Harvard Dropout: “Hard Work Is Dead!” Use AI to CHEAT And Get Rich | Roy Lee

0:00 Today's guest built a $120 million company at 21 years old. I left college less than 6 months ago and uh since then I've been like literally making millions of dollars. Like I'm spearheading something that will probably change the way companies are built like into eternity. Kicked out of both Colombia and Harvard. He's now the CEO of Clooney, an AI software that lets you cheat on everything. You should be able to cheat on everything. And I think people do want to cheat on everything. And the only reason not to is because you're stuck worrying about what society tells you. Whether you're trying to ace a job interview or close a deal, this AI feeds you answers in real time by listening and watching your screen. Right now, people in tech think that the biggest opportunity of our lifetime is an AI. And I disagree. I think the biggest opportunity of our lifetime is In this episode, we'll explore the broken state of modern education, unveil his unorthodox ways of going viral, and

0:45 see just what it takes to turn controversy into a billion dollar company. I really don't think I'm that controversial. And there's a lot of tech people who think, man, this is the craziest kid ever. He's definitely going to jail. But in reality, like I'm doing nothing illegally. Boy Lee, welcome to the Jack Neil podcast. Bro, that was such a tough intro. This is sick. Yeah. And I I thought you said there were going to be some Korean girls here. I flew all the way down from San Francisco, but uh the only way I could get him out here was there's a bunch of Korean chicks. Um Yeah, it's all good. It's all good. Just the two of us is fine. So Roy, um you built an AI that lets you cheat on dates, meetings, and job interviews. Yeah. Tell me about that. Yeah, I guess uh the whole like I feel like the world doesn't realize how much the younger generation is is AI native. You go to school like

1:30 Columbia, this must be like the [ __ ] smartest kids in the world, like one of the top Ivy League colleges, prestigious. Every single [ __ ] there is using AI to write their essays, do their homework assignments. If they can sneak it into a test, they'll do it on a test. Like you would not [ __ ] believe the the the the future leaders of tomorrow's generation, every single one of them is using AI. And um right now there's like the status quo push back against like oh you can't use AI in the essays like like like people are trying to silence it but it's pretty obvious where the future is trending if you are like Gen Z and and and you see what what's happening around you and and where that is trending is everyone is going to use AI and the status quo and incumbents are going to push back and say this is cheating you should not do it. And I feel like if society is obviously going to say that using AI and a bunch of [ __ ] that you shouldn't traditionally don't use AI for is cheating then you might as well embrace it and be in the eye of the storm. This is the only way that we'll get out without significant public backlash. I think a lot of people will look at us

2:16 and say like, "Man, are you not scared of the controversy?" What I'm more scared of is the controversy of us saying, "We're not a cheating app." And then [ __ ] half our users are using it for tools that like like for like cheating purposes, quote unquote, and people say like, "Wow, this app is like all they do is cheat." Like, [ __ ] it. Like, if you're going to say that in the first place, then we might as well embrace it and just uh ride the marketing marketing edge that it gives us. Wow, man. You are wired today. I'm I'm like this 24/7, bro. Yeah, it's fire. Um, okay. So, what does the app do? Essentially, Cloey is it shows up as this like pane of glass application that shows up over everything else on your computer. It sees what you're doing. It knows things about you. It hears your conversations and rather than you having to ask it questions like you go to chachbt.com, ask it like Italian places near me. Rather than that, Clue will predict what you want. When you're on

3:02 Chinese food, and you sort of like ask ask Cle, hey, help me out here, then it will predict like maybe you're looking for calories, maybe maybe you're looking for XYZ. And it'll give you information that it thinks you need to know. So, what have been the most like interesting use cases of Cluey that you've seen in the past few months? Imagine I'm in a meeting with you and imagine this is digital. Like like soon Cluey will be in real life, but for now it's on the computer only. Imagine we're in a digital meeting and I don't know who Jack Neil is and I just hop on the Zoom with you and Cluey just tells me like this is Jack Neil and it does a deep research on everything you've done in your life ever. And it gives me sort of like like the bullet point of everything everything you've accomplished. I can tell like it's it's as if I I studied you beforehand and it's as if I knew like exactly how big your TikTok was before this. I knew your earliest video like like everything there is about you on the internet that might be relevant for the conversation it would be relevant. Now let's say on this on the flip side you're you're meeting me I say hey I'm Roy last summer

3:49 I worked at this company called nucle genomics and it really changed the way I I thought about the world do you know what nuclear probably not but it'd be really helpful if you could just pull up definitions of what exactly is nucleomics and this sort of extrapolates everywhere to everything like everything you've done in your life and everything on the internet can now just be like surfaced in conversations uh as you need them. Um, if you're a consultant, you're talking about talking to someone about like, I don't know, the sales tax or the the import tax of fish in California. All of a sudden, that number gets pulled up and it's like you uh it's like you know everything about it. All of a sudden, the amount of information you have in your head during a conversation is not restricted to what's in your head. It's restricted to the entire corpus of human knowledge that's ever been put on the internet. I kind of understand some of the more general use cases, but what else have people been using it for? Like I've seen dating, job interviews. Uh, is there anything? Yeah, there there's a strong

4:35 pull to job interviews. I mean, this is uh this is how I like I rose to fame. Um, but there's also a really strong enterprise pull. You can there's there's like these Filipino Indian call centers and all these people like these people traditionally they'll read like a 600page textbook sort of memorizing everything about the enterprise uh uh about the company and the product and uh you know if someone has a login issue then you go to this website you click on these three buttons and um they literally sit there and memorize that and now you sort of don't have to do that you can just dump it all into an AI and sort of have them like feed you information live in a meeting. chat voice mode can't do this because um it doesn't truly interact with the meeting and like there's no other tool that just like listens in, knows what you're doing at the time and has like this context about everything. So essentially we've um we are the closest thing to automating at every single like Filipino Indian call center. Now imagine you're like a sales rep and um you're selling

5:21 like a technical product. Uh you I I imagine you don't know much about programming and you're I like suppose you're trying to sell me an API based service. Uh you don't know what a rate limit is and I ask you hey how does your rate limiting compare to competitors? All of a sudden now you know exactly what a rate limit is and you can talk as if you knew. On APIs actually uh what APIs do you guys have access to and like is plugged into Cluey? Yeah we have a few different integrations um a lot in pipeline right now and a lot that we're internally only only rolling out to enterprise users things like uh Google Drive notion Salesforce um a lot of enterprise use cases but but for consumers right now it's text audio only and if you want the deeper integrations you're just going to have to talk to the team. Got it. But okay, so explain this to me like I'm a dumb ass who isn't in the AI software tech world, has no background in programming. Like is CLY plugged into

6:09 chat GBT, open AI's API and then what does API stand for? Yeah, I actually don't even know what the actual term for API stands for. Something like application processing. Yeah. But but but an API is the way that you will communicate with different servers and computers like that are floating around the internet. Um um so so if I have a service that I run on my computer here and you want to speak to it then you'll speak to it like via an API and uh if you want to if you want to interact with a chatbt uh you want to make like a chatbot of your own then you will interact with the chatbt API. Right now we use a variety of different uh models including some open source models. Um like like the the thing with clue is chacht when you ask it a question um for example find me I know hairspots near like hair haircuts near me. Um it will only know exactly what you put in that text box. Uh but

6:56 obviously these models are smart enough to reason over images. They can reason over audio. They can reason over pretty much anything. Now um so clearly the the interesting technical challenge here we have to face is um how can we stitch all the input um in a way that a model can actually understand and reason over and um that's the interesting challenge here. So the models right now like like for for 80% of our use cases we just use the latest state-of-the-art model whether that's gropt 4.1 like like what whatever it is but but a lot of the work happens in context stitching. What do you think is like the best model right now? best model as of today probably Grock 4. Yeah. What makes that better than like Clue's uh amalgamation of all these other models? It it's it's quite different because because for Grock like like what we're doing is we're manipulating the input.

7:42 So we're using the same model but we're all all the work that we do on the Culie engineering team is like determining how do we compress all the input in the way that this model can understand. Um, so for example, like you might say you you might for chacht take a screenshot of your door dash um and then say and then and then put it put like paste that in chacht and then ask it hey here's a picture of my door dash can you find me the best calorie options um or like like the best calorie protein ratio options for me to order. Um but that's not actually the most intelligent way for chatbt to reason over that. It'd be much better if you could uh take that image, you could process it, convert it into optimal text tokens, i.e. this user is looking at a screen, there's Chinese food, here are the options, and you turn that image into text that I can reason over much quicker. And uh and then you say like like your your mission is to give him the lowest calorie. You know,

8:28 there's like various ways that you can various things you can do with the input to make it uh easier for the model to understand. Um and uh so I'd say like like we're we're not better than Grock for at all. The only thing we do better is we are better than you, the user, at prompting. You're also uh you made a promo video and I know the tech doesn't exist yet, but you're making glasses that let you cheat on dates. Yeah. I mean, the the intention wasn't even like the only reason we had glass in the video was to make me look nerdier, but the the the the intention was to have like this real-time HUD overlay. Right now, I think nobody in the world knows yet what the future interface of AI is going to look like. Meta is making the big bet on glasses. There's a lot of people that are making bets on AirPods, you know, pendants. There there's like all sorts of things. Um and in the eventuality I think everyone knows the final state of AI is like a brain chip. Um you will have a

9:14 brain chip plugged into your head and it will use AI to help you think. And this is what Neurolink is building. This is what pretty much like everyone agrees to be the final form factor. Um whether there will even be an intermediate form factor between the chatbot and the brain chip like that's still up to like that's still up for society to decide. I I'm not sure. But when hardware comes, the best people who are best positioned to distribute hardware will be those who distribute software best. Um, and that that's my belief. If we if we become the default way that people use AI on the computer, then we will be in a really good spot to be the to build the thing that is the default way that people use AI in the real world like hardware. Do you see embodied AI or brain chips prior? Um, like like in in embodied as in as an AI like robotics, humanoids. Oh yeah, I mean almost certainly these things will

10:00 be developed in parallel. Uh like I cannot imagine there's a world with cyborgs that does not also have like fully autonomous robots. Got it. So let's give people a bit of your backstory. I do want to dive deeper into just like your thoughts on AI from a macro and micro perspective with uh Culy itself. But so you got kicked out of Colombia for cheating on a job interview. Yes. Um yeah. So I think wait let's let's take it way back probably the story starts in high school. Uh, so like growing up all my life I've been like this super controversial kid. I'm like always sharing my thoughts and I always pretty much do spontaneously whatever the [ __ ] I want to do. I do pretty well in school and October 2022 like like I I 2021 I'm accepted to Harvard early. I think like damn this is so sick. I'm going to do cocaine hookers and I'm going to be an investment banker and I'm

10:45 going to go work at Goldman Sachs like as every Harvard student does. Uh but then later late later later that year like I have a suspension that I got for sneaking out during a school field trip um ju just to like hook up with some randoms or try to hook up with some randoms. Um I got I got an inschool suspension for that for a couple days and I didn't report on my application and then later it got reported to Harvard and uh Harvard says like, "Oh, okay. So you can't come to school anymore." So now I'm like kicked out of Harvard after that. I think you know what? This is a sign from God. I've always been pretty [ __ ] out there. I've always thought I wanted to start a company. This is my sign. I'm going to go start a company. So, I spent a year like locked up in in in my house and programming. I'm trying to build a company. Um, and then my mom's like sort of on my ass saying like, "Ry, you got to go to college. You got to go to college." And uh like the compromise here is, "All right, you know, [ __ ] it. I'll go to college, but I'll only go to college in California." So, I I move I moved to California. I go to community

11:31 college like here, like right by San Francisco. And then I spend I spend like a full year trying to build companies. Nothing really works out. But later that year, I do end up getting into Colombia. And my mom says, you know, like Columbia is an Ivy League. You have to go at least for a little bit. So, I go to Columbia and the only thing I'm thinking like, man, the only reason I'm here is to find a co-founder and a wife. Uh, I think people really, really undervalue the social lubricant that college is. You will never ever if you're a normal person, you will never in your life ever again have the experience of a next semester I'll go to French 4 and I'll meet the same girl and we'll be forced to interact in the same room with each other. there's no other situ like scenario like place in the world where you where you get this environment unless you are uh part of like a close group of highly successful people um or working on a startup. Um either way I go to college I'm thinking like damn where's my wife where's my co-founder

12:16 but at the time you know like I'm not leveled up I'm kind of fat kind of ugly. So so the the the first part doesn't happen but but my co-founder I find after I I literally talk to like 50 people I say to like every smart looking kid I meet I like hey are you a CS major? Do you want to start a company with me? And 49 of them say no but one of them says yes and that ends up being Neil. So Neil and I get together and uh I'm like convincing Neil like, "Yo, like college is not it. Like we we we got to leave and like do our own thing and and but but we don't really have anything to leave for." So we spend the whole semester and a half like like [ __ ] around on projects that don't work. We built like you know like AI date matcher, AI noteaker, AI lecture like like a bunch of [ __ ] Is this the first like uh set of projects that you ever built? Uh no I never or did you build things like in high school, middle school? Yeah, I've been building things like my entire life and I've been like trying to do companies my entire life. like middle

13:02 school, I used to sell sneakers, drop ship, do clothes, and like Amazon, FBA. Um, but I never really made any meaningful money doing that. Um, even like elementary school, I used to like flip snacks. Um, but then and then high school like I I I tried like I did like a tutoring tutoring business that did like six figures. Um, but but nothing that really scaled and had the potential to be like a trillion dollar company. Um, until college. and then college we put out this demo of another side project that we do and it's for a tool called interview coder and this tool essentially it's really really simple all it does is it's a transparent translucent pane of glass over all your other applications that can see your screen and give you answers to leak code questions so all you do is like you take a screenshot and then you ask hey help me solve this leak code problem and it will feed you the solution um and we

13:49 make this first viral viral demo and then it sort of gains some traction and the second it does I'm like okay we're lowkey on to something Um, and then we just decide, [ __ ] it. Like, let's full send this. So, I record myself using this on my Amazon interview to get an internship at Amazon. I successfully like like dupe the entire process. Like, I record myself using interview coder to cheat on literally the entire process. I record it. I post it online and uh it starts going viral. An Amazon executive sees the video, they they're like they're pissed and then they send Colombia a letter saying, "Hey, uh, your kid just could just just just cheated on our entire interview process. Posted online. Now he's selling the software. What the [ __ ] You need to expel him or we're never going to hire from Colombia anymore." and uh Colombia sends that letter to me. They loot me up in a bunch of [ __ ] disciplinary hearings. Um and it's just like some some [ __ ] thing and um I end up getting kicked out of Colombia. Uh and but but by the time

14:37 I do that, I'm posting everything on Twitter and um there's like so many VCs in the valley that I already saw and we're like like I raised like $5 million in like 24 hours and um we're off to the races with Was it intentional to get kicked out of Columbia? I think some people like a lot of people will look at like the viral [ __ ] that happens of like people quote unquote like ruining their lives and they'll think like man this guy's so stupid. He doesn't know what he's doing. Like usually these people know and are are are like trying to they're they're doing it for a reason. And and for me like the second the first post got like a thousand likes, I knew, okay, this has viral potential. I'm pretty much [ __ ] They didn't put me in a disciplinary hearing just to like scold me. Like I'm pro I'm probably getting kicked out of school and if that happens like like this is my exit. Uh, I need to like keep pushing pushing through here. And um like like for like when when Andrew Tate

15:22 started going viral for the first time. Like his was very obviously more intentional, but there are so many people saying like, man, this guy's his reputation is ruined for life. But you just keep doubling down. At one point, you find yourself in like a completely different spot. Like like it's what happened with the [ __ ] Paul brothers, Charlie, Lily, when he said the N word. Like you just keep Well, if he did keep doubling down, but I feel like every time you have a moment of controversy, everyone knows like, okay, you're you're in a weird spot. Uh, and and I I guess all this to say I knew very early on and I just kept doubling down and very early on that was the intention, but at the very very start it was not the intention. So, did you kind of look at Tate for some of the controversy breeds? I looked at a lot of people and Yeah. Yeah. Tate I think Tate is like sort of the the only person to have really caught the internet by storm um through just podcast and clipping and like like like he's the guys. Say what you will about Andrew

16:08 Tate. I've had him on the show, but every single guest who's young and makes over like 500K a month, they all [ __ ] with Tate and they all saw him as big inspiration, which is just interesting. Say what you will about him, but like there's something about the master class he put on on marketing and the influence he's had on clipping and just going on podcasts like he he definitely made a big impact on things if not the election itself, you know. Yeah, I agree. But yes, so Colombia, I saw some tweet. Uh well, okay. One, there's a Poly Market bet. Have you seen this about you? There's a couple. Yeah, there's like a Poly Market bet, which is like a site where you can place like bets on real life social events. This is not sponsored. Um it's like, "Will Roy Lee be in prison by 2026?"

16:54 It's like 20% say yes, which is like wild. But I mean, do you plan on keeping up the controversy or do you think you've got to a point where you're satisfied with what you have? I'm going to be honest. I think like what I do is not really that controversial. If you go on Instagram reels or like TikTok, like there's people literally that commit felonies. They'll pull the fire alarm at like shopping malls and go run out. Like like there's people doing crazy crazy [ __ ] And what I do, you know, post like a pretty controversial video of me like using AI on a date. Like that's not really that controversial. The reason why people like call me like this controversial figure is because uh there's a very big split in today's society between professionals and like people who are above the age of say like 27 and people who are younger and like Gen Z brain rotted. And if you are younger, if you spend any more than [ __ ] like 5 hours on Instagram reels, you understand that crazy things are

17:41 just normal now. It's why you see sports like [ __ ] sperm racing come out and they raise like ridiculous valuations. They wanted me to be in that or invest in that [ __ ] Yeah, exactly. Exactly. They raise millions of dollars to raise sperm. would never ever have been allowed to happen like [ __ ] 10 years ago. But now you just accepted crazy things are normal now. Um, and I I really don't think I'm that controversial. And there's a lot of tech people who think, man, this is the craziest kid ever. He's definitely going to jail. But in reality, like like I'm doing nothing illegal, you know? This is like this is there's people on Instagram reels [ __ ] like like terrorizing people. And so do you think the strat to build a successful company, this might be like not viralized uh conversation, but it's just interesting to me. Uh the strat to build a huge company today is find a product that's inherently controversial and viral uh and it's basically a Trojan

18:29 horse for you to raise money and build other uh adjacent products in the future that are more mainstream. Like you're kind of taking this cheating on interviews and eventually want to compete with open AI. Like do you think that's like the strategy? 99% of the advice that anybody will give you ever is probably wrong. 99% of what you know about society and what they tell you to be true is probably not true. The only person you can ever trust is yourself. And among every single company that's been successful, there have been so many edge cases and so many cases where people have just been wrong about the dogma that they preach. The only two common threads of any successful company and any any successful product ever are one, people take risks and two, people work hard. That said, I do have a take on this and

19:14 I think right now people in tech think that the biggest opportunity of our lifetime is in AI and I disagree. I think the biggest opportunity of our lifetime is in short form content. Genuine question. Have you guys ever listened to this podcast and thought, "Wow, this guy must do a lot of research on his guest. That seems super timeconuming." Well, it's kind of true. We listen to every podcast they've ever been on, watch every video they've ever made, read every tweet or article they've written. But all of this got easier when we started using Poppy AI. Guys, when they reached out to sponsor the show, I honestly didn't want to give it away because the tool is so powerful. But after seeing several large creators already using Poppy and realizing some of you guys don't have employees to help you with research, I decided it was probably best to put it out there so everyone has a chance to make good quality content. With Poppy AI, you can drag and drop YouTube videos, voice

20:00 recordings, documents, and connect all these to a chatbot that processes everything all at once. Imagine if Chat JBT could take a YouTube link, transcribe it, and you could just ask questions about it. Now, imagine if you could take three YouTube videos, four Tik Toks, and five Instagram reels, combine them all together, and use it to write a script. But if you guys want to check it out, go to start.getpoppy.ai/jackneilodcast. They have a 30-day money back guarantee, so if it doesn't give you value, it's completely free. Anyway, guys, back to the podcast. Five years ago, probably the single biggest societal shift ever is when [ __ ] Tik Tok overtook YouTube in terms of uh amount of people using it. Um and now all of a sudden you can pick up your phone, speak to it for 10 seconds and get this seen by 100 million people. This enables you to do so so so much. And I think it is pretty

20:47 universally agreed that if you want to build a big like truly extraordinary outlier company and again this is like not advice you should take as dogma but if you want to do something with outlier results you should find the one thing that you are an outlier at your superpower and just keep doubling down and see where it takes you. What I don't think people really understand is what your superpower is is a lot broader. Like you should be thinking a lot broader about what your superpower is than you actually are. Like it doesn't have to be I'm a great operationalist or I'm great at finance or I'm great at coding. Like your superpower could literally be I'm great at making viral Reddit subway surfer videos and like like like Subway like Reddit story overlaid on top of I'm I'm the [ __ ] goat at that and any video I make can do 5 million views. That singularly can result in like a [ __ ] billion dollar company that of its own. That can be the singular growth channel that that that takes like like a shitty like a [ __ ] perfume company to a million to a billion dollar valuation. Um I think you

21:33 should be a lot broader about what you are really really good at. Um that sounds like you could be more specific. No. Well, what? Well, like if you should shouldn't say I'm good at short form content, you should say I'm really good at this specific uh subway surfer video. Is that Oh, I guess the the idea space of things that you could have a superpower at, don't restrict them to like business tech. That makes sense. You need to be thinking about other avenues. Okay, got it. But but also, yeah, I think if you want to do something like generational like like you don't go out and get, bro, this is not like you should always take advice with a grain of salt. I don't know what the [ __ ] I'm talking about. Sam Alman doesn't know what the [ __ ] he's talking about. only you know you are the only correct person in the entire world. Everyone else is like a [ __ ] like like NPC just do not listen only trust yourself. Um but generally if you want to do something that is super horizontal and like truly changes the world you need to start small. Uh you

22:20 have to plant and expand whether that's with content whether that's with uh um startups or companies or like like anything you need to start somewhere that works and just expand out. So would you say Cluey is a Trojan horse like for like it was do you know what I mean? Yeah. Yeah. Yeah. I mean on on on on both the marketing and product front I think uh and like the these are sort of the same like clearly you'll download clearly like oh this is dope I'm going to cheat on my interview with this or oh this is dope I'm going to cheat on my exam with it and then you use it and you realize like oh this is a lot more than cheating this is like a fully different modality of AI and uh this this is like our full that this is like our growth channel like how people learn about Cluey and how they start using Cluey is uh they'll like figure out like oh I'm going to use this to cheat on an interview or something I'm going to use this try to use this for like a sales call because that's what they they've been saying cheat on sales calls, right?

23:06 And then they'll use it and they'll realize like, oh, this is dope for a lot of other things. It's just a like a SAS, like a website. There's no mobile app. There's a there's a it's it's a desktop app specifically. But um yeah, we're developing a mobile app. Okay, that'd be super helpful. I'm imagining having it in interviews and I'm just like it's like listening to everything and giving me like perfect questions to ask be super cool. Okay, so why didn't you guys do mobile first? You thought the use case was better with calls. like like right now the what what we're trying to build here is the AI that can like is always with you and if we want to do mobile first for that then uh like Apple's never going to let us put clearly in like like let clearly see your Snapchat or your Instagram or what you're doing on the phone. You give free hinge premium to all your employees. Yes, bro. I feel like uh usually people are

23:51 the same like like no matter who you meet whether they're a [ __ ] nerdy engineer whether you're like I don't know like like a professional athlete or whether you're an influencer generally people have the same core desires and one of them is intrinsic to humans it's a desire to not be [ __ ] lonely and jerking off in bed alone at night like every human wants to do this and maybe there there's some like asexual outliers but generally I know when I look at an engineer and he's coding all day he doesn't do that cuz he like all he wants to do is code all day like he probably deep down somewhere he wants a girlfriend when when I sit down and talk to these like nerdy engineers, it is always true that they're like, "Oh yeah, I'd like to have a girlfriend, but I'm so busy with work or like ah like like a million other cope reasons, but deep down the desire is there. I want to have a girlfriend." Um, and I think if if the desire is there to not want to be lonely and we're all stuck in the office 12 hours a day building the company and like like what am I going to I can't like send them out to cafes and have them dig around cold approaching girls. Like I got to I got to buy them hinge

24:36 premium. It's the best thing for the company. Do you think it's uh going to grow your company faster? I mean, yeah. Like if like are people getting in more relationships? You were like a big advocate of like getting a girlfriend at your company just in general, right? Getting a wife. I want kids. I want kids like yesterday. I wish I had like a son. You'd have kids right now. Right now. Right this second, bro. Like like there is no greater fulfilling joy than having a child and like reproducing like like this is my core biological driver. This is everything like billions and billions of years of evolution have led us to this. And uh there like have you ever like when's the last time you've interacted with like a kid like like a three-year-old kid? Uh like a month ago. A month ago. It's like hella lit. Like I I like it would be so tough. Imagine imagine you and I are like imagine we're like working at the same company and um there's like a bunch of little kids running around and you don't really

25:22 interact with them too much, but uh every once once a week you give like Uncle Jack gives a seminar and you'll teach them, "Okay kids, today we're going to learn how to make I don't know like documentary videos and all the kid like all the [ __ ] 10-year-old kids are sitting there listening to you and and then you pass them off. Tomorrow Uncle Royy's going to teach you about [ __ ] HTML CSS." Like how sick would that be? It would be so sick. Okay, so you're single right now though, unfortunately. Is that distracting being single? Yeah. Yeah. Yeah. It's pretty. I mean, it's it's only distracting cuz like uh you know, like I'm thinking like, damn, it'd be it'd be tough if I had a girlfriend, but I don't have a girlfriend. You celebrated your $120 million valuation with a lab dance. Okay. So, we had um we we we had some ladies over for the purposes of shooting a commercial. Um but, you know, some behind the scenes footage of the commercial got leaked. And I feel like it's not all that crazy actually to be

26:09 filming like a music video commercial sort of thing and have like, you know, like ladies in the background. And um that's what we did. We're filming we were filming like this video commercial sort of thing and some behind the scenes footage got leaked. Uh and um it it went pretty viral. But I think I think it's actually not not all that crazy what we're trying to and I can show you after what what a preview of the uh the video would have looked like. So it wasn't like a dancer or like a stripper. It was just a you know like different professions have different names. I'm I'm I'm not exactly sure what what the name was, but yeah, it was one of the ladies. H interesting. Yeah, I feel like the tech world needs some more spice, man. I I think I think it just needs more transparency. That's what I like about having you on this podcast right now is you have like this fire to life and this like [ __ ] you energy that Andrew Tate kind of has. All these guys are like, "Yeah, just sit in

26:55 my room all day, work 10 hours, and like have a family." Like I I don't know. I feel like you got some like spice to you right now. Yeah. I I think it's um I don't know, bro. It's like I'm living the most interesting life ever probably. Like I I I left college less than six months ago. And since then, I've been like literally making millions of dollars, more money than I've ever seen in my entire life. I'm getting to talk with probably the most interesting people in the world. And at at one point, I was like the and I probably still am like the center of the tech universe. This is um like I'm spearheading uh something that will probably change the way companies are built like into eternity. There's literally VCs that uh and like I like like it's crazy. There's there's VCs that describe the world of launch videos as like pre and post Cleuly like like like like clearly like like all the viral stunts that we do. They've like

27:41 meaningfully changed uh human society and the way that startups are built and uh that's like crazy, bro. Like of course I got a fire, bro. Like this is like the sickest [ __ ] ever. Interesting. But you're still locked in to be fair. Uh you seem like you work long hours at the end of the day. Let me ask you this. Uh Roy, you're 21 years old. If Zuckerberg offered you $1 billion for your company, would you sell it right now? If it was completely up to me, like hell no. But like why would I what would I do with more money than uh like say like $67,000 of disposable income income a month? Like I feel like there's not really that many great things in the world to buy, especially if all you do is like spend all day locked up in an office. Right now my life is so [ __ ] fun. I get to be the leader. I'm pretty much like Luffy from One Piece and I'm like leading a crew and I feel like I'm

28:27 part of the new gen. I'm like fighting against the Yonko. Like how how fried would you be? How how upset would you be if you were watching a show all of a sudden the main character [ __ ] left to go join some bigger villains like army? Like what the [ __ ] Like bro, whose show am I watching here? Like uh if if it is true that um you know you only live once and I would like to be the master of my boat then like [ __ ] no I'm not selling my company. But at the same time, you know, like I'd be I'd feel pretty bad if I walked into the office and said like, "Sorry guys, you guys could have been all millionaires." But I said no. Do you publicly disclose like the equity splits? Um, we're extremely generous with equity. Uh, like like generally the the first half dozen employees got like singledigit percent equity, which is like very generous. At this stage, we have a 20% option pool. Um, this is probably like for tech people only. We have a 20% option pool uh early stage. And um this is pretty much like like

29:13 outlier outlier like that that means 20% of the company belongs to the employees right now and this option pool generally grows as the company expands. Usually companies throughout their entire lifetime will not see a 20% option pool and right now at this stage having a 20% option pool you probably wouldn't be able to find another company that is as generous with equity to its employees. And how much did you give up to VCs at VCs? Uh right now we're still maintaining like founders founders are still maintaining uh majority ownership but we've probably given us somewhere between like 30 and 40%. And then when did you raise like the last round? Um this happened probably like two months to a month and a half ago or wait it was it was it was like April or [ __ ] probably the timelines are [ __ ] blowing in my head. It was probably like a month and a half ago. Yeah, month and a half ago. Okay, gotcha. It

29:59 is currently uh I think July 20th or something like that. 23rd, maybe 24th. Um, okay. So, that's how much you raised. And then like what do you even do with $120 million? Like right now? I think uh companies right now are growing so [ __ ] quick. Like right now is such a unique moment in human history. You blink and empires are built. Cursor hits a $100 million run rate in 11 months and then after that they [ __ ] quintuple that in in the matter in the span of 3 months. Lovable hits $100 million in 8 months. Like like these companies are getting built faster than ever. And um right now I think if you are projecting out like one year later we're going to you know like we're going to save all the money and we're going to do something big one year later like like you're just you're [ __ ] like gambling with the future. Like nobody knows when the next Chinese guy is going to come up with the next05 or GPT GPT like like like like research on

30:45 like nobody knows. You can't predict it. Um and if you can't predict that then you might as well just do what is important for the next six months. So right now we're planning on spending all the money over the next six months on what expenses hiring engineers part like it's probably going to be half split between engineers and and marketing. There's so many marketing stunts that you could do if you gave Mr. Beast like $10 million back when he was just starting out. Like he he would have got to where he is right now much much faster. We are the first case study in history of like a ventureback company being um or like venture venture capital like giving money to someone for the purpose of like going viral. And to that end, to answer your question, finally, uh, so some things that we're doing is we're throwing a gigantic rave. So, it's probably gonna be like a $2 million rave with like some crazy headliners. We're try we're gonna try and let like like pack up with like everybody in the [ __ ] world. We're going to give every single software engineer in the Bay Area

31:31 like a baptizing experience and have them leave the night think like, man, Culie is the sickest [ __ ] company ever. And it'll just be a viral moment. We're going to sponsor a full anime. You're throwing a $2 million party? Yeah. Yeah. How How long is it? Is it just one day? Just one day. Yeah. Okay. Okay. Tell me more about what's going to be at the party. Um, it'll just Bro, can I get an invite? Of course. I'll learn CS right now. Yeah. I mean, it's it's it's just going to be a rave. But but the whole point is that this is going to be like a the first tech company throwing a rave and b we're going to get every software engineer in the world like a like like a true night to remember. And everyone is going to go home. Everyone in the tech universe is going to go home thinking like this was the sickest [ __ ] ever. And uh like like how many how many nerdy engineers do you know that like frequently go to raise? Like they don't. If you go to raise a fun, you're probably a bad engineer. Um, that's

32:17 that's the intention. We can really shake up the the world like like the tech world here by by bring sprinkling in bits of culture from the outside. Are you inviting creators to document it? Of course, we're going to try and have like all of right now like probably all of Asian America is going to going to be there. What percent of your company is Asian? Probably like 60 to 70 something like that. Yeah, I had a quote from you here. It was like, yeah, why are Asian engineers better? Okay. Uh well well the thing is like like about over half the company is just like my friends and um in college I just hung out with a lot of Asian people like like like so I found the company with my with my best friend who was Indian so I guess Asian and also my roommate who was also Chinese like Asian um and then the first few employees are all my friends from community college um who are like much better engineers than the people in Colombia. Um and they're like all Asian

33:04 too. And then after that, I I decide, okay, now we're gonna hire like a marketing team, and it's all going to be influencers that I know. And uh, of course, I'm Asian, so I'm scrolling through and I see a bunch of Asian influencers, so I reach out to them and I try to hire them. And now the whole marketing team is like Asian. And they bring on their friends, and they happen to be Asian. And um, the engineers actually is like the most diverse diverse group, but it is still pretty pretty majority Asian. Like, uh, do you think Asian engineers are better? Me? I don't think that's too controversial. If you put a gun to my head and told me you can either like someone needs to ace this programming assignment and you can either have an Asian person do it or a non-Asian person do it and if you're wrong then your mom will die. I'd probably pick the Asian person is is what I'll say honestly. Yeah. But we we we do not have like a racial hiring preference for sure. I just had to ask about it. That's crazy man. So wait are the

33:50 creator is a majority Asian too that you work with, right? Yeah. Yeah. Yeah. Interesting. I mean, just from my perspective, I feel like there's something about I mean, I I've done lots of short form content. I used to have an agency and like run ads for different companies, especially a lot of mobile apps. And it's like if you work with a black creator, it hits a black audience. If you work with a white old grandma, like it hits like a white old grandma. Like uh do you have you seen that the majority of your users are Asian typically? Yeah, I think that's like more so a function of us just being like a tech tech company like early stage tech and generally like early tech adopters are just like nerdy engineers and those are usually all Asian. Um but also say like with with everything I I think you need to like plant and expand and um with content too like I feel like the mo content houses usually don't work because uh they try and get together 10

34:36 people with completely different audiences and they try and merge them together and it's just cringe and nobody wants to watch cringe. But if you plant in like like a really really niche subgroup like everyone loves an interesting um like like like an interesting collab between two creators who have like like overlapping audiences. Everyone loves that. Why does everyone love that? But nobody nobody likes content houses and every content house is cringe. I think it's because you cannot capture like a really big market super super quick. Um you must plant somewhere and you must be cool to like a subset of people before you can be cool to a broader subset. 100%. That's why I assumed you were doing it because it's like you have that cult like uh audience. Was there a specific moment in your childhood where you had to outsmart the system to survive? Um so early early middle school and high school uh well this is like the first

35:21 time I'm I I share the story. Um my mom like my family runs like this after school tutoring tutoring academy. And essentially what that means is like my mom will go out to all the other Korean moms in the area and she'll convince them, hey um come study for the SAT, have your kids study for the SAT at my place. This will be like a daycare/tutoring center. Um and for a year my mom is really really sick. Uh but nobody knows because she's like the public face of the company. And uh pretty much for a year I uh role play as my mom and um I take calls like I I I have tutors come in and like I I run the company when I'm like early in high school um in in place of my mom and uh like nobody knows for an entire year that my mom is essentially like fully out of commission because I have her phone. I'm texting as if I'm her. I'm like calling as if I'm her. I'm I'm like inviting people over as if I'm her. And I just say, "Hey, I have to be in this

36:07 meeting. Like I'm I'm out of the meeting, but let's let me just like I communicate everything over text and essentially keep the business running for like a full year." Um, and just How old were you? I was probably like 15. So you were 15 running your mom's company. Oh, like the family company, but yeah. Okay. Was it It's like a pretty big company or at this point like a localized Yeah. Yeah. Maybe like 1 to 200k year. Is she better now? Yeah. Yeah. Yeah. She's better. It was It was just just the one year, but it was a pretty crazy year for us. Yeah. Yeah. That's W's son. Um, do you have siblings? I do. I have an older brother. Older brother. What's he up to? He's in dental school right now. He's like doing the uh he's he's doing like the traditional Asian route. So, I get to do whatever the [ __ ] I want cuz he already he already made it. Bro, I have an older sister who's not the traditional Asian route, just the traditional like

36:52 school route more. So, she's like doing like neuroscience at Wake Forest or some [ __ ] Um and she's going to be a professor. I'm like, what's like your just general thoughts on school right now? Do you think it's irrelevant in like few years? Yeah, generally I think um the only reason like I think I did school like the best way possible. I went to school with the intention of finding a co-founder and a wife and I made the most out of the the social lubricant that school was. Then as soon as I found the things I was looking for like I like [ __ ] booked it. I think um in life you have in life you can always get exactly what you want as long as you do the two things of take big risks and work hard. If you want to be jacked like like like and and truly you can play this thought experiment up with like anything you want like suppose you want to be really tall and jacked and

37:38 handsome. You can get plastic surgery, take steroids and get limb lengthening surgery and if you need money for all those things like you can make a lot of money to do that and those are all like big risks that that require like hard work to do. And if you if you are willing to do that, you can literally have anything if you want to be rich. Like you can successfully start a company and like like like scale it. All you have to do is uh pick a niche, take a risk, build a product, and work hard. And I guarantee you, you will not fail. And this is like this is like the one thing that um I think uh people just everyone wants to say like life and success is all luck. Uh but in reality, if you really really work hard and you really really take risks, it is impossible for you to fail. you will make it and and you you will get exactly the outcome that you are working for. Which is why it is so important to have a guiding northstar and know exactly what you are looking for. Even if it's

38:24 like [ __ ] embarrassing, people call you cringe, arrogant, whatever. When I go to Columbia and I say like from day one, I want to start a company. I want to be the next Google. Like 90% of people think I'm cringe, arrogant, and narcissistic. And that's probably like still the sentiment of Colombia around me is that this guy's some arrogant kid. But if you do not have this dream and if you're not working hard towards the dream, you will never achieve it. But if you do, I'm certain you can get anything you want in life. Guys, we're going to be doing like a live Zoom meeting. We're gonna throw them on a date. Uh, which will be fun. So, stick around at the end of the podcast to watch that. I do want to ask you, are there any features that you're working with right now on the product or are you just thinking so much bigger and you're trying to like build infrastructure for like different things? Yeah, I think uh right now this product is really interesting like um I'm I'm I'm I'm almost trying to work solely on the product right now like like it's

39:10 very hard transparently. the models are not at a level of reasoning where you can sort of like know everything you've done on your computer for the last 10 years. Like just like imagine that they were imagine that the context window isn't just like a 100,000 text tokens. It's like 100 billion. Imagine that the models could take in like 1,000 times more text. Like how much more could you do and how much smarter like how much better will these models be? Like they'd be so much better. I think nobody knows where the models are headed in just like a few short months. like like the slightest tweak. That tweak being like an expanded input size that literally gives you the option to have like a a model that can um like predict what you want to eat for dinner every single day like like design every single aspect of your life, tell you optimally what time you need to wake up. Like like there's so many things that a more a slightly more intelligent model can do. And um

39:55 like like the amount of products and the the the amount of like worldchanging technology that's going to be developed as a result of like a 10% smarter model is just like it's it's it's [ __ ] unreal. And right now clearly is um we're at a point of infancy in AI like like the state of AI right now these are the worst models you will ever use in your life and the rate at which these are going to get quicker is going to be [ __ ] unbelievable. Um all all this to say right now like we need to focus on the product because the models are not super duper duper smart. They will be soon but in the interim we just need to be fully focused on product. What kind of holds back the models from taking in more uh inputs? Yeah, I guess um like it's it's all well right right now the general consensus is just compute. We need more data centers. We need more GPUs. Right now Nvidia is like literally the most valuable company like $4 trillion market cap on Nvidia worth

40:41 more than like almost worth more than Apple and Google com Apple and Microsoft combined. It's like insane how valuable it is and it's because they are the only people that can provide GPUs that can perform the uh compute necessary to train these models. uh you see like literally there's like billions and billions of dollars of uh just just money being thrown into getting more compute and right now you can't train bigger models. So to to put this in perspective like like like the first few models the first few iterations of GPT had a few million parameters and um and now we're at the latest models that have like 70 billion parameters and the latest models but right right now the current state-of-the-art models have something like 300 billion parameters and um all of this is strictly a function of how much compute is available and um we know for certain that uh there's sort of like this scaling law of compute where as more

41:27 like like the models will get exponentially better as the the the more compute you throw in And um if we can get to a point of I don't know say like a 10 quadrillion parameter model that's strictly a function of how many data centers can we build out like holy [ __ ] nobody's ready for how smart that model is going to be. Nobody. Yeah. You said that we'll have to give up our privacy to win the AI race with China. Does that has something to do with the amount of data they have. Yes. I mean chi like Xi Jinping is out there recording everyone doing everything [ __ ] imaginable. There's like there's like street light cameras. There's cameras in your home. There's cameras in your [ __ ] bathroom. Probably like in 5 years, China will have the data centers necessary to train models on data on literally every [ __ ] Chinese citizen. And when they have that data that like they'll be able to train not just reasoning models, but like world models that will literally be

42:13 able to understand the laws of physics, predict every single thing that happens like like a sufficiently with enough all they need is data centers. Once China gets enough data centers, they will be able to train a model that will be able to look at this room, predict based on like the [ __ ] like like the tiniest pixels like where my arm is going to move next. And all this is strictly a function of compute. Um compute and data like you need data to train and China right now has the data. We don't know when the compute is going to get there, who's going to be providing the comput if it will get there. But if it does, China will have the data and that should [ __ ] scare everybody. Um, if America does not have the data and we don't because there's these [ __ ] for 40 50 60 year old congressmen who can't see past like two feet past their nose like they are trying to restrict all the laws and say like you know AI is dangerous like we can't like bro if you don't wake up and realize that we need data to be China then we're all going to be Chinese

42:59 in like [ __ ] 10 years and nobody wants that. Well the rest of us don't want that. Yeah. Yeah. Except for China. So, so I like all this to say like I really do think um concerns around data privacy they they're not just going to go down, they have to go down unless you want to trying to take over everything. Part of yourself do you not show on the internet? I am pretty transparent on the internet but not fully open. There is a slice of my life that I show and that slice is representative of me. It's not like I don't act like this in real life but I don't act only like this in real life. Um I think uh there's a lot of thinking that goes behind the things that I do and the things that I say and very oft like you will see like the last 5% of like a very long project for example um

43:46 one of the most viral things we did like like like the launch video like like it takes months and months of thinking and tweaking or the launch video where I use Cluey to to to score a date with a girl um that takes months and months of thinking like iterating um deciding which things will like land right, which things will land wrong, how do we iterate on controversy the best and it's like it's like usually pretty quiet. Um it's like just like quiet deep work and then and then at the very end like I'll come out and be like ah [ __ ] like I'm using cheat on day cheat on everything like blah blah blah but in reality there's like uh so much work that goes into that. Um yeah I'd say uh we we try to be pretty intentional with with the with the content that we put out and just what's overall like your mission right now with [ __ ] with your life with all this? I think it's probably true that um we are living in some sort of computer simulation and uh I am plugged into a brain somewhere and

44:32 everything else around me is just like really advanced pixels and NPCs. Um and if that is the case, I only live once. I'm the main character and there's some alien on universe B37 that's like watching me like watching the life of Roy and this is a movie. Um then I can only imagine two things. First of all, my life will be more interesting than any other human humans in history because I'm living in a video game, a simulation. And the alien will make it such that this is an interesting life. I strongly think this is why I am 21. Right when this AI boom is happening, I'm at the most productive. My brain is most plastic. Right at the opportunity of a [ __ ] lifetime. I think that cannot be a coincidence. If I was born 20 years earlier, I would have been too old to participate in the race. And if I was born 20 years later, everything would have happened. I am I somehow got planted at this unique moment in human history where I have the opportunity to [ __ ] change everything. Um, and I

45:17 think if it is the case that I was, you know, like it cannot be a coincidence that I'm born in the most interesting time in human history and uh I want to do like the biggest possible thing at this point in human history. Um, and for me that is uh building like the AI brain chip like like I think it's it's it's it's pretty obvious that this is the final form factor of AI. Everything is converging to the one super company that can build the brain chip that you plug in and lets you use AI to think. And this is where everything is headed. And um right now I just feel like that is if if this is a movie, if I was the main character and I wanted to make it the most interesting movie possible, I'd be trying to build the brain chip. And right now everything I do is like, you know, like trying to trying to move us towards that world. When do you think we'll see a brain chip in humans? Uh probably like two to three decades. If you died in 5 years, what unfinished

46:04 thing would haunt you? I guess this is a correlary to to the whole brain chip idea. But the reason I want to build a brain ship is because I think it's obvious to me it feels obvious that in a few short years companies are going to converge. You might have antitrust laws but like like in reality companies are already more powerful than the government. And you see like Google owns the data centers. They own all the data. They own all of YouTube's proprietary data. They can train the best models. And it's really hard to compete with like the big [ __ ] players. and you can be open AI and you you can like rush to build ChadBt, but in the end of the day like like the the big guys win and you sort of see they're eating up all the smaller players and this is happening at a rate that you've never seen. Acquisitions right now are happening [ __ ] left, right, center like it is crazy how everything in the world is converging towards these big companies and I think this extrapolated over just a few short months to years

46:50 means that there will be ultimately probably like one, two, maybe three super companies and um why is Zuckerberg poaching employees for $100 million? Yeah, I mean like the age of AI you're uh everyone is so hyper leveraged if you are really the single person who is capable of building the next big model or coming up with the next research development like all you have to do is justify what like a.1% increase in meta's value to be worth $100 million and if you come up with the next like GPT6 like of like no no [ __ ] right now one person is capable of more than ever before in human history and I think I think Zuckerberg realizes this I think 100 mil is like it's it's a discount for the for a researcher that will come up with super intelligence. Really? Do you think that researchers alive today? Yeah, almost certainly. Like super intelligence is coming in our lifetimes.

47:36 Um it'll probably become coming before 2030. And are you trying to attract those types of people to your team? Currently, we're looking for the best people in the world like but we are also filtering for pretty specific uh skill sets at the time. I think even like if we had the best researcher capable of developing super intelligence, I think they would be missed working um on the things that we're working on right now. But there are people of that caliber. For example, if you're on an Olympics, if you're the coach of an Olympic sports team or like an Olympic swim team, you probably wouldn't hire LeBron James even if you know it's like the [ __ ] go to go to

…