Guide to Scaling A Startup with Zero Funding | Kata 1

A lot of the early team here have figured out through a lot of pain, if you don't have distribution, it's not going to work. You need to generate and share wealth with enough people because without incentives, you lose your best talent. You're constantly losing your best talent. You've taken completely unorthodox bets that other people haven't taken. And hopefully I'm able to show you some of the vision that we had three years ago. And then I'll also tell you the vision for the next, you know, few years. Because we have so many challenges, we've just started drowning in opportunities. I remember I sent an email to Sam Altman when I was... Just out of college and I was running my first startup, I had sent him an email in 2015. Sam was one more of those rejections. It's easy to believe you are right about everything when you're very young and you're untested. But once you've been beyond the test and you run companies, some have worked, some have not worked, it's very hard to believe anything is permanent. So anyway, I have been thinking very deeply about what I should do, right? Considering now that the business has gotten fairly large, we have many different verticals, we're working on so many things at once. I've been thinking very deeply about what I should do and I thought I should go back to what I do best and that is to write. So I went back and I said what have I learnt over the last 3-4 years and what of those are repeatable principles. Like I've changed my mind on a lot of things in the last 3-4 years as I've got more experience, more exposure, met different different people and I feel like All of the things I've learned in the last 3-4 years, if I went back and told 21-year-old Varun Maya, here's the things you've learned, I think it would have been very, very valuable to me. Because as I've said before, I think social media is mostly very young people talking to each other because by the time people get older, they stop using social the same way young people use it. And the main reason is a lot of the opinions that older people would have, and I'm 32 now, but a lot of opinions older people would have, young people would not really agree with. So older people say, why go through the Janjat? Let me just say things that everyone agrees with and end it there. So I feel like I wanted to do a series on training the best talent that we already have here. And that's why this is not the entire company. This is a segment of the entire company. But training the best talent and seeing what of what I've learned over the last three, four years. And not just from things I've learned myself, but also things I've learned from other people, from books, from everything that I can now relate to. And what can I grant you so that you can do better in your careers? Because actually there's this Venn diagram, right, of your career and what we do as a company. And I think if we can make that overlap as hard as possible, ideally as close as possible, then both you and the company can grow. And I've now seen one cycle of this, right, with the early people at EOS, who have now grown quite a bit. So, the 16 cut hours, I have 16 of these videos, okay? where I'm going to be teaching you everything I've learned about managing people, about people, about technology, about ideas, about content. And I hope that this is useful to everyone watching, not just, you know, in real life, but also, you know, on the screen. There's going to be some things that I'd say that will make a lot of sense to people in the real world here, which might not make sense to people on screen and vice versa. So please bear with me. This is going to be the easiest kata. You know what a kata means? Anyone know what a kata means? Yeah, it's like a karate series of steps that you practice, right? So there are 16 of these. The way I've structured it is every fourth video is going to be highly produced or we're going to do some real world experiment. And the other, like every one, two, three set is going to be very information dense, very information packed. This is not the kind of content we do on reels or short form or anything. This is very, very dense and some of you might not be able to, you know, understand. So please feel free to put your hands up. And I will hopefully stop everything for you and answer those questions. The end goal of this is you take back something in your life that you can actually use. And I think that at this point, my role is to make as many more of the people standing on that side as possible. And that's why we're doing this. So this cutter is actually called how to build a company. It's actually a little bit misleading because we're not actually going to be... teaching you how to build a company in this one. But across the 16 cutouts, maybe you'll learn it. In this one, I'm going to tell you our journey and our story. What were the insights and the ideas that led to getting us where we are? We're close to about 400 employees now. And we have close to 40,000 square foot of office space. We are completely bootstrapped. And we have been profitable from almost the first month, right? So we have been profitable now for three years. Yeah, close to about three years we've been profitable. Whatever we've done has worked. And we've taken completely unorthodox bets that other people haven't taken or would find very strange. And hopefully I'm able to show you some of the vision that we had three years ago because I think we've stuck to that and we've reached where we wanted to reach. And then I'll also tell you the vision for the next few years. Awesome. Let's go to the next slide. Three years ago, I'd made a deck and we were in a house. I'd made a deck for the small core team that we had and I said, I really want to be the best company in video. And I think we'll have to just lock the door from now. But yeah, I want to be the best company in video and later on in video games. And yes, this was part of the deck three years ago. And we want to do it 100% bootstrapped. And the reason we want to do it 100% bootstrapped is it was just such a weird vision. And I knew that nobody would... Three years ago, when I made a video, I made a video called Yes, AI will take your creative job. Sure, it was a little bit of, you know... slightly more like titles are always supposed to be like that but the video is very balanced about how yes AI can be creative and at that point nobody believed AI could be creative this was pre the era of mid journeys and everything we have today right so we had a vision then and I think it played out almost precisely how we thought it would play out we really believed in distribution three years ago now I think everyone has woken up to the idea of distribution so hopefully I'm able to show you the play-by-play next slide I just want to show you a small fraction of our customer base today. This is across advertising, this is across the different businesses we have, some are enterprise clients, some work with us, you know, year-long deals, multi-year deals. This is just a very, very small fraction. And, you know, now I think it's like hundreds of companies we work with. So it's been phenomenal, but, you know, it's, everyone asks the question, how do you get the first one? How did the first one convert? How did you get the second one? How did you know you actually have enough going for you to go from the 10th one to the 20th one? So we'll talk about all of that. Next slide. I want to give you like, you know what a bento box is? Anyone know what a bento box is? Yeah, it's like a box with like, it's a cute, you get these now. Like you can order food and you'll get like, you know, sambar in one side, you'll get like rice in the other side, you'll get a little section for chapati. We have a bento box version of EOS. as an ecosystem. So it started with talent. We have two things that generate talent for us, that create talent for us. One is AVTV, which all of you are familiar with. Now AVTV has a channel. AVTV has a video editing school. So the video editing school is pretty large now. This number is a little bit old, but it's a pretty large school right now. And it helps us produce a lot of editing talent for ourselves. I'll tell you where in a second. the other side is 100x engineers which is not actually a core business but started off as an investment we invested in 100x engineers the company and on the services side as you all know we have yas right which does youtube as a service where we work with brands all the way from zoho to clear trip to zero to vodafone and on the services side we have another entity called aos labs which is the labs division which was the one that figured out the avatars in the early days built all the apps that we built in the early days and now we've we've laser focused on we build technology and workflows but we build it for content right that's our sort of video and content is sort of our forte and of course labs has clients like amazon prime and rcb and a bunch of others we'll talk about all of this in a bit on the other hand we have distribution all of you at some point came into the company or know about the company because of the distribution we have so We have the Varanmaya channel, we have AVTV, the channel side, we have Breakdown, we have Overpowered, we have Full Disclosure, 100x Engineers, we have many more, we have Mr. Nerf, so many other channels, but I'm just, I didn't want to make this a very, very large bento box. Then we have apps. In the early days, we built a lot of apps, so we had built God in the Box when GPT's API first came out. That went very viral. Then we have Alpha CTR to make YouTube thumbnails, and then we had Video Vault, which we've recently launched, and I'll talk more about Video Vault in a bit. And then we have another emerging segment, right? Which is potentially could be very, very large, but we have to do the groundwork that we did three years ago, which is build the team, build the infra. So we have EOS games, which is very much in the proving stage. And then we have Headshot, which is the mobile gaming controller, which again, not run by us. We're investors in the company, but Headshot is a very, very long shot, right? If it works, you know, you could potentially change how consoles. are consumed in the country because this is a country which can't all afford a PS5. Anyway, next slide. You all have seen the office, but I think people on the internet are probably seeing it for the first time, but we have a really killer office. Sometimes I'll walk into the cafeteria and people will be playing like guitar and stuff, which is very, very cool. But you know, two years ago I had so much more control over all of this. Two years ago I'd be like, oh yeah, we'll do a guitar night on Saturday. And it's so weird for me. to walk in a room and be like, oh, you guys are doing this. Like the first time I saw, there's this board you have, right, in the yard section where, you know, there's a timing for Batman 10, there's a timing for squash, bunch of other games, right? I was like, wait, they did this by themselves. So it's very, very cool to watch a company evolve, at least from my eyes, because I remember when it was a tiny little baby. Anyway, next slide. So. Coming back to the insights, why did we start what we started? Why did we do what we do? We kind of did the opposite of what Reddit and Twitter were saying back then, back three or four years ago. And very few people remember the narrative, but I remember it very clearly, right? How many of you remember Don't Build Rappers? Okay, a lot of people believe, remember that? A rapper is basically something where... You use an existing GPT model or Claude model or whatever and you build something on top of it, right? You basically build some app that people can use but you're not building the foundational models. So I said don't build wrappers, we build wrappers. Lot of people told us at that time if you want to build a big company, don't do services. We built two services companies. So clearly, you know, we're stupid. Then lots of people said moat is in the product. How many of you remember that? Remember? and if you have a product you know you can keep it defensible from everybody else we realized that the moat is in access to talent and distribution right we knew that if you get good talent they will build great products right if you if you get good talent and enable them then uh you know um again product is the most important we we sort of figured out and actually we a lot of the early team here have figured out through a lot of pain through previous companies that if you don't have distribution it's not going to work And this is a shocking realization for us because all our life, all the books we've read, all the stories that have been told about successful people, it's like, oh, they built the greatest product ever. That doesn't mean you ignore product. You still build a great product, but you focus on the distribution and getting in the hands of enough people. How many of you have been told, do a single thing at once? A lot of people tell you that, but it's actually true. And there's a very important nuance here of doing a single thing versus doing multiple things. We'll talk about it. But we did. several small parallel bets with very high failure tolerance, right? If it fails, it's okay. Next, we'll do the next thing. So we had the ability to do multiple things, but also if something failed, we didn't beat ourselves up too much about it because we're like, this is the natural process of learning, right? Next, a lot of people told us raise money and I've raised money in the past, but this time I wanted to do it bootstrap. I just believe that there was such a big shift that I didn't know how to convince anyone on like anyone outside on what we were trying to build. People hadn't realized yet at that time that distribution was important. So I stopped trying to convince people. I was just like, we'll do it, we'll prove it. And maybe when we prove it, we'll be able to raise money. But it turns out we didn't need money because if we prove it, then what do you really need money for, especially in a company like ours. The last one is interesting because every young kid in Bangalore today who wants to work in these worlds, tech or content, wants to work out of HSR. I have no idea why we built in North Bangalore. I mean, I do have an idea. We were not doing, I mean, we didn't have that much money back then. So this felt like a cheaper place to move to Hebal. Suddenly, Hebal became expensive. I have no idea why. But we wanted to be away from the unprofitable startup mentality, right? Like, you might think, who wants to be an unprofitable startup? Like, who wants to work in an unprofitable startup? we now have people who for 10 years of their lives have worked in unprofitable startups going from one to the other and it's almost like I think some of the talent has figured out that a lot of these unprofitable startups just spend a lot of money so it's better for me to be here so you know I'm able to make lots of money quickly and by the time the company dies I go to the next one right so we saw a lot of that behavior in HSR because I was also friends with some of these people I'll be like that's like why don't you work hard to make the company grow but no one's really interested in doing that, right? Like they were like, this is the game, this is the meta game, you have to play it. And we were just like, let's just get out of this mentality, right? Let's just go build first principles. Next slide. And that's what I'm coming to, right? Which is first principles thinking, which is remember that all the cool things that we were supposed to do back then, we sort of did the opposite of it. And it comes down ultimately to something called first principles thinking, which is we questioned every known idea on how to build a company. Every known idea on everything, how to do content. Like, people told us you can't build 50 channels. And you know what? If someone three years ago had told me I'll build 50 channels, I wouldn't have believed them. Like, how? Do one first properly. Right? But it was first principles thinking. Like, whatever we believed would be possible, like, let's break it down to its individual components and then figure out can we solve each component one by one. And you have to have the tolerance to fail, right? And I had failed enough times in my life where I was like, who cares if one or two more things fail? and you have to be willing to relearn and discard ideas that don't work. Not all ideas work and some of the ideas that could work are overdone. Like for example, you know why lots of kids want to build like a social network like me many years ago? Because they all watched the social network movie, right? And they're like, I want to do that. But here's the thing, how old is that movie now? How many years has it been? How many people have tried after that? So just goes to show you that the world shifts very often. And you need to know when the world is shifting to be able to sit there and be like, okay, I think the world is shifting. Everything is going to be different. Can I re-see everything from scratch? We call that first principles thinking. And in some ways, first principles thinking is the opposite of status thinking. Status thinking is, what I will do, right? And the exact opposite of it is, forget about what other people think. And when we were in Hebbal, we were so disconnected from the world. I actually, I have a joke, right? I just made sure my core team couldn't go to Indranagar. and HSR and that solved the problem. So, yeah, and here's one last thing which people really underappreciate, which is we had zero funding. Everything you see around you is completely, totally bootstrapped. So we were fully aware every day that nobody was coming to save us if we were wrong. If we were wrong, I knew that we were dead. Like there was no choice. That's why we had to keep our costs low. That's why we never overhired. It's just like you do all the right things, right? Like if the goal is to if you're going to die next week, if you're not able to make money. Next slide. Everything's broken down into four acts. And then there's a future act, which is what are we doing now? I'll tell you about act one, which is AV. We figured out an insight, which is that the average person was spending five hours a day on video. I would meet friends and they were like, they're spending five hours on Instagram, YouTube. That's ridiculous. Like five hours a day on video. And it's not like text. They were not reading blog posts like five, 10 years ago. They were watching videos. And maybe this is because of Jio, maybe this is just like a behavioral shift, phones got better, internet got cheaper. But they were spending a lot of time on video. And I used to often ask these people, because I'm an engineer by training, right? I'm a software engineer by training. I studied computer science and engineering in college. And people would, you know, in 2015, people would ask each other, which new app did you download? Oh, Tinder. Oh, Hinge. And there'd be new versions of these apps, right, in every category. But... I was going to people and I was like, what new app have you downloaded? This is like three years ago, three, four years ago, right? And people couldn't give me a concrete answer. People were like, you know what? I've not downloaded so many apps. And that's because I'll tell you why. There's what I call category saturation. Like now that you use either Spotify or YouTube Music or Apple Music, why do you need another? So that category is filled. And Spotify can easily, you know, if there are micro categories, if there are like audio podcasts, let's say, they can do all of those things, right? Same with food. Have you downloaded any new food delivery apps outside of Zomato and Swiggy? No, until Quick Commerce came in. Until, like, there was a big shift, nobody, like, everyone stays happy with their one or two apps, and most people have made choices, either it's Swiggy or it's Zomato. And it's happened in every industry. And I was like, oh, people are not downloading more apps, then why are we, like, why should I make more apps? Then I asked the question, how many new YouTubers or Instagram accounts have you followed? And that was insightful. Because people would show me like a new YouTuber that they had suddenly started following. They'd be sharing videos with each other, some new kind of content. And I was also noticing I was watching videos. I was like, wait, this is a new, this is, fundamentally it's not a new behavior because watching videos has been around for a while. But watching so much video and following people. Do you know one thing? 10 years ago, I thought it was cringe to follow an influencer. Like, Achina is a good example. My wife is a good example, right? She's like, why are you following an influencer? So stupid. So this behavioral shift had happened where people had started following other people and not just because they were friends, because in Facebook era, people used to follow people because they were friends. But now they were following creators and influencers and they were spending a lot of time. So I'm like, what company produces for this time? And if you ask the question, what company is producing specifically for video? The answers at that time were either it's like these young creators who are trying and at this time, you know, lots of content creators in India would just start picking up. Pandemic, you know, had just, you know, we were in the sort of early days of the pandemic. But on the other side, on the professional side, there were news media and they were very TV focused. They were very like, hey, TV, because we know advertisers are there and we don't want to experiment with this new platform. God knows if they're going to be advertised on YouTube and the CPMs on YouTube are very low. Even now they're very low, but it wasn't interesting for a brand to go into. But first principle thinking was like, what is happening on TV should happen on YouTube, right? It makes sense. If advertisers are coming there, of course advertisers are going to come here. It makes total sense. Maybe we are early. So fine, we'll take an earlier bet on this than everybody else. Next slide. What is the scarce resource in a content gold rush? What is the scarce resource? Exactly. It was the video editor. And you know what? It was so dumb, we didn't know this. We started making videos. We wanted video editors. That's when we found out. Like, I would go to Okus and I'd be like, hey, do you have video editor friends? And he'd be like, he'd give me like one contact, right? And then he'd give me like two or three contacts and then he'd be like, I'm out. And you'd put up a LinkedIn post, but you'd get a very, like, nobody wanted to be an editor this time. Editors would pay like 5-10k a month. It's not a serious career. Nobody cared about it. So I think here's where... You know, Achina really played like a good move because this is her brainchild, but she said, I'm going to make a great channel. And then once I have that channel, I'm going to make sure the channel has the best editing. The best editing. Okay, and because Ocus was involved and Martin was involved, Ocus and Martin said, we know editing. We'll edit this like very, very well. So they started editing these videos superbly well. And it became sort of AV's brand, right? And you should see old AV videos and now, right? Like it's evolved so much. But in the early days, they would edit themselves, they would make these great videos. And when we were hiring, we would struggle so much that one day, Okus and Martin and Achina were chatting, and Achina was just like, hey, I think we should teach video editing. It was not a thought-through decision. It shows you how much of life is, you just get started, and then some idea comes, and then you work on it. It's never these planned journeys that, oh, you know, this will happen, this, that. Because all my plans for what EOS should be started after we figured out this insight. Right? So, next slide. We built the world's largest video editing school. The world's largest, sitting out of Hebbal. You know why? It was the uncool thing to do. I often joke that for two years, we had no competition because nobody thought it was valuable to go train video editors, even though you were consuming so much content. Somebody has to make it, right, on the other side. But you were consuming content and most people didn't have the foresight to know that somebody has to produce this. And I saw there was a total gap there, right? So we started focusing on this. Go to the next slide. As of today, we've done nine cohorts, maybe 10. I don't have the exact number. It's 300 to 400 students a cohort. It's actually, this number keeps changing. I keep changing this on the deck every few days, but it's now 3000 plus video editors. We have exceptional placements. And again, placements are like, it's fully supply demand. Everyone started their own podcast. Everyone, every brand wanted to start doing content. creators sprang up, creators started making money. And that allowed us to have this constant pipeline of talent going, right? And it started with very low numbers on salaries. Now it's like super competitive. I'm sure most of you have realized this, right? Once you join this company, there'll be like 10 different companies pinging you on LinkedIn saying, please join us. Because it's become competitive now. It's a space, right? Because everyone wants to make content. But yeah, I think this was Achina's insight. And I feel it's one of the smartest things we've done. And also the second smartest thing we did here is not to make it like an EdTech. Not to say, that was never part of the plan. I was like, I don't want to become an EdTech. We have learned enough from Baiju and all the others. Let's just do a small, tight thing where we give the best quality, where we only focus on outcomes as much as possible. And then we'll think of other ideas. We'll build a media network instead. Let's not just be like, you know, just think in one dimension. Next slide. And of course, today we have thousands of potential employers slash recruiters hiring from us. Next slide. But then the next piece came in, and I think this is the fundamental difference between us and anyone else who would be running a startup or raising money. If they'd raise money, the investor would say, do more of this. Tisa banao, 300,000 banao, 10 million banao. And I was like, that doesn't make sense. We should own more along the stack. So we built our own channels. We said, if we have the... this scarce resource in a gold rush. Let's just build as many of our own channels as possible. Let's build our own media network because I had fully bought into the idea of distribution. In my past journeys, I had learned how building a great product, spending years building that product, and then no one caring about it, and then my competitor running 10 times as much ads as me. I was just like, we had lost. So we're like, we will build our own distribution. So we built our own distribution. We built, you know, AVTV first, and then my channel. And these are all old numbers, right? My channels maybe had 9 lakh subscribers now. Now we're at over a million on Instagram. We built AVTV, we built AVTVDaily, we built AVVideoSchool, we built Overpowered. We invested in 100X, but we taught them how to do content. We did Breakdown, which Rahul Mathur now does. And we just said, let's just build these different verticals, right? So we built full disclosure for careers. In every vertical, we built a channel, both long form as well as short form. There is this Charlie Munger book called Poor Charlie's Almanac, and we'll talk a lot about this book, right? Where he says all the benefits of a company or an individual come at scale. Because once you have so many channels pulling so many views, your inboxes are flooded with opportunities. This is the thing that people don't get. When you get to a million followers, you probably get like 30, 40, 50 emails a day. And all over, some will be trash. Okay, some will be some... hello, sir, and then nothing beyond that. But some will be like the US Embassy reaching out and saying, hey, do you want to do something together? So we just, because we had so many channels, we've just started drowning in opportunities. And some of those opportunities, we just picked up on and built businesses out of it. Like, I'll give you an example. One of those opportunities, by the way, if you can get to scale, is when ChatGPT or OpenAI launch images in India, GPT 4.0 images in India, they'll do a collab reel with you. So really, we got to... millions of views on this one and we do it routinely now but but for me this was the moment right where I remember I'd sent an email to Sam Altman when I was just out of college and I was running my first startup I'd send him an email in 2015 he had replied to me it was all generic I mean I was I wanted money at that time I was trying to raise money I'd emailed like 70 investors Sam was one more of those rejections in a way but to see that with scale now you get those same opportunities So that was very cool for us. Next slide. But Act 2 is where I think the opportunities start materializing. I told you a lot of these inbound opportunities come in YouTube by itself or Instagram by itself doesn't make you that much money unless you get a massive scale. But Yast was where we finally saw the opportunity to do something cool. Next slide. So some of you might have seen these channels on your feed. Some of you haven't, but let me tell you they all come from the same place. This is Atlas for the Visa company. This is Clearly Tripping for travel. This is Builder Central for no code. So we started making these channels for our customers, right? Like Zoho had reached out to us and they said, hey, can you do a channel for us? Can you experiment? And when we did the channel, we were able to get it to scale. And we all spent a lot of time at that point, right? From an editing perspective, from a writing perspective, from a thinking perspective, from a channel management perspective, writing all our secrets down. And because of that, once you start writing these things down, you can scale them. Person number two can do it. Person number three can do it. And then you hire well. Right? Beyond this, once you hire well and you offer them the playbook and you have this ecosystem or this environment, people are able to look next to each other and be like, oh, that worked on that video. Fine. You know, now I will also copy that technique. Start working and go to the next slide. And then of course, you know, we start doing many different types of channels in many different forms, right? Markets by Zerodha, we do editing for it on long form. NRI Shala, which is for NRIs. Vitamin Pop for Vodafone. Fincredibles for one of the banks. So just like suddenly the, like YAS became very, very large, right? And our biggest advantage here was, of course, we had a very, very large video editing school. So we were able to take the best video editors, put them into YAS and say, hey, we now have like very, now our placement rate obviously shot up, right? Because we needed, the amount of editors we needed, even the school couldn't produce enough, right? Because remember, we had artificially locked the school to be smaller. for the quality purposes. But it just became really awesome that we could take talent, put them here, combine them with a good writer, good channel manager, and it was scaling so quickly. Next slide. And of course, this is just a small sample of the people we work with in YAS. Close to 40 channels now on YAS. And it's been a phenomenal run. The company went from being like 50, 60 people to now close to 400. And of course, the space requirements of... of having so many editors, of building that culture, it's been crazy. Next slide. Now look, if we had done so far, if we had done this much so far, I think all would be great, right? Like we would have become a decent company and we would have got somewhere, but I was having an existential crisis. Like, dude, I'm an engineer. I studied computer science in college. I should be doing something a little more technical or at least find a way in this new ecosystem that I've built to infuse technology. And I feel this is actually also a very strong moat because most people who do well at content are not as technical. So they never think, can I use technology to make content better? But look, I tried DALI the day it came out, like when I got early access. I knew that at some point there will be this collision, right, of our field in technology. And technology enters everything. The one thing you need to know about technology is technology in that way is... cannot keep its hands off any field. So, technology comes everywhere and technology came into this. So, what we did is we set up an entity called Labs under AOS. We called it AOS Labs. Harshan Tejas, you know, said, fine, we'll run it. And Harshan Tejas just set it up as an experimental hub. Like, we'll try everything. We'll fine tune models. We'll make our own model. We'll make our own model for thumbnails. We will, at the same time, put GPT on WhatsApp. We will just have fun, right? And because Tejas is an engineer, and I'm an engineer, and Harsha is an engineer, it's easy for us to come up with new MVPs. To come up with new projects and say, okay, does this work, does this not work? So, next slide. So this is the video I made. It's now been three years since this video, but can you imagine going out and making a video saying, yes, AI will take your creative job, but the contents of the video were very balanced, that AI is coming, but person that uses AI will do better. I got absolutely thrashed in this video. Because before you have ever seen any generative AI tool, I had seen DALI. That's the only thing that changed my mind, right? Like I get early access, so I'd seen it, I'm like, oh, this is coming. But other people hadn't tried or seen it yet. It's a very strange position to be in, where you know something, you've tried it, you're like, guys, this is good. And everyone else is like, no. But I don't think this company would have been possible if the AI tools hadn't come out. The unit economics just wouldn't make sense. Next slide. Feb 18, 2023, one of the experiments that labs did, was this experiment where we were trying to do something called script to creator. I have written a script. Can we convert this into a video and audio? So we used an open source model that time called Koki for the audio and on the video we used something called wave to lip. And I had taken, for this particular video I had taken a Telugu, some Telugu saying that the first thing I found on YouTube and I stitched it together with, you know, this thing that modifies my lips and I had an output. That was very bad. Like some of the people that I showed it to, like laughed at me. They're like, you can't make videos like this. You look like a clown making videos like this. But life is strange, no? And when tech comes out, especially new technology comes out, it evolves very fast. So, next slide. A year later, we were crushing it. We were absolutely crushing it. And I was finding that I wasn't shooting videos. I was like, wait, we are productized being a creator. Which is a very very new thing by the way nobody had seen this and I've been making videos for nine years nine ten years now So I'd never seen this new thing where it was even today. It's surreal sometimes I'll tell a China. I can't believe this is my voice I Can't believe this is my video and you know what the number one thing that people told us at that time Nobody will watch an AI generated presenter remember this wasn't an AI generated video It's an AI generated presenter. How many of you believe nobody would watch AI generated presenters five years ago be honest So everyone else thought, people watch AI generated presenters. Everyone else believed it. I didn't believe it. I was like, come on, it's going to look a little bit inauthentic, but we were trying anyway. And it got to such scale, and I told you, all the benefits are in scale. So we got to such scale on this, that it became, like for us, it became sort of the thing that we got good at, right? Because we had trained, discarded, trained, discarded, we tried so many different models. And then we finally figured out a workflow that even today, Maybe today it's gotten easier, right? But two years ago, it was like nobody believed it would be possible. Next slide. Today, and this was in December, I don't know about now, I went to Davos. Jan, actually, I went to Davos, the World Economic Forum. It's my first time ever being invited. I had heard so much about the World Economic Forum. And in Davos, YouTube has a book, YouTube's Guide to Davos. And in there, they mention that we are the world's largest AI avatar. So imagine random team in Hebal trying random shit. And then one of those things like the world, how the world works, right? It's called the power law. One of those things just skyrockets so much and I was like, this is the simplest thing we worked on. But it skyrocketed so much and it only works because we were clever enough to edit over the parts that didn't make sense. And we were only able to do that because we had a video editing school next door. I don't know how much, what the rankings are now and I don't even know if, you know, there's a serious ranking around this, but we still do hundreds of millions of views a month on just my channel. Except now we do so many channels. using this technique where the script is still human-written, there is a little bit of, you know, some of the channels might use AI, but mostly it's human-written, combine it with AI avatars, combine it with our video editing school, and you have this productized service that just scales so well. Next slide. It's an example we did for Unacademy, right, where we compared and, you know, the problem statement at that time was, hey, can you make this virtual presenter for long form? This is almost a year point something ago, right? where we compared the non-AI and AI views and there's no difference. Views, likes, comments, there's no difference. So we just, for us, it's like imagine having a thesis that maybe we don't know if people watch AI generated avatars, we can try. But suddenly it works out and the views are the same and people can't tell and like so many things change for us, right? Because when you know this and the world doesn't believe it yet, even though we had proof and we had these comments and the world would see it and I'd show people, right? And they'd be like, I can't Even they were like, I can't believe this works. Next slide. And then of course, because we built a brand around this, everyone knew what we were doing. We branched, especially labs branched out to all the deep faking, advertising, call use cases with full belief that the models themselves are commoditized. If there's a good model, because we now have experience of this, right? Like a model will come out and people will be like, this is the greatest model ever. And another model will come out. And then we'd be like, Both of them will drop their prices. So there was no moat in building the model anymore. And we thought very hard in the early days should we build our own model. We'd raise money, we'd hire people, we'd build our own model. But we couldn't find a use case, a value of it. Even today when VO3 came out, we said such an amazing video model. And within a few weeks, Kling came out with something that's reasonably good, right? So we knew that the models were commoditizing even back then. We've seen this pattern hundreds of times. And we branched around to all the use cases, right? Next slide. On Amazon Prime, we did the face swapping with MBCS. We did Alpha CTR for thumbnails. Next slide. I don't know if you guys have seen the Mirzapur campaign, the face swap campaign, where on WhatsApp, you send your face and we have options, which is either use Buddhi or use Bal. We did that. We did the technical implementation of it. So Labs was this team, which was focused on content, but using technology. Very weird combination. Next slide. And now we built something called Video Vault, where, you know, We do thousands of videos a month across YAS and the media network, right? So we're like, how do we handle that? How do we give comments? How do we give feedback? Now, there are tools online called Frame.io. But Frame.io is online. Imagine the average podcast now is hundreds of GBs. How do you upload it online, then download it? It's a pain, right? So it should be happening in our office. We were using a NAS in our office. It was a network-attached storage in our office. We said we should build our own software for it. So Labs built that. So Labs is the experimentation hub. It also has a commercial entity because without a commercial entity, you know, as a company, they'll be like, oh, I'm just running experiments. Young people are fine with it, but as you get older, you're like, where's my career growth? So you have to do the commercial stuff to be able to continue to, you know, afford better talent or keep your talent. We've today done stuff for Bangalore Police. We're actually, we're going to reveal a project we've done for Bangalore Police, which is outside of avatars for them. It's actually a calling use case. We've done the RCB stuff. All the Bangalore police avatars that you've seen, we have played a role in. Next slide. And then finally, you know, the question I asked at that time is, if AVTV provides talent to YAS, and that keeps the engine running, what is providing talent to labs? Because hiring is a problem, right? Like, you don't find good talent, just go out there and find good talent. And we were not well known, we had not raised 100 millions of dollars. We needed... to create our own sort of talent pool for this and that's when we invested in 100x engineers. I know the team very well because they were you know part of my old company and I said I really believe in you guys you should do it but you should do it with the same philosophy small highest high quality and you know ideally maybe in their case it's little different because the people joining are older so less placements more on can they build projects can they build can they can they get a promotion right whatever the outcome for them can we make that happen. I think 100x has worked very hard on that and we know that in this space especially there's a lot of like it's a space where everyone's trying to make quick money so like what are the best signals how can we do the best work how can we partner with the best companies and remember I told you when you get to scale you get opportunities some of those opportunities really materialize next slide 100x has worked very closely with Meta to develop their B2B curriculum right So this is generative AI for marketers and LLMs for developers. Both those are actually driven by 100x. Next slide. They've done the OpenAI Academy. They work very closely with OpenAI Academy. So this is actually from OpenAI and OpenAI considers in India at least 100x as community partners, which is really cool, right? To see this team especially, right? Because the 100x team is really... the best grinders out there. They really care. Like, Sidhantham really cares about student outcomes. And to see him get there and to see the team get there for me is like, it's very special. Next slide. And of course, we wrote a book, which I totally forgot about. Ashina keeps reminding me we've written a book. But I want to show you a very cool video that somebody else had shared with me. And to me, it really made me think for a second. Can you play this video? That's someone's Explore feed. And the question I asked is, why isn't every video in that answer? So to me, that was special because I feel like the advantage of, sure, you can think of a distribution, right? But now I think we have that. I think it's really special being able to have people care and listen to what you say, both on long form as well as on short form, because it allows you to make the world the way you think the world should be. And I don't think there's a right or wrong here. I just feel like... All the things I've complained about all my life, I now have the opportunity to fix them because I have everyone's attention. And I think that's important. And I think we use our time and energy as much as we can to fix some of that. There are lots more cool things that we've done. And probably I'll reveal it over time. But we're most likely getting involved in a large fund as well to invest in cool companies that we see that we also do distribution for. So a lot of interesting things along the way. Next slide. But I think the fundamental difference was everyone was thinking of point solutions at that time. I will build this app and then I will solve it. Which I think is still very valuable. I think there's still use cases for those and those scale much easier I think. But we want to build an entire ecosystem. We want to build the best video ecosystem at that point. And if you see an AVTV video on day one versus now, it's iteratively we've gotten so much better. Because you know what? I have learnt that the best way people learn is from comments. If you do something in life, whatever it is. And you get comments on the video. Good, bad, doesn't matter. You will rapidly get better. Because you want to prove the comments wrong, right? You want to do a better job. You want to improve. You want their feedback to be better. So, I just really believe that we saw from the comments, we need this. People will reach out to us and say, hey, can you do these services for us? And we just said yes to what we thought was in our domain. And I think that has really led to a lot of growth here, right? Because now all these four entities, and five entities if you consider the media properties, are very profitable. Imagine doing five things at once and being profitable on all five. And this was not possible because of just me here or Rachina here, right? It was possible because all these entities are structured very differently from other startups. All of them have their own CEOs, all of them have their own co-founding teams. And EOS is one of those co-founders among the co-founding teams. EOS has Rohit, Laveena, Ronit, and then EOS sitting there. Same with Labs. Harsha, Tejas, and then we are there. Right same with all these other entities and of course, you know when we make an investment slightly different we are less hands-on like with 100x well much less hands-on But But it's been a phenomenal ride. I don't think I'd be able to do it without many of them So for everything that everyone says about the lone wolf cake engineer bed KIA, you know Admi bed cake but a company built correct. I don't think that's possible I think you only build big things when you have teams with you helping you out, who are also as motivated. And as long as you align everyone's incentives, I think it works very well. People have this misunderstanding of business that in order to make money in business, you have to screw someone else. But they're actually not true. You need to generate and share wealth with enough people because without incentives, you lose your best talent. You're constantly losing your best talent. We looked at, sorry, go up. Yeah, we looked at content, talent, software, infra, and ideas the same. They're all ingredients of a bigger ecosystem. So we're like, we'll build the individual units, get really good at each individual unit, and then together the outcome will be better because you get more shots at goal. That's always been our, the AOS philosophy. And whenever we do something, we're rapidly iterating and getting better. Now you've heard this story of us going from zero to one, right? The ideas we have then, the little bit of evidence we saw, and then we converted that evidence into a full business of videos are awesome. I want to tell you, I really have empathy for the people three years ago listening to me say all this because there was no proof at that time. We were working towards a destination where I think we have gotten to and now we have to maintain that which is an even bigger problem. But that was the first part of this journey. I want to tell you now myth-making around the next three years. Around how I think things will evolve. In a slightly different space but you can see how I do it. And I might be wrong. No one can really predict the future. Charlie Munger has this thing called Lollapalooza effects. Which is these multiple things going in the same direction. You should bet on anything where multiple things are going in the same direction for it. Because sometimes when 4-5 things are going in the same direction, the end result is massive. Or it happens much faster than you think. What were the things going in the direction for us that made AOS work? AI avatars were getting better. There was more demand for content, more people watching content. Lots of billionaires that had started podcasts. Right? What else? What is the last piece of this? What is the other thing that was going in our direction? Come on. Yeah. Lots of people want, there's an inherent need that I want to be in the content space. Sometimes, the closest analogy someone told me, one of the people told me was like, being here is like being behind the scenes at WWE in the early days, during Attitude Era. You get to see all these people pulling off, like we had this AV versus VM fight, I don't know if you guys remember the video that we made. A lot of people involved in this, this feels like WWE. Yesterday we did a video where we had students and teachers talk about how they're using AI. And both sides are scamming each other. Okay. Students are faking AI with AI assignment and students think they're very smart. Student thinks, if I take this assignment and put it through a humanizer, spin bot or quill bot, then the teacher can't tell. Teacher's like, dude, I know you for five years. I know you can't write like this. You're not some Einstein. So it's hilarious. But we get to see all of this behind the scenes. really cool thing, right? So there was this aspiration among people that I want to be on this stage, on the stage of YouTube and Instagram. And it made hiring easy for us. But at the same time, I think if we were to run a garbage collection company or something like that, it would be so hard to attract good talent. Because it's not aspirational for anyone to work behind the scenes of a company like that. Next slide. So I want to tell you the future where I think I've seen Lollapalooza effects. And I will talk for three minutes about a thesis that almost everyone online and people here might disagree with. But you hear me out, it might make sense to you. Ready? Awesome. Next slide. So as you know, we're in the proving stages. This was, I mean, a lot of people thought this was a game trailer. This is more like we wanted to prove photogrammetry. It's like the first video we ever did, right? Like we want to prove photogrammetry. I think we proved it. People forgot the memo that it's an early alpha. And I think when the early alpha goes into a first trailer that people can actually see, which is coming soon, I think people are going to be pleasantly surprised. But mainly it helped us hire people. Mainly it helped us tell people we're trying something. Does anyone want to come help us? And we got some really good talent since then. So you'll see the results of that talent soon. And separately we invested in Headshot, which is a mobile gaming controller. I'll tell you why we did it. And hear my thesis out. Feel free to tell me I'm an idiot in the comments or in real life also you can say it. Have you guys seen the Nintendo Switch 2? All of you have heard of the Switch 2. The Switch 2 made a claim. The Switch 2 claimed it was 10 times faster than the Switch 1. Some number they had, but I think it was 10 times. Do you remember the claim? It's a misleading claim. Or kind of misleading. Because the Switch 2 is 10 times faster than the Switch 1, but not because the chip is faster. The chip is obviously faster, but it's not 10 times faster. It's because of a technology called upscaling. You've all heard of DLSS upscaling? Computers come packaged with it today. Laptops come packaged saying, hey, we can do DLSS 4 now. So upscaling now is giving results we never thought possible. Because I've been tracking upscaling for many years. We've made videos around upscaling in the past. So with DLSS 4, upscaling has gotten so good. Like there's a GeForce, there's a 4060 box here. I bet you it has DLSS somewhere in the box. So if you go from 30fps to 300fps, which is what upscaling kind of promises, roughly those numbers, you suddenly have this thing where everyone's racing to make now higher fidelity games and also to retroactively make games work on lower end devices. There's one more thing that's happening with phones. I don't know if you've noticed. Have you noticed that phones are now coming with beefier and beefier GPUs? This has been a trend, but now it's accelerated. Why? It's not because of gaming. It's because of this neighboring thing that I happen to also kind of be involved in, which is generative AI. You have the technology on your phone that allows you to erase a person from your phone, right? Remove a person, put me in this thing, whatever. Where do you think that computation runs? It's not happening on the cloud. It's happening locally on your phone. Now we have the entire concept, not just of GPUs, but also NPUs, right? So because I work very closely with Intel and we worked with Qualcomm in the past and many other companies like this, we worked with Nvidia in the past. We know that these companies are very bullish on the phone as a much better computing device. And it's going to keep getting better. So if you take the lullapool as a effect of DLS is getting better. And at the same time, phone device, like the actual GPUs in the phone is getting better. At the same time, there is a third thing that's going on, which is these folks at Unreal Engine are pissed. Unreal Engine, guys, if you make one more video about optimization, about Unreal Engine being unoptimized, we are going to jump off a bridge. So, Unreal is taking it to heart to optimize the engine. So, if you look at all these three things going in the same direction, which is phones are getting better, DLSS is getting better, upscaling is getting better, Unreal is spending all that time and energy optimizing. There's also a fourth thing that's going on, which is text-to-3D is getting very good. It's not there yet, but if you've tried Spark 3D, or if you've tried the new Hunyaon models, they've gotten very good. They've gotten very, very good to the point where now 3D artists are in the phase where, you know, photographers were three years ago, which is, this is a threat, you know, it's bad to do AI-generated models. But you can't fight technology, pointless fighting technology. It'll keep getting better. Today, almost every resume we get, there's GPT in it, right? Like they would have used GPT to write it. You can't. It's the natural evolution of things. So if you see all these effects going in the same direction, and there's actually a fifth effect also. How many of you have bought Xbox Game Pass? Why do you buy Xbox Game Pass versus buying an entire game? Like it makes total sense that there should be a Netflix for games, right? Like why do you buy each new individual game? Why can't the business model be that I pay one subscription and I get access to whatever, three new games, four new games a month. And it's existed, but if you look at some of those games, like PlayStation has this pack, right? But if you see some of those games, those are, they're not, you know, they're not like the top rated titles. You know, they're giving you the, you know, the sidey games. That's because the economics don't work out. If a game costs, if a high quality game costs You know, 100 million dollars. By the way, I hate the word AAA. And we never used the word AAA in our last video at all. Right? Because the word AAA is a marketing term. Right? It means nothing. There are now AAA games that are made by indie studios. So, it's not like either or. So, I think everyone's wrong about that. But the idea is that if you assume that Texture 3D is getting better and people want subscriptions, okay, the cost of a game has to go down. And this is an awesome game called Claire Obscure Expedition 33 that came out recently. It was made by 30 people. 30 people. And it's probably going to be game of the year. And guess what? It was made with Unreal Engine. So all the complaints kind of... Like, I don't understand the complaints sometimes. Game of the year is probably going to be an Unreal Engine game. Right? So, made by 30 people. Small team. And there was actually an inexperienced team that made it. But they made it with a totally different viewpoint. They didn't come from traditional games. They had some people from Ubisoft and stuff like that. But they had a totally different viewpoint of where they wanted to go. Right? And I think that's a good template because the minute you can make a game with 30 people and with Text2 3D, I think you'll make it and not just Text2 3D, right? We have Quixel, we have Text2 3D, we have 11 labs now helping with background music and things like that, background audio and things like that. There's just like all this, it's not ready yet, but it's moving in the direction where when we did Avatar's version 0.1, when people laughed at us and it got better and then people just keep quiet. They don't even like, they don't respond and say I was wrong. They just keep quiet, right? When they were wrong. So I just feel that All these effects are going in the same direction. Games need to be cheaper for new titles to be available on subscription. The world is going to move to subscription. Makes no sense to pay 5K for every game. And phones are getting better. So that's why we took these two parallel bets. We said, a Nintendo Switch, instead of buying a Nintendo Switch, if your phone is going to get powerful, why not power your phone with this controller where you can charge while playing? And we'll build our own software. Labs will build our own software for this. And of course, we don't, I mean, Karan runs the company, but We are happy to help by building the software. And we'll build our own version of the launcher. And then on the other side, and these are two separate things, right? Because you don't know which of these will work and which not. Like, you have to, like, you try things in parallel and you be merciless about the things that don't work. On the parallel side, we keep making games. And no, we won't make indie games. Or, indie is the wrong word. We won't make these. I mean, people disuse the word indie a lot. You can make a really high quality indie game now. But we won't make simple games. Because we feel that, and the market will tell you this, right? On Twitter, you'll see screenshots of Death Stranding. People saying, wow, such a beautiful game. You don't see it very often with, you know, 2D games. It's just human behavior, right? People, if there's a tired bored person, and especially people that can afford games, they are slightly older. And they want to enjoy themselves over the weekend. So, those sort of people, we think that we have to... first go for that audience because when we go for that audience and build these games and as technology gets better, those same games will be retroactively played on mobile phones for five years later. And that then opens the market in India. People keep saying, oh, if you make a game and use India, it's using nationalism to sell it. Like India doesn't buy games. India pirates games. Unless it's mobile. Just so you know, mobile games, if you look at your BGMI in India, they're north of $200 million in revenue. Like compared to that, any PC game we've ever made, even the Raji, it's like a fraction of a fraction of a fraction. So you need to tap the mobile audience somehow. But we don't want to make mobile games. I have like 14,000 hours on Dota. I don't want to make a mobile game. That's also the thing that I want to do, right? So we want to make PC games, but we believe that the market will expand as this happens. So we're taking these parallel bets in two different directions, and we'll see which works out. And if it doesn't work out, again, we're building the infra, right? You know, if you look at the last trailer that we did, If you look at the next trailer that we're going to do, the animations have gone up 100x. Why? Because in the last trailer, we're using indie suits. In the new one, after we got all that feedback, we upgraded to the OptiTrack, which is proper production-grade mocap. And it makes a world of difference. So we just believe that... you build these units first because Optotrack is a one-time investment. If you have plans to make games for the next 10, 20 years, you're going to reuse that again and again and again. So it'll pay for itself over time. It's very expensive, but it'll pay for itself over time, right? If you have the intent to be here in the long term. And right now, there is no market for buying games in India. So our intent here is let's build out the team, let's attract good talent, and let's tell good stories, and let's try to do it with the cost economics of a lower number like we did with content three years ago. And if people say whatever they want to say, that's up to them. right but we have seen them being proved wrong once and hopefully we can do it a second time next slide yeah this is everything that i was talking about there's a new model now called um so if you look at this that's uni rig that allows you to rig all kinds of models uh meta taylor now allows you to uh do cloth sim without actually having to you know do it manually uh that's the optitrack systems but apart from the optitracks move ai is also available now but it's not as good as the optitrack and then You can see how DLS has grown from DLSS to DLSS 2 to DLSS 3.5 to DLSS 4. You don't believe this is going to continue. You'll be still complaining about optimization performance when everyone has taken up on their heart to optimize everything. And NVIDIA here has a ridiculous amount of money in their bank to actually solve this problem. They are going to solve this problem. And of course, separate from this is also cloud gaming that's really kicking in. It's launching in India shortly. Next slide. I have a question for you guys. Do you think I'm happy? Why? Huh? So if you ask other people, am I happy? You're not happy. Next slide. I started to live in permanent fear. Have you seen my t-shirt? It says, it says, trying my best and failing often. I post this every two months on Twitter. I just post it. Because I feel like it's easy to believe you know how the world will play out or easy to believe that you are right about everything when you're very young and you're untested. But once you've been beyond the test and you run companies, some have worked, some have not worked, and you see these arcs of up, down, up, down, I think it's very hard to believe anything is permanent. I'm very scared of being wrong. I'm very scared of automation being too close. Like the kind of progress AI has made in three years, and it's only getting fast. It's not like, we thought it would slow down, but it's only getting faster. of being wiped out, of a new competitor coming and doing things in a different way and then knocking us out, and I'm always paying attention to that. Therefore, I spend the time on the quest for these market truths, for distribution. Distribution is a good mode. For experimentation, can we try the new project, can we try the new thing, even if it doesn't work, but keep trying. And predictive power, because when you try enough experiments, you can start a sort of, like, I now think I've started to show this ability, at least to myself, which is, I think I'm usually right about how, you know, a product will go from V0.1 to 1 and how audiences will start using it, right? Or how their behavior will change. And I think bootstrap is true hard mode. When you're bootstrapped, and that's why I'm so thankful for the early team that's been with me, right? Because when you're bootstrap, you can't be wrong. Your error margin is zero. When you join a funded company and most of you don't realize this, the salary that comes to you comes many a time in India from a VC, a venture investor. That money flows through the company and then flows to you without knowing whether the company is successful or not. It's like a promise that the company will be successful. Here, money comes because we provide products and services to the world and they pay us and they only pay us while it's of value to them. The world is brutal. Okay, and what they say and what they do is totally different things. What they say about what they will pay for and what they actually pay for are two different things, right? Like we've learned this the hard way. So, bootstrap is true hard mode. It's like almost like I have to listen to what everyone says and then only pick out what I think they're really saying because it's true, because the market actually believes that. But if you look at their buying behavior, it's totally different. Next slide. Now I'll show you some pictures. of the early days of where we started to where we are now. I want to show you some pictures. That was one of the first meetings I had with the 100X team. You can see Sid with this very strange haircut he has in this video. He's blending into the board. Tejas is very confused. Sidhan is also very confused. And you can see Harsha's leg. This is labs and 100X chatting, I think. This is one of the first ever videos AVTV ever made. We made a video about the Museum of Future. And we had gone to Dubai, we had shot some, you know, thing in front of the museum future and like idiots, we had lost the footage. So we reshot it on our screens. It's one of our first videos. You can see we have put the screen in a, there's a 1,200 square foot house in Prestige Westwoods. Next slide. Now is our first Prestige Westwoods office. Okay. And if you see that ball there, right? That ball is there so that we get exercise. Work-life balance. This was Embassy Grove. Our second sort of temporary thing. Little bit better but still constrained. Next slide. So I'll show you. First picture is of policeman Martin. He's wearing a police costume and doing some reel. Second one is actually not safe for work because it's Sidhan's ass. But that's Sidhan's contagious coding in the middle of the night. Now this is Siddhant having kicked Tejas out continuing to code because Siddhant likes that. That I think is Ocus or Devansh, one of the two. Ocus editing a video again in the middle of the night because all our sleep cycles are ruined. Next slide. That is Ronit actually in house two. We had gone from one house in Grove to two houses in Grove. That's Ronit reviewing someone's edit I think because I don't think Ronit is editing the video. This is actually us cutting a cake on 100k, one of our 100k celebrations. I think AV 100k celebration. Next slide. That first one is me in VR. I'm wearing a VR headset but staring at the screen because we were doing this physicswala video and I need to do it outside but I couldn't because they kicked me out so I said I'll do it in VR. I mean I'll do it in VR with virtual production in front. It's a really weird combination. This is Martin trying to convince Ocus of something and Ocus not getting convinced. This is one of the first outings we did in Prestige Westwoods with the entire team. And Sid was somehow there, I don't know why. Next slide. That is Achina doing a meeting with the early AV team. And I just want to tell you one very interesting thing. How many people in that early team there can you recognize? Sumit, Devansh. That's it, right? Everyone else is gone. Everyone else took a 10% increase in salary or this or that. We don't know what's happened to those companies, right? Are they still around? Are these people still in video? But the people that stuck around have done very well in life. At every stage, you'll see some of the early pictures. There are lots of people who were with us and then left thinking they'll either start their own company or do this and fully supportive. But now that I've seen two, three years of their journey, I'm like, when something is going very, very fast, you should probably stay on board. It's just... It's very hard to find those slipstreams in life. It's very hard to find something with that kind of acceleration. Just sit there. This is one of the first videos we made on K-pop. Maybe first 10th video or 12th, 11th video. First set of videos. See how far AV has come from then. But in the early days, most of them would wear shorts. So the upper half would be very professional and the lower half would be shorts. And that was an AV and EOS classic back in the day. Next slide. That's our BMW. So that's the BMW. We had told some engineering students had said we had made a car. I was like, bring it, we'll make a video on it. And then we actually made a video about F1. Where in the background, I just want to tell you the creative genius of what we have done in the early days. In the back, on the screen, the screen is like a big monitor for those of you who've been in our studio, right? Someone in the back was playing F1 in, I think, VR mode. Actually, no, we were playing F1 in one of the camera modes. and he would be driving the same way Priyal would drive. So the guy would be like, Priyal left, and he would turn left, and then Priyal would turn the car left. And he would do this backward. Yes, this is the important part. He would do this backward. Because for some reason, I don't know why, but it was happening backward. That is Tejas' house actually in Embassy Grove. His room, so all of us would live in our rooms and then work was downstairs, in that small hall type thing, right? Tez is the only one of us, apart from Achina, of course, because she's a girl, but Tez is the only one of us who kept his room tidy. If you walked into Sid's room... Because he was never there. If you walked into Sid's room, it would smell like cat shit all the time. And then the cat ran away. I was like, why did you put up with all of this? So Tez is the only room where it was decently kept. Siddhant didn't even have a room. Siddhant was in one room without a bathroom. So... Aquarium. Sidhan lived in an aquarium. I don't know what that means. He says he lived in an aquarium. But I don't know where he stayed. He was just up here. Anyway, next slide. That was us launching, I think, either Alpha CTR or Autocode. One of the products we were launching there. In the first one. This is Ivan giving gyan on something. This is Sidhan sitting on the floor. I think that's Sidhan's house in the first year of... That is Rohit, who now runs Yas. As you can see, he's wearing shorts. So, fully suited up, but underneath, chaddi. Next slide. That's the first time we ever deepfaked anything. So, we made a model of me using a technique called Phase A, using Deep Face Lab, and that was me on Achina's face. Hideous. Correct, it's Achina's body and my face deepfaked on top and his life deepfaked. That is me. reaching 100k followers. We're at close to a million now but I was at 100k. This is the three jokers. Sagar, Okas and Sunmit and Okas' dog. Okas' dog would pee and shit all over the place. Next slide. This is the team. On the left side is the early AV team sitting and eating on the floor. dosa idli I think or this is a cafeteria service if I'm not wrong. Early cafeteria service. So yesterday someone was telling me, I think Shriyansh was telling me that in the cafeteria some guy who was I think from Bengal was saying I want Bengali food in the cafeteria. I'm like these guys would eat, cardboard also they would eat sometimes. So this is their editor's room. They had a nice pool view though. This is Sagar, he had broken his leg in the middle of the night. And in the middle of the night all of us went and scared him even more. But there's a small hospital here. But we all went and spent some time with him, making sure he was fine. So I was very scared. He was like, my leg is gone. Next slide. I have no idea. This was actually the 100X team teaching the rest of the editors how to use AI in the very early days. This is our old Prestige Westwood studio on the other side to see how small it was. You've all seen our studio now, right? And you've seen studio 2, which is even bigger. This is where it all started. Go down. I want to play a video. This video is from 5 years ago. Okay. And today, Ronit is doing spectacularly well. Both financially as well as not. I think he's doing spectacularly well. He runs a team of close to 300 people now. But most people don't know that Ronit used to be part of my Avalon community on Discord. When we were a Discord community. I want to play that video. Right? Because... Yes, yes. We didn't know how to edit videos back then. Hey guys, this is Ronald Nani from Avalon Army and in this video I'm going to share how Avalon Army, the community which started from a simple telegram group and which is expanding like crazy, has impacted my life and how it can impact yours. I'll tell you the before and after, like what was my life like before Avalon Army was started and what my life is like after Avalon Army has started. and the meetups are happening and all of that. So without any further ado, let's get started. So basically, I'll tell you Ronit's template, okay? Ronit was a very ambitious young kid who really wanted to do something and didn't have friends around him who wanted to do the same thing. And actually that's a good identification of my audience too. My audience is mostly people like that, right? Like people who are sitting at home, the reason they watch me and the content and they want to work with us is because they're like, shit, I really want to work on all these things. I really want to try all these things, but who will do it with me? And that's it. The thing I wish I could go back and tell myself is, bro, this Avalon community, you shouldn't have built products for them. You should have built products with them. You should have built products and services with them. That core Avalon, all those people, right? Martin, Ronit, even Lakshia recently joined us. She was like, hey, I was from the Avalon community. So I would say I'm still reaping the rewards of getting a few ambitious... people together who want to do something cool and say, let's just experiment. Of course, I had to get better. I had to understand how markets work. I had to understand where there's an opportunity and how to really not listen to other people, right? To sort of think first principles. And I think this is a great video because he says before and after. I think today is the after. It's been five years since that video and he was a kid back then. But it's been five years. And today, if you look at the after, I think, I don't think somebody like Aroneth, I don't think any other community would have given him the sort of platform to do what he's done here. It also shows that anyone, anywhere, any part of EOS can rise to the top. And we don't care about your age, this, that. Can you just be a leader? Can you lead? Can it come from inside? Because you know what? Like I said in the beginning, right? Like that, what everyone's doing versus what we did was the opposite today everyone wants to do like it's not like they want to do four things at once they have these four options in mind they won't take action to the first one either towards any of those four it's too easy today to detail your opinions like every week i'll see a video about somebody going on youtube and giving their opinion about something cool but can you take that serious now can you convert that into a career can you do can you take any of those opinions and do it seriously can you go down that path And I feel like EOS is this one place where I'm lucky enough to be surrounded by people who take what they do very seriously, who want to prove themselves, who are still as ambitious as they were five years ago when they were children. And my only job is to find the next in this audience. Despite the fact that everyone here is paid a salary, I still want to find 30 or 40 new people who can rise to the top. And there's unlimited space. There's unlimited space. the company can keep growing. Like we don't see a limit to the size of the market for content in India right now. So we think a lot of you can rise to the top just like Ronit has. Next slide. So in the early days, I spent time and energy building up talent from scratch. I think our company will only survive if we make more leaders. The 16 cut hours, this is the first episode, this is the intro episode. But as I go deeper into how we do specific things, as I go into the tangibles, as I talk about, like for example, when we talk about sales, One of the things I want to talk about, one of the things I've learned in the last few years is actually to really sell, you need to understand another person's viewpoint, right? Like if you're selling to somebody with a thousand crore net worth, you need to know where they're investing their money and what is the biggest problems for them. Without that, your sale is not going to work. Similarly, if you're building technology, if you're building a product, you need to understand who among the people that are going to download your product are your core audience because those guys will talk the most about it. There's so many things that we have learned over the years. that I wanted to put in these slide decks and I think you will really enjoy being a part of it. I have one goal. The 16 cut hours are my attempt to teach you as much as I can about leadership, AI use and management to help us build the company together. I have one last thing to say. I think this is all right? There's no more slides. Yeah. I have one last thing to say and this is for the people online. We are known to do slightly different and sort of crazier things. I want you to talk, if you want to, with your public name on YouTube, Instagram, Twitter, whatever it is, about the 16 katas. I want to see all your comments, okay? And we are going to take, and both negative and positive, say whatever you want to. But I'm going to do something really cool with those comments. It's not going to be all the comments, it's going to be some of the comments, and I'm obviously going to favor some of the positive ones over the negative ones. But we're going to do something really cool. I told you, every fourth episode is something experiential. So in the 16th episode, which is the last cutter, we're going to do something crazy. I can't tell you what it is, but it's going to be really crazy. I'm going to title it SaaS. Yes, you can go home thinking, oh, some app is going to be built, but that's not what we're going to do. S-A-A... No, sorry. S-A-A-S. Okay, that's the title of the 16th episode. I can't tell you what it is. And even the people that are here, I'm going to leave like a small box or a glass bowl where you will all put in one sheet of after you finish the 16 cutters, how you feel. And I'm going to do something crazy with it. Cool? Awesome. So that's it. And for the audience online, that's it for me. Bye.

Welcome to the first session of a 16-Kata series where I break down everything I have learned about starting and scaling companies over the past decade.

In today’s dense session, we discuss how AEOS was built, how it was scaled from zero to a profitable organization without any external funding, the decisions, mistakes, and contrarian bets that shaped its trajectory, and what I learned along the way. The session also explains why distribution matters more than product, why talent is the real moat, and why running multiple small bets with high failure tolerance works better than chasing one perfect plan.

If you are building, operating, or trying to understand how bootstrapped companies really scale, this session may have what you are looking for.


00:00 - Introduction 
03:00 - The 16-Kata Framework: How This Series Works
04:24 - Company Origin Story & Ecosystem
10:46 - First Principles & Unconventional Decisions
17:03 - Act 1: The Video & Content Opportunity
20:41 - Building the Video Editing School
27:30 - Act 2 : YouTube as a Service (YAAS)
30:08 - Act 3 : Technology Integration (Labs & AI Avatars)
39:25 - Talent Pipeline (100x Engineers)
48:08 - Act 4 : The Future Gaming Thesis
58:55 - Personal Reflections on Fear & Bootstrapping
1:02:15 - The Early Days
1:16:50 - Closing Thoughts