Possible: Amjad Masad on vibe coding, AI agents, and the end of boilerplate

24:08Aria Finger:You're like, ah, just organize my camp schedule. Right, exactly. Or I want this educational thing. I want to be able to do this fun thing with my kids. And then talk to it like you would talk to a person. Just say, well, I have an idea. And just don't worry about the grammars, all that. You'll get better at prompting. But initially, just really get in with any kind of idea. If you want some sample ideas, I would just say basic things. Just say, I want to build a collaborative whiteboard. And then just imagine if you've done improv. Improv is a good way to do it. Just like word association. I want to collaborate with whiteboard and then look around the room and you'll see a plant that allows me to create a garden or something like that.

24:50So really just be generative and you'll get an experience and you'll get a feeling of it.

24:55Aria Finger:Do you know who is the typical Replit user and what kinds of things are they building already? Is it 80 % of people are building whatever, SaaS software? Yeah, yeah, yeah. I think most of it is AI software, which is why we built this recent thing. It used to be that you have to go figure out how to get an API key from OpenAI or others. But people have AI ideas. Like they see AI and like a lot of people will be like, oh, I wish ChatGPT did this one thing. Right. And then they want to build AI experiences like chatbots, like ReadAI is a great example. Obviously ReadAI is very advanced in that it has video and all of that.

25:33but you can build like you know i'm jot ai like go like tell replet to like go look at your personal blog or paste in your notes and just say i want to create a chat experience that sounds like me based on these notes and so ai is a really big one another one is like people want to build businesses and this is you know as i started thinking about our business more there's two aspects of our business i would say maybe three there is the uh consumer aspect which is you know building family stuff and building personal things. And then there's the enterprise, which is like automating things at work, line of business applications, rapid prototyping, product development inside businesses, larger businesses, and then entrepreneurs.

26:16And the way I started thinking about entrepreneurs is that everyone has domain knowledge in their head that is not monetized just yet. Similar to how Airbnb found that everyone has space in the room that's not monetized. There are people all over the world that have niche-specific problems that they've run into that no one else has solved. I was talking to an entrepreneur on Repli the other day. His wife is a yoga teacher in rural England, and she does these pop-up yoga events, and most of the communication is happening over WhatsApp. Payment is really hard to do. They use PayPal, all these different things, and so he wanted to build a platform for her, but other people like her.

26:57So she doesn't have a yoga studio that you can go to every day, but she does these events at backyards and things like that. So he's built a platform and he has users and he has other instructors. And so every niche in our life, there are people that are experts in that previously were not entrepreneurs because they didn't have hundreds of thousands of dollars to go pay developers to do it. Now they can do it.

27:19Aria Finger:I love that example because obviously there's entrepreneurs that you're serving that are building million-dollar, billion-dollar businesses that serve hundreds of thousands of customers, millions of customers. But I think that like solopreneur, the woman who has a yoga studio, like I think of my mom who's a piano teacher and she's always trying to figure out how do I get my schedule straight? How do I get payment? How do I figure this out? And so there's so many, like she's an entrepreneur, the person who runs a yoga studio, the person who has like a bakery outside their house. Like those are the people who would never be able to afford fancy coders, a fancy website, but Replit can empower them to do so much more.

27:55Aria Finger:So I love that. And there's something about it that makes me really excited about the world to come. As much as I love Silicon Valley, we tend to flatten the world. We just build one size fit all. And that means SaaS software tends to have all these buttons and configuration just to fit every use case. But instead, we're going to explode that and decentralize it around the world. And every community might have a special business that does one specific niche thing. That diversity, I think, is exciting. And as part of that diversity and being fun, we'd like to play for you a short clip of one of the ways people have used Replit to build something important to them.

28:37Okay, great.

28:38Aria Finger:What if you turned Boba into a game? I prompted a basic app and sketched out some Boba. I made three different drinks, taro, milk tea, and matcha. I added logic so the pearls fall from the sky. If you miss three Bobas, you lose. But the game was way too easy, so let's make it difficult and add some bombs and a scoreboard. I noticed the game was a bit glitchy and I wanted to try it on mobile. So I'm at my favorite cafe. I added more colors, different fonts, and harder levels. That's the best thing. People can code on their phones right now, which is amazing. Exactly. What other features should I add?

29:08That's fantastic. And we didn't just select it because she added it to her bio. We first started working on the phone. People said we're just like crazy. Because back then there was no AI. and we had to build a lot of things to make coding on the phone a little bit easier. But I always imagined, because back in school, I remember that I had a Symbian phone. I was like, I wish I could program it under my desk. So it's really cool to see people doing that now. So one of the things that as you dig into AI, it becomes, one, the earlier question, which is how do you broaden your imagination? But also now, what's a good way of optimizing?

29:52For example, obviously you can do things that have huge token counts, and that's kind of like, doesn't build a good business, et cetera. It might be fun as a one-off thing, but that kind of thing. What is, as it were, in the game dynamics, the next level of compute thinking? Because it's like, okay, how does this work within the operational envelope of AI enabling it, being able to figure out what kind of interface elements? Right. What's some of the things you've seen there in terms of, you know, kind of context optimization, managing the AI token outputs, you know, helping people with that kind of compute thinking?

30:28Yeah. I mean, there are two aspects of this. One is strategy. You want to be able to do things that the big labs are not going to do or eat your lunch and like actually be able to build a good business. and then the other thing what can LLMs do right now and what scaffolding do they need around them the scaffolding part is really important because this is where you can build the value on top of the LLMs so the way I think about our business right now we were in the business of training models now where we use all the different commercial models and I see our businesses to build the milieu the environment around LLMs if you think about LLMs as a person you want to create the habitat for them.

31:11So we built the best habitat, I think, for programming. So when we first did this pivot from the human being of the programmer to the AI being of the programmer, I took the same teams at the company and I told them, now you're no longer built for humans, you're building for the AI. So the editor team built editing tools for the AI. The cloud infrastructure teams that have built deployment tools for humans, they now build it for the AI. So I would look at any different vertical and I would think this way, if you want to build a financial agent, just think about what kind of environment you can build for that agent.

31:46Now, I see a lot of people, for example, trying to put LMs into Excel sheets, whatever, which I think works okay, but can you build an Excel interface for the AI? In a way, it's textual, it can use it based on that you want to hit the right distribution of the LMs. You can pick any vertical and if you want to build a really great business, just think about building your first users of the LLM. Think about building that habitat for LLM to live in to create amazing experiences. Amazing things with the product. The unlock of creativity, the unlocks of what does that mean for more personalization, things that work for individuals, businesses.

32:24Creative thinking is great. So let's talk a little bit about also some of the business attributes. One of them is, you just mentioned that part of what you've done as opposed to homegrown in your own models, you're now using a fleet of models. But that obviously makes you somewhat dependent on them, makes you more copyable, et cetera, business attributes. So one of the things that the people who listen to our podcast a lot are entrepreneurs, you're navigating a business challenge that they need to learn from. So how do you think about that? What are the things that you're doing to build unique and enduring business?

32:57What are the kinds of considerations that go into that? First thing I would say is user obsession. There's a massive platform called LMs, right? That platform is useful on its own. People can use it directly. But for you to build a very valuable business on top of what kind of value you're creating, one is if you're really obsessive over certain types of users, OpenAI can't build for every type of user in the world. So pick a user that you know very well. Hopefully it's you. You've had that experience. You're a salesperson. You were going to build the best LM for salespeople, right? That goes into UI, UX, all of that.

33:30But second, and this is my comment. I'll go a little deeper on the comment about the environment around it. So I said that people using Replicate should feel like they're in a safe environment. It should feel like anything is reversible. Now that turns out to be a huge technical challenge. Now you could easily kind of hack around it or say, we'll just restart the app or do whatever. But what we did is we built a transactional file system that allows undo. Like we have proprietary file system that we spent two years building that every action is like an immutable part of a ledger that anything can be rolled back and rolled forward.

34:11That ends up being like a time travel system. And now we're finding more use cases for it. For example, one way to make models better and more performant is by doing sampling because models are stochastic. Every time you're on the model, you might get better results or you might get worse results. And if there's a way to verify whether the result is good or bad, you can sample from the model. So what we do is because the system is immutable, has this characteristics of being very cheap to fork. So you can take a file system, you can fork it a hundred times, run the same prompt with different parameters on these different forks, and then pick the best one.

34:52Do it repeatedly. Obviously that costs quite a bit of money, but for the advanced users, they're willing to pay it, they will get the extra money. So do hard technical things on top of the models in the environment where the models are sitting, right? You want to build technical advantage. It doesn't have to be in the model itself because that costs billions of dollars. There are huge industries around it. You don't want to be training models, at least not yet when you're initially building a business. But you can do a lot of other hard technical challenges on top of it to make the models work a lot better.

35:25That's the second thing. Third is reaching scale. Once you reach scale, you actually have negotiation advantages. You can negotiate really large discounts with providers, similar to cloud. And you can pit them against each other and so on and so forth. So venture capital is an amazing system to give you that initial money that allows you to get to scale quickly. And then you can enter more of an optimization phase of the business.

35:55Aria Finger:So you have this great metaphor, which is cathedrals built from bazaars. And so you want to describe the tension between open source and closed product design. And the bazaar is this messy, collaborative, maybe fun, bright, and that's where open source lives. And then you have the cathedral, and that is the top down. That's sort of Apple, where they live. And you said that you want to build cathedrals from bazaars. Tell us more about what that means to you. Yeah, so I think that comes from Eric Raymond or some of the original Linux hackers, cathedrals versus bazaars. And there was a huge discussion early on in the world of software.

36:33And so I gave a twist on it. It's cathedrals made of bazaars. Now, in the early 2010s, when this idea of no code kind of emerged, there are a lot of different approaches. There are amazing businesses built there, obviously. Zapier, Airtable, Notion, you can all think of them as no code builders. Now, our approach, which for a while didn't work until AI kind of really made it work, the amazing, hyper-competitive, diverse landscape of open source, of people building all these open source packages and components, what's happening there is bizarre. bizarre. It's messy, but it's very innovative. Can you build a facade on top of all of this that makes it feel as good as a user experience as, say, a notion?

37:22It wasn't obvious that you can do that, but the kind of abstractions you have to build on top of that, and you can still hook into the latest and greatest. When a new package comes in open source, Repl.it has it on day one because we built all this abstraction to be able to go get the packages and install them and run them on the system. And so we're riding a huge wave. We're riding multiple waves. LM's getting better. We're also riding open source getting better. New programming language comes up, we can run it on day one. And that acquired a lot of investments upfront. And there are a lot of competitors that are not doing this, and it might be short term easier for them to do.

38:00But I know how the landscape programming works. Things will shift. Like the other day, we saw a new JSON format called Tune that is more efficient. It's, I think, token-optimized format that is more efficient for LLMs. And so again, programming will continue to improve, and we want to be the platform where you can program anything with LLMs. Let's think a little bit also about how the future of work is going to evolve. Most people's experience right now is with chatbots, so they put in something, they get out something, they put in something, get out something. So they're kind of directing the work in some ways, but it's kind of iteratively from what they're doing.

38:40So how do you imagine this trend, Replit, others, in this kind of everyone working with kind of that power of code? Because you need a combination of that individual creativity, imagination, broadened imagination drive, but also standardization for work processes. It's one of the reasons why the classic industrialization, I felt this form and then I hand it from you, and da-da-da-da as we're doing it. So I'm certain that you've been thinking about what this means for the future of corporate work. What are some of the observations you might share? Yeah. When I first came to the corporate world, one thing that I found extremely depressing is how atomized and siloed people are.

39:23So, you know, over time, you know, with capitalism sort of emerging, you have the original, I think, Adam Smith, like, how do you make a pencil, right? Like, someone make an eraser. And so people become sort of cogs in the machine. And I don't think that is conducive to the human spirit. Like I think humans are fundamentally creative. I think anytime we assume the shape of a machine, I think people are actually not happy in their lives. I'm not a Marxist, but I've read some Marx. And Marx has the theory of alienation. I think Marx was right on a lot of the critique of capitalism, maybe not the solutions.

40:02definitely not the solution one of his critiques is alienation is that you don't see the fruits of your work you build one part of a large supply chain of things and that doesn't make people proud like I said when I was a kid when I built this entire thing my brother used it that's amazing that's what being human is about and so I think now Now there's an opportunity in the workplace to a lot of that standardization stuff to be taken care by agents, by work process automation and all that stuff. And what's left for people, yes, we'll have broad swim lanes. Like I'm generally in design, but I can code.

40:49I can write a PRD using chat GPT or whatever. I can even have a marketing idea. and I can use all the tools and like, hey, I built this marketing package. And that to me is a lot more exciting. So we have, for example, one of our customers, large enterprise. One person had an idea. It's a real estate marketplace where they connect real estate buyers with real estate agents. Had an idea on how to optimize that algorithm and couldn't get engineering resources. Now, previously, it was just dead. The idea is dead. Although he's in the business, he understands it. He talks to real estate agents all the time.

41:28Instead, he went to Replit. He built a new routing algorithm that drove tens of millions of dollars, if not$100 million to the business. And that person got promoted again and again and again and again. And now he's sitting with the board members guiding their AI strategy. Entrepreneurship, I think, in my opinion, is not just about building businesses. You can be an activist entrepreneur. You can be a humanitarian entrepreneur. You can be an entrepreneur inside a large business. And I promise you, your life will be a lot more interesting, exciting. You'll meet a lot more exciting people. You'll have a better reputation.

42:02You'll have a halo around you. So my hope for the future of corporate work, but humanity in general, is that as the boring things, the automated things, actually taken care of by machines, we're left to be more entrepreneurial and more creative. Obviously. Totally hope that's the case. It goes back to my very first book, The Startup of You. So it's like, it's a definite strong, because it's like, it's not that we all have to start businesses, but that kind of creativity, entrepreneurship, engagement with it, the exact passion that you had with your brother. One micro thing that I think is an important thing for people to focus on, because it's part of the general lack of context awareness that current agents have.

42:42What do you see as the kind of the longest unsupervised work that agents can currently do? I think it's more possible than our tools allow it to be. So I think state-of-the-art LMs, especially Gemini has done a great job at a million contacts, million token contacts, but others have 200 ,000 or somewhere there. The easy thing is to fit all history inside the contacts, but that's not feasible. There are few components. If you put them together, then you can run agents a lot longer. One component is verification. The work needs to be verified. It's a form of supervision. If the AI agent is going on a path and doesn't have feedback from the environment, from the work, from the human, for someone, then we're very likely straight off the path and do something really, like if you make one mistake, it'll just compound over time.

43:34So Replit Agent 3 came out in September. We now boot up a browser for the agent, have the agent use it to test the application for you. and we have another agent that's adversarial that's like reviewing the code and giving you feedback and because of that we went from agent one ran for two minutes unsupervised after that it didn't work very well agent two is 20 minutes agent three 200 minutes we're seeing 200 minute runs that are actually really really useful there's still mistakes and all of that now we've introduced an autonomy selector inside Replit So depending on your risk appetite, you can go all the way high autonomy.

44:17And some people run it for 10 hours. Now, I don't recommend it to everyone. Now, if you have a lot of disposable income, you should definitely do it. But we have medium autonomy and it can run for 100 minutes, high autonomy. It can run for 200, 300 minutes. So I think it's possible. And in terms of the context, so the verification is very important. So you pass from one agent to another. And so one agent does a unique work. another agent gets a prompt and says, test the app. You don't have to give it the entire thing. So it's a multi-agent system. And then after the testing is done, pass back.

44:49But you don't pass it back using all the existing contacts. You kind of summarize and give the next. It's almost like passing the baton to the next one. Again, ideally, we would put everything in the agent, but I'm not sure we're going to be there anytime soon.

45:03Aria Finger:So I think this sort of relates back to what we were saying, how the difference between programmers and non-programmers is collapsing. because if everyone has a Replit agent, you don't need to be a programmer. And I think so many people are talking about, and you said yourself, you worked at Code Academy the last 15, 20 years. We were telling everyone, like, parents, teach your kids to program. Learning to code is the future. You recently said you don't think people should learn how to code. And so I'm all for the democratization of this for people who are not coders now to be able to use Replit to do amazing things.

45:37Aria Finger:Yet at least right now, I feel like the people who are most adept at using AI for anything are actually the people who have learned how to code. So I would say to you, do you think we're at a moment in time or will that continue to be true? You're a parent. Will you want your kid to learn how to code in the traditional sense or that's going to be a skill that's not important soon? Look, similar to what Grace Hopper said, you know, 75 years ago, she said specialists will exist, but we want this technology to be. So my advice would apply to 99 % of people. Like, you know, I don't think you, most people who want to make things with software and computers don't need to learn how to code.

46:17It's going to be a huge cost on them. It's going to take a lot of time, but instead just jump into making things. And if it's your cup of tea, if you're looking at the Asian traces and you're seeing the code and you're like, okay, I'd like to learn what it's doing, then go learn it. I think it will take you further for sure this time along. But if you believe in AI, as I do, and we're going to continue to make progress on AI agents and all the labs, billions of dollars, perhaps trillions of dollars are going towards optimizing code specifically, then we're going to get to a point where it's, again, specialists exist.

46:52We need them. If we're sending people to the moon, I don't want them to be on a Vibe-coded computer. But for the vast majority of applications, I think Vibe-coders will be able to make them. Now, the question about my kids, I've struggled with this a little bit. I've never wanted to admit that there's a temperament to coding, but there is. There is a certain temperament to coding. Do my kids have that temperament? I'm not entirely sure.

47:14Aria Finger:What do you think is the temperament that a coder needs? Basic thing, like the ability to sit in front of the computer for 12 hours. Sure. I think coding is actually bad on your health. And if you're okay with that, like you really, I mean, I don't know any programmer that didn't go through periods of life where they're like really spending a lot of time in front of their computer with diet codes or whatever and like just like coding all the time and like learning every aspect of it do you enjoy that a lot of people don't and that's fine like again diversity like did people come with all sorts of temperaments and some people want visual things want to interact with visual things some people want to work with other people all the time but if you're someone who is comfortable like I was, and I am, I am like totally, like my wife is not the same.

48:00Like I have an insane ability to be alone. Like I can be alone for a year, but other people can. And you know, that's fine. We should be okay with people of different temperaments.

48:11Aria Finger:That is actually really illustrative to me because I love math. I love logic. I loved AP computer science in high school, but I'm an extrovert. I couldn't, I can't do that. I can't sit with code for 12 hours. Okay, so people like me are okay in the future. They'll find other jobs. Soft skills are important. Exactly, soft skills are important. Well, so I feel like we're in this time where some of us see Replit and see AI and we're like, oh my God, it is so empowering. We're helping small business owners. We're helping young people. But when you actually look at the numbers, it's like more Americans are nervous about AI or scared about AI than are excited about it.

48:50Aria Finger:And so if you're talking to someone like that, like, what do you say to them? Like, what's your sort of go-to line to be like, no, this is why we have excitement coming forward and in the future, even though we understand that there are risks. I think there's like a various distribution of intent. Some people just want to get famous by writing, like sci-fi sounding, like, you know, just to be concrete about it. Like, I don't think anything that Eliezer Ryzkowski read, for example, Eliezer Ryzkowski, big AI, maybe he doesn't describe himself as a doomer. I would say AI doomer, wrote a lot about how AI could end the world and kill literally everyone.

49:24Like I can't, like, you know, every time you try to read some of those, like there's insane loops in logic and really it doesn't compute for me like how that actually ends up happening. And I think it's like a fundamental misunderstanding of the world. Anyone who's tried to build a business, understand there are bottlenecks that are not just intelligence and there are regulatory bottlenecks that are all sorts of things. So without getting too deep into that, I think some of it we just have to counter intellectually. and the AI optimists have to be out there arguing with these people. I don't have time to do that.

49:59I did that for a little bit. But it's important. I know they're out there in D.C. kind of scaring politicians and all of that. And I think that's really bad because we can get to a point where AI is regulated in a way that doesn't allow the U.S. to be competitive. So there's an aspect of it that I think we really need to battle. There are aspects of it that I think is on the creative art film front. Sure. Like, you know, everyone thinks about Skynet and Terminator and things like that. But who's producing like really optimistic visions about sci-fi? I love reading mid-century sci-fi because it's a lot more optimistic.

50:42Something happened in our culture and we don't have really good sci-fi anymore. So that's another part. Now, third part, now this is the more pragmatic. I'm worried about my job. I'm worried about my kids.

50:54Aria Finger:These are just everyday concerns that people have that aren't sort of fantastical that maybe are sort of based in reality. I worry about it for my kids a lot. Like the idea of, it's really hard to go out in the world and make friends and have people like you. It's really easy to have an AI like you. I mean, and humans tend to take the shortest, the easiest path. And like, I worry about this AI friend thing becoming a really problem. My eternal optimist is always like, well, look, you know, we had a period of history where, you know, eating junk food was acceptable. And like, but I think there's like societal antibodies.

51:34I think healthy societies have antibodies around, you know, smoking. You know, smoking is like, you know, when I grew up in Jordan, everyone smoked. I smoked. I came to the U.S. None of my friends smoked. And I just felt like an outcast. And I stopped smoking. I think society is like, you know.

51:52Aria Finger:It's like truly one of the greatest public health battles of our time. And to your point, a lot of it was we solved it with culture. Yes, culture. Culture. I think religion used to play a role as well. And I don't know whether it will in the future or not. But like government is such a blunt instrument. You can't do everything. So I think it's leadership. It is culture. it is incumbent on all of us to just say like, no, chat GPT is not your friend. Like it's friendly and you can use it for therapy. You can use it for all sorts of things. And those are fraud as well. But like, if, if someone is saying to you, it's like, I'm spending all my time talking to GPT, hey, hey man, that's not really healthy.

52:34You know? So I think, I think a lot of it is about culture and society. Well, I do think that like, there's like two points that I think are very important. One is in the friends point, completely agree. part of the reason why like with you know inflections pie you know you kind of engage you're my best friend it's like no no no i'm your companion i'm here to help you have you talked to your friend recently you know what what's like with him or are you you catch up and i think that that that helping especially for young people yes where they might be encountering difficulty and so i think it's like uh tying you back into the community tying you back into connecting with other human beings because by the way part of actual real friendship is not just saying oh oh, Amjad, you're so great.

53:14Oh, Amjad, I was waiting for all year to talk to you, etc. It's like, no, it's that thing of like, where I might show up thinking, oh, I've had a bad day, I want to talk to Amjad about it. And then I hear about your day, I'm like, oh, we're talking about Amjad's day. That is actually part of how you evolve, part of how you elevate, part of how you learn compassion, wisdom, empathy. And it's really important. And that's part of the reason I also share with you, it's not impossible to make AI friends, but it's not what's happening right now. And it's not, you know, anything that is illusory that way is actually, in fact, destructive.

53:49That's why it's causing psychosis and things like that. Like, you know, there's high profile cases where people have psychotic breaks because they have like delusions of grandeur and things like that. And ChatGPT is like reinforcing that. And that's like really bad. And I think it's incumbent on the labs as well to like start a program pushback into these machines as well, right? Yeah. There was this funny thing that a friend of mine showed me, which is by only listening to, you know, AI agents, you end up in a desert by yourself, drinking a glass of milk, thinking the whole world's out to get you.

54:23Right. You know, so it's like, that's the wrong, long path. Now, back to the jobs thing. People ideally love to say, I'd like to have an elevation of my life, but I'd like to have no strain. And unfortunately, those aren't the times we're in. We're in the similar times to the Industrial Revolution. Now there's a cognitive industrial revolution. And so many, many jobs will change, right? Like if you said, my job is I'm a form filler for digital marketing. Like, yeah, that's not really going to be. You better start thinking about other things. Now, AI can help you with that. It's like, okay, well, what can I learn?

54:58What can AI help me do, et cetera? What we have to embrace is that kind of entrepreneurial mindset. People say, well, I don't want to. I just want to keep doing what I'm doing. It's like, well, so did the horse groomer back when there were lots of horses and that was all the transport. That's not the way it works. Right. And by the way, if you try to lock that in, you're essentially going to be stealing your children's future because the other society will move ahead. And then yours will now have put them in a very disadvantaged position. So the important thing, like almost like that Adam Smith specialization thing you were talking about, the important thing is say, look, I understand they're going to be painful.

55:34me, but I'm going to own this, this change because I am helping my children, my community's children, the future generations. And by the way, like own it as an adventure, right? Yes. Like sometimes you don't choose to go on the adventure, but sometimes you just find yourself on it and own it as the adventure. And how do you shape it and make that happen? And so in this vein, one of the things that, you know, companies are starting to look for is what is AI native talent? What kind of pointers, heuristics, thoughts, lenses might you offer to people to say, here's how you can think about becoming AI native talent.

56:08Here's what companies should look for in hiring AI native talent. As you know, you live in AI nativity. On the previous point briefly, like I say version that you said very articulately advice to people on the YouTube comments that I unfortunately sometimes read. Hey, be nice. People are, it's easy for you to say you're rich, right? And I think that Silicon Valley has a responsibility to do something. I don't know if it is some kind of like nonprofit where we have like, you know, retraining and upscaling. Maybe it's a government thing. I'm not entirely sure. But I think there's a responsibility to do that because it is hard for people.

56:56People who have done marketing form filling for 30 years, it is hard to break out of that. But I think there are potential to create AI education programs. Now, on the question of how to become AI native, we were at a CEO event earlier. And even the CEOs are struggling with it themselves. And really amazing CEOs built$100 billion plus companies. that I take a week every few months to learn AI, but I'm constantly behind. And what I said, if you're a programmer, have technical ability, go really deep. Go read the papers. Go build a fundamental understanding of how LLMs work. Maybe even build an LLM yourself.

57:41Like Karpathy has this really nice GitHub repo that you can go explore and try to train your own LLM. Now, this is conditional on the fact that you have some technical background. If you don't, I would say using tools like Replit to build LLM experiences. So kind of think about ideas for how LLMs could like improve your job and go try your hand at building them. That's one. Two, just like train your algorithms and feeds to give you AI news. One of the first things that I did when I just like committed our company to AI is like I went and followed all these like different AI people. My YouTube is all AI content.

58:26There's a wealth of information out there and so many of us consume content passively. You can be consuming information and learning. And this is, and by the way, TikTok can be certainly destructive. I see it as a skill to actually like train my TikTok algorithm. So like they try to bait me with like, you know, different content and I'll just scroll really fast. And I pitched this to Elon and I think he responded to me at once. So I said, you should be able to prompt your feed. I want only AI content today to be able to consume. But this passive consumption that we all do, you can fill it with useful information.

59:00Aria Finger:And so this next question, Reid, I do want you to weigh in on it. But you and Replit recently passed 100 million ARR. Congratulations. Huge milestone. And you guys have been building for a long time. But right now in the Valley, there's so many companies that are getting to, both enormous valuations, but also enormous revenue very quickly. So my question for you is, what do you think about this time where it seems like everyone has to get to insane revenue numbers immediately? Is this good? Is that true? And like, what is sort of the new kind of era we're in, sort of building in an AI era? So one of the reasons, there are a few reasons that happened that made it so that these revenue ramps are really high.

59:41One obvious one is the ROI is really clear. When we make an application for someone and then we put a paywall, they really like the application they wanted. They're much more likely to pay. But also, just like credit card penetration across the world has reached certain heights. Internet stability and connectivity. There's all these confluence of factors that got us to a point where you can grow revenue fast. That being said, AI revenue can go up as easy as it goes down. We've seen businesses, I think Jasper is now like an enterprise business, but at some point they had like a really amazing consumer business that they lost all of it to chat GPT, right?

1:00:20Like, so if you're not an AI business, if you're a SaaS business and your investor's like, oh, why don't you have AI revenue? Well, I mean, the other thing is like, you know, easy come, easy go, you know? So, and we spent like eight, nine years building all this infrastructure to get to a point where the company took off, but we're also still very paranoid. Like Google just released another Vibe Coding thing today and they're a close partner of ours and obviously they're going to do it. But so many companies are interested in the Vibe coding space right now. And so it's very, very competitive.

1:00:52And so if you can build a business that is lasting, that grows slower and compounds over time, I think there's still a lot of value in it. Now the question is like, are there a lot of investors that want to put a lot of capital into that? Well, maybe you don't need a lot of capital anymore. And so yes, a lot of the capital is going to AI, but you can build SaaS businesses today that are cash flowing, that can grow over time.

1:01:17Aria Finger:What do you think your moat is? Is it your brand? Is it your existing consumers? How do you make sure that you're not subject to that? I think there are less moats these days than before. I mean, one of my favorite books is Seven Powers by Hamilton Hamler, and it just talks about there's seven moats, really. Network facts, economies of scale, so on. I think that we have a technology lead in that we built all this technology for so many years. We built a hyper-competitive, amazing team that will continue to produce these technology advancements. We launch a new breakthrough version of agent every few months.

1:02:03Elon said this a long time ago. The only moat is continued innovation, rapid progress. And I think that's true in AI today. I think at some point when you reach certain scales, there are certain decisions that you can make that really create structural moats. We're not there yet, but I think, you know, I would say the big labs achieved economies of scale, obviously, right? And that's a moat for them now. It's very competitive between like the five of them. But I think it is a game that we'll play over time. So I was just, as CEO, just constantly thinking about what are the degrees of freedom that we have in order to position our business in a way to create lasting advantage.

1:02:41Your earlier question was almost like the classic blitzscaling question. And part of the reason why, now I'm trying to remember how many years ago I wrote blitzscaling, it was kind of the lessons from Silicon Valley and China about how the technology companies of the future are built. If anything, it's 5x, 10x true as what I did. It's like the trend of where it's going to go into. Now, that doesn't mean that everyone has to go to, you know amazing revenue 10x employee growth in like six months 12 months etc the thing in blitzscaling which you know chris and i wrote about was it depends a lot on what the nature of the market is what the nature of competition is and if you're going up against blitzscalers it's very hard to succeed without also blitzscaling because the blitzscaling will set the market it may blow up but we'll set the market if it blows up i'll probably blow you up with the market like Like a scooter market.

1:03:31Yes, exactly. And it doesn't matter that you were kind of, you know, minding your operating margins and, you know, it just gets swamped. And so you have to look at what's happening in the market. You have to look at what happened with competition. There are definitely opportunities where you can build more, like, you know, like one of the ones from the Greylock portfolio is Roblox, where like for many years, it's kind of like, yeah, yeah, yeah, out in the wilderness. And then boom, all of a sudden it's the amazing new thing. and you can build towards that and that kind of phase can come so you have to kind of make an evaluation of where you're at now if you're going kind of where the zeitgeist is where everyone currently recognizes it you'd better be moving pretty fast because everybody is recognizing it and so so that's a little bit of kind of the the reason like with like with monist ai we could say hey we can actually take a look some time to think about exactly and and build iteratively to what the AI pieces are for doing for the drug discovery because most of the people don't realize what are the things you need to do.

1:04:30And so you can take more time to build that. You don't have to go, you know, like it's basically do extreme uncertainty risks trying to capture a market. So there's a nuanced answer there. But there's a number of markets where that speed and that speed revenue scale, by the way, the speed revenue scale is not just revenue feud, not just what comes in as capital. It also comes into what happens with hiring employees and a bunch of get to that so it's you you have to make a kind of strategic decision by the way you may have to change your decision as you're going one thing on that is that what's interesting is once you hit an infliction point inflection point then other people pay attention and now it's split scaling because they notice a revenue and they want to go after it so they see you being successful and all of a sudden that becomes the hot industry and so you have to change your stance yes and that's the other thing that's what happened with us right like you know it's like your replet agent was kind of the first vibe coding thing that anyone could use.

1:05:26And suddenly everyone poured into it. And this is one of the things that people don't understand about the blitzscaling side, which is it's not that you start blitzscaling from day one, and it's not that you blitzscale forever. By the way, if you don't end blitzscaling at some point, I think that the seven powers, because I, you know, great book, great call out, maintain. I also think that the configurations of how business models come together will still actually get to different kinds of moats. Now, we have speculation of what those moats are. Are the moats only scale now? Are the moats brand?

1:05:58Are the moats network effects? You know, are the moats data now? Like, it'll be in different combinations of different things in terms of - But you have to be okay with the uncertainty of it in the moment. And you're going to be iterating through it. And you should be thinking about what those moats are. And you can't go, well, we, you know, because Silicon Valley went for maybe even two decades and I was one of the architects of this, you know, gaga over network effects. as the mode. And by the way, network effects are still very valuable, but they're changing in the shape and how they establish and what happens and so forth.

1:06:29And that reinvention, like people tend to think, oh, the most central thing is you reinvent your product or service. And by the way, of course it's really important. By the way, it's also reinvent your go-to-market. Oh, and by the way, it's also reinvent your business model. Oh, and by the way, it's also reinvent what the modes are behind this. And it's all of these kinds of different things. Like Google search doesn't have network effects from its search engine. It has scale effects, but it's AdWords is where the network effects. And then what it does is it buys distribution from Apple and other people through that is the way that it kind of, and so how do these things kind of come together?

1:07:06Aria Finger:And actually sort of per our conversation in the car, Amjad, one of the things I've really appreciated is that your willingness to speak out about things that matter in our society, whether it's Gaza, human rights, et cetera. And it can be difficult for people to speak out, especially when they're CEOs of companies, you're fighting for talent. Why do you think it's important for people in the space, specifically in technology, to speak out about human rights and things like that? I don't do it because of this. I do it because of principles and because, you know, we also have a family history. And I understand what people in Gaza are going through right now.

1:07:39And I really care about that. But I also think that the corporate world, the market has changed and that when you stand for something, I think there's actually an effect where maybe you might be certain people don't want to work with you or don't want to invest in you. But there are other people that might be attracted to that. right so i think the you know people are no longer sort of attracted to these you know bland corporate brands and the faces behind these companies i mean thanks to linkedin and things like that are becoming more important than the brands themselves and as humans we care about things and if you're if you just have no opinions and you're just no one has no opinions right and so uh i think a lot of people come to work on repli because they know i care about these things and people want to invest in Replit because they know we stand for certain things.

1:08:33And I think Silicon Valley is no longer the underdog as well. Like Silicon Valley is now the mainstream. People look to us for leadership, but they also, when we do certain things that are not conducive to society, they're going to come and try to burn us at the stake. And so it's important for us to kind of take that, the role that we're playing seriously. Yeah, I completely agree. And I think actually part of how you think of this Think of humanity first, and that's part of the principles. Alleviate suffering, economic enablement. And part of what I love about Replit is economic enablement across the entire world.

1:09:10Then society. What's the organization on which we live together? And then industry. By the way, industry is super important. It's been my entire life. But thinking about those, even when you're an industry leader, you should not abandon the fact that you're also a humanist and you're also a member of society. That's right. Absolutely. On that basis of frame, what are some of the really important things we should do? Leaders in Silicon Valley and the tech industry to make sure that technology is spreading compassion too. One of the problems with some of the social networks is it's rage and engagement, clickbaiting and so on.

1:09:46But yet that is not who we aspire to become as individuals in society. So what do we do to grow empathy, grow compassion? What are some of the ideas that you might have kicked around on this? One thing I try to do on our business is align our business model with something that allows us to be good in the world. And so Replit benefits when people's lives improves through that software that they're building or the business they're creating or how they're improving their business. And so I've always been really inspired by businesses that create like a game theoretic win-win-win situation. when you think about platforms like Shopify is an amazing business where the company's winning the entrepreneurs are winning and consumers are winning not every business is like that there's a lot of businesses where consumers are losing or someone's getting exploited but by the way the best businesses are the win-win-win businesses so I think one of the speaking of modes and things like that one of the early strategic decisions that you can make is what is the business model and try to try to actually kind of write down, it's usually, you know, different constituents, usually the company, the users, maybe the developers, whatever kind of business that you have, what situations or what business models allow you to create that joint winning?

1:11:12So now we're moving to our rapid fire. That doesn't mean you have to answer rapid fire. There's this questions we ask all of our guests. Is there a movie, song or book that fills you with optimism for the future? You know, a book that I think is amazing, that it pulls together everything that I love. I really love learning about theory of consciousness, theory of mind, the philosophy of that. I like computers and math and physics and all that. I like the human aspect of human stories. I Am a Strange Loop by Douglas Hofstadter is an amazing book. It's partly a book about his wife dying and their kids and how his personal story in that.

1:11:57It's also a book about him struggling with the concept of soul and kind of trying to understand it in a more secular way, in a mathematical way. It goes through vignettes about Gerdel's incompleteness theorem or what that says about how consciousness is potentially a loop. And it's an amazing book that tells you that really everything is interconnected and we can't be technical or innovators without actually caring about all the other things that happens in our world.

1:12:29Aria Finger:What is the question that you wish people would ask you more often? I mean, this question is really fascinating. I mean, I don't get to talk a lot about, you know, literary or philosophical inspiration. So that's a great one. Yeah, one, the strange loop is awesome. So where do you see progress or momentum outside of your industry, tech, AI, vibe coding applications that inspires you? This is very fresh on my mind, but I took a demo, a tour of a plane and I was so surprised that they got to full automation. This is a TBM 960, a turboprop plane. And everything's automated, including landing. And there's not one neural network in it.

1:13:19and so it's surprising because like you know in silicon valley we think that we sort of tapped out on classical deterministic systems in terms of what you can do in terms of automation but actually there are industries that are just slogging along they don't just don't want to use your networks there's like too probabilistic to so that was really fascinating and it's like there's there's a lot more to go even in the classical systems and the traditional technology that we have.

1:13:44Aria Finger:So always our final question. Can you leave us with a final thought on what you think is possible if everything goes humanity's way in the next 15 years? And what's the first step to get there? The thing I think a lot about, and this is driven by my history, my family, is that anyone who has ideas should potentially be wealthy. Anyone who has ideas and can follow through and has the grit and has the moral upstanding and can do the right things, I think that's the true promise of capitalism. And I think it's very easy to imagine the world where it's certain corporates that are controlling everything and the business models are not aligned with humanity in such a way that it's all based on spying on us and selling us the worst things that we desire.

1:14:33but instead in a world where the tools create the decentralized power, ability to generate wealth for everyone in the world, the entire world is sort of interconnected. I think people waking up in the morning and feeling like tap dancing to work, like Jeff Bezos, like most, 99 % of people don't tap dance to work, right? Like, so including me, like I love my job a lot, but I hate email. Like why hasn't anyone solved that, right? The technology is there, but the tools are not there just yet. So rewarding work, work that is more human, allows us to relate with one another more creatively and less siloing and less politics at work, but also the ability for anyone to generate ideas and build businesses.

1:15:24That's something I'd be really excited about. Amen. Yeah, Anjad, it's a great pleasure. We could have gone on for hours. Thanks for being here. Thank you for having me. Possible. is produced by Pallet Media. It's hosted by Ari Finger and me, Reid Hoffman. Our showrunner is Sean Young. Possible is produced by Tanasi Delos, Katie Sanders, Spencer Strasmoor, Imozu, Trent Barboza, and Tafadzwa Niemorundwe.

1:15:47Aria Finger:Special thanks to Surya Yalamanchili, Sayida Sepieva, Ian Alice, Greg Beato, Parth Patil, and Ben Rallis. And a big thanks to Ben Kaznoka and the Village Global team, Philip Smith and Action House, Lorenzo Davis and Will Whiteley at Static and Caitlin Norod and the team at Replit.