a16z Podcast - Replit's CEO on Vibe Coding, Wealth Building, and What Most People Get Wrong About AI

Episode Date: April 15, 2026

Jack Neel speaks with Amjad Masad, CEO at Replit, about how AI is making it easier than ever to build and ship software without a technical background. They discuss Replit's rise from a browser-based coding tool to a platform generating $250 million in annual revenue, why Masad turned down a $1 billion acquisition offer, and his case for why AI represents empowerment rather than existential risk. This episode originally aired on The Jack Neel Podcast. Follow Amjad Masad on X: https://twitter.com/amasad  Follow Jack Neel on X: https://twitter.com/jackhneel Listen to Jack Neel: https://www.youtube.com/jackneel   Stay Updated:Find a16z on YouTube: YouTubeFind a16z on XFind a16z on LinkedInListen to the a16z Show on SpotifyListen to the a16z Show on Apple PodcastsFollow our host: https://twitter.com/eriktorenberg Please note that the content here is for informational purposes only; should NOT be taken as legal, business, tax, or investment advice or be used to evaluate any investment or security; and is not directed at any investors or potential investors in any a16z fund. a16z and its affiliates may maintain investments in the companies discussed. For more details please see a16z.com/disclosures. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

The world was built by people that are not much smarter than you. Your job is to find the way of doing things that's most aligned where the world is headed. I think it's the easiest time to get rich in the history of capitalism, but certainly in the history of internet. Growing up in Jordan, today's guess was fascinated by programming, but couldn't afford a computer, which inspired him to make coding accessible for everyone. You can cast almost any problem in life as a coding problem. And I thought, okay, I'm going to hack into school and change my grades.

But when his company hit a billion dollar valuation, he refused every offer to sell, doubling down on his mission to turn the tech industry from a monopoly into a democracy. How much were you offered to sell your company for? When we're very small, not a lot of people, I think six people were offered a billion dollars. And why did you say no?

Because I think I can build a trillion dollar company. In this episode, we'll give his exact blueprint to build a million dollar app in minutes. Explore why the most powerful tech companies tried to kill his vision and question whether AI will enslave us. or empower everyone to escape the rat race. Why do you think AI isn't going to kill us all?

For most of the internet era, building software required learning to code. That bottleneck shaped who got funded, who got hired, and who got rich. Replit was built to break it. In 2011, Amjad Masad posted a simple idea to Hacker News. When any programming language in your browser, no installation required. That became Replitt. Today, Replitt's AI agent produces a working app in under an hour, and the company's revenue went from $2.5 million to $250 million in just over a year.

When a competitor offered to buy the company for $1 billion at six employees, Masad said no, because he thinks he can build a trillion-dollar one. His argument, not having a coding background, is becoming an advantage. The people who win now are. the ones closest to the problem, not the ones who know the syntax. In this conversation previously aired on the Jack Neal podcast, Jack Neal speaks with Amjad Massad, CEO at Repplet. I'm JobMissad. Welcome to the Jack Neal podcast. Thank you. Amjad. You built a billion-dollar company that makes apps just by talking to AI. If you wanted to build a million-dollar app in five minutes,

how would you do it? It depends on my context. I would look around for problems to solve. So no matter where you are in life, you're in college, you're at work, there are people dealing with problems all around you. One of my very recent startups that came out of Rapplet is a finance guy. And he was on a plane. And next to him was sitting in investment banker,

just spending a lot of time building something. spreadsheets and building decks for clients. And he had an idea to automate a big part of that. And he told him, you know, I have an app for you. And he didn't have an app. Can I come push it to you tomorrow? He said, yeah, I mean, if you can solve this problem, if he can make it faster for us to get to our clients,

he went home. And he's been using Replit just personally, just for fun websites, things like that. he spent the night working on the app next day went and pitched it left there with half a million dollars of letter of intent and did this a few more times

with other bankers and he's just raising at a $35 million dollar valuation right now so it's more than $1 million, it's $35 that because he already has a lot of contracts right out of the door so it's very contextual We have a educator that's like a more of a story from two years ago because the company is pretty big right now, half a billion dollars worth. But he's a teacher during COVID. I think he just left his school and started playing around with AI, playing around with Rupplet, going in and trying his hand at coding with AI.

And because he knows the problem space deeply, he was able to build a lot of tools for teachers. for grading students, for creating assignments with AI. And education is one of like the hardest markets, but AI has this amazing ability to sell itself. And it quickly grew the company to 10 million annual revenue, 20 million annual revenue. And now it's like half a billion dollars worth company. But there are a lot of smaller ones too.

The other day I was on Twitter. I saw this guy who created an app quickly with Replit. to generate brands kits and brand design material, logos, all of that. Do you enter your product name? You go through a simple flow. It's called anymark.co. And you pay, I think, 40 bucks or something like that, and you get an entire brand kit generated with AI.

And so I see this stories every day. And it's typically someone who has some domain knowledge in a certain thing. You know, they realize there's a problem around them. Maybe, you know, I see all my friends trying to. start companies, but brands are really hard to build. So let me automate that. So just like look around you in the world and just see what, what are the problems that people are dealing with that are willing to pay for it.

And it's so easy to try things because it's so easy and cheap to try things. You can iterate really quickly and arrive at an idea. I think it's the easiest time to get rich online. And I think it's the easiest time we get rich in the history of capitalism. but certainly into history of internet. So if I'm sold with zero dev coding experience, what problem did you guys solve here? And like, what are people missing about where AI's at?

Like, you can kind of just speak apps into existence at this point. Like, how much editing of it do you have to do? Like, how much of this really is just, it's done. Month over month is improving. I will say at this point, we have an automated software engineer that is as good as a mid-level software engineer. It would get a job at Facebook or Google.

Like, it is really good. Like, you don't have to look at the code at all. Actually, you know, initially, Reppler started as like, let's make coding easier. So they're still coding in the interface. Increasingly, it was just removing the coding features because you don't need to code anymore. Actually, even professional software engineers

are not coding anymore. So code is almost fully, automated. I mean, it depends on the specialty and the language and there's some nuance to it. But for the most part, people are not coding anymore. It's become a more higher level thing. So engineers still do some kind of engineering systems, engineering, whatever. But if you're a product builder, all you have to care about is who's the customer is,

what the problem you're solving is, what's your core differentiator? What do you understand about the world that other people don't? And can you put that into an app? So if you go to our outlet right now, there's a prompt box like chat chiptie. You type in your idea, like I want to create, you know, brand kit generator. It'll go through a planning process. I'll tell you, here's what I understand you want to build. You can go back and forth on the plan.

And then you tell it to go. It'll work for 10 minutes. It'll get you a minimum viable product. Obviously, that's probably not ready to ship yet. So you're going to iterate on it. you're going to ask it for adjustments. And there's a preview there and there's the chat box there. And you're going back and forth.

So you're telling it something. It's going writing the code, fixing the bugs, testing. It even, we give it a browser. So it can start a browser. It can look at the app itself. It can test it. It can go to the internet, fetch information. It can integrate other AI models like image generation and things like that.

And you go through this process. And I think within an hour or two, most people have an app that they're ready to put in front of a user. It depends on the idea, obviously. But for a lot of ideas, you can get something done that you can share with a friend

or a target user and get feedback from them. You don't need any development experience. You need grit, and you need to be like a fast learner. You need to be, like I will say, if you're like a good gamer, if you can, like, jump in a game and figure it out really quickly, you're really good at this.

But even if you're not a good gamer, you'll figure it out eventually. But people who grew up with technology or like fast learners are now like the best out of this. I will venture to say that not having a coding experience is becoming an advantage

because coders get lost in the details. Product people, people who are focused on solving a problem on making money, they're going to be focused on marketing, they're going to be focused on user interface. They're going to be focused on all the right things. So at some point, I think this year it's going to flip and I think not having a quoting background

is going to be more advantageous for the entrepreneur. Now when the company grows and you're getting a lot of revenue coming in, at some point you'll hire engineers just to make sure the security and the infrastructure is scaling. and Rappler continues to help with that as well. But getting to market and generating a revenue, you should be able to do it in a matter of days. So I guess just to zoom in a little bit

and give people practical steps, as someone with multi-billion dollar company, if you wanted to build an app that could get a million downloads in six months, like what are the five major steps you need if you could distill it. A unique idea. An idea that is not like an exact copy of something out there.

Because that thing exists. You need a spin on that idea that's interesting. How do you find ideas? I think that's actually the core skill in the AIH. I think if you want to work on a skill, it's going to be about idea generation. Because the cost of implementation of those ideas is going down rapidly. It's going to go to zero at some point.

So the bottleneck becomes, how fast can you generate ideas? And that skill is about one perception, like just looking around you in the world and seeing what's happening, what are the trends? Are you plugged in on social media? What are people talking about?

What is, like, the most interesting thing that's happening? And is there a market for that? Maybe we'll get to that later on, but we're discussing this idea of a looks maxing app, right? I mean, it's your idea, and I think it's a great idea because it's something in the ether right now. People are discussing, people are interested in it.

Can you build an app that gives you feedback on your looks maxing progress? That's really great. I saw an app today that judges, like, gives you, allows you to track your hairline progress. allows you to like, you know, it gives you interventions to make in terms of like what medications to take and things like that.

And unless you take like a scan if you're head and like obviously it's past time for me. But if you're dealing with that early on, it can help you track your progress. So that's very important. Like I think a lot of young people now, your generation and younger,

care a lot more about this than say my generation. Right. I'd say that's the big. the big trend of the past few months. So being plugged in is super important. And so, you know, a lot of the vices that my, like, older generations think our vices might actually become advantages. So if you're brain-rotted, you know, terminally aligned person,

that might be advantage because you know what's happening in the world. if you're someone who's also just like ADHD, really interested in novelty, want to try a lot of different things, that's actually an advantage because AI really benefits people who can try a lot of things really quickly. Obviously, you need to get things to completion.

You need to have some grid at some point once you've got some validation, but trying a lot of ideas is important. So back to your question, I don't feel like I answered. How do you generate ideas? it's practice. It's a skill. Like, generate ideas, put them out there.

I really like to use Twitter. I used to do with that a lot where I talk about ideas. I kind of share them and see what the feedback early on. But now just make the app and see what the feedback is and learn from that and go from there. And think about ideas all the time. I think it's a muscle. Like just continuously thinking about, okay, what if I have I? I built this, what would happen.

And so there's all these different ways to be a better idea generator. I do want to ask you some ideas you have specifically, but maybe we'll get on that later in the interview. So step by step process, first one, get a good idea. Get a good idea. Typically tied to a trend. Second, break that idea down as much as you can. into like say a paragraph with a bunch of bullets.

Like the app, the looks maxing app should, you know, have a camera with an AI integrated and it should be able to take a photo of my face on my phone or my laptop. And it should kind of draw, I'm just making this up, it should draw lines. Like get specific. Like imagine the user interface and get specific about that. You don't have to write more than a paragraph because what's going to be. happen then it'll give you the initial implementation and you'll react to that and see if it got it right or wrong and then you can like sort of nudge it in in different directions you're looking for

your key use case yes yes you're looking for the main use case don't overcomplicate it don't like add a bunch of features you're just like what is the core experience and how can you get really quickly to value users consumers these days just don't have a lot of attention span so they're They're going to give your app like five minutes. And so figure out what's the user journey to get to like a five minute value, a ha moment. And then once you do that, you got your MVP, go try it on someone. Go try it on a classmate, on a friend.

Before that, though, so let's just assume for the sake of your company is Replit, use Replit or tool like Replit. And then how do you build it? Like how simple is it? Yeah, yeah, you just put in the prompt. You hit start building. It'll start the working environment. It'll work for 10 minutes.

It'll show you a preview. You can, if you want to use it on your phone, you can open that preview on the phone. It will give your QR code. And then test the app. If the app is not exactly what you wanted, most likely it's not. The first iteration is not.

Go type to the AI, tell it exactly what it got wrong. It's like you misunderstood this. And don't try to overcome it. You just talk to it like you would talk to a person. Just be as specific as you can. So you misunderstood this. I meant the I should work in this or this other way

and give it feedback. Go through that iteration cycle a few times. And then go test the app. It's really that simple. Like, you know, you just need to be able to explain ideas well. I think that's super helpful. I don't think the marketing aspect is as important in this particular discussion, but because good products market themselves.

Well, I would say it is important. Like, yeah, good products are very important, especially if you're creating something totally awful. But figuring out how to promote this is going to be important. The easiest places to do it is find communities, like on Reddit and other communities. Like if you have a Lox maxing app, there's probably a LuxMaxing Reddit, just like posted there, try to get some early users through there, go to the various discords.

So that's like the early users. That's how to get like 100 users. At some point you need to scale that. And that's where in Instagram and TikTok, comes into play if you're someone who's good at that you have a superpower

if you're someone who's able to like create these short clips and talk about products you have a superpower but if you're not you can still go reach out to influencers

and you can cut a deal with them you can get them you know part of the subscription part of the revenue you can just pay them and and then and then you go from there

But I would say that's like the first few weeks. That's really interesting about the communities. That's also a way to find ideas in general is just scour Reddit, scour TikTok hashtags and see like what's a community of people. Let's say example, example, like a specific sports team or like sports gambling, something like that, not promoting gambling. But you would look at the community, look at the problems they have and then kind of get ideas for what you could build. It's really fascinating. That's exactly right. I mean, the original, when I posted Replit on Hacker News,

that's how I got my first users. Hacker News was still kind of a very popular place for programmers, technical people in general. Before Replaglet became more possible for non-technical people to use, it was more of a technical product. Do you remember your headline? Yeah. I said, try Python, JavaScript, all these different programming.

in your browser without installing anything, right? And so it's like gets to the core value proposition. At the time, to write any piece of code, you had to download insane amount of software. Maybe you try to take a coding class in college. It's really ridiculous how much you have to deal with in terms of just like writing any kind of code. And so I was like, why shouldn't I be able to code in my browser?

like my browser, you know, I can do my email, I can write my docs there, should be able to code there. And I also like scouring of these forums, I'd already seen interest in that. And those are small examples that weren't really working of people trying to build that. So it was like if I got it working, I can really wow these people. And just like there's a, there's a demo effect, right? Like if you're building something novel, like if you can make your looks maxing app,

like a truly cool experience, then people would just be attracted to that as well and they're going to promote it and tell their friends about it. So in terms of the headline, yeah, just like, what is the core value proposition? Right. And for me at the time, it's about coding in the browser. That's so fascinating about what you've built

because something I discussed with a lot of young entrepreneurs that built these multi-million dollar AI apps or startups was like this concept of making short form content before you even build out your MVP or like your minimum viable product and then just seeing if there's interest in it. But with your company and where AI is at, just build it instantly.

And you don't really have to worry about the moat of, oh, this thing does work. Let's build it really fast before someone steals it. It really changed. It really changed like even the past few months. That advice of, it's like called the lean startup advice I think started in like

mid 2000s there's a book called Lean Startups and the idea is that it was following from this Toyota way of building cars where they they

they would you know cars used to be manufactured in like a heavy industrial way where they kind of know exactly what they're going to build and like they'll You know, they have all the schematics and they'll, like, create the factory pipelines. And it's very inflexible.

They can't react to demand. They can't react to changes. They can't react to, like, recalls or customer feedback. And so Toyota kind of developed a way of, like, called lean manufacturing. And so it was, like, borrowed from that as, like, how can you de-risk startups and, like, make it more iterative? But we're at such a different place. Like, execution is no longer bottleneck. if you have an idea, just like make the app

and then go figure out of the demand because it's a lot easier than just like talking about it. I do want to ask, as far as your mission goes and what you've built today, what's the most important part or moment of your childhood

to help us understand who you are and what you're working on? I grew up in, Jordan kind of lower middle class family

we don't have a lot of money I was lucky enough to get good education but also the hard thing about it

is because I was in a class with kids that had a lot more money than us and so I saw their PlayStation their Air Jordans

their like all the cool stuff they had that it didn't have and that really fucking sucked and I I was always

motivated to like make money. A lot of Silicon Valley entrepreneurs be like, oh, it's all about connecting the world or whatever. Yes, I care about all of that stuff, but my journey starts with, I love programming.

I've always loved programming and understand like how computers work. But at the same time, I was also always motivated to make money with computers. And I felt like we're at a moment in time where you can build

you can build businesses really quickly. ever since, like, I found the internet. And in the, like, 90s, I wanted to, like, build websites to, like, make money. And there was, like, a bunch of ideas. There's a bunch of examples of people making money really quickly. There was this idea of, like, a million dollar homepage where someone was, like, selling pixels and you can, like, be Microsoft. You can buy, like, 10 pixels and write Microsoft's logo, whatever, 100 pixels. And so there was this media, even at the time, it was like, okay, you can make money online.

And I was going to the internet cafe because we don't have internet at home. So I would like, you know, scrap whatever money we had of finding couches or whatever to go pay for like an hour of internet. And I would sit there and like look at different websites, look at what people are doing, different games, different ways. I could like build something that could like make money. And then the main idea was sitting right under my nose, which is the internet cafe itself. I was like, well, internet cafes make a lot of money. But they don't use software to manage their business. Like their business is you go in there, you pay, you know, a buck or two, you get an hour on the computer.

And when you're sitting on the computer, they have to keep track of you. They're like every hour, they look at the clock and they're like, okay, your time is up. They tap you on the shoulder. Also, when you're on the computer, you can like install a virus. You can like delete files. They can do whatever. So I was like, okay, what if I built something that secures the system, makes it so that there's like usernames and passwords,

makes it that so that internet cafes need less employees because once you run out of time, it'll just like boot you out. So I built this internet cafe management system. And because I knew it. And this is the, I keep coming back to this idea of the way to build businesses to build things you know about, either trends you're plugged in or problems around you in the world. So I was 12 or 13 and I spent like, you know, a year or two

trying to build this software for another cafe and then went out and sold it and made quite a bit of money from it. At the time, McDonald's was just opening in Jordan and I took my entire class to McDonald's and no one had McDonald's at the time. How much money did you make from? I made like 500 bucks which was amazing for me

you know salaries and Jordan are not nowhere near the U.S. And so it's like you know so yeah I felt rich

and it felt good to make money and it felt even better to make money from something that I love doing and later on you know as I found more ways make money with the computers and programming.

I bought an Xbox. I was so happy because I'm able to buy things that, you know, my friends' parents buy for them, but I'm able to buy it myself, even as a teenager. So it felt empowering to make your own money. But as part of that, like a lot of the painful part, which is like the tools. Like I was talent programmer and I was fighting with the tools all the time. And you just, you have an idea. you have the right market, you have the right product idea,

and skills to build that, but you're still dealing with a lot of the problems with the tools. So I was always motivated to make better programming and coding tools in order to, in order for me to build easier, more startups, more companies, more products, and for other people as well. And I thought that the internet could be this great wealth, equalizer and generator. Like, you know, there's no reason

that Silicon Valley capture most of the wealth on the internet because it is the most distributed, decentralized, accessible technology in the world. I think part of the reason is because the tools are hard. And maybe there's like a intentional part that the tools are harder than they should be.

And so like the guiding mission for Replit became how do you make coding tools so easy that you don't even need to be a quota to use them? And what kind of world does that create in terms of accessibility to wealth generation, wealth creation? I want to hone in on that point. How much were you offered to sell your company for?

At some point, I was offered when we're very small, not a lot of people, I think six people, were offered a billion dollars. And why did you say no? Because I think I can build a trillion-dollar company. Do you think Silicon Valley or whoever was offering, do you think the people offering to pay a billion dollars for your company were just going to kill it?

Yeah, it was a competitor. And implicit in any acquisition offer is a threat. If you don't sell, we're going to build something to compete with you. Because they're interested in the product, they're interested in the market. You have to make a bet on yourself. That if I'm not going to sell, they're going to go buy some other company or are they going to go build themselves and they're going to compete with me. And do I first have to self-belief that I can go against the juggernauts,

go against the big incumbents and win? Ultimately, for me, it came down to what I would, regret more. Would I regret not selling or would I regret more not achieving my and the company's potential? And I thought that would have been a bigger regret.

Yes, I would be rich but I would be yet another rich asshole. There's a lot of them. But building something meaningful that creates meaning not just in your life, but in other people's lives, whether it's like the team we built, the customers we serve,

the entrepreneurs that we're creating, the worldview we're putting out, the mission we're putting out, the influence that comes from that, I thought it would be a loss not to try. Do you think big tech in general is threatened by your mission? Yes. You know, what's the saying?

First, like they laugh at you, then they try to fight you, then they join you or something like that. they're saying it like that. We're at the like they're joining us right now. Like they initially it was like, oh, rep let us this toy. You know, kids use it.

Yeah, we have a lot of kids using it. We're proud of it. It's so easy that kids use it. Yeah, and it's like, why is that a negative thing? It's probably a good sign. It's a good sign, yeah. And then and then and then when they felt threatened by it, they would like try to buy it.

or like try to compete on the margins or try to like, you know, compete in the press or PR, but not really on the product because they still don't believe in it, but they want to counter the narrative that actually programming can be coding or making software can be something anyone can do.

And finally, when we've, we've grown so much, like our revenue, like 100 X, then they paid attention. And they're like, okay, this is a real thing and people want it and now they're building actual competing products. Do you see what you're doing as,

I mean essentially you see as the democratization of building software, do you think we'll see something similar with tech that we saw with social media where, or with media in general, where the barrier to entry was so low that instead of, I mean, there's two versions of this, there's the big media like CNN, Fox, ABC, to influencers. And then there's influencers to micro-influencers.

And there's tons of people making 10K month with social media. Do you think that's what we're moving in tech where, I don't know, someone like my parents in Kentucky can build an app, make a nice couple thousand dollars a month. And then that's just their job. Yeah, I think that's like an age-old thing that's been happening with technology. since the dawn of time. So thinking about literacy, right?

Like during medieval Europe, like reading and writing was regulated, was only like priests were able to do it. And priests were like controlling the population, right? Before the Protestant Reformation. And so there's like the priesthood control over the ability to read and write. And then we had the Gutenberg printing press. And that was decentralized the ability to print books. and have books available to the population.

People started learning how to read and then learning how to write. And that caused massive changes in the world. Like leading up to revolutions and different religions and democracies and all of that happening because of reading. Like none of the world that we have today would not have happened without mass literacy, right?

We have the big media companies, but you can have a substantive. and there are people making millions of dollars in substack. There are people making 10K on substack. There are people making 2K on substack. And so, you know, Instagram is the same thing. You had to be a professional photographer at some point. And only the professional photographers could make some money.

And then later on, anyone could like take photos and make money. Influencers is the same thing. Really like every skill and every professional. goes through this, where it's usually an elite minority and their gatekeeping, they don't want other people to do it. And certainly programmers in BigTac felt threatened by Rapplet is because they make a lot of money. Like, we were in this like really strange world over the past, you know, 15, 20 years where if you go to computer science school for four years, you're guaranteed, like, the best life and the best, you know, the upper echelon of income and, you know, the upper echelon of income.

And if you get a job at Google, you're set for life. And, you know, in four or five years, you're pulling a couple million dollars. And it's certainly threatening when, like, that skill can be democratized and anyone can do it. Anyone can build it. It's also threatening to VCs and the way Silicon Valley works because Silicon Valley needed to build all this infrastructure in order to and raise all this capital in order to. to like fund, you know, companies need a lot of capital to build the software and market it and all of that. But with social media with tools like Rapplet, you don't need as much capital. Right.

And so I think every piece of technology in history has had this decentralizing and democratizing effect. And every time there's like a pushback from some group of people that are benefiting from the gatekeeping. But then the dam breaks and we live in a fundamentally new world, a much better world. And I think it's true of media. Like the democratization of media has been really good. I think that I, you know, I get, I get my news more from social media than I do. And it has its challenges, obviously, but it's much better than getting propaganda. That's very true. That's super interesting comparison about the Gutenberg press with how religions evolved over time, because, I mean, it really was,

I mean, it would be ignorant of me to say it was the big three, but I would say they were much more dominant and there wasn't so many branches of, I'm trying to think of the dating, but it's a very apt comparison. I'm talking, what do you think everyone is getting wrong about AI and jobs? AI is seen as a replacement,

as opposed to a tool that can be wielded by the most creative, by the most ambitious people who want to make a fundamental change in their lives, in their communities, and their companies. Obviously, AI can automate a lot of work, but to automate that work, someone needs to build that system. Someone needs to be observant enough

to look around their company or look around their school, whatever community they exist in. So we're doing a lot of BS things. There's a lot of, yeah, bullshit work. There's actually a book called Bullshit Work, and it's by David Graber.

And it talks about how, like, most of the economy is kind of bullshit work. I mean, if you think about it, you know, post-industrial revolution, we created a world where machines were doing a lot, but not enough. And humans had to step in and act like machines. And so if you look at any office,

job, most of the time, they're doing things that are easily automatable. They're putting data entry in Excel sheets. They're repeating the same process time and time again, and they know it's automatable, or they're sending the same sort of email, or they're doing the same kind of marketing. And you could say, okay, you know, we can get an AI system to, like, replace that person. But there's going to be more work to be done, not less, because, you know, you get more revenue, get more users,

you're going to have different challenges, different operational things you should do. So it's a hamster wheel. But we can go much faster now. And the way to look at AI and jobs is, how can I upgrade my workforce for them to become generalist business people that can wield AI for the benefit of our customers

on our bottom line. And we see it in a lot of our enterprise customers where the most ambitious people are creating billions of dollars of value of their company. So there's every company out there, every tech company for sure, there are people that are closest to the customer

that feel that they have ideas that could make more. revenue for the business, but they're often blocked by engineering. Engineering don't think the ideas is worth it, or they have their own roadmap, they don't care about it. And now a lot of them are bringing in Replit to work. They're building that idea, and they're making the money for their company, and then they're getting promoted, and then they're giving more power.

And a lot of them go out and hire more people like them. I'm going to build a team of Vipequarters that are going to go around the company and find all the inefficiencies and go solve them. So we have a new role of this like generalist automator. And they're less parochial than the engineer. The engineers like really focused on systems and engineering. They're not as focused on one particular domain like sales and marketing. They can hold the context of the entire business in their head.

And they can find ways to increase efficiency or add revenue to the business. What's the most obvious AI automation that people should be doing today? in the context of companies, we see a lot of just like copy pacing of data. Like you have data in Salesforce and you want to get it into an Excel sheet somewhere or you want to get it into some data lake or something like that. And we see a lot of data entry work that is like just copy pacing

because they don't know how to write code. Engineers are busy. IT is busy. So just like anytime you find yourself copy-pacing things from one place and another and you do it regularly, like it's an obvious automation. Go to Rapplin just say go pull the data from Salesforce and put it in Snowflake, right? So that's in the context of companies. I think there's a lot of data movement that happens manually and that really blows my mind. And even if there's like some process in between that you're doing some transformation in the data that you're doing,

You can also have the AI do that. So for example, like, every company has this person that's called like the deal desk person, every, you know, B2B company especially. And the deal desk person is responsible for generating quotes to give, like, order forms to give salespeople. So the salespeople put in the deal content in Salesforce, the deal desk person pulls it from salespeople. Salesforce puts it in a PDF and then like post it to like Slack or some other chat group. They talk about it. They send it to the client. The client gives them feedback.

They go. They regenerate that. A lot of companies end up hiring a lot of people doing just that. Like, you know, there's also entire complicated software systems that people pay hundreds of thousands of dollars for quote configurators, right? So we hired the steel desk person and she automated all of that. Every time we get a new deal in CRM, there's an AI that generates the order form directly, posted in Slack. The salespeople take it like they send it to the client, they get feedback.

In Slack, they give the bot feedback on the thing. It kind of updates it and they send it back. So a lot of that used to be manual copy-pacing things and putting that from Excel to Word to CRM. So anyone listening to you to this, working at a company, If you want to get a promotion, just look around you and see what people are copy-pacing data around. And you could build a bot, or you could build an AI or can build a piece of software that can automate that. In personal lives, there's a lot of things that, like, if you're someone who's interested in health and do a lot of tracking of health, like, I like to track my sleep. And at some point it has like some sleep issues.

and my doctor gave me like a sheet to track like you know what I ate what medications I took if any if I exercise and then what time I slept what time I woke up any sleep interruptions and like I would wake up in the morning and like fill the sheet and most of the time I'm lazy I'm like I don't want to do it I forget whatever I'm like you know I can automate this I took a photo of that sheet put in a Rapplet you can just skip Rapplet screenshot and just say like making it into an app, made it on an app, and now it got a little easier to, like, put in the, you know, all this information. Then I was like, I still forget what I ate or what medications I took or, you know,

whatever I've done that day. So I was like, why don't, why can I, like, take photos of these things as I'm doing them instead of, like, forgetting and, like, doing it later? So throughout the day, I can, like, take photos of whatever I ate or have done. and then that AI will like generate the content for for the next morning. And then I'm like, okay, but like I'm putting my time to go to sleep and what time I woke up and then interruptions manually. But like my mattress, I use eight sleep, for example,

or you can have a wearable, have all that information. So there's just like total replica. Just like pull that information from there. It asked me for some login information. I did that. So now the entire process is seamless and automated. And finally I can commit to it. So I think in people's lives, there's so many things that are manual and they, like, get lazy and they don't want to do.

And one thing I would say about the age of AI, and this is back to this idea of, like, how do you generate ideas? If you're lazy, that's going to be a virtue. And I don't mean it in a way that you don't want to complete your work or the, but like, if you're naturally just, like, don't want to do manual work, you're going to go about your life and you're going to see all these places. that are just like boring and you should be doing manually and just like go build an app for that. And maybe that becomes a consumer app that I, you could like the app that I'm building for myself, I could potentially put it on the app store at some point

because it is probably useful for other people who want to track their sleep. You've heard of Calli Eye, right? No, what is that? Really? Oh, that's, we had the founder on. I mean, they do like, I think he made like 25 million from it in high school, maybe that was the valuation.

It was just take a photo of food and tracks calories automatically. Oh. It just sounds like you built that for yourself as a personal. These things are easy. And it's not just that. It is, you know, it does sleep and does some of the other stuff. Yeah, yeah, yeah.

I was like, that sounds more integrated. That's really interesting. Huh. That's a cool space to think about, like, all the wearable tech and, like, all the things that are tracking data and building apps that combine all these things in general. But what do you think separates someone who's bad at prompting AI? from someone who's really good at it. You know, I kind of hate the word prompting AI

because AI researchers spend a lot of time figuring out how to make AI respond to human queries. I would say what separates those people are the people who are good communicators. You need to just be good communicator in general. Like prompting humans responding AI. Are you someone who can manage an intern well? If you're someone who can manage an internal model, you can manage an AI well.

Think about it AI as an employee or an intern. Instead of thinking about, oh, how do I talk its language? No. Just communicate well. Just take an idea, break it down to its parts. If its parts are still complicated, break them down further. It's all about breaking down things into individual components. And then giving precise feedback.

And giving as much context and feedback as possible, taking screenshots and taking screenshots, thoughts, taking images. If you want to get really good at just communication in general, I would suggest like doing public speaking. Right. I had actually stage fright growing up. And I would still do it because I just don't like to be held back by my fears. But when I first came to the U.S. and I felt like I want to be able to start startups and be on

podcasts, be a leader, I thought that I need to get past that. And so I thought, like, what is the hardest place I could put myself in? It's like throw myself in the fire to get past that fear. So deprogram the fear by desensitization, right? It's a popular, like, psychology method. I think, you know, whereas in such a soft world right now that like psychiatrists, psychologists, don't treat things like that. But it's always been the case. Like if you're afraid of water, if your kid is afraid of water,

throw them in the fucking water. That's how my dad did it. So I took up improv classes in New York. And improv is, if you're bad at public speaking or thinking on your feet, you're going to do really bad at it. And then I took up storytelling classes. And I did a storytelling show

where I told a story telling show where I told the story. about how I hacked into my school to change my grades. And so I'll have to hear that story, but you continue your thought. So over time, I just got better at talking to people and like trying to communicate complex ideas simply because I think that's what a really good communication is about. And if you're good at that, then you're good at talking to AI and prompting it. My last girlfriend, she was like,

Jack, I'm such a bad commutator. Like, how do I work on this? I was like, just going on an improv class. She didn't really love it, but like I said, I think it helped her a bit. Understand like the basic techniques. Yeah. I think it's super helpful. But it is.

So you hacked into your school to change your grades as a kid? Yeah. So I was going to university in Jordan, Princess of my university for technology. Good computer science school. But, you know, I don't know how to pay attention to things that I don't like, like if I'm taking like a history class or econ class and I feel like whatever they're teaching me is not going to be useful for me in life or it's just not interesting the way they

teach it. I just can't. It's like incredibly painful for me. Let's just sit there and just listen. So I would skip class and I would like go work on my programming, go take a job. I always had jobs during school. And then I would study enough to just pass. And I almost always pass. And some cases in like math and computer science, I would like get good grades. The problem in my school is, in a lot of schools around the world,

is that if you, they count your attendance. So if you missed class too many times in a row, you get, you get disqualified. And so I felt like it was, was kind of deep injustice, especially for people like me who are like creative, want to be doing things, they want to be sitting in class, which I think is like a very outdated way of learning in general. I think we're going to move past especially with AI. So all my friends were graduating. I'm like five years in, almost six years in school. School should have been three, four years.

and I feel like life is passing me by like, you know, I had all these big dreams and ambitions I want to build companies, make a lot of money, all of that, but I was like stuck. And so I thought, I'd use my skills to get over this problem. You can cast almost any problem life as a coding problem. And I thought, okay, I'm going to hack into school and change my grades

from when they fail me for attendance to like a passing grade. So I spent two, three weeks in my parents' home in the basement trying to like break into the systems there.

I would like work on like scripts and do network scans and all of that and I implemented polyphasic sleep Do you know that? That sounds similar. What is that? Michelangelo, you know, the inventor artists,

was like one of the most highest output human beings in history probably ever to come. Would sleep 15 minutes every four hours because he was working nonstop. Tesla did a similar thing I remember. Tesla's probably, yeah, did that as well. And so I was like, yeah, I want to be able to like, you know, work a lot and really hard, be really efficient. So I did that as well. So for two weeks, just madness just going and like trying to map out of the system and figure all that out. So I finally found a vulnerability, a way to get into to the system. But I was worried about maybe I didn't have the right systems. And so I had a

neighbor who was going to the same school as me. I was like, hey, you know, I have this thing. I can change grades. Can I test it on you? And so we test it, we changed the grades, but actually turns out that I was doing that in a replica database, not on the master database. So when he went to kind of pull his grades, his transcripts, it wasn't reflected there. So I spent another two weeks and then found a way or the actual system. And then I was confident that I could like change by grades and actually get to graduate. And I still felt like it was the just thing to do.

It was a fair thing to do, right? So I changed my grades. Calculus and other things that I didn't want to go to school, but I passed. I know, you know, but they failed me anyways for attendance, so I changed my grades. And I bought my gown, went to the graduation parties, took photos, I was just graduating.

One day I'm at home and the telephone rings and I pick it up and it's like the person responsible for the college network and the registration system. He's like, look, the entire network and the diversity is down. And whenever we look at the code, we see

there's a bug in your record. There's a problem where it says you're disqualified from the final exam, but you're also passing. It turns out they had another field that I didn't know about that was like true or false, whether you're disqualified for attention.

And they had all these bad names for the field. So I think they're bad programming on their part. But still, I kind of missed something. So I had a choice. You lie and you kind of live with a lie. Or you come clean. And I just didn't want to lie. And I told them, look, I have something to tell you.

I'm going to come tomorrow's school, kind of share it with you. So I went the next day. and it was a much bigger deal than I expected. All the different deans were sitting there, the computer science dean, the engineering dean. They've been racking their brain a whole night, stayed up all night trying to figure out what's the problem. And so I'm like, this is my opportunity to make it a tech talk

as opposed to an interrogation and pull out the whiteboard and explain all the different problems and all the systems they had and how I hacked into it and really like pulled out all the, all the sobs and all the charisma to kind of try to impress them. And they were really impressed. So impressed and so kind of in the element of the technical discussion that I'm like, all right, now that I've said, I'm headed home.

I'll see you guys later. And I opened the door. I was like, wait, where are you going? Like, we got to deal with you. You just did this thing to our school. We just hacked into the school. We can't let you go. Luckily, the president of the university was someone who was like really enlightened and saw the talent that I had.

And they could have easily derailed my life, you know, but I, you know, I explained like my reasoning for why I did this. And he was understanding. And he gave me at the time the Spider-Man line, it's like, you have a great talent, great power. But with power comes responsibility. I was very sincere about it. He said, we're going to let you go, but you're going to have to spend the summer working for the university trying to fix the, for vulnerabilities and the security issues. And so I was like, that's no problem.

I would love to do that. So I show up in the summer and the programmers there hate my guts because I hacked into their system. And they wouldn't let me in. Like I would go knock at the door where they work and I can see them and they wouldn't let me in. So they didn't give me a chance to contribute and fix the system. And by the way, everyone was cutting me slack.

I think all the professors just was worried I'm going to hack into their emails. And so I had this like implicit power. Hey, it was like notorious. It was really cool as well. I was getting recognition everywhere. And this computer science dean comes and tells me like, hey, you know, I helped you a lot to get past this problem. So for your graduation project, you need to work with me. And what we're going to do is we're going to hack into the school again.

I'm like, bro, I'm not, I'm past that. I'm not even working on security and hacking. I'm just like, you know, I'm going to build companies and things like that. is like, no, you have to do it. Or, you know, I won't graduate you. It was like, okay, I'll do it. And so I'm like, okay, but this time I'm going to build a system. I'm not going to go hack directly into it.

So I built a piece of software that scans any system for vulnerabilities. And sure enough, I found some vulnerabilities in the system. So my graduation defends. So you're sitting in front of the different deans and you're kind of getting a presentation. And I told them how I built the security scanner. And by the way, I'm going to run out of the university system, and yet we have a bunch of foreign abilities here. And then different dean of engineering was like, you're lying.

And I was surprised that he said that. I'm like, why do you think I'm lying? And he's like, because we fix the system. I was responsible for fixing the system and we fix the system. I was like, clearly you didn't. It was like, prove it. I'm like, okay. I also built like something in my system where you can,

automatically execute the vulnerabilities. So I get a shell access to the database. I'm like, okay, what do you want us to find your password or your, or your salary? Or how do you want us to prove? It was like, okay, get my password. So with a query, I get his password. And it was some embarrassing password. I forgot what it was right now, but people have all sorts of embarrassing passwords.

And his face turned red and he got really angry and he got up, shook my hand, and left. Well, turns out I was like a pawn and this rivalry between the two different deans. So the other dean, like one dean was given the responsibility to secure the system. The other one wanted to kind of make sure I rehack the system to prove it. But anyways, at the end, I graduate. And I now have like really great relationship with the university. That's a great story, genuinely. it seems like some type of commonality with really successful founders in your space is like this kind of hacker, black hat, white hat background.

Just on this concept of learning in general cheating hacking the system, how do you think we should be using AI to become smarter? And how do you do it personally? I think that there's a mindset here. there's a mindset of first of all the world was built by people that are not much smarter than you

and there are a lot of rules that are accidental that are there for historical reasons and your job is to find a path to the future find the way of doing things

that's most aligned where the world is headed. Because the advice your teachers, your parents are going to give you are just not valid anymore. My generation, like millennials, the advice that the boomers would give us is that you work hard,

you get into good university, you check this box, this other box, you get a job, you buy a house and a car, get a family, that's it, you know, a few years in, like,

you'll you know 20 years later you're a millionaire because your house property went up value and you retire and live happily ever after that's that doesn't exist that does like existed for like 20 30 years post world war two and then it's gone so it's not about cheating it's about figuring out how the system is rigged against you and how do you actually hack it how do you get past it, how do you do things in a way that is more native and true to where the world is headed today. AI is a great tool for that. Just the fact, if you're someone who just understands how these tools work and understand how you have an amazing time advantage, you can do things a lot faster. Again, there are a lot of programmers that are still doing programming the old way and they're

going very, very slow, and they're going to continue to go very slow until they get obsolete. But there are a lot of new builders coming up today, and they're like, I'm going to just, like, make things with, with Rapplet or tools like that, and I'm going to go just a lot faster. I'm going to iterate faster, make better ideas, build better companies and systems. So I would reemphasize this virtue of laziness. like I think a lot of what we're taught in school is it's not worth it's not worth our time and my approach to school was like how do you get around it like I needed a college education

to get to the United States otherwise I would have dropped out if I was in the U.S. I would have dropped out and it's it's probably a good idea for a lot of people to drop out and not not go to college um And so, but I needed to do that in order to get a visa to get here, right? And so think about your circumstances and think about ways around them. And there's probably ways in which AI could help you get there faster and do things better. And realize that if you truly understand how this technology work, if you're up to speed on it, If you can do things better and faster than other people, you have an advantage and just lean into that.

I think that's super helpful. What is it? Hormosey says it's volume times feedback loops and AI helps you build things faster, iterate faster. On the concept of jobs, what do you think will be the highest paying jobs in the age of AI? Entrepreneur. Entrepreneur. Like, my mindset has always been about building wealth as opposed to getting salary. Right.

There are a lot of different ways to build wealth. But all of them revolve around ownership as opposed to getting salaries. When I got my first job in the U.S. Working for a Code Academy, a company that sold for half a billion dollars, I told them you can just like pay me enough to eat, just give me as much equity as I can give me. I was paid $70,000 in New York City. You know how painful, though? I was living in a studio with other people.

But who cares? Like, you know, if you're young and scrappy, like, you know, you can eat anything, you can sleep a few hours, you can sleep on the ground. Who cares? You know, you're very resilient when you're in your 20s and, you know, early 30s. I think it's get a little worse after that, but you're very resilient. and your job is to build equity.

And the best way to build equity is to start a business. The second best way to build equity is to join a business that someone else started to get equity in it. And the third is if you already have capitals to invest and build equity that way. And I've done all these different things. I started by joining a startup that was growing fast, got as much equity in as possible,

worked as if I was a founder. then I started a business, got some liquidity from that, started investing in other businesses. By the way, a lot of early Repliton employees are already rich. A lot of early Replit employees already got some liquidity out. We let them sell some shares. Some early Replit investors sold some shares as well.

If you invested in Replit in when you were like, you know, six million valuation and we were like many of, order magnetages more than that right now. You made a lot of money, right? So if you're surrounded by really smart people, you don't have to have the best ideas. You can join their ideas. You can also figure out a way to invest in your friends

and join in their venture. Or if you have a certain skill, let's say you're an influencer and you want to partner with an entrepreneur, try to get equity. You can get a lot of cash for sure, but what are you going to do? you're going to get taxed 50% on that cash.

You're going to spend it on nonsense instead try to get equity and build that equity up. It seems like the only two ways to get rich are build a cash flow business and move to Puerto Rico or Dubai and revoke your citizenship, hate your life for a few years and sell a company. We don't have to hate your life.

Like you can have fun building a business. Right. That's fair. I guess I hate your life in terms of not have excess cash to spend on stuff, just kind of like living below your means to an extent. But if you build the business properly, I guess you would have access cash. It would have access to cash at some point. Like, it's a lot more fun later in life to like pulling a six figure salary.

Like you're not going to, it's not an exciting life. I think having wealth is a lot more. just the calmness on your nervous system that you've escaped the rat brace that itself is worth it. And so I don't know, I've never, like yeah, I've went out and partied and experienced all of that,

and you can do that with not a lot of money. Back in Jordan, we used to, like, drive to Beirut and, like, go clubbing for, like, you know, 20 bucks. You know, it's like, you can have fun with not that much money. But then there's like a huge dead zone in terms of like how much quality of life and how much hedonism or fun. If that's why you're interested in up until you get to operational on of wealth.

Like there isn't that much difference between like a student going out. Like some of the best fun you could do just like going out to clubs or traveling or whatever, going to low income places and like having fun there. And then it's not that much better when you're pulling. you know, a few hundred grand a year. It's like you're kind of stuck at that point. But what you want to focus on is building enough equity and up until to the point where you're just like starting to build a family because that's when you're going to have to

like sell down a little bit. You're not going to have as much time and so on and so forth. So I would just like focus on building equity. That's super helpful. Yeah, the first startup I was at, I, had equity. I did the same thing as you. I took a lower salary

to try to get as much equity as possible. But I think I felt victim to the sentiment like, oh, better on yourself. I didn't even like pay the strike price to vest my shares. And I looked at it the other day and the company had sold. And I was like, shit.

I didn't know how much for, but yeah, if someone gives you the opportunity for equity, it's usually worth it. If you think they have a good chance of selling, especially if people are building to sell. It feels like some people are building to sell quickly versus some people are building to sell on a longer time horizon.

I do want to ask you, what do billionaires know about money that people just starting out in entrepreneurship don't realize? What's something you wish you knew about the way money works itself or things you can do with your money like two years ago, four years ago? Money is money,

like cash like dollars are worthless. They are depreciate fast depreciating assets. Like they depreciate faster than your 1976 Honda Civic, right? Like it's like don't hold cash. Like again, that's why I emphasize equity. It's like assets, like buying assets. Like all the assets go up like gold right now is just ripping like a shit coin. So assets go up.

um cash doesn't and do stocks really go up though like if you're just doing s&P or is it just kind of like a wealth maintenance thing from your perspective okay i i mean over the past couple years the smp has been dominated by a few companies then they've gone up a lot like if you invest in an smp what is it like 20 30 percent in 2025 that's huge is it 20 30 percent i'm not entirely sure i think the average is 10 percent with a accounts for inflation and be like 5% returns. Oh, really? That's it? Okay. I don't know. I do have some money with like a money manager that is like just like trying

to maintain wealth. But I also do buy stock. I buy I invest in startups. I buy stock in the companies that I really like. I bought Tesla like a long time ago because I really like the product. It's like the grog brain approach to to investing. is like buy stocks you like. I don't know. I don't want to give like too much of financial advice because like a lot of this is like very, very selective. But like the main idea here is assets,

like understanding inflation, how inflation works and how the dollar is like continuously being printed. Really every currency out there is like continuously being printed. And if you're optimizing for cash, you're really losing wealth and you're not really building wealth. wealth is through accumulation of assets. And those assets tend to, there's certain assets that tend to compound.

I bought Bitcoin. When I left Facebook to start Replit, I sold my Facebook shares, which I could have held on them and though they've done really well. But I bought Bitcoin and put the rest in my company. And so understanding inflation is core and like go research that.

And really understand how the system is working on debt and quantitative easing and all that stuff. You don't have to understand that that much in depth, but the main lesson is the economy is run in inflation. And the rich don't hold cash. They hold assets. I don't know. I mean, I don't really think about money all that much.

Like, and I don't say it in a way to like, you know, to brag or anything. It's just like, for me, just like build cool shit and money will come. Right. like build poor things hold ownership in those things that you build or invest in like important things and products and services that you think are just the future try to predict where the future is headed be plugged enough to figure out where the future is headed like follow the tech news follow the trends and form a prediction about the future and then and they bet on that and then evaluate that over time and continue to like

figure out. Like when you hear about a new thing, don't be cynical about it. Like you hear about like, you know, a lot of my generation like heard about Bitcoin. A lot of them is like, what is this? But if you pay a little bit of attention, you're like, okay, and you understand the system, you understand that this is an alternative asset to cash. And like, you know, a lot of people are looking for alternative assets. And so you have some hypothesis about the future. So I have some predictions and some beliefs about the future and try to bet behind them. And again, most importantly, build things. I think that's the best way to gain wealth

is to build useful things. You're one of the few AI billionaires that has a, it's funny that it's even a contrarian perspective on the future of AI, but why do you think AI isn't going to kill us all? I, a lot of my peers in Silicon Valley

have this like very mechanistic view of life. They think of humans as, you know, meat robots. And I think fundamentally, I think there's something more than that. There's something, there's some spark about consciousness, about humanity that's different and that's special and that life is important.

Maybe it's a religious view, but even without like a very concrete religious view, you can arrive at the same answer by just being perceptive, like just looking at the world around you, like how can how can this all be an accident? Like just like even through science

like when you look at, you know, I see these videos of like how DNAs work and just like this insane engineer, micro like nanoengineering. It just like it doesn't feel like an accident. So there's something about the mystery of life, the mystery of creation,

the mystery of why we're here that I try to start. with and I think a lot of people in Silicon Valley, a lot of people in tech have this very dry physicalist view. Like we're automaton, it's input output. And therefore, the natural conclusion is that humans are not very special. We're going to build a more intelligent being. That intelligent being is going to be so much more intelligent that again, treat us like ants in the same way we treat ants. But no, I think there's something special about him. humans. And I believe in, like, I'm, like, I'll tell you something about my personal life that I think is

very interesting. I think a lot of success comes, comes down to, to like, mindset and visualization. Why does that work? Why does visualization work? It does work. A lot of people swear by it, like, if you visualize yourself as successful, I should, like, visualize the cars that I want to drive and invariably I would get those cars somewhere in another I would like if I want to meet someone I would like have this intent about meeting this person

and eventually I will meet them somehow and it's like random circumstance um like whether it was like people in tech that are respected with us Paul Graham Mark and Jensen um I always wanted to

meet Paul Graham uh I always had this intent of like I would like imagine the conversation we would have program is the founder of Y combinator um and like he also founded the site hacker news and I was reading that a lot and he wrote a lot about programming and things like that I really respected how he thought um and you know one one morning obviously like I do things in the world I don't sit there just visualize like you act one morning in 2017 I get a message from Sam Altman on on on on Twitter. And it's like, hey, I'm Sam. I run, I see, dude, I know who you are.

So Sam before Open AI, I was running Y Combinator, the premier startup accelerator. And it's like, I'd like you to come meet me. And so I went and met him and he gave me this address. I'm like, where is this? This is not Y Combinator. I know my Y Combinator is mine view. And I walked in this place and it was called Open AI. And next to it was NeurLink. It was like a small Elon thiefdom in the mission. And in San Francisco, when Elon was still involved in opening eye. So I met Sam there. And when we sat down, he said,

he had his laptop front of him and turning him around. He's like, read this email. There's an email from Paul Graham. And he says, like, there's this site replica. It's like really cool. It's something I've thought about for a long time where it can like make it really easy to code and host applications.

And we should like reach out to them and get them into the YC batch. And we had like try to apply to YC many, many times before. But it like had this intention. that like I'm going to meet Paul Graham and he's going to like know about a replica that we're going to discuss it and he's going to get really excited about it

and that happened and like two years ago I was like totally different person, totally different industry so it'd be really cool to meet Tucker Carlson I respect him a lot or respect how brave and courageous he is and how he can

speak his mind and how inspiring is like view of America is and what it could be. And through a series of accident, I ended up meeting him. Do you know Martin Schrelli, farmer bro? That sounds a little familiar, but no. You should definitely look up Martin Screlli. Martin Screlli is a farmer entrepreneur that went to prison for some like financial reasons, but he became very notorious for a few things. One is like he bought a Wu-Tang album. that was like one in production.

And like he'd become hated because like he he hoarded that. But like also he just like was like the initial like social media troll and like got into under everyone's nerves and all of that. But anyways, I used to watch his stream because he would like put out a lot of really cool financial material. We talked about financial literacy and I learned a lot from him. And when he got out of prison one day, he was like coding a line. It was quoting a replica. And so like I sent him a message. I'm like, hey, like I saw you're quitting a replica,

it would be called, like, I sent him a tweet. So he responded, we got on a phone, whatever. I ended up investing in his new startup. And, you know, you know, introduced him to a few other people in Silicon Valley. And then one day he's like, you know, I owe you, man. You know, it's really cool. Have you been able to help me get back into the startup game? What can I do for you?

I'm like, you know, you were on Tucker the other day, invited you on the show. love to meet him. And then, and then he made that connection. We got on a call. I have a very similar story with Joe Rogan as well. I got on Joe Rogan's show. So, I mean, I'm kind of digressing a lot, but all I'm saying is that there's like more mystery to life. And I've experienced a fair amount of affecting the world in a very indirect sort of mental way. And there's something to that. a lot of people just don't struggle with the fact that there's more to life

than there's just like mechanistic physics of it all. So when you start from this premise, and I can talk about all the different, I can have a technical debate about AI as well. But just like at rock bottom, I feel like there's something special about humans. There's something about the mystery of the world that we haven't really figured out.

By that same token, just to clarify, because I think this is fairly high-level thinking, but I understand what you're saying about it, is basically you feel as though you're the type of individual that thinks there's magic to life and you will things into existence. And this is something that it is your logical view as well as it is the view that you want to have for humanity.

But do you think that these other individuals that have this AI, Doom thesis, are trying to will that reality into existence? I think there is, a sense in which intentions matter and you can get into a self-fulfilling theories of the world, right? And we've seen that throughout history. Predicting doom and bringing about crisis happens a lot. And so I think it is important to be optimists.

And I think if you have it like a dumer mindset and it is dangerous for that reason for sure. And so I do think like people who are naturally depressed and like not very optimistic do kind of gravitate towards like a more doom or mindset. It's so fascinating because

with you in particular I spoke with another AI founder about this. I was like, it's so interesting that Amjad's thesis is like AI won't kill us all. And like you would assume that would be the thesis of all the billionaire AI founders like, oh, it's not going to kill us on. It's like, oh, shit, it might actually.

But it's really like they're all saying it is, but it's actually not. And I think the only other person that's really in the space with these individuals that has a similar thesis to use, Naval Ravi Kant. There's probably a few others that I just haven't. consume their content. But what's the main logical reason for it? Just to give people some hope, because I've had a lot of Dumers on my show

and I haven't had anyone in your position. So think about how machine learning models are trained. Right now, machine learning models are trained by consuming super large corpus of content, essentially all of content on the internet. and then simulating creating algorithms

so they would learn by creating internal algorithms essentially that's a good way to understand machine learning to simulate how a human might respond to a query

you can go really far with that it's amazing that it works but if you give a query to an AI for something it doesn't have an algorithm for it will fail at.

And there's a bunch of research showing that it's called like out of distribution queries or context. So, you know, if a machine learning model is not trained on a certain like language or flavor of math or something like that,

it will struggle to give you like a reasonable output and you end up in hallucination end up with all sorts of problems. The reason AI is really got a coding because coding has a binary outcome. It's either true or false. Anything that is soft, more requires,

like reasoning through the problems not based on prior material. AI right now is not doing very well at. And AI companies are having to buy more and more per pressure data to get better at these things.

So the way their training models is they'll buy, you know, they'll go into an industry or let's target accounting. We're going to go buy all the data that exists

out there. We're going to scrape all the data that exist out there. We're going to buy all the data. We're also going to get accountants in Africa that are going to sit down

all day and they're going to do spreadsheets and they're going to like talk over how they're doing spreadsheets. And they did that in coding as well. And they're targeting

bi-tac, accounting, science. everything in order to do that. And that's a very repeatable process. And you can create AIs that are very good at doing jobs that people are currently doing. But that is not like a general intelligence that you can drop it like you would drop a human into any sort of environment and they can like learn efficiently. So self-recursive improvement is kind of a fallacy to you like when it comes to things that

aren't binary? I think we need to struggle with the question of consciousness because consciousness I think is an important ingredient of generality in how humans reflect on their decisions and their questions. And, you know, we come up with these ideas that seemingly come out of nowhere. We call them inspiration, uses, whatever people call them. It's like, oh, suddenly I'm struck by this idea like Eureka, right? And we don't know how it works in order to build an intellectual.

to machines. All of like science come, a lot of the scientific, big scientific discoveries come from these Eureka moments. That is not based on prior, on strict prior training. That is a fundamental paradigm shift, a huge jump that comes from somewhere.

We don't know where it is. And by the way, all the original scientists in antiquity and history talked about this in spiritual terms. You know, I remember reading Bertrand Russell book about like the history of like Western philosophy and things like that. And he was talking about Pythagoras. Like Pythagoras was running a religious cult. It was not running a mathematics club.

And they came up with all these different different theories. And if you look at the current state of humanity, we're actually not having that much scientific breakthroughs. And I think counterintuitively, because we've become so mechanistic in our thinking, unless spiritual, less open to mystery and mysticism, that degraded the quality of science. Because science became an industrial process, it became a bureaucracy,

it became government funded, it became all of these different things as opposed to, as opposed to people seeking true knowledge or original knowledge and trying to seek in all these different ways and arriving at it and all sorts of random and interesting and mystical ways. Science used to be more spiritual, is what you're saying?

Yes. I mean, Pythagoras, correct me if I'm wrong, he basically invented like this whole religion of numerology, right? That's probably true. I think Newton spent most of his life, like Newtonian physics was a side project, most of his life was spent in studying a religious text

and doing things like alchemy and random things like that. Tesla said that all of his ideas came from dreams. Yeah. Like spirit world. Einstein too. Like Einstein was like

sitting his chair and like dreaming at the whole time. And so we kind of lose something about what is essential to us as humans and what is true discovery and inspiration when you think about the world in purely mechanistic ways.

And I doubt that without struggling with these questions, that we're going to be able to replicate that into machines and have machines going to go discover new discoveries. They're going to do amazing things for science. Like protein folding is an exact kind of problem that you want to direct an AI on. It's a computational intractable problem.

It's really hard for us to solve in many different explicit ways. It is a large data problem, big data problem. And so we're going to use AI to advance science at all sorts of exciting ways. those original paradigm shifting ideas I think are human in a way that that is hard to describe. Is there any spiritual practice you do

to come up with ideas, be more creative, or just do better in business? I do cold plunges. Anyway, I do a cold plunge. I, it shuts down this like constant, thinking, talking mind that I have, and suddenly I become more receptive to inspiration and ideas outside of me. I don't know.

I think you can replicate some without meditation. I haven't done that very well, but I find physical shock to my body tends to create stillness. And so cold plunge is one of the big ways. you can get it with exercise as well. Like exercise to exhaustion to weigh like you can't really think a lot and a lot of people find that.

But I think you can replicate in different religious practices and things like that. But for me, like I found that cold plunge just forces my mind in a certain way and I've, and it's just like

that creates this clarity and sometimes there's like really exciting ideas that comes out of that. It's all about creating state changes and kind of split testing, like what ideas

or I guess even personalities communication styles like come out of those. But do you feel like I gave you a good answer for like the Dumerism question? I really did want to touch on the cult. But I'll say this.

The aspect of Amjad's thesis on AI Dumerism that I found most fascinating. was in his interview with Tucker Carlson where he talked about the cult, the Silicon Valley of these AI Dumers.

And when I realized that there's some propaganda involved, maybe some weird spiritual shit, I was like, yeah, I don't really entirely believe in this AI Dume thesis. So I'd definitely recommend checking that out. By the way, part of the reason why I didn't focus on that

is because I think that influence is waning. When I did that interview, it was at the height of the influence of the effective altruisms. And I think luckily through various things that happened, including like, you know, tech people responding to it, their influence is sort of winning. But certainly, I mean, in that interview, I talked at length about the self-serving nature of their thesis

and how they use that to like manipulate people and to gain all sorts of like favors and weird things that are happening there. But luckily, they're not as important. Theoretically, do you suspect intelligence agencies or the military have some involvement with AI development? The cool thing about large-langs models is that it doesn't really matter

because it is a very simple process. being at the frontier is very hard, like creating GPD 5.3 or Opus 4.5 or whatever. GPT2 at some point was the frontier. Now you can train GPT2 on your phone. So there isn't enough of a gap that they can create

to have proprietary technology. This technology is decentralizing and becoming more accessible really, really quickly on order of months. Now we have open source models from China. coming out, and maybe there's involved in the government there, but

coming out that are as good as like the models that came out three months ago from Anthropic and Open AI. So maybe, but it doesn't really matter because once we learn what they've done in the latest generation of things, and as

compute gets cheaper, anyone can replicate these things at home. And that's the cool thing about what's happening in the world today. These companies are trying to create enough of a mode around data and computer and things like that in order to block competitors and create more proprietary technology and create an oligopoly around it. But so far they haven't been successful, luckily.

So you would suspect that the top models are at most a few months ahead of what's publicly accessible. Okay. That's fair. I think a lot of people have suspicions that, oh, they're gatekeeping this technology that's 10 years ahead. You can download Kemi 2.5. on your computer and run it.

And you would get a model that is as good as, you know, GPD5 when it first came out. I'm John, you've built a billion-dollar company by making tools more accessible to the everyday person. If you could distill everything you've learned, what's the one piece of advice you would give to make sure anyone listening could find success? I was sorry with intention,

with intention and focus and not and and and perseverance like if you're really intent on finding success you're going to find success no matter who you are like I think if you if you put if you just like visualize it you put your mind into it you have the right mindset you don't have limiting beliefs I can meet whoever like I can there's no difference between me and some other you know billionaire. I can learn all the skills needed. No one is like,

we're better than me. Start with these beliefs and you're going to be unstoppable. I think at some point, the hard thing about scaling a company from like a billion dollars to

$100 billion to $1 trillion dollars is the game changes fundamentally. And that's what I'm going through right now, you know, you become resource rich but time poor. That's like the worst feeling. Like you don't have enough time to do things manually and you have to like delegate a lot. You have to learn how to like build a team and hire and do all these different things. Those are all very important. But I think getting the initial set of success, it's all about being really good at believing yourself,

not quitting, not taking no for an answer, keep going, keep going, never quit. I think those are really the only necessary ingredients in order to achieve enough success to potentially retire or like escape the rat race. I love that. I've been wanting someone to give that. advice for a while. It's kind of like this idea that people say you get what you deserve,

but it's really like you get what you want. And to find a success, you have to really genuinely want it. And if you don't have it, then you actually don't want it. Yeah. Yeah, exactly. That's right. You have to really, really want it. Here's the irony of life. At some point, you're going to get to a position where you're like, do I really want it? And that's when, you know, you have to grow and change your mind. And that's what I'm saying is scaling past that is a different thing. Like, there isn't any material thing that I really want right now. I'm not that liquid of a rich person, right?

Like, on paper, I have large wealth. Like, I can't buy everything that I want. But I also bought enough things to know that that's not what life is about. And I wish that for everyone. I wish for everyone to get to a point that they find that the material world is just, like, not all that worth it. I think money is great in so many ways, but you buy a car and you're like, all right, that's cool.

It's like it's not going to really at some point, you're like, okay, what do I really want? And that question is actually a lot harder. But if you're starting from zero, you know, fix your mind on whatever it is, a car, success, status, whatever it is, and drive really hard at it and you're going to get it. Beautiful.

Well, everyone, this has been your guest. I'm John Mossad. This is the Jack Neal podcast. I appreciate you coming on. Thank you. Thanks for listening to this episode of the A16Z podcast. If you like this episode, be sure to like, comment, subscribe, leave us a rating or review, and share it with your friends and family. For more episodes, go to YouTube, Apple Podcast, and Spotify.

Follow us on X at A16Z and subscribe to our substack at A16Z. Thanks again for listening and I'll see you in the next episode. This information is for educational purposes only and is not a recommendation to buy, hold, or sell any investment or financial product. This podcast has been produced by a third party and may include paid promotional advertisements, other company references, and individuals unaffiliated with A16Z. Such advertisements, companies, and individuals are not endorsed by AH Capital Management LLC, A16Z, or any of its affiliates.

Information is from sources deemed reliable on the data publication, but A6, A16 Z does not guarantee its accuracy.