Replit CEO Amjad Masad: agents drove 500% metric gains, net dollar retention above 100% on consumer plan

it's Jack. It's Jack Ma with uh with an Well, speaking of someone as handsome as Jack Ma, we got I'm John Msad in the building. Welcome to the show. There he is. Crazy intro. How you doing? What's going on? How are you? Good. I just dunked my head in a a bowl of ice. We got the Saratoga here, too.

I hope you use Saratoga, man. I hope you got your banana on there, too. The banana is key. You You can't get You can't start your morning without banana on your face. I know. Just some banana all over, you know. Are you still lifting like crazy?

Last time I saw you, you were deadlifting like 500 lbs or something like that. I lost a lot of weight, you know? It's just like you get you get to a point where it's unhealthy. It's it's not exercise. It's it's it's killing you. It gets it gets a little risky.

The risk of injury gets up there, but uh you got to be able to, you know, break a sweat every once in a while. Do you see uh Daniel growing Daniel's tweet is like if you want to look like and feel like start powerlifting.

Yeah, because you just, oh well, if I want to get the number to go up, I better bulk more and then I'm just put eating so much. Yeah, exactly. I still maintain some of the muscle, but but you know, just trying to work out for longevity and health. Yeah. Anyway, uh thanks for joining.

Can you give us just like a brief overview of like where Replet is, what you're working on, how's it going? Yeah. So, uh, Replet kind of like the, uh, genesis of it, uh, has always been how do you make it so that, you know, programming is more accessible to more people. It's had a profound effect on on my life.

like uh you know coming from uh from Jordan uh growing up there you know lower middle-ass family being able to come to the US come to New York uh initially and then move to the Bay Area and be able to build a billion dollar company that's outrageous uh and it shouldn't be possible but uh the reality is that you know we have this ma massive wealth engine machine that's called the internet and for a long time you know hasn't really been uh possible for a lot of people to to kind of take part in that.

There's the I think Peter Teal said like the paradox of the internet. Uh it was supposed to sort of decentralize wealth yet somehow concentrated into into Silicon Valley. But I think what we're seeing now is finally the tools of creation are uh becoming more accessible and that's really been my my mission.

And so initially reput was about let's get people learning how to code and um I think that was a good pursuit but people don't want to code. The reality is I spent months uh whatever working at code academy and replet. I mean it's good.

I think we taught a lot of people how to code and at replet we built a ton of infrastructure around how do you set set up a virtual machine the runtime install packages automating all of that and all the way to deployment. But the really the big unlock was the agents. You know putting an agent on top of all of that.

We were the first agent in the market to be able to configure uh services, databases, deployment and go all the way to uh to deploying uh deploying an app and it kind of transformed our business just in the past seven months.

Can you talk a little bit about the the benefit of knowing how to code and how that collides with vibe coding? I see a lot of people that are like, "Oh, I vibe coded and it didn't really work. " And then Andre Karpath is like, I vibe I vibe coded it and it worked great.

And I'm like, there's probably a little bit of a difference between when Andre Karpathy is vibe coding and when someone who's completely non-technical is vibe coding. Like is there still value to learning to code in 2025?

I I I I think there I think there is, but I mean it depends, John, about what is your prediction about where AI is headed, right?

If you know in the up case like you know what Dario just said recently all code will be I generated you know I assume this optimization path we're we're on where agents are going to get better and better and better you know the answer would be different the answer would be no it would be a waste of time to to learn learn how to code um but like you know you could have different predictions and I think different people will will make different assumptions I I I'm at this point like sort agents pled.

You know, I I the the progress we're seeing with agents, we just we just are rolling out V2 agent. I've never seen a sort of like an AB test as we're rolling out. I've never seen metrics in my entire life. I worked at Facebook, other places.

We're moving metrics like crazy just by improving the agent like 100% 500% on certain metrics on on success that users are having.

And so uh I think where vibe coding goes wrong is when a a beginner is using cursor some other platform great product but they don't give them guardrails like you know that there's the guy that lost six months of work on on Reddit there's the guy on Twitter that uh leaked his API keys like onlet we do get commit for you don't have to worry about that we hide the API keys in like a secure vault and so I think a lot of it goes to like hey these beginners should not be using a lot of these tools and instead they should come to places uh like like replet uh but that being said I'm I'm very bullish like you know I sort of changed my answer even like a year ago I would say kind of learn a bit of coding I would say learn how to think learn how to break down problems right learn how to uh communicate clearly uh you know with you know as you would with humans but also with machines.

Can you tell me some stories about like success stories from Replet maybe like the the what's possible as a solo dev today? I mean we're seeing Peter Level's io do crazy stuff, but what are you seeing from the from the most uh successful Replet users?

Let me start with the with the business cases because I think those are are kind of really profound because we've always had like micro entrepreneurs like make money and you know it's certainly accelerated massively but 100 plus year old company uh Sears um I don't know about you I didn't know they they continue to exist but they they have this um uh spin-off called Sears Home Services.

About a year ago, they hired a team to like a more trendy uh sort of uh tech team to get them off of Cobalt, which is whatever 100-y old programming language. And they were able to do it very quickly.

And then after they done that, you know, basically most most companies as they're modernizing, they go adopt a bunch of SAS tools and no code tools and low code tools and to become like a modern company.

They actually skipped all of that and went to went to AI and their operations team which manages the kind the kind of field workers that go out and every day kind of do house repairs and all that. Uh they start they started building AI tools.

So one AI tool that they have is like the worker would would go and would talk to the AI and it would optimize it their route and would you know that made it so that they're making a lot more money and now that team everyone in that team started building AI tools and so you have this ops team that's highly levered that are able to move business metrics uh and and they leaprogged an entire era of computing and and landed on on AI agents and how they these are the generalists that are building the business and I think this is a view into into the future where firms would not be made of like you know SDR and like account whatever it's going to be like I'm a sales generalist right I can do all that stuff because I can spin up AI agents that can do all these things that makes sense how how do you think about um balancing the sort of needs and demands of the different types of people uh and customers of the of your platform, right?

You have somebody that comes on uh they're sort of early in their sort of like entrepreneurial journey. They build something, but then suddenly there's this moment where they become a business and once you become a business and you're sort of like servicing.

So like you have the challenge of having so many people on your platform and they and some of them can end up wanting different things. Uh what what is Yeah.

just what does it look like in terms of you know you want to be able to continue to service these um users over time while still making sure that replet is the best place for somebody to sort of like start their journey we we've taken u a a a difficult approach approach and perhaps not not the smartest right the reason I've been working on it for my entire life is because uh you know what I try to do is like um build a an end to-end experience where we infuse a lot of taste and decisions and choices on behalf of customers.

You actually you had that in the past where like the Apple ecosystem kind of makes a lot of decisions for developers and even Windows as much of as we we uh you know look down on it but you know Visual Basic was a great product.

It made a ton of decisions for you and like you know a lot more people were making things at the time.

And so then you had the open source movement and the open source was was this radical disruptive you know bottoms up thing where uh there's 100 competing libraries and 100 computing ways of doing things and that created a lot of mess and um created a lot of power but now getting started and doing things programming became a lot more difficult like imagine setting up a modern web uh you know app, let alone trying to set up like a Python environment.

Like if you mess up your Python environment, the best thing you could do is throw your laptop in the garbage and get a new one. I've been there. It's miserable. It's terrible. So, you know, so so you had this, you know, bottoms up innovation kind of like the early hacker lingo. You had the cathedral versus the bizarre.

So the Windows ecosystem is the cathedral. The open source ecosystem is the bazaar. And so in my mind, well build let's build cathedrals ma made of bizaars. Like I think open source is awesome, but let's actually build tasteful experiences on top of that. So uh my work at Facebook, I was part of the early React team.

Uh you know, I was working on the tooling there and especi especially like React Native, we made it so that in five minutes you can spin up a hello world app as if you're trying to do it using Apple and iOS. that would take you, you know, hours perhaps.

So, you know, now fast forward uh to to Replet and we built basically that you all the way from nuts and bolts to the UX of what it means to to make software. And so that's been uh a lot harder, but but I think you end up with this general purpose horizontal tool that can satisfy uh a lot of needs.

Um the challenge of course is like positioning and marketing this thing. You guys know as entrepreneurs like people like to do to pick one product on the market.

But look back on the history of computing typically the the the best products that end up being a core part of our lives are the general purpose horizontal products.

Uh there's kind of a narrative that's floating out right out there right now that companies that are aligned with the AI growth trend are seeing extremely fast rates of adoption. You're seeing companies get to 100 million ARR faster than ever before.

The risk that people are kind of worried about is that maybe churn rates are going up as customers test a new AI tool and then kind of bounce around. They're using cursor one week and then claude code the next week and then they're bouncing to this one.

um what are you seeing and what do you think's happening and do you think there's any uh any merit to that fear in the market or on X right now that maybe the AI revenues that we're seeing are not as sticky as previous SAS installation cycles is certainly suspicious uh I mean it anyone who tells you differently is sort of and I was a little worried about our growth like is it is it going going to go away.

What's going on? Like this is like a little too fast. Um and I think uh you know we really start focusing on the on the metrics that really matter which is the revenue quality retention things that more advanced companies are thinking about like net dollar retention.

we actually have above 100% net dollar retention even for our consumer plan which is probably unheard of but um uh you know I I think that uh I think maybe ARR is becoming a bit of a vanity metric like how fast you can grow the ARR I think you want to make sure you're buying an awesome service and also make sure that people just want to stay on your platform I think especially on the VS Code Wars the switching cost being so low is is gonna be a challenge for them.

You can go from copilot to wsurf to cursor in like five minutes, same project. Yeah. Quite like it's one click. You got to open the project. And so they're going to have to find a way to differentiate. The great thing about Replet, we provide this really great convenience. And so if ain't broke, don't fix it.

you deploy something on replet and people people really tend to stay. That makes sense. Uh internally I'm I'm sure the entire team is using replet all you know all day long.

Uh h is there still a place for traditional prototypes or are you is everyone across the team sort of like doing rapid prototyping of features which is obviously something that your end customers use or is there still like hey let's just you know make this in in Figma and kind of a line around it uh first.

There's certainly uh a place for for Figma.

Um I think there are uh first of all there are things that are tough to prototype and there are things that you want to do very very quickly even even faster than than than sort of any kind of software prototype and it depends on how people think like I think designers still would want to would want a canvas to to think I think products might end up growing in that direction where maybe you go to rep you initially have some kind of canvas and that kind of results in a in a piece of in an application.

But right now I I still think there's a there's a place for Figma.

But the way I think about what Replet is doing to to our organization, I hope more and more companies is like everyone is highly levered like um we just built you know agents are new and part of the exciting things is that no one knows how to build agents.

Yeah, there's no tracing tool, you know, like an APM, like a New Relic or data dog for agents. So, we built our own internally and, you know, a a couple uh people on our team spent like a week using Replet and VIP coding uh this really a great infrastructure tool that has saved many many hours of engineering.

We have even more exciting story. or we have someone in HR that couldn't find an org chart tool that really fit what she wanted to do or it was too expensive. She spent three days she spun up this org chart tool with like uh connection to ADP and and and sort of versioning history and it's amazing.

She never coded, right? So I think the way to think about it is like um how to increase the productivity in your company and less about like oh what should we displace you know I think I think these these tools still have their place. Yeah.

How do you um how do you think about sort of buying software as someone that enables the creation of software when you know you've obviously been building replet for a very long time.

10 years from now you guys are in market for a new CRM you're talking to the big players is is building it yourself just become uh a part of that sort of like RFP process where the team is like okay well we can build this like this is you know what we think is going to it's going to cost to you know basically build it maintain it used the biggest bare signal if a startup was building their own CRM it was like oh no you're way over your skis this is going to get terribly ly wrong and you're going to be maintaining this thing for years and you actually don't have a special you're not a special snowflake in this case just use Salesforce or something but as an like NIH not invented here syndrome Yep totally so um yeah I agree I so I think I I I thought quite a bit like the future of SAS I think that the the question is like uh if you can conjure up a SAS product can like is there even a future there and we do see this every time I wake up on on Twitter, I see someone using Replet and replacing a piece of software that costs them 15,000 100,000.

I just retweeted one this morning. Um and uh so there's a lot of I think vertical niche specific SAS tools that are generatable very easily even with the today's technology let alone like a year from now.

I think you'll be able to kind of generate them especially on a platform like Replet which gives you all the cloud services involved. I think there are ecosystem SAS products that are fairly hard. So I think think about Ripink how embedded into your company as like the u what do they call the the like system of record.

Yeah. Yeah. And and Salesforce think about the ecosystem and the developers that they have. Yep. So I think if you're starting a SAS business today, you want to go against the Silicon Valley dogma of focus. You want to build a Chinese style company, which is I think what Rippling is doing.

And I think uh I think this idea of like uh being generative and m you know having larger pieces of software I think is going to become more important in the future. So I think I'd be really worried about vertical specific SAS but I think I would bet on those horizontal SAS companies. Yeah.

Well, we got to ask about the Verscel drama. We talked to GMO earlier this show. What's going on there? What's your take? Can you break it down for us? Well, you know, I uh someone asked me um do you do you plan on supporting Nexus and and Replet agent?

I said no because uh we want to bet on something like Vit that is open like community built uh and the way Nexjs has been trending is that it's optimized for Versell they disagree but it is actually not a controversial statement in the larger community there's a project called open next that makes it so that you know uh nextjs is deployable on more platform Matt the CEO of Netlifi commented uh agreeing with me.

The larger community feels like, you know, Nex. js is is perhaps more tight of herself than they would like to admit. And I think they overreacted to my comments and and and they and they striened my my comments. I think it would have been a le less of a a story if they, you know, didn't overreact to it.

And and look uh uh Burcell is an amazing product. It's an amazing brand. They've done a great job managing a lot of open source projects, not just Next. js.

And for me, water under bridge like you know I uh you know I think you know emotions get get high on on Twitter and you start name calling and you know I wish them really the best. Yeah. H how is how is Next different than the React relationship with Facebook?

I mean with all these big main core frameworks you're kind of if you're in Angular you're aligned with Google. If you're with React, you're aligned with Facebook, but uh has has Facebook or Meta done a better job of kind of releasing that into the community. Well, they're not trying to monetize it as aggressively.

Well, not at all. Not not monetize. Look, I worked I exclusively worked on open source uh at Meta and the the advantage and disadvantage of working in open source at Meta is that there's no direct tie-in to to to the bottom line, right? So, uh we were not trying to monetize uh React.

Um so far they haven't monetized Llama in terms of like selling an API. Sure. Um and so generally as long as they are um sort of they're using it internally they're going to maintain it and support it and get some value out of the community and mind share support.

Now the disadvantage of that we actually weren't ever able to like make a complete product because like if you go to Google uh and you know use now flutter I think that thing is actually much more polished of an experience because they have a direct uh incentive to advance the Android ecosystem for example.

Um and so Facebook we built really cool pieces of infrastructure but the UX around them were bad. So you have companies like Expo building on React Native to make it a better experience. Companies like Versell building React to make it a better experience.

And these are amazing companies and I think that's how it should be. I think where it gets tricky is when you end up also maintaining the the core core software and really tying them together.

Um, what do you think about kind of the Ben Thompson aggregation theory around there might be a new company to emerge from the AI boom that is the hub for these vibecoded apps and websites and games?

Obviously, it for a long time it's felt like web development services, if you could call it that, were very much it was bifurcated. like there was Shopify which is a power law winner but then there was also Squarespace and Weebly and like dozens of other companies that did quite well.

Um but if there's a new aggregator emerging and Nikita Beer's written about this like maybe there will be one company to really break out and be the the hub for vibecoded N64 level games that would be very valuable. It would also be more of a winner take all environment.

How do you think that that plays out in your category? Look, I I don't really that's not the model for for for our business. Like the the way I think about it in the early computing era, you had the mainframe.

It was the domain of the expert and then you had the PC and it became possible for anyone to use computers and that ended up eating our economy essentially because that's how disruptive technology work.

And so in you know the modern era the modern software engineering is also the domain of the expert and now it's become so that anyone can can make software.

Initially it looks like a toy vip coding game games and things like that but I think it will fundamentally change and transform the um uh the economy and and my model for what we're building it's let's build and excel. It should be fun. You know, we do a lot of things uh that you know make it fun.

We actually build a 3D gaming stack, but but the model is like can a billion people use it to um you part of their their everyday work uh to to build businesses to to really be a core piece of the infrastructure in the economy. And the idea that you know it's it's going to be Roblox I don't see that as as interesting.

I mean, Roblox is a great company. You know, maybe Roblox is the aggregator for for uh vibe coded games, but I think it's really not the the the big prize. It's like you want to build a trillion dollar company, you got to think a little bigger than that. I love it. Uh yeah.

Do you do you have any do you have any words for your doubters? Something you said is like the next big thing looks like a toy. It's been this idea within Silicon Valley that you can apply to you could have applied it to crypto. This was Chris Dixon's, I think, original thesis.

It still feels like, you know, Riplet is this billiondollar company now, but I still think your doubters and your haters would say, "Oh, it's just a Oh, it's just a toy. Like, it's, you know, kids making apps or whatever. " But, uh, do you do you like that that people still kind of underestimate this potential?

Is that part of what creates the sort of uh the the opportunity for the next 10 years is people not realizing like the tsunami that's kind of about to hit uh the industry. Yeah.

Uh I think I it used to bother me but but now I think it's it's actually an advantage and I think um you know um when you know when we're at the height of like AI trendiness and replet obviously was early you know we were the first outside of core pilot to train a code model we're like this company in the city and we're growing fast and you know hired a bunch of executives that I think are part of the the sort of the hype kind of Silicon Valley.

they jump from one company to another and um you know last year you know I did the founder mode I uh cut the team in half wow now like 65 people I moved us out of the city we're now in in Foster city literally there's no other sort of there's actually nothing out there sort of a desert um and and so we have this nice sort of uh kind of office that we're turning into a campus we have we have food we have all that and um and and the team here is sort of a tight-knit uh team, there's a selection effect on who joins Replet.

If you are smart enough to figure out that Replet has this amazing potential in the future, uh then we would want you on the team. If you think uh if you really think it's a it's a toy, it's not worth uh it's not worth your time, then then great. And so I I think these things that just sort themselves out. Awesome.

Well, thanks for joining. Love having you. We'll have to have you back soon. Yeah, it's I'm I'm incredibly excited to watch what you guys do for for the next year, the next two, the next five, the next 10. Uh and uh yeah, I would love to have you on again in the future sometime soon. Yeah, this is fantastic.

Thank you for for doing this show. It's it's been it's been really cool. Thanks so much for joining the man. Cheers. Awesome. Do we uh do we have another guest? We have a surprise guest for surprise guest, folks. Is he ready? Uh I just sent him the link. He's jumping in in just a minute.

Um, and uh, he should be all queued up. We'll we'll we'll go through some more posts while we wait. Um, oh, Bryce Roberts bought two cars. We got to talk to him. We got to have him on the show, figure out what he what he bought. Bryce Roberts, uh, over at INDVC says, "Bought two cars today.

One dealership wanted me to come down to the lot, get on a bunch of calls, and refused to send me basic info over text. The other was willing to send me anything I needed and requested o requested over text and email. Can you guess which one got the sale? And it's not a Tesla.

He went he he said he's done being a Tesla customer. Not a political statement, just that his lease is up in his Model X was absolute trash according to him. I wonder what he did to the car, but uh we'll have to talk to him about what car he chose. I'm hoping it's a Dodge Demon and a Lamborghini LM2.

I think that would be a perfect fit twocar garage for Bryce, but we'll have to talk to him about it. But it really is a an interesting take. I We talked to the car dealership guy, and I've wanted a car dealership that Yeah.

I mean, sometimes I just want to buy a car and just want to click a button and I don't really care. Jordy likes to go to the dealership and take a look at everything. I completely disagree. I want to just push a button and have it arrive. Uh, which is maybe stupid, but I don't know. I I certainly enjoy it.

And let's see if I'm ready when you are. So, we are going back into the 23 andMe news and we'll give some more context on this, but I like that we're able to be responsive. It's actually crazy that Keon singlehandedly bankrupt.

Yeah, he was barely in market for Well, I mean, it doesn't it's not a surprise because he did the TBPN interview and I'm sure I'm sure the market saw it very well and said, "Hey, you know, if he's going to be in the temple of technology every week, let's just get out of the