All right. So, Amjad, I planned to start this conversation with the sort of history of Replit for context for everyone, but I asked Paul Graham what the one question that I had to ask Amjad was. And I expected this to be some sort of like philosophical question about, I don't know, building companies or something nuanced in classic PG fashion, but he said you have to ask him about drifting in Jordan. So, we're going to start there. And then maybe we'll get into Replit, but let's hear the Amjad origin story.
Yeah, like actually drifting cars, and not just drifting me personally, but it's um, yeah, I've always been into the thrill of things. I've always liked sort of most extreme things like I think a lot of entrepreneurs. So, I used to mod cars back in Jordan and race them, and then I got into drifting. And drifting is this weird thing where you have to drive the car sideways for as long as you can. It's like an art aspect of it where you're trying to create some beautiful symmetry in the kind of circles and angles that you're trying to create. Um, but it also got me in a lot of trouble with law and things like that. But um, it was fun. It was I think a great childhood in many ways in that I was able to take a lot of risks in a relatively safe environment.
And so, you grew up in Jordan, so take us from there to the sort of the start of Replit.
Yeah, so I started coding at a young age. I was lucky enough that my father thought that getting a computer was an important thing. He didn't know why, he just had an intuition. So, in the early '90s, we got a computer, first computer I've seen in my life, first computer in my neighborhood, in my school. And I just watched my father struggle with using it. And I just found pretty early on that I had a gift for it. Like I could understand how computers think. And this idea of like sitting in front of MS-DOS and being able to type commands sort of reminds me a little bit like chat or AI in a way. Like CLI commands is kind of the proto prompting. And you could like tell the computer your ideas and it will execute them exactly the way you want. That was like a fascinating thing for me. So, I started coding. And uh, the first business I built was like internet cafe management system. So, I used to go to a lot of internet LAN gaming cafes and play Counter-Strike. And they weren't using any computers, and so I was like I need software to manage their store. So, I built that. It took me like 2 years to build that. I remember making so much money that McDonald's had just opened up. I took my entire class to McDonald's. And um, and then like fast forward to actually the building what becomes Replit started back in Jordan as well. I was going to school, and every time I wanted to do a little bit programming, I would have to spend a lot of time downloading gigabytes worth of IDEs and dependencies, and I was like you got to be able to just open a browser window and start coding. Coding should not be this boring of a task. It should be fun. And so, that was the original spark, and I was able to build an open source project that was the first to put coding in the browser. And that was very interesting, especially for education companies, and this is really what got me into education early on. Codecademy had just launched in the US, Y Combinator company. And they used my open source work, so they hired me. I came from Jordan to New York. I worked at Codecademy for a few years. But the vision that I started crystallizing for me is not about learning code. It's about enabling people to build things, to become builders, and become entrepreneurs. I started hearing stories like, oh, I, you know, fitness instructor, I learned a little bit of coding and now have a coding app that's making me $100,000. Like, wow, this is the real thing. That you need to build an entire platform to help people go from an idea to a business. And Replit was born.
Yes. So, after I left Codecademy, worked at Facebook for a while, and then Replit was as a company was born 10 years ago to the date. So, 2016, Replit came out as basically the easiest place to get started with coding in the browser. And a lot of the applications that I saw were so sort of educational because again, coding was so difficult to get started with. But again, those stories that we kept hearing of, you know, I learned a little bit of coding, I was able to build a business, I was able to automate a big part of my job. It just always felt like we're on the precipice of this explosion of entrepreneurial activity. And the dream just kept getting bigger. And um, you know, although we didn't have AI at the time, like we don't have LLMs, it just felt like we can keep making coding easier and get more and more people building. And that became the sort of the vision for the company is to create a billion software creators.
Yeah, so you, I remember we first talked about the billion software creators many years ago. That's been the North Star from day one, I think. But that was, you're one of the only companies in the space that's sort of operated in the before times and sort of post LLMs, post agents. And so, maybe talk us through that, that's sort of like extremely unique to Replit that you had this vision. I think most people would have said it was sort of highly contrarian, probably impossible, borderline insane 10 years ago when you started. Now it looks completely inevitable. And the best companies obviously go from contrarian to consensus. But you had to make this bet before LLMs and agents and all these things made this obvious, but then you were sort of perfectly positioned for that moment. Talk us through what that was like internally at Replit when all this started really come together.
You know, one of the most fun things to reflect on is to go look at your inbox and see all the rejections from VCs. My favorite one was a VC saying, you know, this is a great thing for the world, and I really think it should exist, but I'm a capitalist, and I don't think it'll ever make any money. If I was a nonprofit, if I was giving money out, I would totally give you some money to go run it. And I was like, no, we're planning to make money. It's like that's not a nonprofit. So, I think the hardest thing about it is having a vision and being in the desert and not really finding a business model for 8 years requires a lot of selling internally. Not just VCs, not just externally, but inside the company, I had to go up every day, every week and re-narrate the vision of what's possible. Now, the cool thing is we always had these individual stories that would come up every month or two of someone discovering Replit and it changing their lives. And now the stories happen like a thousand times a month. But at the time, it would happen once a quarter, and I'm able to tell that story. And so, a lot of people that worked at Replit was really attracted to this mission of making it so that anyone who has good ideas and willing to work hard to be able to get rich. And the technical work was also a lot of fun. So, we had a lot of engineers that were just attracted to that. Can work on AI, can work on infrastructure, I can work on code editors, I can work on all this stuff. And so it just took a lot of culture creation in order to sustain a team that is for all intents and purposes failing for 8 years until the inflection happened. And it all came together.
Yes. So, maybe before we move on to kind of the moment we're in in a more macro perspective, I think time is like one of the last great sources of alpha in the world. And um, in order to take advantage of it, you have to have this sort of maniacal focus on this big vision, a lot of relentless determination, but you also have to be willing to sort of have a very good pulse on where things are going, make hard decisions, probably kill things that are very painful, double down on things that are working. Maybe take us through a couple of the moments at Replit where over those 8 years where I felt like you were wandering in the wilderness and you started to kind of, you know, see some signs of life. Hard decisions, or maybe decisions that weren't even hard in the moment, but in retrospect were like completely pivotal for the company. Um, there was a big emphasis on education early on, that shifted more to develop. Like just maybe talk us through a few of those moments.
Yeah, I think the thing that attracted me to Silicon Valley, which when I arrived turned out is kind of rare, is that people who actually want to change the world and make the world a better place. Turns out it's actually harder to fund these companies and harder to sustain them. But you know, the reason I met you and a lot of other education founders is that there are few founders in the valley that actually are willing to spend 10 years working and like Primer is a good example of that where you had to change US regulation as a seed startup in order to succeed. And so, for me, it was like, you know, we're willing to wait. We're willing to wait for the right moment. We're willing to build all the right technology in order to get ahead. So, how do you actually navigate that? We built a lot of different business models along the way. And, you know, one of them, for example, we were selling to schools. And at some point it was making us a couple million dollars a year, which was like good enough to cover burn. At some point we were selling API for code execution, which by the way now is like a multi-million dollar industry. We navigate a lot of these different pivots. But they didn't feel off track for us because they were all monetization of the same technology that we're building. Ultimately, Replit is about reducing barrier to entry for building. And you can monetize that by, you know, making it easy for students to do that. You can monetize that by, you know, providing services and APIs. I think in 2023, I stood up on a stage like this at TED and I made the prediction that AI agents are coming and they're going to solve software engineering. We're on the precipice of LLMs being able to do the entire, not just write code, debug code, provision databases, and deploy it. And that was incredibly controversial at the time. But I think one thing I'd like to do, which I don't think is as hard as it sounds, is like try to predict where the technology is headed. And if you sit down and like really try to think hard about the trend line where LLMs go from being able to do auto complete to being able to do chat, to be able to do a little bit of tool calling. I mean, obviously, the next step is to be able to combine chat with tool calling with code and run for longer. Clearly, agents were coming. It was just a question of is it 2 years? Is it a year and a half? Is it 3 years? So, 2023-24, I felt comfortable to do that pivot, shut down all the other distractions that we've had. We had to do a layoff because some people weren't just aligned with where we're headed and we needed runway. Now, the problems with layoffs is that I didn't anticipate it's that it kills the dream. Like you're constantly selling this dream. And the moment you let go of, you know, people who've been part of the team and all of that, everyone's like, 'Okay, this company's not working.' And so, we actually lost a full half of our team through attrition and other things.
This is 24. Wow. And the irony of it is 3 months later not only we changed the company, we changed the industry. Replit agent came out in September 2024, first coding agent on the market. And from there we go from 2.5 million ARR a year later to 150 million ARR, 100x in 1 year. Not only that, we inspired a whole generation of AI agents. That was before Claude code and all of that. All of it came after. But it was the right bet to make. It was the right time. And we were lucky to like not quit and have enough talented people that really believes in the vision.
So, you launched Replit agent end of 24. Over the last 18 months, probably everyone in this room has experienced, I think, probably the most rapid moment of technological progress that humans have ever encountered. And everyone's thinking about how they build teams differently, thinking about go-to-market, sales, how do you build products. Um, I'm curious, maybe we get like a little bit of an anthology of the last year or so from your perspective, from inside Replit, but also what you've seen. I mean, you all sort of kickstarted all of this. And now the world will never be the same. And in some ways it's sort of dizzying, I think, even for those of us that sort of like watched it all happen up close. Like how have you experienced the last 18 months? How would you sort of describe what's happening? And then as you think about kind of what's coming next, where do you get really excited?
You know, we had an inclination that if you give LLMs a coding environment, it supercharges them and they become more general. They become sort of AGI light or AGI like. And it was just like an intuition. It wasn't entirely clear. Clearly, the reason to use coding is to create software. But then we started seeing these use cases around, like I said, marketing, around recruiting, around sales. And it's like people making apps to automate some of these things and agents to automate them. And then it just sort of struck us. It's like, 'Okay, this is a general purpose work tool.' And then, you know, a lot of others in the industry created great products, obviously. And I think the consensus now, wasn't as obvious back then, is coding agents are general agents. They change how we operate and they change the future of work itself. So, if you're not a coder, coding agents have actually more impact on your work than it has on engineers. Like engineers will get by. Engineers will use whatever tool available. But if you're any kind of knowledge worker, you can use coding agents to automate just medial tasks. You can automate entire workflows. And you can just fundamentally increase your leverage. You can become a 10x marketer, a 10x sales person by using tools like Replit. So, I think the moment we're in, I think the engineering part is fully understood. I think the thing that is yet to be understood by enterprises, by everyone else, is that coding agents change the future of work.
So, how do you think about how that connects to building teams? What I've seen what's happened at Primer is you have these moments where you give everyone access to these tools and then there's sort of this explosion of things you could have never predicted on how they use them. How do you think about this change? I mean, the obvious thing is like each person can get more done. They can work more cross-functionally. You can have, you know, operations people that are shipping new products. But how do you think as people here thinking about building teams, especially over the next year, how do you think about the sort of concrete ways that this impacts how they think about building their own teams?
I think this moment changes how you should screen for people. And you should really screen for entrepreneurial. We always said in Silicon Valley, hire the smartest people. I don't think that's true anymore. I don't think raw IQ matters. I think what matters is being able to get things done and being able to be resourceful. And so, screening for people who would want maximum impact. No matter what their pedigree is, what their background is, how smart they are. Is like, are they hungry enough? Do they want to get things done? Because, like you said, these tools are filtering into various startups and enterprises. We see a power law. We don't see everyone adopting them. They'll dabble with it. But you'll see one person create millions of dollars of value. Um, you know, at one of our customers, Zillow, one product manager increased top line revenue by tens of millions, perhaps 100 million dollars, by increasing conversion of a certain part of their business. And so, it always existed this power law. Steve Jobs talked about the 10x engineer, right? He said, 'Engineering is unlike, I think his analogy was like taxi driver. You can be 20% better taxi driver. Engineering, there's no limit to the multiplier of how impactful you can get.' And now that translates to everyone in knowledge work. And so, I would say hire entrepreneurial people. And it really doesn't matter what their exact background and experiences. And if you do a good job at that, first of all, you're just going to have a lot of fun at work. It's just like exciting when everyone's building, everyone's presenting what they're building. And you're going to make a lot of progress. We had one of our recruiting leads used Replit. We have, you know, sort of these high priority searches for candidates and for specific roles. And he used Replit and he built this system that basically takes the job description, takes everything we're doing at Primer, and then pulls in that candidate's background, and then weaves together this website that essentially personalizes why the role is a great fit for their background in very specific terms. Like you worked at this organization, you had this impact, and you had this outcome, and that's exactly what we're looking for at Primer. And it takes like 6 minutes to spin these up. And so, when we use these for executive roles or leadership roles, he told me this morning that the conversion to response rate and opting into a conversation is double. Doubled. Before when it was just a normal cold outbound you would send to any candidate. And no one, he just had that idea on his own. And so, it really feels true that agency is extremely underrated. By the way, it also works on me. Some people apply for jobs at Replit and they create a Replit app with a resume and why they... And you go, 'Wow, that's so cool. Let's interview you.' Pro tip.
