Replit's CEO: "You Don't Need to Code Anymore"

He runs a $9 billion company whose engineers almost tripled their code output this year thanks to agents. So can Replit run itself? That's this week on Platformer. This episode is brought to you by Jira by Atlassian, where teams and coding agents get the context they need to do the right work. Try it free at jira.dev. That's J-I-R-A dot D-E-V. Welcome to Platformer. I'm Casey Newton. This week, our guest is Omjad Masad, co-founder and CEO of Replit, which started as a website where beginners could learn to code in a browser and is now one of the biggest names in vibe coding. But it turns out he hates that term. We'll get into it. But first, here to give us some data on the state of AI and jobs is Platformer fellow and Gen Z AI correspondent Ella Marchianos. Ella, how are you today? I'm doing very well. How are you? I am good. Now, when it comes to vibe coding, have you ever used Replit or a similar tool, or what is your vibe coding tool of choice these days? I'm only doing stuff in Claude, to be honest. The platform's locked me in, and I'm never doing—I think the thing I use it for the most is creating apps that waste my time. Interesting. Such as? I was actually recently hanging out with a friend of mine who I discovered had, like, independently developed the same addiction as me, which is building apps that let you create, like, Elo rankings of things you like. We're like— Okay. Yeah, it's like the ranking system in chess, and so you just, like, swipe between two things you prefer. I do it for dresses if I'm online shopping, so it gives me, like, an extra kind of, like, recycling of the online shopping experience. My friend actually did this for all of his Spotify liked songs to get an additional metric on, like, what songs he prefers the most. And so, like, you do this long enough, and then at the end you essentially have, like, a ranked playlist of the songs in the order that you actually prefer them. Yeah. Well, listen, I've heard of worse apps, I'll say that much. Well, yeah, this seems like a pretty fun way to spend your time. And speaking of spending time, it is time to find out once again what's going on in AI and jobs. Not my best segue I've ever done, but it was what I had today. I appreciate that, Casey. Thank you. So basically, the World Bank has a new report out about AI in developing countries, and it's kind of a change on the narrative that I've at least been seeing a lot on X.com, a website which I unfortunately still have a lot of contact with, where people are talking a lot about, like, AI sovereignty, as in, like, countries who aren't at the frontier developing their own AI models, and about how, like, if they aren't at the frontier, they're going to be in the permanent underclass. They're going to get washed by OpenAI and become irrelevant. That's the narrative that the World Bank really disagrees with. In particular, the World Bank thinks that currently developing countries are better set up to benefit from AI development because fewer jobs in developing countries are susceptible to AI automation. Okay, so tell us a little bit about why they think that. Yeah. So, like, the headline estimate that they have here is that they've kind of put together an index of, like, jobs that are very at risk from AI versus jobs that stand to see productivity boosted but won't necessarily be eliminated by AI. And for jobs that are at risk from AI, in developed countries, they say 14.2% of jobs are at risk from AI. I might put that a little higher. But in developing countries, they say that only 4.5% of existing jobs are at risk from AI, basically because developed countries have, like, more professional services jobs, more IT and communications jobs, and they think computer-based work is more vulnerable. But then the amount, yeah, the amount we expect they expect to see benefits is, like, about 16 to 18% both in developing and developed countries. And what do you mean by benefit? Like, is the idea that these jobs will become more valuable and so wages will go up, or what? Yeah, the idea is the jobs will become more valuable. So, like, employment will stay the same and, like, hopefully wages go up. So, yeah. So, I mean, look, this is a—we always love to feature optimism on the show, and I'm somewhat heartened to hear that the World Bank has drawn these conclusions. At the same time, when I hear that, like, only 14% of American jobs could be subject to automation, that seems sort of surprising to me, given everything else that we've ever talked about on this show. So, are there any nits that you would pick with this particular report or its methodology? Yeah, to be clear, it's 14% in like an index of developed countries, not just American. I see. Yeah. But anyway, yeah. So one thing that I just want to impress in the mind of our listeners is when economists say estimates to you about what percentage of jobs are exposed to automation, by default, do not trust them. Okay. So the methodology behind this particular thing is, like, somewhat complex, so I want to be, like, a little bit more careful with this one. First, I'll say for background, the studies that are worse than this World Bank report that, like, some other economists rely on, of which there are, like, dozens, are studies where people literally, like professional economists, will literally ask the most recent ChatGPT, Do you think you can do this task? Is this task exposed to AI automation? And then they will make their index based off of that. And please don't do that. If you're listening, please don't do that. Yeah. Thankfully, the people who the World Bank is, like, basing this analysis off of have, like, created a methodological innovation off of this, which is that they have this kind of complex pipeline where, like, they survey people about a certain task, they survey experts about a certain task, and they survey LLMs about a certain task, and then they kind of, like, ensemble these things to try to get a reasonable balanced picture. But I will say that while this is, like, a 2026 analysis based on their, like, idea of what tasks will be automated, the original study that tries to figure out what tasks are vulnerable to automation doesn't use LLMs after GPT-4o. GPT-4o? I was using that in elementary school. I'm sure. I'm sure. Yeah. I mean, all kidding aside, that is like a very old model. And I think, like, if you haven't updated your sense of what jobs might be prone to automation since GPT-4o, like, I don't really think you're part of today's conversation. Yeah, this is like, yeah, this is like really the problem is, like, I can't even necessarily blame the people doing this analysis because, like, it takes a really long time to, like, make these studies. They had to, like, wait on survey data from, like, many thousands of people. They had to go through peer review. They had to, like, ensemble these things, do a bunch of, like, complicated statistical analysis. And so, like, they just don't have time to make an analysis of the quality that they're making and have it be the most recent model. But, like, this is the problem is, like, even if you're listening to the best economists, this is what's going to happen to you because it just, like, takes them so much time to, like, put together these types of studies. Yeah. Something else that might, you know, make your statistical analysis take longer is if you're using GPT-4o. So I hope that they've at least upgraded to, you know, o5.6 since this work began. Last question on this, Ellen, something I've always wondered: does the World Bank have ATMs? To be honest, I did not find that out while researching for this segment, and I still don't know. Well, listen, if any listeners know, we're curious because I need to get out a little money before I head to the airport. But before we do that, Ella—and Ella, thank you very much. Very interesting study and data, as always. A lot more to say in the conversation that is coming up about the potential for AI to automate jobs away. Somebody with a very interesting perspective on that, the CEO of Replit, Amjad Masad, coming up right after the break. Jira by Atlassian is where agent speed meets team intelligence. Assign any work item to your favorite coding agent and see exactly what it's doing without leaving your flow. When something stalls, you're not digging through logs. You can see what's running, unblock what's stuck, and stay in control at scale. Don't let your agents start cold. The Teamwork graph feeds them context from across your entire stack, delivering 44% more accurate results with 48% less token usage. So the work they pick up is the right work, done right the first time. Try it free at jira.dev. That's J-I-R-A dot D-E-V. My guest today is Amjad Masad, co-founder and CEO of Replit. If 2026 is the year that vibe coding went mainstream, Replit is a big reason why. The company shipped one of the first commercially available AI coding agents in September 2024 and has ridden the surge of interest in coding tools to become a real leader in the space. Amjad grew up in Amman, Jordan, in a family with Palestinian roots. He was writing software as a teenager and selling it to local internet cafes. He came to the U.S. to be a founding engineer at Codecademy, a company that teaches people how to code, before going on to run JavaScript infrastructure at Facebook. In 2016, he founded Replit with his wife, who is the company's design lead, and also works with his brother. The product started as a place to run code in a browser without having to install anything. Today, it uses the more familiar chatbot model. You type a sentence and an agent designs, writes, tests, deploys, and hosts an application. In March, the company raised $400 million, led by Georgian, at a $9 billion valuation, which is triple what it was worth six months earlier. Replit says more than 50 million people build on it, including 85% of the Fortune 500, and that it's on track for a billion dollars in ARR by the end of this year. That's up by $2.8 million in revenue in all of 2024. And in July, the company published a really striking claim, a blog post that said its engineers nearly tripled the amount of code each person ships just over the past six months, and said the quality of the code stayed consistent, and that the company can now close its customer support tickets 60% faster than it used to, all because agents are now working alongside engineers at every step. That's why they're calling it the self-driving company, and they say that their people don't feel automated, they feel promoted. The company has also had to fight its share of battles against competitors like Linear and Cursor, and also against Apple, which for a time blocked Replit from being hosted on the App Store. And since you could argue that they compete with Anthropic as well, here I offer my usual disclosure that my fiancée... My fiance works there. So there's lots to talk about. Here is my conversation with Amjad Masad. Amjad Masad, welcome to Platformer. Thank you. I'm a big fan of the show, so I'm happy to be here. Thank you. Well, so I want to start with a vocabulary term, because Replit, I think it's fair to say, is one of the biggest names in vibe coding. But I know that you don't like that term. So tell us a little bit about why that term grates for you. Well, look, I mean, it was never about coding. It's always about solving problems. It's about, you know, creating things, creation, making software, making apps, making exciting experiences, making games, making all of that. And the whole mission of Replit was always to make it so that anyone can get access to this technology. This technology has always been accessible to 0.5% of the human population, those who could afford a computer science degree. And when you put coding in the name, you're immediately, you know, you're kind of signaling to people that they have to learn something complicated. Of course, you put vibe in front of it, and maybe that makes it a little more digestible. But still, I think I talk to a lot of people, and they're just like, Yeah, I'm not into coding. Well, have you tried vibe coding? Well, yeah, it is coding. I'm not a coder. And the other thing is, like, all of AI is vibe coding at this point. Like, AI, you know, any work app you use, cowork, all these things, they're writing code under the hood. And so what we realized at Replit early on is that coding is the substrate, is the infrastructure that makes AI really powerful. And so we'd rather think about Replit as a general problem solver, as a creativity engine. That makes sense. I feel like if we're going to defeat vibe coding, we do need an alternative term. And I wonder if folks have suggested any to you, or if you've tried out a few on your own. Um, I have not. I, you know, I hear all sorts of kind of suggestions from people. Nothing really stuck with me. But I think it's one of those things that will kind of disappear into the background and just become about creation. Like, you know, I want to make something, and, you know, you think of Replit, you think of apps like that. That makes sense. I mean, you know, it's sort of on my mind as a non-technical person who now has two apps on my Mac that I made myself by typing into a box. It feels really cool. I almost want to go in the opposite direction and just tell people I'm a 10x software engineer, just like, albeit a non-technical one. So, interesting you're saying that. I think not only that, the thing is not only a 10x software engineer. I think in the future you're going to be a 10x mathematician. I think you're going to be a 10x biologist. I think you'd be a 10x machine learning researcher. So take me, for example. Like, I've always been into AI theory and, like, how AI works and the future of AI and all the sci-fi around AI. But I was never a researcher. I wasn't trained to be a researcher. You know, I don't know how to construct a neural network from scratch. I hired great people who could do that, but I myself, I lack that skill. And so I was like, okay, you know, AI has gotten really good at software. Can it do other things? And I'm seeing people use it for math and all sorts of things like that. And so I was getting into— And so I was getting into chess around the same time, which is a very old ripe age, 38, of getting into chess. You know, you get into chess when you're six years old. I've been getting humiliated by, you know, three- and four-year-olds online, and I was like, man, you know, this is not going well. So let me actually do what I do best, and, like, I created a chess engine. And the challenge I wanted to give myself was, can you create a fully neural network-based chess engine? Because most chess engines are hybrid, like AlphaGo, AlphaZero. It's sort of like part neural network, part, like, sort of traditional programming and search and things like that. And if you look at the literature out there, and I kind of used ChatGPT, whatever, to look at all the literature, it's actually kind of an unsolved problem. And I was like, okay, let me take on this challenge. And I was able to train a chess engine, and I put it on Leechess, a platform kind of like Chess.com, and it's now playing against humans, playing against other bots, and it's got an ELO score of 1,300, and when it's playing against humans, actually 1,700. And so I'm like, wow. It's sort of like in The Matrix when Neo is like, they ask him, like, do you know how to fly an airplane? And he kind of, like, lags a little bit. I was like, I know how to fly a helicopter. I feel like I have the same power. Like, whatever you get, like, I have this, you know, genius intern, but that's also really sloppy and bad at ideas and things like that, and I have to direct it in a certain way, but I can do anything. It's really fun, and I feel like I've had similar moments myself as I've dabbled here. But let me go back in time a little bit, because you've said publicly that you think it basically no longer makes sense for most people to learn how to code. And I think that's interesting because you were a founding engineer at Codecademy, a company that taught people exactly that, right, how to code. So tell me about your journey from teaching people how to code to deciding that they shouldn't bother. What was the moment that you said this old way of doing things is over? Yeah, I mean, my journey in kind of... Thinking about coding education and all of that starts really young. Like when I was, you know, I learned a bit of programming when I was seven or eight years old. My father got us a computer when I was—it was very unusual. I grew up in Jordan in the '90s. There wasn't any computers around me, so it was like one of the first computers in my neighborhood. And so I built a business when I was 13 years old. I was selling software to internet cafes. You know, internet was expensive. You had to go to a place to use the internet, and they didn't use software to kind of have accounts or protect the machines. People could download, you know, malware and things like that. So I built, like, this entire management software. And the thing that really annoyed me is, like, dealing with the minutia of programming. Like, how do I—like, I made this great application. How do I get it into someone else's machine? So deploying it, like running into, like, basic errors. And I was like, programming is really fun and could be really a great way to become an entrepreneur if you're someone who's interested in impact and making money and all of that. But also there's all these annoying aspects of it that should kind of be automated. And so that's kind of the spark of the idea. And then I ran into another problem in college. I didn't have a laptop, so every time I go learn a little bit of programming, do some homework, I have to set up the programming environment over and over again. And so luckily you started programming after AI, and AI could do most of the setup. But, like, back then, downloading Python and Java and all of this stuff, and you have a missing DLL or whatever, it was a huge, a huge headache. So I was like, okay. Why can't I open a web page and start coding? And so that was kind of initial idea for what would become Replit. So I built the first open source coding engine that could run in the browser, and that went viral on the internet. And that was like an insane moment for me because here I am, a kid in Jordan hacking in my parents' basement, and I'm looking at Hacker News, and it's the top story in Hacker News. People are talking about it at conferences. And then there's like this little YC company. Literally before Demo Day, they went viral. They were on Hacker News, they were on Twitter, they were on TechCrunch, all these different places. And I look at it and I'm like, man, that kind of looks familiar. Well, it turns out they're using a lot of the software that I'd written. And so they gave me an O-1 visa. I become the founding engineer. I moved to New York and we start working on this company. And the kind of stories we hear were amazing. It's the kind of stories that I hear every day now, but back then we would hear them once a quarter, which is I learned a little bit of coding in Codecademy, and now I used to be a fitness instructor, and now I built my fitness app, and now I'm making like, you know, $10,000 a month, and I feel like I'm financially independent and all of that. I was like, wow, this could be really impactful. But still, like, the coding aspect of it kind of fits a certain personality. You have to have a lot of time. You have to have a high tolerance for pain. And no matter how much we made Codecademy easier and coding easier, it was still like not attracting the general population. And so when I went to start Replit, the idea was that let's kind of blur the distinction between learning and building. Like, you should be learning as you're building, because learning on its own is kind of boring. I was never a textbook kind of guy. I actually did kind of poorly in school. I always learned by doing projects. And so Replit's core thesis was that there shouldn't be a learning step and then a building step. You should just go straight to building. Now, without AI, that's actually incredibly hard. We did a lot of, like, more classical automations around generating code for you using wizards and things like that. But I always thought that the key to solving this was NLP. No one says NLP anymore, but LLMs came out of NLP, natural language programming kind of processing field of research. And so when I saw GPT-2 in 2019, I thought it was like, oh, we're getting close. Like, we have these new machines that could understand human language. Surely we can translate that into code. And so we started building on top of that. And I wrote an article, and actually it's on a blog in 2020, and I said, we're on the verge of solving programming. I think for the foreseeable future, from my vintage point at the time, you still need to learn to code, but you have to learn less and less of it. And I actually came up with a—yeah, I was on the YC's podcast. They called it Amjad's Law, but I said, like, the returns on learning to program is doubling month over month, which was true up to a point. Now, in 2025, when we started getting, like, Opus and the new Claude models, I was like, okay, we're getting to a point where I need to be honest with myself. Like, my entire life mission around teaching people coding and making coding easier is changing. You don't need to code anymore. And I started building things without looking at the code. And I was like, okay, this was before Karpathy even coined vibe coding. And so I was on TBPN, and I said, You know, it was like actually kind of a painful moment for me because here you are, you dedicated your life. I'm like, you know, 37. I've been working on making coding easier for literally, you know, 15 years, like, you know, most of my adult life. And I'm saying everything that I worked towards is kind of moot. And the way forward is that we make programming so easy that anyone could do it and you don't need to learn to code anymore. And so I said that, and that was hugely controversial. A lot of engineers on Twitter and Hacker News kind of attacked me, like, How could you say this? Like, you know, it's not true. You're, like, going to get people in trouble because they're going to build apps that are insecure. I was like, Well, you know, any problem AI could create, AI could solve. Like, we could create a security agent. So none of their concerns were actually unsolvable. And lo and behold, here we are mid-2026, and most engineers, not just vibe coders like you and I, most engineers don't look at the code. Like, I talk to engineers at our company, they kind of only look at the code when it's about to get deployed. They do a code review, and increasingly we're automating the code review process as well. And so it was like this big, big kind of painful moment. But I think as an entrepreneur, you have to be honest with yourself. Otherwise the company would die and you would not be aligned with the future. Totally. I mean, it is amazing to me, sort of like through that entire story, how much has happened just in the last 12 months, right? Like, because I can remember, you know, when Claude Code first came out and thinking, should I get this running on my laptop? And disclosure, my fiancé works at Anthropic, and he basically said, It's not worth it. It's going to take too long. Like, it's just kind of a whole pain, you know? Like, it's probably not, like, ready for you yet. And then you fast forward to November, and all of a sudden you basically just type in the box, Let's go, and all of a sudden, you know, I can skip the last 15 years of your life. So I can understand why that would be super painful for you, but it's also great for me. And obviously, you know, you guys are thinking really hard about how to turn that into business opportunities for yourself. And speaking of that, and at the risk of getting ahead of ourselves, I wonder how much of this conversation applies to design. In June, I interviewed Figma CEO Dylan Field about AI and design. According to Dylan, designers have never had a brighter future. You all just released Replit Design, which promises to automate a lot of the design process. So is design about to hit that same moment that coding did? And what do you think it means for people who are designers today? Arguably, it already did. Like, when I look at how designers work at Replit, like when we released Agent 4, most of the design work happened inside Replit. Like, why would I make a static mock if I can make something dynamic that we can touch and feel? And, you know, one of my favorite kind of books to read is about Apple. I've read all the different books about Apple. Like, you know, obviously the iSixth book is a great one. It talks a lot about this. But there are ones that are really in the process of, like, how Apple used to work. I don't think it's the case anymore. But Apple had this, like, huge— Demo culture, where even just like the original iPod, even before they wrote any software or manufactured anything, like Jony Ive created like a little kind of prototype you could touch, and like the scroll wheel, like you could interact with it and it gives you haptic feedback. It didn't do anything, but Steve Jobs would like show up to these meetings and would use the thing, would touch it, feel it, and see how it works. And they would do it on a weekly basis, and that was their invention kind of engine. And like, for example, they talk about a lot in that book, I'll remember in a second, about how they invented the keyboard. Actually, you know, if you have a standard keyboard on your iPhone, it wouldn't work because everyone has, well, like, you know, fat finger, all the different... But autocorrect made it work. And the way they arrived at autocorrect is through this iteration process, and every week they demoed to Steve, and Steve would tell them, This is shit, and the next week they would come back with a new prototype. Eventually they got it. And so I think design has always been about prototyping. Somehow we landed in this world where we have designers on one end creating these static images on a website, and then we have programmers that are trying to turn these static images, and this is like kind of the factory model. I think, you know, with technology, we tried to kind of cast it into the shape of what humans already know, which is manufacturing and the factory model, the industrial revolution. But fundamentally, the information revolution, the AI revolution is different. And now we're finally getting to a place where, you know, design and engineering is kind of blurring. And so we actually have a group at Replit called design engineering, which is like people are good at both, and we actually combine it with PM as well. So PMs are... Can do all three. Can design, engineer, and kind of the roles inside the organization is kind of blurring. And so I agree that design has a bright future, but it's not the same process. Like, you can go straight to code, and you can create something that you can touch and feel from the first go, and that doesn't mean you're moving your mouse and kind of manipulating pixels on the screen. Yeah, I mean, I have to say, you know, I was playing with the Claude design product, and basically just one afternoon said, what would, like, a complete redesign of Platformer look like, and made it in my browser in the course of 10 minutes. And, you know, on one hand, I'm like, this is incredible that I can do this, and on the other hand, if I were a designer, I think I would be so irritated because if nothing else, it would make me feel like my job was about to change a lot. And it sounds like that's already starting to happen at Replit. Yeah, this brings a good point about jobs and change. I think that we all, all of us in the economy, need to be very adaptable. And that kind of the moment that I talked about, that painful moment to admit that your life's work is obsolete, we all have to go through that and know that the skills will change and we need to go to the next thing. Now, there are first principle and fundamental things that will never change. Design thinking is something that people could do when they're kind of designing a grocery store or designing a house or designing a website. And so this fundamental UX thinking is a skill that, you know, machines are not good at yet. I don't know when they'll get good at it. But so there are fundamental skills that don't change, but how you apply those skills will constantly change. You know, you go from moving pixels on the screen, now you type a few words and you'll see a lot of different options. The cool thing about Replit that's different than Claude Design is something we call ambient intelligence. When you click on an artifact you created, we actually give you a bunch of options, create variations, you know, do this differently, whatever. So we're actually prompting for you. And so a lot of what, you know, designing with Replit is, you don't even have to articulate what you want, because I think one of the things that's hard about design is I don't have the language for design. I can't say, oh, make this a neo-brutalist... list kind of way, I, you know, which I learned recently, but I don't have the language, but I can, like, click through the different options until I find something that I really like. And then maybe I can ask the agent, What do you call this? so I can learn that style. So kind of you're learning as you're going. The other thing that's different about Replit is the model choice. If you haven't played with Kimi for design, it's actually With Kimi for design is actually really good because one thing they've done differently with Claude, Claude jumps straight to code. Kimi actually reasons about design in the same way it reasons about software engineering. And so I think Kimi is actually now a better designer than Claude, and we have it as an option on Replit. That's really interesting. And I like the way that you're talking about this. Like one of the things that I wanted to do with this mini series was give people tools that they could use to make them more productive or just kind of like put them in a better position for whatever might come. And I think a tool that is prompting people as they're building is essential because as people, you know, as AI enables people to explore new fields, they're not going to have the language for the design that they might want or the app that they might want. So I think for somebody, you know, like you and with your company, building tools that sort of hold people's hands through that process is going to be really important. Yeah, 100%. Like I think that's like a big thing for me with image generation. Like when Midjourney first came out and, you know, some of my designer artist friends were getting amazing results, I was getting this dogshit results. I was like, What's going on? It's like I don't know how to prompt, and I don't know why the image generation companies haven't really solved this problem yet, but there's a lot of UX pattern you can use to kind of hone in on your design or aesthetic more visually. Yeah. Let me ask you about this post that you all did about becoming a self-driving company. You wrote that in the first six months of this year, the lines of code contributed rose by 5.8x among your engineers. Basically, your engineers have almost tripled the amount of code that they are shipping. Tell me about the practical effect of this. Like, is it the case that shipping more is better? And if so, where is that showing up at Replit? Right. So a few things about that. One is a lot of what we do at Replit is about inventing the future of work. That is both in our product, but also in how we operate as a business. You know, I, for one, in many ways I hate work. I, you know, this, like, you know, this new Silicon Valley mindset mentality, 996, whatever, it sounds lame as hell, you know? And it's like, oh, we want to replicate, you know, what the Communist Chinese Party wants to do. Right, right. It's like, no, that's not the future. The future is actually having fun at work. Work feels effortless. Like, the best times in my life, the most fulfilling times in my life working, I didn't feel it was work. Like when I invented Replit and the technology that was based on Replit, I didn't work a second. It was just pure invention, creation, having fun and getting excited about the future. And so, okay, how do we work as a company and live that ethos? So the idea of a self-driving company is how do you automate the bureaucracy? How do you remove the bureaucracy? How do you kind of make the running of the company more happening automatically in the background so that we all can focus on the creative thing? As CEO, I don't have to do a lot of, like, a lot of what you do as CEO is routing. Really, CEO is a kind of glorified router. You're like, you kind of, you say, oh, this person is working on this. Make sure you're talking to that person. Let me introduce you to this person. Let me move information from this place to another. So, okay, first thing is how do you make information universally accessible, right? Like, you should be able to tag a bot. We have a bot internally as part of the blog. We talk about it. It became sort of like the brain of the company. You can ask it any question about the business, the metrics, where the business is headed. The cool thing about it is, and we just published another blog about kind of a self-driving data semantic layer that connects and understands all the different data sources that we have. So when you ask a question to that bot, it's not going to just one data source. It's going to our Notion pages. It's going to our Postgres database. It's going to the code base, and it can synthesize all that information into a report that it can give you. So you sort of eliminate— The hierarchy as a mechanism of information routing. And you go one by one, and you sort of, I think you get to a point where the company looks less like a hierarchy and more like a network, like an open-source project of sorts, where there isn't this strict hierarchy. Now, when it comes to code, I think that, you know, volume of code is not always better, but it depends on the period and time that the company is in. I think AI really is a hits game. So Replit Agent invented the coding agent in September 2024, and Claude Code and all these others copied us, and they would tell you that as well. But we kind of invented that, and the reason we were able to do that is because we have high velocity as a company. And so now we feel like we're in another period. You know, vibe coding is a thing. Replit is a business successful, continues to grow. We're one of the fastest-growing companies in the world. But we want to invent the next thing. It's not just about, you know, continuing selling the current thing, which we're doing, but what comes next. And to invent the next thing, you need volume, you need velocity, you need many shots on goal to be able to do the next big thing. So I want to see how concrete we can make this for our listeners. If you're an engineer, a builder at Replit, you're shipping three times as much as you used to. Like, what does that mean? Does that mean that the velocity of features you guys are shipping has increased? Like, you know, I imagine that factored into the release of this Replit design product that you just released. Talk to me, like, really concretely about, like, what it means that people ship three times faster. Yeah. So feature completeness is definitely important. We at Replit don't want to create a suite of products and you have to switch between them. Like, I'm already confused by ChatGPT. Like, when should I use Codex versus ChatGPT? When should I, you know? And so same thing with Claude. It's sort of they're reinventing the Microsoft Office. And so, yeah, Replit should be one product, and it should be, it takes you from an idea to design to code to deployment to growth, all in one place. And to do that, we have to build a lot of things. Now, you have to also just keep it simple. And so we spend a lot of time designing, refactoring, changing things around, doing A/B tests, making sure, doing UX research studies, and that generates a lot of different potential changes we can make to the product. And so the faster we can move, the more likely the product will hit its growth metrics and goals, and it'll feel better for people. We also want to be everywhere. We have a mobile app. We have one of the first mobile apps that allows you to code. And we have a desktop app now. And so it's the same app. It just has a lot of different sort of interfaces, and you can use it from anywhere. And so it's very important because, like the other day I was at the beach and I was, like, shipping software from my mobile phone, and that's great. You know, I like this idea of, like, work-life harmony. Like, I'm obviously enjoying the beach and everything, but I get a notification from Replit. It's like, Hey, I need an answer from you to do this thing, the Replit agent. I was like, Okay, do X, Y, and Z. It's almost like getting a Slack message from someone. You can just respond really quickly. And then the actual experience of being an engineer at Replit, you are trying a lot of things. You are ideating very quickly, whether it's you and the designers and the PMs. People spend a lot of time on whiteboards. So if you walk around Replit, You know, two years ago, everyone is hands on keyboard. You hear clicking all the time. And now actually it's a lot more vibrant environment. People debate a lot, they discuss a lot, they spend a lot of time on the whiteboard. And then that's interrupted by agents asking questions or them sending the next prompt. In Slack, when someone reports a bug, someone tags our AI autonomous engineer that we built, and it says, Hey, you know, fix that bug. You know, it's in this X, Y, and Z place, whatever, and it will go all the way to a pull request. So it's a very dynamic environment where you can go from a user feature request, user insight, kind of idea that we just brainstorm to a working product to a shipped thing that we can try with people very, very quickly. That makes sense. I want to ask about how you guys apply your own tools internally. In July, you said that Replit did not renew a seven-figure SaaS contract because an internal app that your team vibe coded was better. So you know I have to ask, what is the contract? I actually don't know that exact one because it's happening all the time. It's literally, you know, every... You know, I also just don't want to talk about companies. I mean, I feel bad for a lot of SaaS companies. But, like, for example, a recent thing we changed entirely, we used to use like three or four different analytics products. And Replit is really good at analytics. Like, we connected, we created a semantic layer that's self-correcting and that creates, allows you to query anything across. Like, I can join our data warehouse with our production Postgres database and create a report from that. Whereas the analytics products either inject themselves into your traffic and store data on their end, which is a privacy and security concern, or they only connect to one data source. Or if you want to connect to multiple data sources, they become really slow and they have to create their own views on top of these data sources. But Replit Agent can write just a Python program and connect all these different sources and present the interface for you. And so we basically canceled or downsized most of our analytics products, and now we kind of use Replit for all of that. So I'm curious how this comes about. Like, do you sort of set a goal of like, hey, like, let's take a look at through some of these contracts and figure out what we think we could build ourselves? Or is it somebody just built something and it worked better than the thing that you were paying for? Like, how does that happen? It's more of the latter. It's like a thousand flowers bloom. Like, I actually just got a Slack message from an engineer at Replit, and it's like, hey, I built a new Git hosting service, and he showed me a demo. And it's like, because GitHub is down a lot these days, and I feel bad for them. It's not their fault. It's like the amount of agents that are committing to GitHub is kind of insane. And so, you know, that might end up being a product we launch, but I don't know. It's probably not. But, you know, every week I see a new invention at Replit. Some of it could be productized. Others could be used as an internal tool. We replace a contract that we're using. And you just have, you just give people enough freedom. You create a culture where it's fun to build things, where that's rewarded. You create a culture where it's not intense all the time. Like we, I'm a big fan of seasons. I think humans work in seasons. Like, you know, our ancestors used to kind of farm in a season and harvest in a season and then, like, eat and sleep and do nothing in another season. And so, or like before that, they would, like, go hunt and hunt the megafauna and, like, store the meat and eat, you know, from that through the winter. And it's the same thing. We have these two or three releases a year where there's, like, three or four weeks of intensity, but there's a lot of time where there isn't a strict roadmap or deadline that everyone is kind of driving towards. And that period of time is where a lot of the invention happens. Well, I'm personally really looking forward to sleep and do nothing season. I may arrive briskly. These questions seem really relevant to this conversation that we were having earlier this year about SaaSpocalypse, right? The idea that these software as a service businesses might be in a lot of trouble as more and more people started to build their own internal tools. So I wonder if you could just tell us, like, how you think about what it makes sense to build internally versus continue to rely on a vendor for. Like, are there areas where you're like, it's never going to make sense for us to, like, build our own HR software? Or are you more in a mindset of, ah, build it, let's see if it's good. More of the latter. Like, I think if you asked me a year ago, I would probably say it's mixed. But now you can really unleash an agent and have it work continuously, and not only create the first version of it, but actually maintain it and take user feedback and iterate on it. You can create an entire autonomous loop for building and maintaining software. And I think this idea of autonomous loop, because a lot of our customers ask us, Hey, like I created this software and people are using it, but like I have a job. I'm now like a full-time software developer and like they're asking for features and things like that. You should be able to create a loop, and that loop kind of collects feedback and you kind of swipe left and right whether you want that feature or not. And you have like a very light-handed management on top of that agent, but that agent is really maintaining that software. So maintenance is no longer an issue. Security is no longer a big issue. I think we're close to a point where agents are better at writing secure software and better at scanning software for security issues. I mean, we're seeing that with Mythos and all of that. So all the excuses, all the things for why to buy Verspell is going away, to be honest. Interesting. All right. So it sounds like maybe if I check back with you in a year, you would have built a lot more things internally, and you'd be buying less. Yes. Yeah. Let me sort of ask a related question, which is about jobs. Like over the course of this mini-series, I've had a chance to ask lots of different leaders about it and get a really wide range of answers. Like sort of on one end, I talked to Eugenia Kuyda at Wabi, who basically said, I'm only hiring star athletes. Like I only have room for like the best, and everything else is going to be something that I outsource maybe to a contractor. And then on the other hand, I talked to, you know, like James Manyika at Google, who says that basically he sees AI as this force that is just going to open up more and more jobs that, you know, we've sort of removed all the bottlenecks on building, and we're just going to need more and more people to kind of manage that. So where do you fall on this question of— On this question of what AI and tools like the ones that you're building are going to be doing to jobs in the short and the medium term. So with regards to Replit hiring, I see that. I see Anthropic, for example, like I think they said they don't hire as many junior people; they just hire senior people. You know, there's this whole trend in Silicon Valley where you hire, like, IOI medalists or, you know, all that hype around that. I'm a normal person. Like, I'm not a, you know, particularly, like, genius at anything. And I, you know, we hire a lot of normal people, and they do extraordinary things. I think it's actually a better story when normal people, ordinary people, do extraordinary things. And so, you know, we don't really look at You know, pedigree and especially fake achievements. Like, I feel like most school achievements are fake. I think I'd rather have someone with a track record of building a startup, even if it failed. That's a lot more interesting to me, or working at a company, even if it failed, but, like, building something extraordinary is very important. And so looking at their track record is important. But young people who don't have a track record but have energy, excitement, you could tell it's infectious when you talk to them about the things they're excited about. They're kind of learning AI inside out, and so on and so forth. I think that that's something we care a lot about. We have a bunch of 16-year-old interns now. You know, one is doing security, others doing other things. We have people who are 18, 19 taking time off of college. Some of them are thinking of dropping out as well. And so we hire across the board, and I think AI does allow for really creative people to do extraordinary things, even though they don't have the skill. Like, I think AI research is getting to a point where it could do vibe research. Like, I'm training models in addition to my chess engine. I'm training models for Replit. Like, I've recently trained a cost estimator. Like, one of the things that I don't like about vibe coding products is you put in a prompt, you never know. You're rolling a dice. It might be $20. It might be $10. It might kind of take your entire credits for the month. And so it'd be nice to kind of know, and so that it's working surprisingly well. As for the, like, rest of the economy, I think that there is a policy question. Like, no doubt, I think there is something that the government needs to pay attention to and worry about. I think it's— Like if I'm giving advice to people, I would give them advice to be adaptable, to be lifelong learners and all of that. But I also think it's kind of unreasonable to tell someone in their fifties that the skill they've done for 20, 30 years is no longer relevant and you need to go back to school or you need to go learn something on the weekends. It's cruel and sort of, you know, remember when people used to tell minors, like, go learn to code or something like that? Or I think there was reporters also, they told them, go learn to code when there was layoffs. They did tell us that, yeah. So I think there's something that the government—I'm not an expert on that—but I think they're not thinking about it. And even if it is training and upskilling, there needs to be a program around that. Right. But the reason that you're saying that is because you do buy the argument that, like, yes, many companies are going to hire fewer people because they're going to be able to do that in an automated way. I think net-net there's going to be more jobs, but also there's going to be more companies, but companies will get smaller and they will do layoffs, 100% sure. Yeah. Right. And if you had to guess about Replit, like three years from now, do you have fewer employees, more employees, about the same? I think more employees. And I think the nature of our business is such that we can go into a lot of different domains. Like, you know, it's part of the thing that makes working at Replit fun, and I've been doing it for 10 years, I'm not bored from it, is that it is very expansive. And our mission is to kind of make software, to democratize software. And there's so many aspects of software we can democratize. You know, you go into programming, you go into design, you go into entrepreneurship, but there's a lot of different areas we can go into and just, like, make computers easier, more fun, more productive to use. Yeah. Sort of related question. A couple weeks ago, I was talking to Thomas Paul Mann, who's the CEO of Raycast. He told me that within three years, 30 to 50 percent of the software on your computer will be self-made, like just tools that you made for yourself. You're sort of in this business. Does that seem like a good prediction to you, or where would you predict? You know, I'll make a prediction that maybe will sound maybe even detrimental to Replit's business, but I think that a lot of software will not be used by humans; it will be used by agents. So right now, you have a problem. You make an app. You use that app to solve that problem. So there's intermediate steps. So why don't we have a problem? You articulate the problem. Problem gets solved. And that's fundamentally what agents do. And the way I use Replit increasingly, I make apps, but oftentimes I just tell it to do the thing. And I'm sure, like a lot of Claude users and ChatGPT users and Codex users are doing that, where instead of creating the intermediate app, it just, the agent does it, and maybe it writes software, writes libraries that it could use to do that. And like the way we're building the Replit agent is that it has this concept of learning. So it will learn, it will, like, store memories, it will create skills for itself. Those skills has software, has API calls, has scripts it could use. And so in three years, I think we'll be using less apps. And yes, the apps that you do use, a lot of it will be self-made, perhaps 50 percent. But I think we're going through this explosion of apps, and actually that will decline. And then agents, computer use is going to get really good over the next two to three months. Agents will be better at using your browser than you are, and so you'll be able to text them on iMessage or ChatGPT or whatever, and it will be using software on your desktop on your behalf. You'll be able to create loops, and that's something that Replit is interested in kind of building as well. You'll be able to create loops that can use and write software to solve problems. So we go from prompting to expressing intent: I want to solve this larger problem, and the agent will figure out what it needs to do in order to go do that. So concrete prediction: in three years, we'll be using less apps, we'll be using more agents, and those agents will be using the apps on our behalf. That seems plausible to me. You know, like when I hear Apple talk about what it wants Siri to eventually be, I think it would look a lot like the thing that you described, right? Where, like, instead of having to go to an app store and download an app, you just are sort of talking to your computer, say, Hey, I want this, or, you know, you ask it some sort of question, and then it just kind of happens for you. My assumption has been that the sort of system that you're describing does want to live at the level of the operating system, or at least that the makers of the operating systems are going to do everything they can to make sure they have a really, really good version of that, because that's going to help them, you know, sell laptops. Do you think that that is where this is going, and if so, like, where does Replit fit into that world? Right. I think so. I think the operating system makers... have not showed enough creativity or invention, especially Apple. I think the technology we have today is already at a place where you should be able to talk to your iPhone, and iPhone is doing a lot of things on your behalf. Like, you should be able to tell Siri, go book something on, you know, whatever, OpenTable. Yeah, it should be able to open that and do it. So I think they'll eventually be successful. Apple is always late to things, but eventually they figure it out. And where does Replit fit into that? We actually have a great partnership with Microsoft, and so I think we're able to exist in the Microsoft ecosystem and add a lot of value there. Apple, we've had some trouble kind of working with them. And we have a great relationship with Google and the Google Play Store, and so there's something to be done there. But also our big opportunity is in the enterprise, because the enterprise, there isn't a single operating system. You need to create that sort of operating system. So the self-driving company thesis that we have is where the future lies for Replit. We can create the operating system for the enterprise, for companies to work as the intelligent operator. Yeah. I also noticed in May that Visa invested in Replit and announced a partnership to let apps built on your platform, like let those apps move real money and eventually let agents transact with other agents. I think you even said that someday your customers might make more revenue than you do. So this is a big platform play. I would be interested to hear a little bit more about how you think that's going to work. Yeah, look, this is the intersection of a few things I'm really passionate about. One is entrepreneurship. Like, I really think that we're at a time and place that's unprecedented, where you need very little capital, very little resources to be able to change your life fundamentally and become an entrepreneur. And then the other thing is the movement of money. You know, I've been in Bitcoin since 2015 and been really interested in how do you actually bring money as native primitive on the internet. It's kind of crazy that, you know, we have communication as a primitive. We have email, HTTP, SMTP, all these things. We have, you know, the programming primitives. We have storage primitives. Why don't we have a native internet money primitive? I thought it was going to be Bitcoin. There's still a chance it could be some kind of crypto thing. But it looks like the existing money rails, the fiat money rails, are interested in solving that problem. And I think they already have the network and infrastructure for it. People trust these services. So we partnered with Visa to figure out what does it mean to create The new, the, like, new money primitive for computers and the internet. And so what that looks like is, A, make it super easy for entrepreneurs to collect payments. That's an obvious one. B, as an entrepreneur, as a programmer, you're asking agents to do certain things, and it should be able to go transact on the web. So, for example, if you need to do some kind of specialized legal thing or security thing, the agent should go talk to, say, Expo as a security agent, pay Expo, and have Expo do a scan on your app. And so outsource work. So your agent is outsourcing work, so it has some kind of wallet and is transacting with other agents to be able to do things on its behalf. It should be able to provision infrastructure as well. Like, you know, let's add email marketing to my app. I don't really know what email I should use or email system. It should go do deep research on the internet. It might pick something like Resend, and it should go to Resend, open an account, pay for a subscription, and start using that. So it just provisioned a piece of infrastructure. And finally, I think people should be able to pay agents as well. So I think ChatGPT and Claude and all these general chatbots will have a lot of information, and, you know, they're kind of crawling and I heard they're eating books now. But I think there's specialized knowledge that certain people have that you should be able to go pay for. So pay for an agent to do specific legal work for you. So there would be an autonomous law firm you should be able to kind of transact with it as well. Yeah. Interesting. Well, I want to close by asking one more question about productivity, because as somebody who loves using tools like the ones that you make to build things, and as somebody who is also trying to make myself more productive, I also know that there is a version of this future where I don't actually get any time back from all of these little productivity tools that I'm building, right? Like agents write more code, but then the code needs more review, which is done by agents, which produces more code that need to be reviewed, and we all just sort of end up exactly as productive as we have always been. How do we avoid a future like that so we can get to the season where we just sleep? You're really looking forward to that. Yeah. I think if you're optimizing for productivity, it's a never-ending optimization function, right? Like, you can always squeeze more productivity from yourself. Why don't we optimize for meaningful work? Why don't we optimize for fun? And I think by doing that, we will get more done for sure, or optimize for creativity. You know, meaningful work, I think will include creativity. And so I think changing the question, changing the thing that we want to do and get out of life, and I think it's the time to do that. You know, the history of capitalism has been like, Oh, let's make this GDP number go up. But what is GDP? I mean, a lot of it is nonsense. And so it's not making our lives better. And so instead, I think the conversation should be about how do we have a meaningful life? And no one is really... Talking about that conversation. A meaningful life will include being productive, will include, you know, having income and financial freedom, but also include being creative. It also include enjoying the people you work with and, you know, being in settings you enjoy, aesthetics, all of the above. And so I would encourage all of us to maybe try to change the optimization target. I like that. I will leave it there. Amjad, thanks so much for joining us. Thank you. Thank you, Casey. Platformer is produced by Lindsay Chu and edited by Fitz Harris at Story and Sound. You can watch this whole episode on YouTube at youtube.com slash Casey Newton. My email is casey@platformer.news, and we'll see you next week. Jira by Atlassian is where your team and your agents work from the same context. Try it free at jira.dev. That's J-I-R-A dot D-E-V.