An Interview With Replit Founder Amjad Masad

Ben Thompson’s Stratechery interview with Amjad Masad, read the same day Kyle hosted his own Replit event.

publish: false — the type default for longform. Nineteen highlights from a much longer interview; these are reading notes rather than a reproduction.

Notes

• One of the challenges in covering startups is the lack of available data. My solution is to go in the opposite direction and interview founders directly, letting them give their subjective overview of their companies, while pressing them on their business model, background, and long-term potential. Contrary Research
• IDE’s are where you write code, but then you have to store the code, deploy it, make it available to other people — Replit wants to do it all, and more.
• There had already been online IDEs. So you had Cloud9 that was acquired by AWS, you had Nitrous.io that was shut down, a few other ones that failed and so it was like a dead-on-arrival space and so immediately that’s the immediate comp. You also have to remember this is pre-GitHub acquisition so dev tools as a space was really nonexistent and it wasn’t getting a lot of funding at the time.
• I should be able to open a browser tab and start writing code, in the same way that I open a browser tab and start writing prose in Google Docs. Everything was moving to the web, and I felt like programming had to also move to the web.
• Across my career you see this thread, and obviously I wasn’t planning this out, which is why I love Steve Jobs’ connecting things backwards speech.
• We’re running a run time, running a container so we’re good at that, so why not also run your production code? You’re in the same place, you can also imagine CI/CD kind of being built on top of these primitives. So we build primitives, and turns out it’s actually easy to mix and match them to create new services and tools.
• So if you look at the PC revolution, the microcomputer, it started with kids and with hobbyists and computer clubs. What did the professionals do? They were using IBM mainframes at their companies and they looked down at everyone that used a PC.
• Apple was mostly a home computer and education computer, Microsoft was also an education company, Microsoft’s first product is BASIC, it’s a beginner programming environment. So a lot of computer revolutions, a lot of these big companies start kind of simple, start like a toy, and a lot of the initial beachhead market is hobbyists, teachers, schools and things like that, and that’s a great place to be in.
• I just tweeted about a kid who he said every line of code he’s ever written, CS student at Waterloo, has been on Replit. He just interned at Apple, and presumably he used Replit as well there. So you’re going to see a whole generation of people growing up on Replit and taking it to their jobs and I think at some point there’s going to be enough of them that also the companies are going to be paying attention
• Collaboration was an emergent property of the architecture of the system just by being in on the web and then collaborative coding — Google Docs sort of popularized this idea of being in real time together, and it was an obvious extension of being able to pass a URL and being able to have access to the same object and the same thing that you’re viewing the same document.
• When you have your entire history of coding on the platform, you have your entire project history, you have your databases on Replit, you have your applications on Replit, your friends on Replit, we have people that make friends on Replit. They don’t have friends in real life just because they’re nerds and are on the computer all the time — their entire social network, their entire friend group is on Replit and is Replit and so that’s the kind of passion that we have on the platform, so you have that going.
• Most of our focus is on the greenfield right now. We want people to build entirely new apps and products on Replit and that’s not just because we think that enterprise is hard to access, I’ll tell you a lot of ways we can access it, this is where we think we can add the most value. This is where we think we can change things, this is where our passion is at.
• I think one of my favorite Steve Jobs moments was at the All Thing Digital conference, I think it’s now called Code. They were pressing him on Flash, “Why haven’t you built Flash?” And what he said was amazing. He said, “We can only have a few big bets and we want to bet on the future of the web and HTML5, and we think Flash is the past.” That’s incredibly prescient.
• But I want to say this — we’re capitalists, we’re going to build a great business, it’s not a charity, I want to make sure people understand that. Actually we think that we’re more capitalist than a lot of other companies because what we want to do is we want to find the best talent in the world and give them opportunity and then share some of that profit and value that we create. So that’s our biggest bet.
• One way to think about Replit is, if Airbnb and Uber, the sharing economy, found idle assets in people’s homes and cars, Replit is finding idle assets in brains.
• The most obvious example of a competitor is obviously GitHub. GitHub has Codespaces, they have Copilot, and I do want to get to GhostWriter, your AI multiplayer companion, for sure. Microsoft obviously is very, very strong in the enterprise so that’s the obvious this is the Replit competitor, makes total sense. You said on Twitter that actually no, the real competitor for Replit is something like Roblox, it’s something where people get started programming. A third possible angle, what I’m hearing from you, is that actually no, our real competitor is AWS, in that people are going to run these massive applications on our infrastructure because they got started there, and they’d never want to go somewhere else.
• In 2013, I read this paper called On The Naturalness Of Software. I actually have it on my website because I love that paper so much. So this paper — I read it fairly early on and it basically says code is like natural language. They actually have this statistical reason for why they think code can be thought of as natural language. And then, they’re like, “Okay, if it is natural language, can you apply NLP (Natural Language Processing) on it?”
• We released explain code before Copilot and before anyone else. We released a bunch of experiments on OpenAI.
• It’s a great company, we love OpenAI, we’re partnered with them but at the end of the day, it just made sense for us to build our own.