h4 id="people-tend-to-ignore-other-parts-of-your-personality">People tend to ignore other parts of your personality&lt;/h4>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/the-side-effects-of-being-studious/image-6.gif"
alt="GIF">&lt;figcaption>
&lt;p>This is what goes on in every fucking person’s mind when I talk with them&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Most of my conversations with old schoolmates start with them making direct/indirect references about how studios I was. It’s like the only image people remember is that we were studios. And well that hurts a bit. Especially when you are now in college and are highly embarrassed by being called studious ( Peer pressure in colleges often makes us ridicules our past-studious habits, though, I now feel that we should continue enjoying our studies if we really love doing so). Heck, even boring studious guys like us have some other interests and we would like for you to remember them and talk about them for once. But alas! they all tend to be overshadowed by the single trait of being a studious nerd. Maybe we are at fault too, we tend to be so overindulged in our academics at school that we forget to talk to other people about stuff that interests us other than books.&lt;/p>
&lt;h4 id="you-tend-to-be-weird-at-socialevents">You tend to be weird at social events&lt;/h4>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/the-side-effects-of-being-studious/image-7.gif"
alt="GIF">&lt;figcaption>
&lt;p>I might also hate people :V&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Studious people tend to lack apt social skills. Maybe it has something to do with most of us being likely to be introverts. Heck, I am writing this blog due to being bored as hell at a wedding, away from all the bands and the ruckus. Still, it’s not that we are not interested, we just find it hard to do the small talk, maybe because we have so much to remember from out studies that we like to cut short to the chase and talk/remember only important stuff. But it’s a bit weird, being well-read in a variety of topics, we tend to be more informed about many things than most people and should be able to carry out interesting conversations. But well, I guess social dynamics are something that just can’t be taught.&lt;/p>
&lt;h4 id="you-tend-to-be-weird-around-the-oppositesex">You tend to be weird around the opposite sex&lt;/h4>
&lt;p>&lt;a href="https://tenor.com/embed/4888416" target="_blank" rel="noopener">(GIF)&lt;/a>&lt;/p>
&lt;p>&lt;em>Does anyone else relate with Chandler?&lt;/em>&lt;/p>
&lt;p>Have you seen the meme that many girls have a crush on the school topper? Let me tell you, its just a meme, a thing to be laughed at, cause the reality is just the opposite. A side effect of being studios is lack of interaction with the opposite gender, with the deficit compounding over your school life, leaving you incapable of even the simplest interaction in the end. The only interaction we usually have with the opposite sex in school is in the form of solving the study related doubts ( and well if you are a no-fucks-given jerk like me, even that probability rules out, leading you totally unaware of the unsaid norms of talking with the opposite gender).&lt;/p>
&lt;h4 id="you-tend-to-likemusic">You tend to like music&lt;/h4>
&lt;p>&lt;em>Ah! The feels 😛&lt;/em>&lt;/p>
&lt;p>Attribute it to some weird scientific correlation between liking classical music and IQ, or to just the need to cut off the noise while studying, I have personally observed that studious people tend to love music. Well, the type of music doesn’t have to classical though, I myself love rock, pop, rap, R&amp;amp;B and a bit of classical (I am listening to Highway to hell by AC/DC at the time of writing this line). Maybe the reason behind it is that it helps us focus and think clearly and gather our thoughts. I also feel that I can remember things better while listening to music, although listening to too much music can sometimes be detrimental also, causing an inability to think properly in its absence ( it also becomes the reason of constant scolding from your parents, due to listening super loud music in your headphones :P)&lt;/p>
&lt;p>Well, whether I like it or not, being studious has been a dominant part of my personality and has had a major impact on my life ( and I guess of many other people like me). And although I am becoming increasingly less studious ( mostly to avoid being ridiculed as the “nerdy geek”), often I have a thought that I should be true to myself and should enjoy studying, since I really love doing that and gathering knowledge about new things. Maybe someday I’ll acknowledge this “studious” part of myself as something to be proud of, instead of trying to hide it and maybe then I’ll recognise the above-listed facts as a part of my personality and come to peace with them. Until then :&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="GIF"
src="https://ayushtues.github.io/blog/the-side-effects-of-being-studious/image-9.gif"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/the-side-effects-of-being-studious-a09fa3ad64e4" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>What I learnt from my love-hate relationship with Social Media</title><link>https://ayushtues.github.io/blog/what-i-learnt-from-my-love-hate-relationship-with-social-media/</link><pubDate>Fri, 26 Jun 2020 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/what-i-learnt-from-my-love-hate-relationship-with-social-media/</guid><description>&lt;p>Just another “normal” morning, slept late last night watching shit content on Facebook and stalking people or looking at old comments on my posts. I put on my glasses, and everything is still fuzzy in my mind. But wait, something’s wrong; something’s missing, it feels like my hands are restless to grab something, oh yeah, my mobile….even though I promised myself that I wouldn’t use my mobile within an hour of waking up. I fail and spend about half an hour checking the seemingly endless amount of notifications that accumulated through the night. Indeed, the guilt due to failing at a personal commitment is just the feeling you want to begin your day with?&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/what-i-learnt-from-my-love-hate-relationship-with-social-media/image-1.gif"
alt="GIF">&lt;figcaption>
&lt;p>Just one more meme, promise :P&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Well, this guilt trip continued to happen on countless mornings, and it made me consider my relationship with social media critically. The honeymoon period of wonder and high levels of dopamine was now over, and it was finally time to put it under a critical lens, to decide if the relationship was becoming toxic. Well, indeed it was! Realising this, led to a metaphorical break up with social media and maybe that lead to this post as a rebound?&lt;/p>
&lt;p>Anyway, before starting this all-out, highly biased rant on social media, let me first explain a bit about my relationship with social media. I won’t go through the details of my social media journey, but for a little perspective, I “officially” came into the world of social media about two years back, fresh out of a self-imposed social isolation of about two years due to my JEE preparation (a highly competitive engineering entrance exam in India) and its suffice to say, from then it hasn’t been anything less of a love-hate relationship, with multiple highs and lows, with me acting like a typical tsundere, both actively criticizing social media and loving it. I have gone from using Facebook the entire day, laughing off memes, chatting with people I didn’t even know well on messenger ( which in hindsight, didn’t result in creating any meaningful connections except a rare few) to running a computer script to unfollow everyone (to make my newsfeed blank) to recently manually unfriending everyone of my 500+ friends and deactivating Facebook forever for good (yep, that was a pain in the ass, performing the same operation 500+ times). Now I just limit myself to WhatsApp, Slack, Hangouts and LinkedIn. Well I know it’s a still a bit hypocritical of me to write this blog, being still a moderate user of social media, but well, whatever.&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/what-i-learnt-from-my-love-hate-relationship-with-social-media/image-2.gif"
alt="GIF">&lt;figcaption>
&lt;p>Pretty sure you must be bringing out your judgement cat on me&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Having experienced the complete cycle of social media usage, I realise that just like all things in the world, it is neither black nor white, its more of a grey slippery slope, where its easy to tumble into the black pit if you are not careful and comparatively difficult to reach the white light at the top. Given that its a complex entity, I first thought about writing something that agrees with the popular opinion of the effects of social media, by researching about it on blogs and watching some TED Talks, but I think that it would be better if I just gave my personal opinions without any external influence, because well, an honest personal opinion is rarer and rarer to find in the current social world. Also though I &lt;em>have&lt;/em> watched a lot of TED talks on this topic in the past, it&amp;rsquo;s around 2 AM in the night when I am writing this and at this time, the things that come in you are what you really think, you can’t lie well even if you tried to :P&lt;/p>
&lt;p>So anyway, I’ll try to explore both the positives and negatives aspects of social media I experienced as well I can, although I’ll admit, being a bit bitter from the metaphorical breakup, I’ll be a bit ( read: highly) biased on the negative aspects:&lt;/p>
&lt;h3 id="positives">Positives&lt;/h3>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/what-i-learnt-from-my-love-hate-relationship-with-social-media/image-3.gif"
alt="GIF">&lt;figcaption>
&lt;p>Wooooohoooo &lt;em>someone&lt;/em> thinks I am worthy!&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;h4 id="acts-as-a-temporary-moodbooster">Acts as a (Temporary) Mood Booster&lt;/h4>
&lt;p>A lot of people, including me use social media as an escape from the real world. The “virtual world” of social media is like a lucid dream, you can control what you see, who you talk to, the world is customised to only show you what you like and it gets better and better at it. It allows you to escape from whatever shit is going on in your real life and indulge in a dream world, obviously boosting your mood. But well, remember, it’s just a dream, you gotta wake up sometime :P&lt;/p>
&lt;h4 id="keeps-you-updated-about-stuff-happening">Keeps you updated about stuff happening&lt;/h4>
&lt;p>Okay, I’ll have to admit this, I started missing out on some important things ( but also on a lot of things that I shouldn’t care about) once I stopped actively using social media. Everyone you know is on social media, hyper-sharing what’s happening around them, what’s cool nowadays, which meme is the most popular right now ( my meme knowledge took a serious hit after quitting Facebook :P) Being an active social media user helps you keep yourself in the loop. You know what your friends are up to and it prevents from the awkward moment when you are the only one who doesn’t know something in a conversation, which is supposedly common knowledge in the social world. Social media is also a great place to get updated on the news, with various news-agencies, political leaders, businessman posting frequent updates about what’s happening in real-time, in a relatively less censored fashion than popular media.&lt;/p>
&lt;h4 id="great-platform-for-personal-branding-digital-marketing">Great platform for Personal Branding/ Digital Marketing&lt;/h4>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="GIF"
src="https://ayushtues.github.io/blog/what-i-learnt-from-my-love-hate-relationship-with-social-media/image-4.gif"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>Personal Branding is important. People like political leaders, actors, entrepreneurs have huge PR teams trying to maintain a consistent social brand of themselves. After all, in this day and age, you are a BRAND!, you need to sell an image of yourself that people approve of and social media is the perfect platform to do so ( although, this creates a sense of alienation from your real self if your social brand is different in value from what you really are). There is also the entire field of Digital marketing, where businesses advertise their products and services through social media ( heck this is how social media websites earn money, by displaying targeted ads to you after auctioning the ad-place amongst various companies).&lt;/p>
&lt;h4 id="great-platform-for-organising-online-fundingdonation-events-petitions-etc">Great platform for organising Online Funding/Donation events, Petitions etc.&lt;/h4>
&lt;p>Well love it or hate it, you gotta give social media this one. There are many instances where collective funding on social media platforms allowed to provide funds to people in dire need for medical funds/donations. ( And I am not talking about the “zombie” money laundering scheme introduced in Breaking Bad..though that was cool :P). Also, social media is a great level playing field. Everyone’s is free to voice their opinions and get it noticed, and when a huge amount of people collectively align on a particular viewpoint against social malpractice, they can use the power virality gives them to voice their collective opinion as an online petition, which does get recognised by the authorities some times.&lt;/p>
&lt;h4 id="gives-a-fake-sense-of-belonging">Gives a (Fake) Sense of Belonging&lt;/h4>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/what-i-learnt-from-my-love-hate-relationship-with-social-media/image-5.gif"
alt="GIF">&lt;figcaption>
&lt;p>Relatable?&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>We all seek social approval. It&amp;rsquo;s rooted deep in our nature due to evolutionary reason. All of us strive for a “sense of belonging”, the feeling that you are connected with people, people who care about you. Social media gives you just that. Doesn’t it feel good when you share something, and people like it on social media? You get instant gratification. You feel that you are connected to your followers/friends, thus satisfying the ever-longing need for belonging. And well, in most cases, it’s good to connect with people you lost contact with due to some reason or the other, but it&amp;rsquo;s also important to realise that at its core this is not a solution, the only way to fill that deep hole of loneliness is to make actually meaningful connections with “real” people.&lt;/p>
&lt;h3 id="negatives">Negatives&lt;/h3>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/what-i-learnt-from-my-love-hate-relationship-with-social-media/image-6.gif"
alt="GIF">&lt;figcaption>
&lt;p>Aah, I have been waiting for this for a long time :P&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;h4 id="it-lowers-your-self-esteem">It lowers your self-esteem&lt;/h4>
&lt;p>Has it ever happened to you, that you opened a social media website and saw people sharing about the new high paying job they got, the perks associated with it, or how they have the most awesome partner, which instantly put you into an existential crisis of what you are doing with your life? Well, I most certainly have experienced that. This is probably one of the major reasons why I quit social media. (I vaguely remember seeing one such post just waking up, and that made me feel depressed for the entire day). Social media only displays the “highlights” of people’s life. People only post the good things that happen in their life, projecting a false image of success. You don’t get to see the real shit that’s happening in their life, the struggle behind the success, the fights in the relationships, you only start comparing your life with the “high points” of other’s lives, leading to increasingly lower self-esteem.&lt;/p>
&lt;h4 id="the-dopamine-addiction-aka-thetwitch">The Dopamine Addiction aka “The Twitch”&lt;/h4>
&lt;p>Everything we do is largely controlled by the feel-good chemical called Dopamine. Heck its also basically how drugs work, by inducing the brain to release huge amounts of dopamine, leading to a “kick”. Social media can ve called a “drug” in a similar fashion, you get a small dopamine kick every time someone likes something you shared, or you get a new notification. Researchers call this “The Twitch”. You get so addicted to the small amounts of dopamine the instant gratification from social media provides you that your brain almost subconsciously “twitches” to engage further in social media to get another kick. Just like drug dealers exploit people, the people behind social media carefully design their apps to get you addicted to it. When you start using it, you get huge amounts of dopamine from it, leading to huge kicks, but soon, being accustomed to these new levels of dopamine, your brain demands an even higher level of dopamine’s to stimulate itself to get the kick, this pushes you to engage even vigorously, trapping you into the voracious trap of dopamine addiction. ( There’s an awesome book called “&lt;em>Hooked: How to Build Habit-Forming Products&lt;/em>” if you want to explore how such products are designed).&lt;/p>
&lt;h4 id="it-tends-to-reduce-focus-and-wastes-productivity">It tends to reduce focus and wastes productivity&lt;/h4>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/what-i-learnt-from-my-love-hate-relationship-with-social-media/image-7.gif"
alt="GIF">&lt;figcaption>
&lt;p>Yep, been there, done that&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>It’s said that the average attention span of a person has reduced to around 5 mins. This means that you can’t focus on a particular thing more than 5 mins, and I believe social media is a major culprit behind this. There is just so much content there, that you keep on switching from one video to another, one meme to another, losing the ability to appreciate simple things. I myself struggle a lot focusing on complex stuff, due to being accustomed to processing information in easy to understand tidbits as is the way on social media ( With tweets being restricted to 140 words or less). Apart from decreasing focus, social media also eats away huge amounts of productive time, you don’t realise when an hour passes by whilst scrolling through Instagram videos or memes. You don’t realise just how much time it&amp;rsquo;s wasting away, once I started reducing my social media consumption, I instantly got more time to work upon some actually productive things, like say writing this blog :P&lt;/p>
&lt;h4 id="its-full-of-shit-content-that-makes-peoplestupid">It’s full of shit content that makes people stupid.&lt;/h4>
&lt;p>Have you ever watched videos of people reacting to, people reacting to, people reacting to something? I mean there’s something seriously wrong about the quality of content that we are wasting our time on social media. People create derogatory videos full of casual sexism, body shaming and profanity just cause these things tend to get more views from users. If you are gonna waste your time, consuming content, then at least watch some decent content. Take memes, for example, they are not even funny most of the times, if not offending. And don’t get me started on TikTok videos. Now I know you must be thinking that these are highly creative and shouldn’t be disregarded as stupid, and well you are entitled to your opinion, but personally, I have never got goosebumps watching any meme or TikTok video, but I did get goosebumps listening to “No time to Die” or watching masterfully directed scenes from Breaking Bad. I personally think that if you are consuming content, then it should be something that makes you think, make you challenge your perspectives and inspires you to do something great.&lt;/p>
&lt;h4 id="promotes-a-fake-sense-of-beauty-and-is-full-of-hate-commentstrolls-promoting-body-shamingsexism">Promotes a fake sense of beauty and is full of hate comments/trolls promoting body shaming, sexism&lt;/h4>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/what-i-learnt-from-my-love-hate-relationship-with-social-media/image-8.png"
alt="Just a gentle reminder to all those trollers">&lt;figcaption>
&lt;p>Just a gentle reminder to all those trollers&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Everyone looks so pretty on social media, people spend hours trying to get that perfect selfie. This intensifies the need for people, especially women to be over-conscious about their appearances developing deep insecurities among them, while in reality most of the people don’t really care about how you look as long as you have a decent personality ( though that doesn’t mean you don’t need to groom yourself). Also, it&amp;rsquo;s full of some A-holes whose only reason for existence is to troll people or spread hate. Its weird, and bit hypocritical when people body shame women on Instagram. I mean its a follower based app, you yourself follow the person to view their photos ( and indeed most of the guys use Insta only for this purpose) and then you yourself body shame the person for their pictures. I mean if you don’t like it, then just unfollow them, they are not asking you to see it.&lt;/p>
&lt;h4 id="makes-it-very-easy-to-influence-people">Makes it very easy to influence people&lt;/h4>
&lt;p>You might have heard about the Cambridge Analytica scandal, which uncovered the adverse extent to which people are influenced by propaganda on social media, which can have huge effects like affecting the presidential results in the US heck, I guess even the Indian general elections were highly influenced by social media (btw there’s a great movie on Netflix called “The Great Hack” on the Cambridge Analytica Scandal). It’s almost scary how much these apps know about us and even scarier that they use this data to subconsciously influence our decisions using targetted ads. Your entire personality is reduced to some numbers, and numbers can be manipulated :P&lt;/p>
&lt;p>Well, like it or not, Social media &lt;em>has&lt;/em> become an inseparable part of our life, and its really hard if not honestly impossible to separate ourselves from it. But still, it&amp;rsquo;s my belief that just like everything, having an idea about how it affects us and its pros and cons can help us improve our relationship with it. I personally feel that I now use social media in a much healthier fashion than before and that is also reflected both in my mental state and personal/professional life. I hope this article at least made you a bit more critical of social media and may improve your own love-hate relationship with it a bit :P&lt;/p>
&lt;p>So, Well until the next post —&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/what-i-learnt-from-my-love-hate-relationship-with-social-media/image-9.gif"
alt="GIF">&lt;figcaption>
&lt;p>And now that Thomas had seen everything, it was time for Thomas to leave :P&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/what-i-learnt-from-my-love-hate-relationship-with-social-media-1d863639303f" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Ah Shit, Here We Go Again!</title><link>https://ayushtues.github.io/blog/ah-shit-here-we-go-again/</link><pubDate>Wed, 10 Jun 2020 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/ah-shit-here-we-go-again/</guid><description>&lt;blockquote>
&lt;p>The past is a puzzle, like a broken mirror. As you piece it together, you cut yourself, your image keeps shifting. And you change with it. It could destroy you, drive you mad. It could set you free. — Max Payne 2&lt;/p>
&lt;/blockquote>
&lt;p>Remember the first day of school whenever a new class-year began. All the teachers made students stand up, introduce themselves and tell everyone about their hobbies. To be honest I legit hated that part since the only possible answer, I had to the question was — “Video Gaming”. Since this was way back in time, before gaming giants like PUBG, CSGO, Apex Legends made gaming famous and mainstream…the teachers straight out used to ridicule me by saying that video gaming was not a hobby ( Lol…I’ll love to hear them say that again :P).&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/ah-shit-here-we-go-again/image-1.png"
alt="Whenever someone asks why you don’t play PUBG/CS GO">&lt;figcaption>
&lt;p>Whenever someone asks why you don’t play PUBG/CS GO&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Still, being a nerd I loved PC video-games, they constituted the entirety of my leisure time as a kid (which also lead to me having shitty eyesight in the current day). I had a PC at my home from as long as I can remember, so I started playing games probably from as little as 1st grade ( Well..to be honest, that mostly used to constitute of me sitting next to my brother and watching him play…but ah well…that was fun).&lt;/p>
&lt;p>I don’t play computer games now, mainly due to two reasons 1.) I use the Ubuntu operating system now which isn’t well suited for games 2.) I am more of a retro-styled gamer who loves single-player games, and as it seems, the current generation of players are hooked on to MMORPG games like WoW, DOTA, PUBG, CS GO which I honestly don’t like (though CS GO is pretty impressive, I played CS 1.6 a lot and Half-Life as a kid).&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/ah-shit-here-we-go-again/image-2.png"
alt="What’s the point of having a gaming PC, when you run Ubuntu on it and can’t play games :/">&lt;figcaption>
&lt;p>What’s the point of having a gaming PC, when you run Ubuntu on it and can’t play games :/&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Despite all this, I thought it would be awesome to write about some of the retro games that I used to love to play as a kid, and which I would still prefer over PUBG/CS GO any day. All of the games I’ll talk about were released before 2002 ( Yeah I know…that’s pretty old and most of you would get cringed by their mediocre graphics…but hey…this is my blog..so fuck it :P). Now I am not a professional game reviewer, so this won’t be a very critical review, so take it with a grain of salt:&lt;/p>
&lt;h3 id="oni-2001">Oni (2001)&lt;/h3>
&lt;blockquote>
&lt;p>“What if a cyber brain could possibly generate its own ghost, and create a soul all by itself? And if it did, just what would be the importance of being human then?”&lt;/p>
&lt;/blockquote>
&lt;blockquote>
&lt;p>— Motoko Kusanagi ( From the mind-bending brainfuck of an anime &lt;a href="https://en.wikipedia.org/wiki/Ghost_in_the_Shell" target="_blank" rel="noopener">Ghost in the Shell&lt;/a>)&lt;/p>
&lt;/blockquote>
&lt;div style="position: relative; padding-bottom: 56.25%; height: 0; overflow: hidden;">
&lt;iframe allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" allowfullscreen="allowfullscreen" loading="eager" referrerpolicy="strict-origin-when-cross-origin" src="https://www.youtube.com/embed/rW4w_7GRUPE?autoplay=0&amp;controls=1&amp;end=0&amp;loop=0&amp;mute=0&amp;start=0" style="position: absolute; top: 0; left: 0; width: 100%; height: 100%; border:0;" title="YouTube video"
>&lt;/iframe>
&lt;/div>
&lt;p>&lt;em>We want an Oni Sequel!&lt;/em>&lt;/p>
&lt;ul>
&lt;li>&lt;strong>Year: 2001&lt;/strong>&lt;/li>
&lt;li>&lt;strong>Genre: Third-Person Shooter, beat ’em up, action, hand to hand combat&lt;/strong>&lt;/li>
&lt;li>&lt;strong>Rating: 7.5/10 (IGN)&lt;/strong>&lt;/li>
&lt;/ul>
&lt;h4 id="overview">Overview:&lt;/h4>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/ah-shit-here-we-go-again/image-3.png"
alt="The first game whose storyline made go wtf.">&lt;figcaption>
&lt;p>The first game whose storyline made go wtf.&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Set in a &lt;a href="https://en.wikipedia.org/wiki/Dystopia" title="Dystopia" target="_blank" rel="noopener">dystopia&lt;/a>n world that is highly inspired by the cult classic cyberpunk anime like &lt;a href="https://en.wikipedia.org/wiki/Ghost_in_the_Shell_%281995_film%29" target="_blank" rel="noopener">Ghost in the Shell&lt;/a> and &lt;a href="https://en.wikipedia.org/wiki/Akira_%281988_film%29" target="_blank" rel="noopener">Akira&lt;/a>, which is so polluted that little of it remains habitable. The government is &lt;a href="https://en.wikipedia.org/wiki/Totalitarianism" title="Totalitarianism" target="_blank" rel="noopener">totalitarian&lt;/a>, telling the populace that what is actually dangerously toxic regions are wilderness preserves, and uses its police forces to suppress opposition. (Kind of reminds me of George Orwell’s 1984). The player character Konoko (who seems to be highly inspired from the Ghost in the Shell’s &lt;a href="https://en.wikipedia.org/wiki/Motoko_Kusanagi" title="Motoko Kusanagi" target="_blank" rel="noopener">Motoko Kusanagi&lt;/a>) learns her employers have been keeping secrets about her past. She turns against them as she embarks on a quest of self-discovery. The player learns more about her family and origins while battling both the TCTF and its greatest enemy; the equally monolithic criminal organization called the Syndicate.&lt;/p>
&lt;p>Now the magic of the game is that it came out in 2001 (just one year after I was born Lol) but the cyberpunk science fiction theme of the games beats the shit out of most of the recent games. The game is an absolute treat to play, with futuristic elements way ahead of its times.&lt;/p>
&lt;p>&lt;strong>Gameplay:&lt;/strong>&lt;/p>
&lt;p>While the game doesn’t boast of high-end graphics, the gameplay is just beautiful. Since the player can carry only one weapon at a time and ammunition is scarce, hand-to-hand combat is the most effective and conventional means of defeating enemies. The player can punch, kick, and throw enemies; progressing into later levels unlocks stronger moves and combos. The game makes you feel like a martial combat guru, having complex animations for each movement (which can sometimes be pretty brutal), the game also features many elements of parkour and typical avoiding of laser-beams (kind of like Prince of Persia)&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/ah-shit-here-we-go-again/image-4.png"
alt="One of the moves in which you had to run towards the enemy and press the punch key and Konoko would just snap the guys head off :P">&lt;figcaption>
&lt;p>One of the moves in which you had to run towards the enemy and press the punch key and Konoko would just snap the guys head off :P&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>There are ten different guns in &lt;em>Oni&lt;/em>, including handguns, rifles, rocket launchers, and energy weapons. The concept behind some of the firearms was highly creative (like the one with the homing spikes or the one-shot one kill guns or the life-sucking biochemical weapon wisp kind of thing). Powerups such as “hyposprays”, which heal damage, and cloaking devices, which render the player invisible, can be found scattered throughout the levels or on corpses.&lt;/p>
&lt;p>There are multiple classes of enemy, each with its style of unarmed combat. Each class is subdivided into tiers of increasing strength. &lt;em>Oni&lt;/em> does not confine the player to fighting small groups of enemies in a small arena; each area is fully open to explore. The fourteen levels are of various sizes, some large enough to comprise an entire building (Bungie hired two architects to design the buildings)&lt;/p>
&lt;p>&lt;strong>Favourite Features:&lt;/strong>&lt;/p>
&lt;ul>
&lt;li>Highly complex plot&lt;/li>
&lt;li>The Dream Drive Level&lt;/li>
&lt;/ul>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/ah-shit-here-we-go-again/image-5.png"
alt="Konko Fighting Herself in a dream in the Dream Drive Level, The game is full of awesome mind-bending levels like this">&lt;figcaption>
&lt;p>Konko Fighting Herself in a dream in the Dream Drive Level, The game is full of awesome mind-bending levels like this&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;ul>
&lt;li>Flexible arm to arm combat and combining that with awesome guns&lt;/li>
&lt;/ul>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/ah-shit-here-we-go-again/image-6.png"
alt="In a particular level, you had to avoid each of these spinning lasers, to prevent yourself from being annihilated by a rogue supercomputer.">&lt;figcaption>
&lt;p>In a particular level, you had to avoid each of these spinning lasers, to prevent yourself from being annihilated by a rogue supercomputer.&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;h3 id="carmageddon-ii-carpocalypse-now1998">Carmageddon II: Carpocalypse Now (1998)&lt;/h3>
&lt;blockquote>
&lt;p>Everyone is a target!&lt;/p>
&lt;/blockquote>
&lt;blockquote>
&lt;p>— good old &lt;a href="https://marvel.fandom.com/wiki/Free_Comic_Book_Day_Vol_2019_Spider-Man/Venom" title="Free Comic Book Day Vol 2019 Spider-Man/Venom" target="_blank" rel="noopener">CARNAGE&lt;/a> from the spiderman comics&lt;/p>
&lt;/blockquote>
&lt;div style="position: relative; padding-bottom: 56.25%; height: 0; overflow: hidden;">
&lt;iframe allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" allowfullscreen="allowfullscreen" loading="eager" referrerpolicy="strict-origin-when-cross-origin" src="https://www.youtube.com/embed/skFN4mFfncA?autoplay=0&amp;controls=1&amp;end=0&amp;loop=0&amp;mute=0&amp;start=0" style="position: absolute; top: 0; left: 0; width: 100%; height: 100%; border:0;" title="YouTube video"
>&lt;/iframe>
&lt;/div>
&lt;p>&lt;em>Sorry for the low quality trailer, Its 1998&lt;/em>&lt;/p>
&lt;ul>
&lt;li>&lt;strong>Year: 1998&lt;/strong>&lt;/li>
&lt;li>&lt;strong>Genre: Racing, Vehicular Combat, Gore&lt;/strong>&lt;/li>
&lt;li>&lt;strong>Rating: 7/10 (IGN)&lt;/strong>&lt;/li>
&lt;/ul>
&lt;h4 id="overview-1">Overview:&lt;/h4>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/ah-shit-here-we-go-again/image-7.png"
alt="I used to play this a lot as a kid…no wonder I ended up being a rash driver :P">&lt;figcaption>
&lt;p>I used to play this a lot as a kid…no wonder I ended up being a rash driver :P&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Now I am no psychopathic schizophreniac (well maybe, yet to be clinically tested for it) but sometimes anarchy and pure Carnage just feels right. Escaping from the “traditional” norms of sanity gives a sense of freedom that’s just wonderful, and that’s just what this game represents. It takes the standard racing format games like Need for Speed and just beats the hell out of it. Being a racing game, it has no storyline per se, but oh well that’s the point, it’s just Madness (btw Madness reminds me of &lt;a href="https://en.wikipedia.org/wiki/Midtown_Madness" target="_blank" rel="noopener">Midtown Madness&lt;/a>…check that out, it was excellent as well). So tune in to the personification of Madness viz Mad Damage and the badass of a car that is the Desert Eagle and wreak havoc!&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/ah-shit-here-we-go-again/image-8.png"
alt="Did come close to killing myself though when I first started driving">&lt;figcaption>
&lt;p>Did come close to killing myself though when I first started driving&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>&lt;strong>Gameplay&lt;/strong>&lt;/p>
&lt;p>Now the primary way to win in a typical race game is to complete a race, but what’s the fun in that? So in this game, there are three ways to finish most levels:&lt;/p>
&lt;ul>
&lt;li>Complete the track, passing through every checkpoint before time runs out — The standard approach in racing games, well arguably no-one plays the game by this standard, at least I never did.&lt;/li>
&lt;li>Waste all the opponents — So yeah, you have to use your car to pulverize opponents, ramming them into traffic poles, walls and whatnot, I can’t describe how fun it is beating the crap out of other vehicles, aiming at them at super-high speed with all types of superpowers infused in your vehicle&lt;/li>
&lt;/ul>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/ah-shit-here-we-go-again/image-9.png"
alt="Desert eagle wasting the shit out of another car ( Lol, check the status — “Twisting In Agony” )">&lt;figcaption>
&lt;p>Desert eagle wasting the shit out of another car ( Lol, check the status — “Twisting In Agony” )&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;ul>
&lt;li>Kill every pedestrian in the level — Well as good old Carnage said, in this game, “EVERYONE IS A TARGET”, so that includes these tiny little human pedestrians inside the game whose purpose of life is just to be either crushed by your car, electrocuted by your vehicle, burnt into minced meat like your car or a combination of these and guess what its fun :P&lt;/li>
&lt;/ul>
&lt;p>The time limits are pretty generous, and the player can kill pedestrians, crash into opponents or perform specific actions to gain extra time. It also featured deformable model removable parts on its cars, making for more realistic dents and crashes. The vehicles can be bent in half, leaving them with fewer wheels on the ground, and can even be sheared in half, causing retirement from the race if the damage is not repaired before the vehicle touches the ground.&lt;/p>
&lt;p>&lt;strong>Favourite Features:&lt;/strong>&lt;/p>
&lt;ul>
&lt;li>Instant Car Repair 😅&lt;/li>
&lt;li>Cool Powerups like Kangaroo Jump, to make your car jump, afterburner for turbo boost and the all favourite Opponent Repulisificator (which shoots springs from your car to the opponents bouncing them away :P)&lt;/li>
&lt;li>Theme Circus Levels&lt;/li>
&lt;li>HELIUM FILLED PEDESTRIANS!&lt;/li>
&lt;/ul>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/ah-shit-here-we-go-again/image-10.png"
alt="Well, you don’t get to see such straightforward descriptions nowadays, do you?">&lt;figcaption>
&lt;p>Well, you don’t get to see such straightforward descriptions nowadays, do you?&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;h3 id="age-of-mythology-2002">Age of Mythology (2002)&lt;/h3>
&lt;blockquote>
&lt;p>Sugar and caffeine. My willpower crumbled&lt;/p>
&lt;/blockquote>
&lt;blockquote>
&lt;p>— Life Lessons from Percy Jackson&lt;/p>
&lt;/blockquote>
&lt;div style="position: relative; padding-bottom: 56.25%; height: 0; overflow: hidden;">
&lt;iframe allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" allowfullscreen="allowfullscreen" loading="eager" referrerpolicy="strict-origin-when-cross-origin" src="https://www.youtube.com/embed/Onik-BD92Ow?autoplay=0&amp;controls=1&amp;end=0&amp;loop=0&amp;mute=0&amp;start=0" style="position: absolute; top: 0; left: 0; width: 100%; height: 100%; border:0;" title="YouTube video"
>&lt;/iframe>
&lt;/div>
&lt;p>&lt;em>Note taken, never touch anything in an old archaeological site :P&lt;/em>&lt;/p>
&lt;ul>
&lt;li>&lt;strong>Year: 2002&lt;/strong>&lt;/li>
&lt;li>&lt;strong>Genre: Real-Time Strategy, Mythology&lt;/strong>&lt;/li>
&lt;li>&lt;strong>Rating: 9.3/10 (IGN)&lt;/strong>&lt;/li>
&lt;/ul>
&lt;h4 id="overview-2">Overview:&lt;/h4>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/ah-shit-here-we-go-again/image-11.png"
alt="This was way before marvel made Thor famous.">&lt;figcaption>
&lt;p>This was way before marvel made Thor famous.&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Gaming isn’t the only thing I love. I love reading for fun, one of my favourite series is the Percy Jackson series, which deals with the adventures of a Greek Demigod son of the Sea God Poseidon called Percy named after the famous Greek hero Perseus, who was well…the Son of Zeus, not Poseidon (weird slip-off by the author Rick Riordan…anyway).&lt;/p>
&lt;p>But that wasn’t the thing that got initially interested me in Greek Mythology. It was this gem of a game, a hugely underrated spinoff of the oh so popular Age of Empires Series (btw &lt;a href="https://en.wikipedia.org/wiki/Age_of_Empires_III" target="_blank" rel="noopener">AOE3&lt;/a> and another similar game &lt;a href="https://en.wikipedia.org/wiki/Rise_of_Nations#Reception" target="_blank" rel="noopener">Rise of Nations&lt;/a> and its spinoff &lt;a href="https://en.wikipedia.org/wiki/Rise_of_Nations:_Rise_of_Legends" target="_blank" rel="noopener">Rise Of Legends&lt;/a> are great as well) based on the exploits of an Atlantean (a member of the mythical greek city of &lt;a href="https://en.wikipedia.org/wiki/Atlantis" target="_blank" rel="noopener">Atlantis&lt;/a>) war hero Arkantos as he hunts down the bad guy (Cyclops?) Gargarensis in his attempt to resurrect the biggest badass out there, the motherfucking Titan of Time big Daddy Kronos, which if succeeded would result in (or as the Norse like to call it Ragnarok).&lt;/p>
&lt;p>Oh and it’s not only the Greek stuff in there, as it so happens, by some dumb coincidence, it also features the same mythologies as the Rick Riordan books namely:-&lt;/p>
&lt;ul>
&lt;li>Greek Gods (&lt;a href="https://en.wikipedia.org/wiki/Percy_Jackson_%26_the_Olympians" target="_blank" rel="noopener">Percy Jackson &amp;amp; the Olympians&lt;/a>)&lt;/li>
&lt;li>Titans (&lt;a href="https://en.wikipedia.org/wiki/The_Heroes_of_Olympus" target="_blank" rel="noopener">The Heroes of Olympus&lt;/a>)&lt;/li>
&lt;li>Egyptian Mythology (&lt;a href="https://en.wikipedia.org/wiki/The_Kane_Chronicles" target="_blank" rel="noopener">The Kane Chronicles&lt;/a>)&lt;/li>
&lt;li>Norse Mythology (&lt;a href="https://en.wikipedia.org/wiki/Magnus_Chase_and_the_Gods_of_Asgard" target="_blank" rel="noopener">Magnus Chase and the Gods of Asgard&lt;/a>)&lt;/li>
&lt;/ul>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/ah-shit-here-we-go-again/image-12.png"
alt="Okay, that will take a while to sink in">&lt;figcaption>
&lt;p>Okay, that will take a while to sink in&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>So fasten your seatbelts as you control a massive army of diverse mythological characters like half-bull half-human minotaurs, multi-headed hydras, fire breathing giants, the world-dragon Nidhogg, one-eyed Cyclops, Egyptian mummies and various famous Greek heroes like Odysseus, Achilles, Ajax, Chiron in this fantasy world.&lt;/p>
&lt;h4 id="gameplay">Gameplay:&lt;/h4>
&lt;p>Typical of RTS games, it has a stage-wise evolution of the civilization. Consisting of 4 stages&lt;/p>
&lt;ol>
&lt;li>Archaic&lt;/li>
&lt;li>Classical&lt;/li>
&lt;li>Heroic&lt;/li>
&lt;li>Mythical&lt;/li>
&lt;/ol>
&lt;p>For advancing into each stage, you needed to build some specific buildings and collect some resources. At each stage you choose between two minor gods to worship, each having its advantage, including their myth units and god powers ( A god power is a one-shot superpower that you can use)&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/ah-shit-here-we-go-again/image-13.png"
alt="I used to go with Hel, though a pack of Fenris Wolfs and the Fimbulwinter power could destroy entire towns">&lt;figcaption>
&lt;p>I used to go with Hel, though a pack of Fenris Wolfs and the Fimbulwinter power could destroy entire towns&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>There are four resources food, wood, gold, favour. Food can be extracted from farms or by herding/hunting; wood is collected by cutting trees (yeah that sounds like captain obvious), gold is obtained from mines and also from trading between markets. Favour is received by worshipping at temples, creating monuments, killing people and stuff and is used to create mythical units.&lt;/p>
&lt;p>The army consists of infantry, cavalry, archers, heroes and mythical units. Each having an advantage over one of the others and having a disadvantage from others. So you had to keep a balanced army to defeat your opponents.&lt;/p>
&lt;p>The campaign storyline is incredible and full of twists, although the difficulty level is not that high, it still requires some patience to complete them.&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/ah-shit-here-we-go-again/image-14.png"
alt="Awww YEAAAAH!">&lt;figcaption>
&lt;p>Awww YEAAAAH!&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>&lt;strong>Favourite Features:&lt;/strong>&lt;/p>
&lt;ul>
&lt;li>Titans :P&lt;/li>
&lt;li>God Powers ( one-shot powers you get from choosing a god, like Zeus’s thunderbolt for one-shotting any unit, earthquake to create well…an earthquake)&lt;/li>
&lt;li>PROSTAGMA!!&lt;/li>
&lt;li>Colossus, Hydra, Giants and Tons of other cool myth units&lt;/li>
&lt;li>All the detailed mythological history in the description of each unit&lt;/li>
&lt;/ul>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/ah-shit-here-we-go-again/image-15.png"
alt="A typical game scene with some cyclops, hydras, minotaurs">&lt;figcaption>
&lt;p>A typical game scene with some cyclops, hydras, minotaurs&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;h4 id="warcraft-iii-reign-of-chaos2002">Warcraft III: Reign of Chaos (2002)&lt;/h4>
&lt;blockquote>
&lt;p>&lt;em>The beginning of wisdom is the statement ‘I do not know.’ The person who cannot make that statement is one who will never learn anything. And I have prided myself on my ability to learn.&lt;/em>”&lt;/p>
&lt;/blockquote>
&lt;blockquote>
&lt;p>— Thrall (the savage looking orc in the left, who is actually quite nice)&lt;/p>
&lt;/blockquote>
&lt;div style="position: relative; padding-bottom: 56.25%; height: 0; overflow: hidden;">
&lt;iframe allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" allowfullscreen="allowfullscreen" loading="eager" referrerpolicy="strict-origin-when-cross-origin" src="https://www.youtube.com/embed/c2vm-QjK2xQ?autoplay=0&amp;controls=1&amp;end=0&amp;loop=0&amp;mute=0&amp;start=0" style="position: absolute; top: 0; left: 0; width: 100%; height: 100%; border:0;" title="YouTube video"
>&lt;/iframe>
&lt;/div>
&lt;p>&lt;em>Why is there a raven in every fantasy series?&lt;/em>&lt;/p>
&lt;ul>
&lt;li>&lt;strong>Year: 2002&lt;/strong>&lt;/li>
&lt;li>&lt;strong>Genre: Real-Time Strategy&lt;/strong>&lt;/li>
&lt;li>&lt;strong>Rating: 9.3/10 (IGN)&lt;/strong>&lt;/li>
&lt;/ul>
&lt;h4 id="overview-3">Overview:&lt;/h4>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/ah-shit-here-we-go-again/image-16.png"
alt="For all the DOTA fans, here’s the one that started it all">&lt;figcaption>
&lt;p>For all the DOTA fans, here’s the one that started it all&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Well the Warcraft platform requires little introduction and honestly, I wasn’t going to cover such a popular game and wanted to focus on less known games, but this one holds a special place in my heart, and I couldn’t control myself to write about it. (As a matter of fact, I am watching someone play the game while writing this :P)&lt;/p>
&lt;p>So, meet the game that gave rise to massive successes like DOTA, WoW, this game is a revolutionary masterpiece that made the genre of Real-Time Strategy popular and also is single-handedly responsible for popularizing the concept of Massively Multiplayer Online Game (MMOG).&lt;/p>
&lt;p>It tells the story of the Burning Legion’s (basically a huge, huge evil army) attempt to conquer the fictional world of Azeroth with the help of an army of the Undead, led by fallen paladin Arthas Menethil. It chronicles the combined efforts of the Human Alliance, Orcish Horde, and Night Elves to stop them before they can corrupt the World Tree.&lt;/p>
&lt;p>The game features four campaigns, each focusing on a different faction (Human, Orc, Undead, Night Elf) namely:&lt;/p>
&lt;ul>
&lt;li>&lt;strong>Human Campaign — The Scourge of Lordaeron&lt;/strong>&lt;/li>
&lt;li>&lt;strong>Undead Campaign — Path of the Damned&lt;/strong>&lt;/li>
&lt;li>&lt;strong>Orc Campaign — The Invasion of Kalimdor&lt;/strong>&lt;/li>
&lt;li>&lt;strong>Night Elf Sentinel Campaign — Eternity’s End&lt;/strong>&lt;/li>
&lt;/ul>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/ah-shit-here-we-go-again/image-17.png"
alt="Sorry for the lame analogy, couldn&amp;rsquo;t help myself :P">&lt;figcaption>
&lt;p>Sorry for the lame analogy, couldn&amp;rsquo;t help myself :P&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;h4 id="gameplay-1">Gameplay&lt;/h4>
&lt;p>Warcraft III, like other RTS games, has a resource-based system. Having three primary resources food, lumber and gold. It features four playable factions: The Human Alliance — a coalition of humans, dwarves, and high elves — and the Orcish Horde — composed of orcs, trolls, and minotaur-inspired Tauren — return from the previous games while the Undead Scourge and the Night Elves were added as two new factions. Each race has a unique set of units, structures, technologies, and base-building methodology.&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/ah-shit-here-we-go-again/image-18.png"
alt="A screenshot from an undead game">&lt;figcaption>
&lt;p>A screenshot from an undead game&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Also, &lt;em>Warcraft III&lt;/em> adds powerful new units called heroes. For each enemy unit killed, a hero will gain experience points, which allow the hero to level-up to a maximum level of 10. Progressing up a level increases the hero’s attributes and also allows the hero to gain new spell options (bringing role-playing video game elements to the series). Individual hero abilities can boost allied units. All heroes can equip items to increase skills, defence, and other skills. At level six, the hero can obtain an “ultimate” skill that is more powerful than the three other spells that the hero possesses. Heroes can also utilize the various natural resources found throughout the map, such as controllable non-player characters, and markets in which the hero can purchase usable items. Often, hero units become the deciding factor in determining a winner.&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/ah-shit-here-we-go-again/image-19.png"
alt="Guilty as charged 😅">&lt;figcaption>
&lt;p>Guilty as charged 😅&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>The game’s multiplayer mode introduced anonymous matchmaking, automatically pairing players for matches based on their skill level and game type preferences, preventing cheating and inflating their records artificially. The sole objective in multiplayer games is to destroy all the buildings of the opposition. In default melee matches, players can pick their heroes, and losing one will not end the game. To make the game proceed more quickly, by default the map is fully revealed but covered in the fog of war.&lt;/p>
&lt;p>Players can also host custom games, using maps either created in the Warcraft III World Editor, or the default multiplayer scenarios. The map editor allows a variety of custom maps to be created, such as several tower defence and multiplayer online battle arena maps, the most notable of which was &lt;a href="https://en.wikipedia.org/wiki/Defense_of_the_Ancients" title="Defense of the Ancients" target="_blank" rel="noopener">&lt;em>Defense of the Ancients&lt;/em>&lt;/a> &lt;em>aka DOTA&lt;/em> (And that peeps is how it DOTA happened)&lt;em>.&lt;/em>&lt;/p>
&lt;p>&lt;strong>Favourite Features&lt;/strong>&lt;/p>
&lt;ul>
&lt;li>In-game Cinematic Cutscenes (It almost feels like a high-budget movie)&lt;/li>
&lt;li>Illidan Stormrage (One word — Badass)&lt;/li>
&lt;li>Thrall’s FOR DOOMHAMMER!!&lt;/li>
&lt;/ul>
&lt;div style="position: relative; padding-bottom: 56.25%; height: 0; overflow: hidden;">
&lt;iframe allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" allowfullscreen="allowfullscreen" loading="eager" referrerpolicy="strict-origin-when-cross-origin" src="https://www.youtube.com/embed/sotg3le_EqU?autoplay=0&amp;controls=1&amp;end=0&amp;loop=0&amp;mute=0&amp;start=0" style="position: absolute; top: 0; left: 0; width: 100%; height: 100%; border:0;" title="YouTube video"
>&lt;/iframe>
&lt;/div>
&lt;p>&lt;em>These are just so freaking cool!&lt;/em>&lt;/p>
&lt;p>Though I only covered these four games in this blog, there are many other awesome games that I loved playing, like:&lt;/p>
&lt;ul>
&lt;li>Metal Fatigue, Netstorm&lt;/li>
&lt;li>Black &amp;amp; White 1,2; Zoo Tycoon&lt;/li>
&lt;li>Total Overdose ( a mixup of PoP, GTA and Max Payne)&lt;/li>
&lt;li>Star Wars Knights of the Old Republic 1,2; Star Wars Battlefront 1,2&lt;/li>
&lt;li>Halo Combat Evolved, PoP series&lt;/li>
&lt;/ul>
&lt;p>And many many more, it would be a pretty long list if I write them all down. Unfortunately, people today don’t remember these classic games, it’s still a treat to play them (whenever there are no compatibility issues).&lt;/p>
&lt;p>So that’s all for this post. Hope you liked it and it made you relive some awesome old memories. Since I am just starting writing blogs, please let me know how I can improve my writing skills in the comments. Until then, So Long!&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/ah-shit-here-we-go-again/image-20.gif"
alt="GIF">&lt;figcaption>
&lt;p>It’s never really over, is it?&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;h3 id="sources">&lt;strong>Sources&lt;/strong>&lt;/h3>
&lt;ul>
&lt;li>Wikipedia&lt;/li>
&lt;li>Countless hours of playing these games :p&lt;/li>
&lt;/ul>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/ah-shit-here-we-go-again-35dae2acf8c4" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item><item><title>Some shit I randomly thought up :P</title><link>https://ayushtues.github.io/blog/some-shit-i-randomly-thought-up-p/</link><pubDate>Thu, 09 Apr 2020 00:00:00 +0000</pubDate><guid>https://ayushtues.github.io/blog/some-shit-i-randomly-thought-up-p/</guid><description>&lt;blockquote>
&lt;p>When you want something, all the universe conspires in helping to achieve it.&lt;/p>
&lt;/blockquote>
&lt;p>Sounds familiar right? It sure did to me. Now I am by no means a fan of SRK, I don’t even watch Bollywood movies anymore (the last one I watched was Chhichhore somewhere around 5 months ago and before that Jolly LLB2 and that was 3 years ago) but before I got a bit older and supposedly became too cool to watch Hindi movies, I had sure as hell watched Om Shanti Om as a child and enjoyed it thoroughly 😅.&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/some-shit-i-randomly-thought-up-p/image-1.png"
alt="#Plagiarised">&lt;figcaption>
&lt;p>#Plagiarised&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>There you go — It was a pretty good dialogue, wonderfully delivered by SRK. But little did I know back then, that I would find the same dialogue, years later in a completely unexpected scenario.&lt;/p>
&lt;blockquote>
&lt;p>Just a little disclaimer — I don’t really write for others to read, I write, cause, well there are many reasons for why I write, maybe I am bored, or I might have too much on my plate and want to switch off from it for a while. So this is not gonna be a well-designed article. It will have a lot of unrelated personal stuff about me. It’s my habit to say whatever I am thinking at the moment to people whether or not it’s related to the talk, so I can’t completely get rid of that habit while writing also*😅*&lt;/p>
&lt;/blockquote>
&lt;h3 id="a-little-bit-of-info-about-meokaymaybe-not-so-littlep">A little bit of info about me…Okay …Maybe not so little :P&lt;/h3>
&lt;blockquote>
&lt;p>Feel free to skip this part if you are someone who already knows me or isn’t interested in knowing me …it’s my views on myself …its gonna be biased by my hubris anyway&lt;/p>
&lt;/blockquote>
&lt;p>I am Ayush Mangal…a normal 19-year-old guy, hailing from a relatively small city called Gwalior in Madhya Pradesh, India, although I am currently living in this beautiful small town called Roorkee in Uttrakhand, India for my college life. I really love small cities because of their slow pace and simpler life, I lived in Banglore for around a month some time ago and yeah …being in a big city is great and all …the thrill of seeing those huge buildings/malls with all those great shops which have anything you want is great and all, but to be honest, I prefer a simple and quiet environment over the hustle and bustle of these large cities …maybe it has something to do with the introvert part of me. But ah well …it seems like I’ll inevitably have to move to a huge city someday …for a better life, whatever that’s supposed to mean.&lt;/p>
&lt;p>Anyway, I usually come about as this extremely boring guy who cares about nothing else but his academics, has little to no hobbies, has a virtually non-existent social life, and is, in a nutshell, a good for nothing except for maybe getting some marks in exams. and no, I am not pessimistic, I have actually asked many of my old friends of how they remember me …that’s what they usually say or mean to say but are reluctant to say it to my face.&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/some-shit-i-randomly-thought-up-p/image-2.gif"
alt="GIF">&lt;figcaption>
&lt;p>Every Fucking person who meets me&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>And yeah that’s true to a large extent, as noticed by many people around me who actually know me well, the single most dominant trait about me is my habit of overthinking, and over-worrying and that habit (which also gave birth to this post) has made me into a person who is lost in thoughts of the future and never really lives the present (yeah I know, pretty cheesy stuff). I think my mind has been conditioned by the older people around me since I was young to believe that the only way to a secure future is through studying the hell out of your childhood and my brain, being the fickle little devil it is, took that as the truth, and although I now know that it’s not true, I am honestly too afraid to come out of my little comfort bubble that has been developed for the last 10–12 years or so.&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/some-shit-i-randomly-thought-up-p/image-3.gif"
alt="GIF">&lt;figcaption>
&lt;p>I went full-on overthinking mode even while selecting this gif&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>As for my social life, I don’t know, I don’t like the question — “Are you an introvert ?”. Something as complex as someone’s personality can’t be binarised into a dichotomy or even into a trichotomy including the third ambivert category. Typical “extroverts” also have some introverted tendencies, and the reverse is true too as well. Maybe I’ll write someday about the complex spectrum of introversion and extroversion. That’s something I have thought about a lot.&lt;/p>
&lt;p>Anyway, I have kind of a simple rule&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/some-shit-i-randomly-thought-up-p/image-4.gif"
alt="GIF">&lt;figcaption>
&lt;p>No …..not this rule…although I wouldn&amp;rsquo;t talk about it if it even existed :P&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>I usually don’t care enough about anyone to take the first step to start a conversation, that often leads to the impression that I am, I don’t know, rude, proud or simply a jerk and that lead to me finding it hard to make friends in a new place like my college, especially with the opposite gender, for which there is supposedly some unspoken law that it’s necessary for me to start the conversation, leading to me having made little to no friends amongst them in my entire life. And although I usually don’t care about it much, it does make me sad sometimes that I must be missing so much because of it, cause, to be honest, I really like talking to people, I just can’t bring myself to disturb someone I don’t know well with a random conversation.&lt;/p>
&lt;p>Also, even though I really am good for nothing except studying, there are still some other things that I love. I love reading a lot. It’s my favourite timepass during the boring lectures at my college or whenever I have to wait somewhere alone. I also love listening to music, and you can always find me with earphones attached when I am at my room (much to the annoyance of my parents)…or sometimes even in the classroom, Also, I do recommend you to try Youtube Music, I usually listen to songs on Spotify, but I tried Youtube music whilst writing this blog and take my word for it, nobody gives recommendations better than youtube, well maybe Google does, but google owns Youtube too now :P&lt;/p>
&lt;p>I am also a self-proclaimed otaku, having watched anime daily since the last 5–6 years, I got this habit back in 9th standard when I joined my coaching classes and met some people who used to watch animes. So I also got into the habit of watching animes ( I started with Inazuma Eleven if I remember correctly) and soon got hooked to them ( there were time’s when my mother caught me watching anime at 6 AM when I woke up early to study😛). It was pretty sad that when I came to college, people didn’t seem to really watch anime over here, and kind of think of it as something childish and cartoonish. Although I was still able to connect with some people who loved anime as much as I did and yeah…that was pretty cool.&lt;/p>
&lt;p>Okay, …that turned out to be quite a lengthy introduction, but I really don’t care, I wrote whatever came in my mind, and you know what …it feels awesome to express your mind through writing. Especially during this lockdown period when there isn’t really anyone to share your thoughts with, Anyway, I know this all feels pretty out of place, its cause, it is. I didn’t really plan this post out very well and this kind of happened on a whim….okay for that matter, let me tell you how this post happened.&lt;/p>
&lt;h3 id="how-this-post-came-out-tobe">How this post came out to be?&lt;/h3>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/some-shit-i-randomly-thought-up-p/image-5.gif"
alt="GIF">&lt;figcaption>
&lt;p>#Day 30 of COVID&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Okay so this is my first blog post ever, and like a lot of other things I have tried out….it sort of happened randomly&lt;/p>
&lt;blockquote>
&lt;p>I wrote this post in sort of in a non-linear fashion time, so there would be stuff like last night, yesterday, tonight, which actually correspond to different days, also talking about non-linearity, checkout this awesome anime called Baccano, it focuses on a bunch of seemingly non-linear stories which all make sense in the end :P&lt;/p>
&lt;/blockquote>
&lt;p>So last night, I had a talk with a really old school friend after a long time ( I sometimes talk with random people…who I haven’t even talked for a long time, it’s fun seriously….reconnected with many peoples through this), she is an awesome person, I have talked with her a few times in the last year and she’s like that type of person who tries everything out in life and has a very exciting life and also is cool enough to converse with a random guy like me who could also be a potential creepy stalker for all she knows :p ( Oh I really hope she doesn’t read this lol….if she is reading this right now….then ah well…fuck it …doesn’t matter 😛). Anyway, I discovered in the convo that she also writes some original stuff ( which is totally awesome btw) and she was also good enough to share some of it with me and yeah, sure enough, she wrote some awesome stuff!&lt;/p>
&lt;p>Now seeing her write all this awesome stuff made me also try writing out, So I decided to write about something. But what should be that something? Well since it’s my first post…and you always remember your firsts….I wanted it to be about something that is special to me….and that would be books/anime/music (Yup I know ….I am a pretty boring guy). I wanted to go with anime…since it’s less represented in blogs and stuff and I really want to improve that situation ….but that turned out to be quite tough to pull out …. so I thought I’ll just write about books&lt;/p>
&lt;p>Now if memory serves well, it was during the midterms examination of my 3rd semester that I came across this small book — The Alchemist — By Paulo Coelho.&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="" srcset="
/blog/some-shit-i-randomly-thought-up-p/image-6_hu9828039085114454480.webp 400w,
/blog/some-shit-i-randomly-thought-up-p/image-6_hu14689778392207179644.webp 760w,
/blog/some-shit-i-randomly-thought-up-p/image-6_hu5292682628317764414.webp 1200w"
src="https://ayushtues.github.io/blog/some-shit-i-randomly-thought-up-p/image-6_hu9828039085114454480.webp"
width="649"
height="464"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>I was going through a phase of my life when I had recently tried to cut off watching animes and quit social media ( I was a total social media addict back then, be it FB or Instagram) and I’ll be honest, I was doing quite well for someone like me who makes a new rule to improve himself every day and then takes the saying that “rules are meant to broken” quite literally 😂. ( Lol …I never seem to improve though…its quarantine time now and I am binge-watching Psycho-Pass currently)&lt;/p>
&lt;p>Anyway, doing so left me with a lot of free time, which I used to try out some new stuff, like learning guitar, hitting the gym regularly( I know I am making it sound like I was being super productive and stuff but, believe me, that’s far from true, I haven’t done anything since the morning and finally came to write this blog since I was bored as hell….but hey…you gotta put up a good first impression, don’t you 😛).&lt;/p>
&lt;p>It was during this time that I re-started my habit of reading books and novels and stuff, something I did a lot during my 9–12th — No, not my course books, I am talking about series like Percy Jackson ( I am a huge-huge Rick Riordan Fan — I have literally read all the similar series written by him), Harry Potter ( I don’t like cheesy epithets like Potterhead, Sherlocked and similar stuff people put on their Insta bios, but yeah I loved the books and also went around reading 4 fanfictions on James Potter as well) and other similar titles.&lt;/p>
&lt;p>But I had left this habit when I came to college — cause well, nobody likes a nerd who likes to read books for fun, and I was already nerdy enough, to begin with. Anyway, I started reading books again when my second year started, cause well — after having lived the first year of my college life in a place with only 5–6 people around me….I had stopped caring about people’s opinions, cause there weren’t many people to give their opinions around me anyway 😛.&lt;/p>
&lt;p>So I had already read some books that semester and was all set to start another one. I googled a bit and found people going crazy about a book called “The Alchemist”. So I decided to give it a try. Now I am by no means a fast-reader. I usually take ages to complete a book and if I remember I had my midterms in a few days. I usually go full-on study mode during that period. But this was different. I got hooked to the book. I have never got so intrigued by any other book before and to my surprise I binged the book and completed it within a day.&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/some-shit-i-randomly-thought-up-p/image-7.gif"
alt="GIF">&lt;figcaption>
&lt;p>Me after reading the book 😛&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>I decided the book to be The Alchemist. To do justice to the review I read the book again, and although I already knew what was about to happen in the book from a plot perspective. I got hooked to it again. Now that’s the beauty of the book. Its plot doesn’t have much importance, it’s the concepts, the lessons in the book that are brilliant. But anyway I’ll give a brief plot summary of the book.&lt;/p>
&lt;h3 id="the-part-that-is-actually-about-the-alchemistp">The part that is actually about the Alchemist :P&lt;/h3>
&lt;p>So the protagonist is Santiago, a young shepherd, and a hopeless romantic who is having a recurrent dream about a treasure stuck somewhere near a pyramid in Egypt. He gets his dream interpreted by an old woman, and uh …meets …an avatar of God? (Wtf, I am an atheist dude, why did I even like this book). Who tells him that it’s his destiny to find the treasure. He says …cool…thanks for the tip about the future dude! He takes the advice of the random stranger, sells away all the sheep he has, goes to Egypt, then meets another man, who tricks him into giving away all his money (the boy wasn’t a bright one, to begin with really). The boy says cool…I’ll just work somewhere to get more money. (Ah how I wish I was this optimistic), He then works in an old crystal shop, gets pretty wealthy. Joins a caravan going to Egypt. Some shit happens and he meets a mysterious guy called the Alchemist, who teaches him some hippy stuff and converts the boy into an alchemist. (People who can turn pretty much anything into anything, like lead into gold, human into the wind and also are immortal due to some liquid they developed). He then reaches the pyramids and realised there that the treasure is at the same place where he was having his dreams.&lt;/p>
&lt;p>Okay, now I know this was a pretty shitty description of the plot. Cause well …the plot doesn’t really matter, its the quotes in the book that are brilliant. Oh and one more thing, I’ll majorly be focusing on the first half of the book, cause after that the book focuses on stuff like love and some other fantasy shit which I find pretty irrelevant 😅&lt;/p>
&lt;p>So what I basically did while re-reading the novel is that I extracted all the quotes that I found interesting in a google doc. I got around 23 quotes, now obviously I can’t cover 23 quotes in a single blog post, and given the current COVID circumstances, I don’t even know if I’ll even live long enough to write a second part.&lt;/p>
&lt;p>I had to somehow compress those 23 quotes, so I clubbed some of them which were based on similar concepts and also asked one of my closest friend, who also happens to be a pretty awesome writer to select the top 7 quotes he found to be best, I finally ended with 5 sets of quotes.&lt;/p>
&lt;blockquote>
&lt;p>Everyone seems to have a clear idea of how other people should lead their lives, but none about his/her own .&lt;/p>
&lt;/blockquote>
&lt;blockquote>
&lt;p>I am like everyone else — I see the world in terms of what I would like to see happen, not what actually does.&lt;/p>
&lt;/blockquote>
&lt;blockquote>
&lt;p>When you possess great treasures within you, and you try to tell others of them, seldom are you believed.&lt;/p>
&lt;/blockquote>
&lt;p>When Galileo tried to convince the world that the earth is heliocentric, the world declared him a heretic, when Elon musk started selling flamethrowers people thought he was just a lunatic ( maybe he is 😛), anyway, the point is all of us are a bit too judgemental.&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/some-shit-i-randomly-thought-up-p/image-8.gif"
alt="GIF">&lt;figcaption>
&lt;p>Okay ….maybe its not such a good idea to believe him 😂&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>I have some friends who recently started working on their own startup, they asked me to join in to help develop it, they really believed in their idea, but to be honest I just straight away ridiculed their idea. Fast forward around 2 months, their idea won an ideathon, got invited to the finals of Hult Prize Scholarship( a 7 crore funding) at Jordan (although that got cancelled due to corona 😅), but the thing is, they found something good, something worth exploring, they tried to tell me about it, but I just saw everything in the way I would like to see. I thought about how the technical aspects seemed primitive but couldn’t see the bigger picture they were focusing upon, cause that’s what I wanted to see, not what they wanted me to show ( I hope they don’t read this though….that would be embarrassing….I used to go to the gym with one of them everyday lol).&lt;/p>
&lt;p>Coming to the first one, whenever someone asks me for some advice…whether it be for academics/coding/general shit or for that matter sometimes even their relationships (which I have no experience in, and thus shouldn’t really comment upon, but still do when people ask), I always think that it’s so obvious what their problem is and that the solution is pretty obvious, but when it comes to my problems, which are often mostly focused around my career, I have no idea what to do about them, I keep stuck on my problems for months to no avail.&lt;/p>
&lt;blockquote>
&lt;p>At a certain point in our lives, we lose control of what’s happening to us, and our lives become controlled by fate. That’s the world’s greatest lie.”&lt;/p>
&lt;/blockquote>
&lt;p>Okay, so this is an idea I highly believe in&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/some-shit-i-randomly-thought-up-p/image-9.gif"
alt="GIF">&lt;figcaption>
&lt;p>Me as an atheist: To me kya Karu, Job Chhod du?&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>No, not this crap, I seriously do believe that whatever happens to me in life, is just a result of my actions and I am wholly responsible for it. Although my friends still always have to listen to me rant about how unfair life is, and how everyone seems to be hellbent on fucking up things for me,….. but hey, that’s just my coping mechanism :P&lt;/p>
&lt;p>Deep down, in my mind, I know that it’s all a result of my own actions, I don’t really believe in the notion that everyone has a pre-destined path made by some oh so almighty God and we’ll just have to accept it, I really hate it when people, like my parents, say “Bhagya ye zyaada aur Samay se Pehle, kisi ko kuch nahi milta” (Nobody gets more than what fate has decided for him/her and before the right time comes). I mean like this is the worst possible philosophy to live your life with, heck even physicists agree that the world is non-deterministic in nature ( Sorry Classical mechanics ….you suck!).&lt;/p>
&lt;p>Though I do believe in all this, I do know that yeah…sometimes it feels like there isn’t really much you can do about the things that are happening to you, I mean I am not really one to talk about it…since I am really at a lot more privileged position than many other people in the world ( for which I am really grateful for actually) and I haven’t actually experienced any real hardship in life per se, but still I have watched interviews or read about so many people ….like really successful and famous people, who all had pretty major setbacks in life, enough to make anyone feel as if their fate doesn’t want them to be successful, that every single one of them fought back against that so-called fate and came out stronger than before cause oh well&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/some-shit-i-randomly-thought-up-p/image-10.gif"
alt="GIF">&lt;figcaption>
&lt;p>And hey …..why would I ever lose control of my life, I can end it anytime I want ( A little bit of dark humour is fine right? 😛)&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;blockquote>
&lt;p>Once again he saw that in that strange land, he was applying the same lessons he had learned with his sheep.- All things are one&lt;/p>
&lt;/blockquote>
&lt;p>All things are one — That’s the basic concept behind alchemy as introduced in the book, they say that everything is the same at its core, you can look at a grain of sand and see in it all the marvels of creation, and since all things are one, lead can be converted to gold, since they are essentially the same thing. Now, this was something from the second half, which didn’t make much sense to me. But there is something in it, as mentioned by Santiago in the above quote that did make sense, that solutions found in problems faced in one aspect of life, could be actually found to be applicable everywhere.&lt;/p>
&lt;p>Let me give an example, most of the research that has happened in my field has been developed by studying nature, by observing nature and applying lesson&amp;rsquo;s learnt from it into developing fucking AI’s ( Ah well I know this might reinforce my already nerdy image…but well …don&amp;rsquo;t really care)&lt;/p>
&lt;p>I have also observed it on a personal level, people who tend to explore things a lot, also tend to be quite good in things than normal people ….lesson’s learnt in one task help a lot in other tasks ( Heck there is even an entire research area called Multi-Task Learning in my filed entirely based on this concept), I have tried to apply some organisational skills learnt during my intern and via contact with some more experienced people to other aspects of my life, which does seem to have improved my productivity a bit.&lt;/p>
&lt;blockquote>
&lt;p>It’s called the principle of favorability. When you play cards the first time,you are almost sure to win. Beginner’s Luck. Why is that? Because there is a force that wants you to realise your personal legend. It whets your appetite with a taste of success.&lt;/p>
&lt;/blockquote>
&lt;blockquote>
&lt;p>Making a decision was only the beginning of things. When someone makes a decision, he is really diving into a strong current that will carry him to places he had never dreamed of when he first made the decision.&lt;/p>
&lt;/blockquote>
&lt;blockquote>
&lt;p>When I have been truly searching for my treasure, I’ve discovered things along the way that I never would have seen had I not had the courage to try things that seemed impossible for a shepherd to achieve&lt;/p>
&lt;/blockquote>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/some-shit-i-randomly-thought-up-p/image-11.gif"
alt="GIF">&lt;figcaption>
&lt;p>Me whenever I win at a new game :P&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>Now I don’t really know if this actually works for everyone, cause, as a matter of fact, I consider myself to be an extremely lucky guy, and no I am not being grateful to God or something for all the happiness in my life and feeling lucky ( I am a strict atheist for that matter, although I studied in a catholic school and was brought up in a Hindu Family 😅…I know, weird flex :P)….I am talking about real blind luck, the type of luck that gives you the confidence to randomly guess 20 questions in one of the most important exams in your life, like you know it’s gonna work out anyway for you since you are so damn lucky 😛 ( I am not kidding, that really happened during my MAINS paper…. I thought there was 1 hr remaining, but the teacher announced there are 5 mins left, and lol, I had around 20 questions unattempted and I just randomly filled them and well its enough to say that…they saved my day :P)&lt;/p>
&lt;p>But, I have felt this many times, whenever I start something new, I get a bit lucky. ( I remember when I first played UNO a few years ago, I got super lucky and got these great set of cards….although I still lost cause well….I am a bit dumb and the other guy ….who is currently in IIT KGP ….was smart as hell) I get some success at the beginning, and that little extra-dopamine makes my brain go crazy and keep on exploring that thing even when the initial dopamine high fades and I start facing the actual excruciating hardness behind the task that requires actually loving the work to complete irrespective of success 😅&lt;/p>
&lt;p>The other two quotes can not be more true, I’ll just give a personal experience, so basically I decided to explore a particular field of computer science a bit when I was back in my second semester, that was nothing new, many people explore a lot of stuff during their first year. (It’s really awesome and scary at the same time that seemingly small things like just a chat with a senior have such a long-lasting impact on our life) But as I got more and more interested in the field, I had to finally make a decision, whether to continue in software development( which is what people usually do at my college) or focus on the not so common field that I was pursuing. I chose the latter cause I loved it, and boy I couldn’t have imagined what would happen because of it. I got to meet a whole new group of awesome people in my college who were interested in the field, got to do a research internship at IISc Bangalore ( Oh btw just yesterday I got selected for a summer internship there again 😛) which was more or less something I always fantasised about even during my 11–12th, during which I also participated in a hackathon by Microsoft in Bangalore that had many senior developers and start-up founders and some big investors and other high profile people I couldn’t even imagine to interact with otherwise. ( Fuck man the guy who was sitting on my table in front of me turned out to be the founder of a startup…oh and he also won the hackathon 😛, there was also this one lecture by the Vice President of Google AI … I mean literally starstruck moment) Just a few days back I had a meeting with a professor in the US, who was a person who had connections with people who were the pioneer of research in my field and people like the CEO of IBM etc. I also have an interview with a PhD at CMU scheduled today night (and yeah I am writing a blog instead of preparing for it, I am too stressed at the moment to study anyway……PS..I got selected :P), Recently a month ago I just casually mailed a lab at IIITD to check if they had some project …..and what do you know ….here I am one month later… all set to publish my first research paper…..I couldn’t have even imagined talking to such awesome people when I started exploring the field just cause I was bored.&lt;/p>
&lt;figure>&lt;img src="https://ayushtues.github.io/blog/some-shit-i-randomly-thought-up-p/image-12.gif"
alt="GIF">&lt;figcaption>
&lt;p>😛&lt;/p>
&lt;/figcaption>
&lt;/figure>
&lt;p>( Also this was not meant to be a paragraph about me bragging about the amazing stuff that I have done, or how successful I am, believe me, there are days when I want to just stop doing everything, never open anything even remotely related to my field ever in my life, go back home and I don’t know, figure something else out because I am so frustrated by the shitty way things were turning out😅….I just wanted to give a personal experience of the amazing things that a seemingly simple decision led me to)&lt;/p>
&lt;blockquote>
&lt;p>I’m afraid that if my dream is realised, I’ll have no reason to go on living. I don’t want to change anything, because I don’t know how to deal with change. I am used to the way I am.&lt;/p>
&lt;/blockquote>
&lt;blockquote>
&lt;p>Every blessing ignored becomes a curse.&lt;/p>
&lt;/blockquote>
&lt;blockquote>
&lt;p>I don’t want anything else in life. But you are forcing me to look at wealth and at horizons I have never known. Now that I have seen them, and now that I see how immense my possibilities are, I’m going to feel worse than I did before you arrived. Because I know the things I should be able to accomplish, and I don’t want to do so.&lt;/p>
&lt;/blockquote>
&lt;p>Okay, so this one hit close to home. You see, I am currently in college, and you know how college is right? The possibilities are limited only by our own imagination. There is always someone around you who is doing things that you think cannot be possibly done by you and you come in contact with these people…you know that type…..who are perfect in time management, who are supposedly effortlessly excelling in everything….the cool guys…no not those jerks who try to act cool…the real cool guys who don’t need to explicitly express how cool they are ( I could give some examples from my personal life, but if there is ever a chance someone from my college reads this, I’ll like to prevent embarrassing myself more than I already have 😅).&lt;/p>
&lt;p>Anyway, the thing is, after meeting people like this I am usually put in a self-conflict. Being the overthinker that I am, I can usually figure out what are the things that they are doing differently that makes them who they are, and I know that if I incorporate those habits I can also become like them. And I am not saying that I don’t try to become like them, after coming to college I have tried to imitate many of such people I have met/read about and yeah I usually quickly see some results ( Beginner’s Luck 😛), but after some time….I just stop and revert back to my original shitty state of life. And then begins the cycle of guilt, I know that I should be able to improve myself and that it would be really good for me, I’ll make new friends, explore new horizons, get in good shape, and above all increase my self-worth but as said in The Alchemist, I don’t want to change, I don’t know why, but I just don’t want to change, I am so used to this shitty life that I can’t deal with anything new. I like being in the small comfort zone that’s my room, with my laptop, binging Animes, seeing weird videos on Youtube that make no sense at all and would make someone believe I have gone crazy if they found out.&lt;/p>
&lt;p>Okay, so that’s it for the post. For anyone reading the post, who made it till the end. Thanks a lot for wasting your time on this utter BS!. If you are someone who I know well, please let me know how you found the article. If you are an old acquaintance and we haven’t talked for a long time, let’s talk again!. All thanks to COVID-19, I have all the time in the world to listen to you 😛&lt;/p>
&lt;p>
&lt;figure >
&lt;div class="flex justify-center ">
&lt;div class="w-100" >&lt;img alt="GIF"
src="https://ayushtues.github.io/blog/some-shit-i-randomly-thought-up-p/image-13.gif"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;hr>
&lt;p>&lt;em>Originally published on &lt;a href="https://medium.com/@ayushtues/some-shit-i-randomly-thought-up-p-fbf0e17eed82" target="_blank" rel="noopener">Medium&lt;/a>.&lt;/em>&lt;/p></description></item></channel></rss>