Skip to content
Lenny's PodcastLenny's Podcast

Amjad Masad: Why Replit makes everyone a full-stack builder

Through Replit agents deploying full-stack apps from a prompt; ideas now bottleneck product, not engineering, hinting at billion-dollar zero-employee firms.

Amjad MasadguestLenny Rachitskyhost
Nov 21, 20241h 4mWatch on YouTube ↗

EVERY SPOKEN WORD

  1. 0:00 – 2:41

    Introduction to Amjad Masad and Replit

    1. AM

      (instrumental music plays) The idea behind Replit is that making software today is very difficult. We want to make it easier. People view this as a developer in their pocket, essentially. We have 34 million, uh, users globally. There's people everywhere learning to code on Replit, building startups, building personal software, personal tools.

    2. LR

      For people building products, say product managers, founders, like, what skills do you see will matter more, matter less?

    3. AM

      Typically, your bottlenecked where your ideas are not fitting in (laughs) because, like, they need to be made and they need to be made quickly, now you open up that bottleneck. So now, like, actually making things is, is a lot easier. Actually, you become limited by how fast you can generate ideas.

    4. LR

      I think people are unaware of just how far things have gone.

    5. AM

      I could imagine, whatever, five years from now, someone running a billion-dollar company with zero employees, where it's like the support is handled by AI, the development is handled by AI, and you're just building and creating this thing.

    6. LR

      Man, the future is wild. (instrumental music plays) Today, my guest is Amjad Masad. Amjad is the co-founder of Replit, an AI-powered software development and deployment platform for building and shipping software. It's one of the fastest-growing developer communities and AI products in the world. There's a lot of talk these days about how AI is changing how products will be built, how product teams are gonna operate, which functions will be more and less valuable over time. But I feel like very few people have actually seen what modern AI tools can do and have fully grasped how much you can get done with very little technical skill now and in the future. And so, I'm gonna do an experiment with this podcast where I'm gonna do a series of Behind the Product episodes, where we go deep on important products that product builders should be aware of and should probably start playing with. In our conversation, Amjad does a demo of what you can do with Replit today, which is gonna blow your mind. And then we spend most of the conversation talking about the implications of this on the future of product development, on the future of product management, and on the future of startups and founders. It's a very exciting time. It's also a very scary and destabilizing time for a lot of people. And my thinking is, the more you are aware of what's possible today and where things are going, the better position you'll be in to thrive in this very wild and crazy future that is coming very fast. If you enjoy this podcast, don't forget to subscribe and follow it in your favorite podcasting app or YouTube. It's the best way to avoid missing future episodes and helps the podcast tremendously. With that, I bring you Amjad Masad. (instrumental music plays)

  2. 2:41 – 6:50

    The vision and challenges of Replit

    1. LR

      Amjad, thank you so much for being here. Welcome to the podcast.

    2. AM

      Oh, it's my pleasure.

    3. LR

      I thought it'd be great to start with just having you explain what is Replit, what's the vision, where is this going, what job does it do for people?

    4. AM

      The idea behind Replit is that, uh, making software today is, is very difficult and, uh, we want to make it easier. Uh, one of the reasons for the difficulty is that, um, it is very fragmented. So, you would need to download, uh, what's called an IDE. That's basically a code editor. You need to download the runtime, basically Python or JavaScript. You need to figure out a package manager to configure your kind of open-source packages. And once you've done all of that, you need to figure out how to deploy it, how to share it, how to... And so, it's, it's a very hard process and that's one of the ways, uh, where people get stuck and never learn how to code because it just feels like this cumbersome IT process. And so, the vision, uh, for Replit has always been is like, okay, making software is fun, is great, more people should do it, but, uh, so, so for more people to do it, it needs to be, um, easier to do, it needs to be in one place, and it needs to be learnable, it's easy to learn. And so, so, so that's the product today. It is, I think, one of the more easier, um, IDEs/environment/deployment environment, uh, on, on the internet. And, uh, and I think we make it really easy for people to just jump in, even without prior experience of coding, especially now with the new AI products that we've built.

    5. LR

      (instrumental music plays) This episode is brought to you by WorkOS. If you're building a SaaS app, at some point, your customers will start asking for enterprise features like SAML authentication and SCIM provisioning. That's where WorkOS comes in, making it fast and painless to add enterprise features to your app. Their APIs are easy to understand so that you can ship quickly and get back to building other features. Today, hundreds of companies are already powered by WorkOS, including ones you probably know, like Vercel, Webflow, and Loom. WorkOS also recently acquired Warrant, the fine-grained authorization service. Warrant's product is based on a groundbreaking authorization system called Zanzibar, which was originally designed for Google to power Google Docs and YouTube. This enables fast authorization checks at enormous scale while maintaining a flexible model that can be adapted to even the most complex use cases. If you're currently looking to build role-based access control or other enterprise features like single sign-on, SCIM, or user management, you should consider WorkOS. It's a drop-in replacement for Auth0 and supports up to one million monthly active users for free. Check it out at workos.com to learn more. That's workos.com. This episode is brought to you by Persona, the adaptable identity platform that helps businesses fight fraud, meet compliance requirements, and build trust. While you're listening to this right now, how do you know that you're really listening to me, Lenny? These days, it's easier than ever for fraudsters to steal PII, faces, and identities. That's where Persona comes in. Persona helps leading companies like LinkedIn, Etsy, and Twilio securely verify individuals and businesses across the world. What sets Persona apart is its configurability. Every company has different needs depending on its industry, use cases, risk tolerance, and user demographics. That's why Persona offers flexible building blocks that allow you to build tailored collection and verification flows that maximize conversion while minimizing risk. Plus, Persona's orchestration tools automate your identity process so that you can fight rapidly shifting fraud and meet new waves of regulation. Whether you're a startup or an enterprise business, Persona has a plan for you. Learn more at withpersona.com/lenny. Again, that's withpersona.com/lenny.

  3. 6:50 – 10:49

    Replit’s growth and user stories

    1. LR

      What's the scale of Replit at this point? How large has this gotten? How many people are using it?

    2. AM

      We have 34 million, uh, users globally. We have a large, uh, global presence. There's, um, you know, people everywhere learning to code on Replit, building startups, um, you know, building personal software, personal tools, or internal tools at companies. More recently, we've been, uh, expanding to companies. We released our kind of B2B package in, in July and that's been growing really fast. It's been really fun to see people bring Replit to work as well.

    3. LR

      Damn. I knew it was popular; I didn't realize it was that large actually. Um, as I was preparing for this podcast episode, there's this tweet that kinda, that went viral where this guy Jevin, who I actually know, I know this guy from Canada, he's awesome, tweeted about how his 11-year-old girl built an app, uh, in Replit. She just, like, had an idea and she built it. And the best part of it is someone in the... like, replied to him and they're like, like, "You have to... To, like, launch an app, you have to, like, host it somewhere. You have to build a database. You have to, like, deploy it. There's no way to do that." And he's like, "No, that's exactly what Replit did." (laughs)

    4. AM

      Yeah, that's what we do. That, th- the, everything that, that, that, uh, commenter was talking about, it... And he's right, right? Like, it's, it's... The, the surprising thing about an 11-year-old building, uh, an app, it's not so much even the coding. It is, like, all the nonsense around it. And so we just, we just abstract a- all that away.

    5. LR

      I love that. And that's, uh, I struggle with that myself when I was an engineer way back in the day.

    6. AM

      Oh, you were an engineer? I didn't know that.

    7. LR

      I was. I was an engineer for 10 years. I was an engineering manager, and then I jumped ship into product, and, um-

    8. AM

      Wow.

    9. LR

      ... I'm happy I did, but I do miss it.

    10. AM

      Yeah.

    11. LR

      I was. I was not an amazing engineer. I was, like, a good enough startup engineer, so this is the kind of stuff that I would've loved to use. So, uh, we're gonna jump into a demo of what this actually looks like. I thought maybe actually before we get into it, there's other tools that people are aware of that help you build stuff, and so to kind of put a finer point on, like, what this does and how it's different from other things you may have heard of, say Cursor comes up a lot these days, just talk about, like, a little bit about the competitive landscape of who else is out there that helps you build product.

    12. AM

      Uh, again, we go, go back to this idea of, like, end-to-end platform for making software, so that's, like, from w- writing code all the way to deploying it and, um, and monetizing it and all of that. Uh, now every step in the process of the software's development life cycle, there are a lot of different tools. So Cursor, uh, is a fork of VS Code that's made, um, that has really awesome, uh, AI tools, but that's an editor. You still need, uh, runtime. You still need a deployment environment. Actually, quite a few users use, use Cursor in tandem with Replit because Replit just simplifies the runtime and, and deployment environment. And so you have products, uh, you know, AI products, uh, in different, uh, places in the software development life cycle, uh, but really what differentiates Replit is that we, we do everything. But, but also, uh, you know, that makes it harder to adopt for, for certain people. Like if you're at a big company, it's very easy to bring a new editor, uh, and start coding with it. It's quite hard to buil- uh, to, to bring, you know, something that's quite opinionated about, about everything, uh, f- from the f- from how the code runs to how the code deploys, but that's a trade-off we're willing to make is like, yeah, we're not gonna get into, uh, into the enterprise, you know, main software development pipeline, but we wanna empower everyone to, to be able to build software. And that means product managers, designers. We, we have, you know, operations people, sales ops, HR ops. We have lawyers, uh, uh, u- using, using Replit, and so it is, it is democratizing the act of software, uh, software engineering.

  4. 10:49 – 16:51

    Demo of Replit’s capabilities

    1. AM

    2. LR

      Amazing, and that's why you're here. Let's do a demo. While you're pulling it up, you're gonna share your screen and show us what this product can do. And the reason I, I'm excited about doing a demo and this is an experiment, kind of a new type of podcast episode I'm doing where we're diving into specific products and what y- they can do, I feel like there's so much talk about AI and what it's doing and people keep reading about, "Oh, AI can do this and AI can do that," and I feel like not many people actually, like, see this stuff in action, especially the most cutting edge stuff. Like, I think people are unaware of just how far things have gone and how much is actually possible, especially when someone that knows what they're doing is using the product. So, I'm excited to show people what is actually possible and especially because this is gonna impact the future of product management and product teams. So I'll turn it over to you. Give us a demo.

    3. AM

      Awesome. So this is, uh, Replit's, uh, homepage. You can create what's called a Repl, which is a project. Uh, we have all sorts of languages you can, you can pick from, uh, really, uh, in the hundreds, but most recently, and this is how Replit became, like, a thousand times easier, is you can just describe what you want to make. So you go onto this homepage. We have this text box and you can, you can write something like, uh, "Make me a cool app," uh, or what have you. But, uh, you know, a, a more descriptive prompt is, is better. And so, uh, I asked, uh, our PM at Replit, Aman Mathur, who's a, who's a fan of the show, to, to tell me what PMs like to build, and so he came up with a prompt. He kind of, uh, really crafted a great prompt, so I'm gonna just put it here. Uh, and basically...... what we're asking for is we want to build a web application. You can, you can say what stack you wanna use or you can leave it up to the AI to decide. Here, we're saying, you know, build it in NodeJS for product managers to track feature requests on a public dashboard. So say, you know, I have a product I'm growing. I have a community, I want that community to engage with building the product. I want them to submit feature requests and vote on them. I wanna be able to manage that. So we were talking here about the features, uh, upvoting system, feature requests.

    4. LR

      Read a few of them just for folks that are watching on YouTube, to give them a sense of some of this, some of the stuff in this prompt.

    5. AM

      Sure. So, so a feature request submission, so allowing the users to add features. Upvoting system, so allowing users to upvote these, uh, features, feature requests. And status tracking, being able to, uh, it's like a Kanban style board with columns, like planned and progress. So that way the admin can, can kind of share with the community what they're building. And we want it to be user-friendly design, so make it modern and all that nice kind of prompty things. And then, uh, admin controls for product managers. So, you know, as a product manager, I wanna be able to kind of really manage this, this community, and use our-

    6. LR

      I love that it builds internal tools too, not just-

    7. AM

      Yes.

    8. LR

      ... the front end.

    9. AM

      Exactly.

    10. LR

      Okay.

    11. AM

      Exactly. All right. So we're gonna start building. Uh, since this is, uh, like a pretty big, big prompt, uh, the co- the, the initial coding might, might take a while. There's different styles of using Replit agents. Uh, I often go with like min- uh, minimalist prompts. That's also how I code as well. Uh, like I have, like, a vague idea for what I, what I wanna build and enter it from there. Other people, you know, product managers like to, like, (laughs) write PRDs and, and, like, more descriptive things, and you can do either of those things. The AI now responded and then said, you know, uh, you know, "I'll build all of that for you. I'm gonna build up the initial prototype and you can tell me how it feels and, and then we can make it better from there." The AI is also suggesting adding comment threads, implementing email notifications. And so I can select those and, you know, it's being creative. It's telling me what else I could build. But for now, I'm just gonna go with a prototype and then we can assess from there. So as you see, as the prototype is starting, you can see this progress pane where we can watch the AI doing its thing. So here, it's created a Postgres database. Obviously, when we're building a full stack application, you need to be able to save things. Uh, so this is one of the cool things about Replit. We have all these services, storage, database. So now it's coding. It's building the database schema. Uh, and now it's building the, um, the home, home page. And it's, it's actually quite fun and edifying to watch, watch it build this because you can, uh, you can really start to learn how to structure web apps. And, you know, if, uh, it runs into a problem and, you know, as things get complicated, it might run into, into a problem, and you wanna be able to, like, help debug and things like that. It's good to be, to be able to have an idea of what's going on. But it's not necessary. I think a, a lot of people just don't, don't care about the code and are, are still able to, to build things. But we wanna make the process transparent. We wanna show people-

    12. LR

      Oh-

    13. AM

      ... exactly what the agent is doing.

    14. LR

      ... like you're basically sitting there behind an engineer on a computer and just watching them code is what the-

    15. AM

      Exactly right.

    16. LR

      ... experience feels like.

    17. AM

      Yeah. And, and actually the way we built it is, is like it's a multiplayer system. So Replit is, um, you know, has real time what we call multiplayer coding. And, uh, we reused this, the multiplayer system to build the agent. So the agent in the code is structured as a-a-another user of the platform. So basically, we're both coding together. So I can go into the files here and that, that's the thing that makes Replit really cool. I think people are familiar with some of the more, like, chat interfaces, uh, like v0 and others where it's purely chat. But this is like a full IDE where you can, like, go and look at the files and, and edit them yourself or ask, ask the AI for an explanation.

  5. 16:51 – 25:04

    Building and iterating with Replit

    1. AM

    2. LR

      What's kind of the limitation of what this can do today? Like, what can't you do? Say you're like, you have zero coding experience. What sorts of products can you not yet build with something like this that might be possible in the future? How far does this take you now?

    3. AM

      You know, you can build, uh, MVPs. I think you can also start to get some, some initial users. I think when it's... when you start iterating on the, the product, like, large iterations, you might run into problems. For example, you know, it's not very good at data, database migrations. And so we're trying to fix that. So, you know, a lot of... When, when you're iterating on the product, a lot of times you're actually, um, you know, changing the, um, the structure of the app and, um, that, that requires database migrations. And so now, like, it might change the database in a way that creates an error that's unrecoverable. Uh, and it... At that point, it... you might, you might get stuck, especially if you don't know how to code.

    4. LR

      Mm-hmm.

    5. AM

      Uh, some people will figure it out by going to ChatGPT and Claude and, like, asking questions. And, and, like, I, I actually re- am really inspired about how persistent some of our users are, which is really amazing. But I think, yeah, that's... Like, you'll get an MVP past the MVP where it's like a product that's working and you need to change it and iterate on it. It's, it's, uh, it's still a struggle now but I expect, uh, you know, over the next, uh, few months, we'll, we'll continue ex- It's like if you think about it, it's like sort of we're building, uh, you know, we're building as you're building. So we're building out the agent so that it can...... continue, uh, getting better, um, as our users are also, uh, building their applications.

    6. LR

      Got it. So what I'm hearing is it's really good at building, like, the first version and helping you get to something that you can even have people use. It's not amazing yet at evolving from there, like using AI to help you make the product better and better and better and iterate.

    7. AM

      Yes.

    8. LR

      But you can get in there if you have- if you know how to code and take it from there, right?

    9. AM

      Yes. Or you can hire someone.

    10. LR

      Mm-hmm.

    11. AM

      Um, we have a feature on the site called Bounties where you can hire human coders to kind of-

    12. LR

      Oh, I love that. (laughs)

    13. AM

      ... help you finish, finish your, finish your part.

    14. LR

      (laughs) That's gonna be the- or job for humans for, uh, like, that'll remain for a while. (laughs)

    15. AM

      You know what we wanna do? We, we wanna get to a point where, uh, the agent can go grab a, uh-

    16. LR

      A human?

    17. AM

      ... a human when, when it (laughs) runs into problem. I think, I think that would be- that would be sick.

    18. LR

      Oh my God. It's like everything's reversed. I love it.

    19. AM

      Yeah.

    20. LR

      Oh, look. I think it might be done. Check that out.

    21. AM

      Yeah. So, so now, um, the agent is asking us, "Is the application running and showing the homepage?"

    22. LR

      Hmm.

    23. AM

      Uh.

    24. LR

      I like that. I like that it's confirming, like, "Let me..."

    25. AM

      So it's, uh, yeah, almost asking us to do QA. I'll just say yes. Uh, so, so it- it found an error. So there's an error here, and it's like, uh, "If there's a DOM warning, I'm gonna fix it." So in the meantime, as it's fixing it... So it's, uh, you know, it- it can be proactive, right? It... because it, you know, looks at all the errors and- and things like that. But in the meantime, we can use it as just, um... create an account. Uh, it's- it's coding. It's... (laughs) Okay.

    26. LR

      It's fixing the bug.

    27. AM

      It re-

    28. LR

      That's cool. We can't wait.

    29. AM

      Yeah. Restart. Okay. Well, we'll wait for it.

    30. LR

      How long would you say it would take an engineer to build this? Like a- you know, like a typical engineer?

  6. 25:04 – 30:13

    Real-world applications and use cases

    1. AM

      all of that.

    2. LR

      So to follow that thread, what are you seeing inside of startups or even big companies in terms of how folks are already using this, knowing this is, this is like the worst it will be and it will only become smarter and better?

    3. AM

      Right.

    4. LR

      These days what have you seen, what, how are people actually using this, say that are, say product managers or just like non-technical people within startups or bigger companies?

    5. AM

      On the SMB side of things, uh, a lot of people are building kind of back-office, uh, tools, right? So we have real estate agents that, you know, uh, have a lot of data, have a lot of, um, uh, things they, they want to manage in their business so that, they are building a lot of these, these tools that they otherwise would have to buy, but typically when you buy it's, it's actually not exactly what you need and that's the, kind of the problem with SaaS is like it's, it's like one size fits all. And so a lot of people are seeing it as sort of a SaaS replacement for in-house tools and, and, and things like that. And, and then wh- when you go to the like bigger companies, um, it's anywhere from, from prototyping to, to, to, to actually production apps, uh, to, to tools as well. So, um, w- we've seen, uh, product managers build, like I said, like a V1 of an app and actually go out and test it with users. Uh, a- and I can't name the company but, uh, you know, I, it's a, you know, there's a public company that, that have used Replit to test a V1 of a, of, of an app, um, a- and obviously a- after, after that sort of works, they, they take it to the engineers and they're like, "Okay, we, we built this thing, we think it's, uh, we think it's a great thing. We tested with some users, uh, l- let's go actually put it on the roadmap." And, and built it, built it into the actual product. So you are sort of unblocking product managers from having to need engineers for everything that they want to build so they can really build the V0 or V1 of the product, uh, and, and, and that's super empowering, uh, for them. We s- we s- saw it also with like marketing departments like, uh, SpotHero has a, uh, marketing, uh, head of marketing that actually can code decently well and but uses Replit to, to buil- build his apps. And they built like a competitive analysis application that looks at, um, competitor's pricing and makes sure that they are being benchmarked, uh, correctly. Uh, and so it's, it's a full-stack app, uses a database and everything and it runs on a continuous fashion. And we see, um, sales engineers, um, uh, u- use Replit to spin up prototypes really quickly so, uh, actually someone at, at X, formerly, uh, Twitter, uh, uh, is on the sort of, uh, partner engineering, uh, side of things and he uses Replit Agent to spin up, uh, applications, uh, and prototypes for, for customers to see how they can use the XAPI.

    6. LR

      I love this. I love these examples. Uh, by the way, the demo, is there anything else you want to share about the demo before we close that out?

    7. AM

      So, so it, it created an admin account.

    8. LR

      Yeah.

    9. AM

      Um, we can ask it what the username password and, and, and kind of go into it and, and, and manage it. But basically that's it, you know, it's, um, the, the app's complete in terms of what we asked for, uh, we can like send it out, I can give you a URL. Uh, we can... Let- let's actually just deploy it really quickly.

    10. LR

      Oh, yeah.

    11. AM

      Just to show people how you deploy.

    12. LR

      And we could maybe, maybe in the show notes we'll link to the app, you could check it out.

    13. AM

      Sounds good.

    14. LR

      Okay, cool. That's amazing. So this is deploying it onto some like cloud provider, I don't know what you use but...

    15. AM

      We use Google Cloud. Yeah.

    16. LR

      Okay. So it's-

    17. AM

      But we a- we, we abstract, we abstract all of that away from you but, but, but we use Google Cloud behind the scenes.

    18. LR

      Imagine a place where you can find all your potential customers and get your message in front of them in a cost-efficient way. If you're a B2B business, that place exists, and it's called LinkedIn. LinkedIn Ads allows you to build the right relationships, drive results, and reach your customers in a respectful environment. Two of my portfolio companies, Webflow and Census, are LinkedIn success stories. Census had a 10X increase in pipeline with a LinkedIn startup team. For Webflow, after ramping up on LinkedIn in Q4, they had the highest marketing source revenue quarter to date. With LinkedIn Ads, you'll have direct access to and can build relationships with decision-makers, including 950 million members, 180 million senior execs, and over 10 million C-level executives. You'll be able to drive results with targeting and measurement tools built specifically for B2B. In tech, LinkedIn generated 2 to 5X higher return on ad spend than any other social media platforms. Audiences on LinkedIn have two times the buying power of the average web audience and you'll work with a partner who respects the B2B world you operate in. Make B2B marketing everything it can be and get $100 credit on your next campaign. Just go to linkedin.com/podlenny to claim your credit. That's linkedin.com/podlenny. Terms and conditions apply.

  7. 30:13 – 33:48

    The technology stack

    1. LR

      Let's go down this thread actually while this is happening.

    2. AM

      Yeah.

    3. LR

      Just like what, what allows for this to be possible technology-wise, like what is the, kind of the stack, whatever you can share that enables this to exist? That-

    4. AM

      Yeah, for sure. First of all is, uh, it's the, uh, u- uh, all the abstractions that we built. So, uh, the way Replit works is, um, you know, at the, you know, very bottom layer it's, um, our runtime. So this is the operating system, this is the package manager, this is the language runtimes.... uh, we built a system that, uh, is able to install packages in any language, including native packages. So the AI anytime, um, it needs a, it needs a package, I can, I can go here and show one of those. By the way, the AI can take screenshots as well so that it checks if it works. So here, uh, you can see, it's- it's- it's taking screenshots to make sure the homepage is rendering. Here you can see, uh, you know, uh... Oh, it wanted to drag and drop library and so it installed, uh, installed that. And so it has access to all- all the packages across all languages including Linux and- and all of that. And then, uh, the layer on top of that is the editor and the infrastructure that runs the editor including what I described as the multiplayer editor. And then we expose all of that infrastructure to the AI. And, um, there's, like, almost like a new discipline, um, called, um, AI computer interfaces, so sort of like HCI is now A- ACI. A- and turns out, like LLMs need interfaces that are- that are actually quite different than- than humans. Uh, they're trying to make them use human interfaces like anthropics computer use, but those are really expensive and you need to kind of process all those images and video. So i- instead we, you know, f- for the shell for example, we give it a, you know, uh, a sort of a text representation of what the shell is doing at a certain increments. Um, for- for package installation, we give it- we give it a certain tool. For editing, we give it a, like a- like an editor tool that, like, when it's writing the code it's getting feedback on whether there are errors or not, similar to what a human sees but it's actually like all text just to make it easier. So that's-

    5. LR

      Mm-hmm.

    6. AM

      ... uh, AI computer interface. And obviously all that's sitting on- on foundation- foundation models so- so the- the improvement in- in foundation models, uh, has- has allowed it- allowed us to- to- to build this. The- the, you know, uh, the most important model that we use is the Sonnet, uh, model fro- from, uh, Claude, uh, from Anthropic. Uh, and, um, it is the best model at- at coding so that's w- the model we use for coding. But we use k- uh, models from, uh, OpenAI as well 'cause it's a multi-agent system and so we have models that are, uh, you know, critiquing. We have manager, editor model, and we have, like, critique model. And different models will have, uh, different, uh, powers. We also train, uh, s- some of our models like the embedding model for- for search, uh, is something we- we trained internally. So, you know, I actually wrote about it back in like '22, uh, I said, uh, "It's gonna be a society of models," like products will be made o- of a lot of different models. And, you know, i- i- i- it's a quite a heavy engineering- engineering project.

    7. LR

      (laughs) To say the

  8. 33:48 – 39:36

    The evolution of Replit and its capabilities

    1. LR

      least.

    2. AM

      Yeah.

    3. LR

      Uh, we were talking offline and you said you've been working on this since 2009 when you first built the f- first idea of Replit, is that right?

    4. AM

      Yes. Yes.

    5. LR

      Oh my God.

    6. AM

      Here's the deployed app. (clears throat) Uh, I can-

    7. LR

      Beautiful.

    8. AM

      ... I can send it to you and you can use it and you can- you can see my- my- my request even on the logged out page so I can register, upload it, and log in as admin and move things around. We can see what's in progress, what's completed.

    9. LR

      Like this looks like a product I could see designers spending like days t- you know, designing.

    10. AM

      (laughs) Yeah.

    11. LR

      Passing it to engineering, PMs, you know, having feedback, engineers taking a few days to build it.

    12. AM

      Yes.

    13. LR

      And here is just a prompt. Here's what I want.

    14. AM

      That's right. That's right.

    15. LR

      Unbelievable.

    16. AM

      And we can iterate on it very easily. We can also iterate on the UI. We can- we can say, you know, "I don't like this or that," and it'll- it'll do- do a good job so we can go here, we can start a new session. Or like a new session to create an entirely new feature here and it'll- it'll just do the right thing.

    17. LR

      And it builds from that code base, it understands here's what you've built.

    18. AM

      Yeah.

    19. LR

      I wanna add this thing.

    20. AM

      Yes.

    21. LR

      Okay.

    22. AM

      And that becomes- that becomes your, sort of, your history, right? Like this was the V1 and now I'm working on this new feature, uh, and, you know, and- and- and, you know, it- it's almost like what engineers do in Git commit messages. By the way, it generates G- Git commit messages for every-

    23. LR

      Mm-hmm.

    24. AM

      ... for everything that it does so you can roll back, uh, as well. And- and so we're trying to, like, make it so that yes, it's- it's for everyone, but we're trying not to abstract too much away. W- we wanna build tools, right, for you to learn to use and so we want power users to be able to understand the full power of- of- of Replit and it- it's really deep product. I think you can spend- you can spend a couple of years to- to kind of master it.

    25. LR

      I wanna talk about implications but I wanna come back to something you mentioned that e- (laughs) uh, is incredible that people may have missed. You basically built a computer specifically designed for the AI agent to use that is a different version of a computer specifically optimized for how AI wants to use a computer.

    26. AM

      Yeah. Yeah, yeah, so, um, uh, you know, uh, there's an entire discipline, uh, called like HCI, right? Like it's like how do I-

    27. LR

      Human-computer interaction.

    28. AM

      Yeah. So- so now there are papers about AI computer, uh, interfaces and the interactions. Um, and- and- and so, you know, eh, large language models are trained on large text corpus from the internet but they're still like kind of alien creatures so they're not like humans, you know, so they- they have different behaviors and like it's unclear like what's the best way to give it an editor. Uh, so there's so many experimentation about like what's the best way to give it a view, on what's editing, how many files can you show it like, you know, uh, before it starts to hallucinate, and- and right now it's like more of an art than science but it's becoming more and more like a science.

    29. LR

      This is insane. So it's a simple way to think about it. There's this foundational model. Here's what I want you to build, and here's a computer to use to build it.

    30. AM

      Yes.

  9. 39:36 – 44:04

    The future of AI in software development

    1. AM

    2. LR

      Okay. This can go in so many directions. I'm gonna bring us back to the implications for people building products, say product managers, founders. How does this change that function, that skillset? Like what skills do you see will matter more, matter less? Which functions are maybe in some danger and they should start thinking about a different career path?

    3. AM

      One, one interesting persona that we're seeing is the CEO, the CEO of a startup, the CEO of, you know, uh, you know, A- Andrew, uh, Wilkinson from, from Tiny is, is a big user.

    4. LR

      Tiny, yeah.

    5. AM

      Um, and so, uh, these people are, are typically, you know, creatives, right? They, they built a company, uh, they hired people. A lot of them like can't code. A lot of them are, are designers or product managers or, or so- so- something else. And, and they... You can imagine a bottleneck, you can imagine a bunch of ideas in their head, uh, and the ideas have to translate through them talking and then someone else listening to them and like assuming that someone else actually understands what they say. And then that's someone else going and trying to build what they wanna buil- what they want built. And also assuming that person has, has time, right? Because a lot of times your engineers are kind of stuck building the current thing, they're not thinking about the future thing. And so, uh, what gets me excited is a lot of these CEOs are building the, the future concept, the next comp- the next product they're gonna build, the next, you know, say company they're gonna build. And so, uh, it unlocks, uh, the creativity and again, sort of unblocks 'em from that. And look, it's, you know, it's a V1 of the product, but it can push things forward, you can touch it, you can feel it, you can say, okay, this is, this really has lags and, and we should, we should work on it. You give it to your engineers and they can, they can improve on it, uh, from there. So that's, uh, that's one persona. Uh, but I'm really excited about it, the CEO/founder. Um, in, uh, in sort of, uh, uh, companies, um, one, one of the things that, that I think is sort of hard about tech, tech companies is, uh, sort, sort of these silos between, um, designers, uh, product managers and, and, and engineers and, you know, everyone feels that pain of kind of we have low bandwidth communication, which is, which is language, which, and text on, on Slack and, and Zoom calls. And it leads to a lot of frustration because it's really easy to misinterpret, uh, people. And again, leads to sort of siloing where, where people working on, on something and then you pass it on to the, to the next team and it's not really what they, what they expect. That happens a lot between designers and engineers. Um, but, but like the common language that, uh, that, um, that everyone shares is code, right? Like ultimately, uh, in software tech companies, uh, everything that we're talking about need to eventually flush out in terms of code. And so, so what if the language becomes actually working prototypes and working applications. Um, you know, for, for example, we have a Figma extension that translate, um, th- uh, you know,... Figma mocks into, into React that runs on, on Replit. So instead of, instead of, um, you know, g- giving, giving the engineers, you know, just, just mocks or screenshots, whatever, you just say, "Oh, here's, here's a bunch of React code. You know, just make sure it runs on our infrastructure, but, like, don't mess with it. Don't move the pixels around," right? (laughs) Um, and, and so I think it, it just, like, opens, uh, opens up, uh, you know, silos of the companies, ma- make communication around, around product a lot more concrete because I can, I can give you a working, uh, prototype. And that'll change how, how people work. Like, if you, if you can imagine that everyone can, can make software, it, it's really kind of a radical re-imagining of not just what tech companies are but really what, what most companies are 'cause, 'cause, you know, everyone can be more, more general.

  10. 44:04 – 47:26

    Skills for the future: generative thinking and coding

    1. LR

      So say you're a PM listening to this, an engineer, a designer. What skills do you think ... If you were one of these folks, if you were in building Replit right now, what kind of skills would you suggest folks focus on more and what you think are just like, "Okay, this can be less valuable in the future. Don't worry about these sorts of things."? And you can either pick one of those three functions or all three.

    2. AM

      I think, uh, uh, a very important skill that, that's, like, perhaps harder to develop but it's worth working on is being generative, being more generative, um, being able to generate, uh, new ideas, uh, quickly because, you know, you can think about it as, as like a factory line, right? Like, so, so you have ideas. Uh, you have the, the production of, of these, of these ideas or, like, the initial kind of production of, of, of these ideas. And then you have, you know, o- o- other people that wanna consume these ideas or work with you on these, on these ideas. And so, t- typically, you're bottlenecked by, by the middle kind of part where your ideas are kind of like ... there are a lot of them and they're not fitting in because, like, they need to be made and they need to be made quickly. And so now you open up that bottleneck. So now, like, actually making things, uh, is, is a lot easier. Actually, you become limited by how fast you can generate ideas. Um, and so, uh ... And, and I find that true of, of myself as well. Like, you know, I, I consider myself, uh, quite generative but, but now I have this tool and I can, like, build, build a lot more and explore a lot more. And I'm, I'm finding that, um ... Well, actually, I'm running out of ideas sometimes. (laughs) And so, and so, so, uh, you know, training that, uh, that muscle, I think, is a, is a good thing. Um, I, I think, like, learning a little bit of coding and, and like not the traditional way of learning coding. Like, when you go ... If, like, if you go to, like, a coding bootcamp, uh, they're gonna start with, like, what is Git? Actually, my co-founder, Haya, uh, who's a designer, uh, when we were first building Replit together, uh, she went to, to, uh, web- WebAssembly, uh, to do, like, a coding course. Um, and the first day, they were like, "Spent this whole time on Git." And she's like, you know, "What is that?" Like, "Well, what does it do?" (laughs) Like, I'm like, "I still don't know what Git exactly does." But, uh, it's, it's like, um, you're, you're inverting the process. Like, you're giving the, the tool before the actual problem. And so I think all of that stuff you don't have to worry about, so things that you don't have to worry about. I think a lot of the, you know, as a PM, as a designer, as someone who's not, like, in your code editor every day, don't worry about all the tooling. And if you learn, uh, a little bit of coding just by f- you know, talking to an AI, doing a little bit of debugging, building something with Replit, you know, running into a problem and trying to fix it, uh, just using AI, you'll learn a bit of coding. And, you know, I have this, um,

  11. 47:26 – 50:36

    Amjad’s law

    1. AM

      I have this. That's been called, uh, not by me, dubbed as Amjad's Law, which is, um, the return on investment for learning to code is doubling every six months. Um, and really just learning a, a little bit of that skill, learning a bit of skill about how to, you know, prompt, uh, AI, how to read code and be able to debug it. Ev- every, every six months, that's netting you, uh, more and more power, uh, because you're, you're gonna be able to create a lot more. You're gonna be able ... It's gonna be easier to create. You're gonna be able to create a lot, uh, you know, uh, a lot more complete, uh, things. Um, so, so that's, that's another, uh, skill that I think, uh, could be, could be necessary.

    2. LR

      This is super interesting. Okay, so this last point you made, Amjad's Law, uh, it's interesting 'cause when people ... Like, as someone's listening to this, I could see them being like, "Engineers are in trouble. Why do you need engineers at this point? They're, these agents are building the code." Your point is specific engineering skills are gonna be incredibly valuable and more and more v- H- how often are they doubling, would you say? Every year, you said?

    3. AM

      No, I, uh, every six months.

    4. LR

      Every six months these specific engineering skills are becoming more valuable.

    5. AM

      Yes.

    6. LR

      And the idea is this, you don't need to, like, know everything. You don't need to know the foundati- like, to build the app as much.

    7. AM

      Yeah.

    8. LR

      It's more to un- unblock the agent and understand the mental model of how this stuff is built-

    9. AM

      Yes.

    10. LR

      ... so that you can move forward fast.

    11. AM

      That's right. That's right. Understanding the basic components of it, I would say. Yes.

    12. LR

      Yeah. So it's like we need new engineering-... schools to teach you these very specific skills-

    13. AM

      Yes.

    14. LR

      ... versus spending years on like algorithm, algorithms and, and games.

    15. AM

      And, and I think, I think no one has done that yet.

    16. LR

      Right.

    17. AM

      And I think this is, this is like a, you know, uh, big business probably, uh, ready like to get built. It's like, uh, AI native coding. It's totally different than, than like traditional coding.

    18. LR

      Yeah.

    19. AM

      Um, that's why on Hacker News, there's so much skepticism about like AI native coding tools because they're like, "Yeah, it's a, it's a glorified auto-complete." And I understand like if you're, you know, writing operating system kernels, you know, it's not really doing that much for you. But if you're building products, it's building it for you (laughs) at this, at this, at this point, right? And so, you know, if, if you're, if you're starting a school to, to teach AI native coding, you would skip so much of computer science and the, and the basic tools. And, uh, you would, you would teach the basic idea of how to structure an app, and then you would teach prompting, uh, and then you would teach, I think, a little bit of debugging. I think debugging is, is quite a, quite a good skill, uh, right now to learn.

    20. LR

      And interestingly, if you wanna be good at debugging, you, like there's a lot you need to understand, which is basically what you're saying, is like that's the subset of things to understand-

    21. AM

      Yeah.

    22. LR

      ... is things that break. And to do that, you have to understand how it all works, what are servers, what are APIs, all these things. Okay.

  12. 50:36 – 1:04:08

    Replit’s new developments and future plans

    1. LR

      How far, so we've been talking about how this is very good right now, building a prototype, building a V1, MVP, people can use it, you can deploy it. You deploy this app, people can start using it, and there's a, like a scale it can reach. Do you see a future where you can build like a Salesforce size business fully Replit or other tools that can scale to hundreds of billions of dollars of value? Or is there just gonna always be some limit of like you need like actual engineers and designers sitting on this thing, building it, thinking it awesome?

    2. AM

      If like my law is like, (laughs) you know, directionally correct, even, even if the months are not, uh, are not exactly correct, the, the duration is correct, you're gonna see a compounding effect of, of the power. Like, it's actually quite hard to convince yourself. But if you really convince yourself that we are on a massive scale of improvement in, in AI, then, then, then the answer is yes. And, and it's, it's like absurd to my engineering mind that I'm saying this, but, you know, Ray Kurzweil with this like, you know, futurist, um, you know... Talks about how exponentials are really hard for humans to, to grasp. And so, actually when we started building the agent, um, you know, I, I told the team it's easy, and we've fallen in this trap before, it's easy to build and optimize for today. You know, in, in '22 we built like, um, you know, co-pilot like thing and auto-complete, we train our own models, we optimize the hell out of them. But at some point, like that modality was kind of, you know, not the right modality, which is like the auto-complete modality. And that, and the right modality is actually this, I think, for now, uh, is being able to chat inside the programming environment and, and for the agent to, to create things for you. But in order for us to, uh, make that bet, you know, a year ago, uh, the models were actually not there. Like the models could not do this, but we were like, "Okay, we're gonna build for the models that are landing in six months." And it w- i- it truly like six months later, the models started l- to land that, that are capable of this, of the reasoning that we need and, and whatever. And so that was just like, you know, sauna for you won, which is, oh wow, like we switch to it and the reasoning improved so much. And six months later you have a sauna for you too. Um, and so it's (laughs) really almost like a six-month cadence. And so i- i- if, if we're really on the, on this trajectory, then, then, you know, I would say, you know, next year you're able to, to scale and maybe, maybe you get, you know, thousands of users paying you. Uh, the AI can do maintenance, you know, we already showed the, the AI doing like SQL queries and doing migrations. Um, so the AI will be able to do maintenance, debugging, things like that. Um, I think where it gets really tough is that, you know, uh, w- when you're hitting scale and you wanna architect a system that is re- resilient. Um, and so that means, uh, you know, y- you would start, you know, charting databases and you would start like using different queue systems and components and, and things like that. And I think, uh, you know, the AI needs to have access to the entire suite of, of tools to be able to do this. And, and, and I think that, that's gonna be the next bottleneck. And I think the AI needs to get, be a lot more reliable, uh, at, at, at doing that. But I could imagine like whatever, five years from now, uh, someone running, you know, a billion dollar company with zero employees where it's like the support is handled by AI, the development is handled by AI. Um, and, uh, and, and, and you're, you're just, you're just, you're just building and, and, and creating this thing that is, you know, that, that people are finding valuable and are, are paying you for it. That being said, it's worth like thinking about the economics of it. Like if the, you know, if the cost of software goes down a lot, like w- uh, then, then what is, uh, you know, what is the, the price that you can charge on, on software, right? So can you actually build the next Salesforce if anyone can generate Salesforce, you know? And then, and then the question is like, what is the, you know... And this is why I emphasize being generative because I think then the, the thing that would make you better is like by being able to iterate and improve the thing really quickly and generate new ideas.

    3. LR

      And stay ahead of all the other people building-

    4. AM

      Yes.

    5. LR

      ... these tools so quickly.

    6. AM

      Yeah.

    7. LR

      (laughs) Oh my god. Um, (clears throat) an interesting other kind of mental model I'm, um, seeing as you talk about this sort of thing, uh, is...Not to offend religious folks, but there's this concept of God of the gaps. I imagine you've heard that.

    8. AM

      Yes.

    9. LR

      Where it's like God explains all the things that we don't yet understand, and over time that kind of space shrinks and God's like... All the things we don't get yet, those gaps? That was God. And that's, that's what-

    10. AM

      Right.

    11. LR

      That, that's, that proves there needs, there needs to be a God. And it feels like right now humans are like the gaps in these tools.

    12. AM

      (laughs) Yes.

    13. LR

      Or these agents you talk about that you can hire within Replit are like fixing these little gaps. And over time, AI will fix these things themselves.

    14. AM

      That's right.

    15. LR

      And these gaps will shrink.

    16. AM

      I mean, un- unless, unless, uh, we hit some fundamental limit in the current-

    17. LR

      Mm-hmm.

    18. AM

      ... regime of, of, of AI, which, you know, I'm not, I'm not an expert about like how far transformers could, could scale. But I, I feel like it is, you know... We found the thing that could, that could scale pretty far, but maybe there are limitations in data or, or o-other things like that that we could be, we could be surprised by. But if there isn't, then we are on a massive trajectory of removing these gaps, uh, quickly.

    19. LR

      Yeah, very true. We have no idea. We keep thinking it's just gonna keep going, but maybe, maybe it'll stop at some point (laughs) . I could keep going and going, uh, but I think we should also let people go play with these things and process all the things we've been talking about. Uh, is there anything else that you think might be helpful for folks to think about or learn or study?

    20. AM

      You know, I'll give advice to, to sort of, um, uh, founders or, or leaders at, at companies. Um, the way we work is, is gonna change rapidly and it's important to be s- sort of resilient to that, to that, that change. One thing that I think is really difficult now is having roadmaps, especially if you're doing anything in AI, but really any- anything that AI could affect. You wanna be able to react to it really quickly. And so, (clears throat) you know, when the, uh, Anthropic dropped the computer use, um, uh, sort of capability, you know, we slot it in our roadmap because we don't really have an (laughs) explicit roadmap. We, we like immediately jumped on it and started building things, and we launched some things around it. Uh, and we're gonna be doing more with it. But like, there's gonna be capabilities that, that are gonna drop and you, you wanna really, uh... I- in some cases, if it really affects your business, you wanna be able to jump on it really, really quickly. So, so being agile, not being, uh, sort of stuck with, with roadmaps, being able to kind of just, just say, "Oh, we're just gonna, we're just gonna switch priorities right away," uh, is gonna be, uh, super important. Not being, uh, you know, uh, like I said with silos, um, at Replit there's so many people that are on, on the scale of like, you know, designer to engineer, designer to product manager. Actually, uh, I mentioned Aman e- earlier. He started as a designer at Replit and now, now as a product manager. We have people who start as, um, uh, designers, become engineers. And we have people in the middle and we're comfortable with that, like design engineers, um, a-and that fit at different parts of the scale. And the, the design engineers go to the design crit meetings and, and d- some designers go to the engineering meetings and... You just got, you gotta be fluid, right? Uh, because, you know, again, when, when designers can code and, uh, and engineers can design, I mean, it's, it's really, really comes... It, it... You can't have a lot of structure around that. So you wanna build a culture and you wanna build, uh, a, an environment or milieu that is like really f- really flexible, um, which is uncomfortable for a lot of people.

    21. LR

      Man, the future is wild (laughs) . Everyone's, everyone's a hybrid person now. Uh, let me just actually double down on what you just said, which I think is really interesting. It's almost like if you're an engineer where your skillset will become most valuable is unblocking these AI tools and knowing debugging and figuring out how to unblock, h- h- allow it to go further and further and further. Within PM and design land, if... Based on what you're describing where the skills will become more valuable is, uh, I... Generating ideas, almost like finding opportunities, discovery, finding what problems need to be solved, and then articulating that as clearly as possible to the AI tooling.

    22. AM

      That's right.

    23. LR

      Super interesting.

    24. AM

      I would, uh... Yeah, this is, this is, uh, a very crisp, uh, sort of advice that, that people can follow today, I think.

    25. LR

      Oh man, what a world. Okay. Uh, Amjad, this is incredible. Uh, my mind is racing. I've gotta go build some apps immediately (laughs) .

    26. AM

      Give us feedback about that.

    27. LR

      Uh, I will do that. So just to leave listeners with a couple things. One is just what should they know? Should... Where do they find you? Where do they... How do they try Replit? Anything else other than just going to replit.com?

    28. AM

      Yeah, just go to replit.com. It's, it's in open beta right now. We're, we're kind of, uh, quickly, um, improving and gonna exit beta, I think, in a, in a, in a few weeks. But if, if you're comfortable testing something that's like not, not perfect, go to replit.com. Uh, if you, uh, subscribe to our, uh, core plan, you should be able to access the agent and, and start using it. Um, and we are... Uh, you know, I think, uh, the place where we're most active is, is Twitter. So Twitter or like X, uh, the handle Replit, R-E-P-L-I-T, or my handle, uh, @amasad.

    29. LR

      Oh yeah. One other thing I wanted to make sure we had a chance to touch on is you're working on something new, something that's coming in the very near future, maybe the day this episode drops. Talk about that.

    30. AM

      All right. So depending on when, when the episode is, is coming out, this could be the fir- first time people hear, hear about it. But, um, we have this product call, called Agent. It is sort of high agency, does everything from setting up the project and, and all of that, right? And so now we ha- we're, we're, uh, we are working on Assistant.So, Assistant is, uh, like, let's say, the- the cousin of Agent. It is a little less powerful, but a lot more controllable. So you can, like, focus on, uh, features or areas of the code that you- you wanna change and y- uh, you- you still don't have to know how to code but it is a lot more manageable and it is a lot faster. So you saw how, like, it took some time to kind of create the project and code some of the things. Assistant i- is in the order of milliseconds and seconds to be able to respond to you. Uh, and so again as- you know, as I talk about the idea of tools, we want people to have as much, uh, power and autonomy as possible and so there- there are certain instances where- where Agent is the best, it's gonna do the debugging for you, it's gonna create the database for you. But if you want more control, Assistant is, uh, is gonna give you that.

Episode duration: 1:04:08

Install uListen for AI-powered chat & search across the full episode — Get Full Transcript

Transcript of episode Bp_h674oIhw

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.

Add to Chrome