Lovable CEO Anton Osika：基础模型现状、Grok 对决 OpenAI，以及 Replit 对决 Bolt

I think university is not the best place to learn. It doesn't matter what you're studying. I'd invest in Grock, and I would short Anthropic.

Why?

I think it's more the slope on the Grock team. They're doing something which I respect a lot, which is hiring missionaries for the data curation part. The morale is super high. OpenAI has gone through all this mess.

Right. Do you think there will be a leading model that has not been created yet?

Yes. From China.

Do you worry about China?

I do think there's a 50/50 chance they will have the best model. We'll be using a Chinese model at some point because it's ready to go.

Anton, dude, I'm so excited to be here with you in person. Thank you so much for joining me on the show.

It's great to see you, and thanks for coming to Stockholm.

1. Is AI an Arms Race… Or Just a Talent War?

Dude, it's great to be in Stockholm. I want to start with the great round you recently raised. We're seeing a lot of money going into the space, and I wanted to start with this: Is it a capital arms race, a case of whoever has the most money wins, or is it something else?

I think it's an arms race to build the best team, and then it's an arms race to build the best brand and trust from your users. Capital can help. For us, it's not a constraint at all.

If you're building something like the best foundation model, it might be a constraint, just because the compute for training and so on is so large. But for us, it's all about moving extremely fast and collecting the best talent.

2. How Does Anton Compete with Zuck’s $100M Packages for Talent

So, if we think about talent as the number one, we've seen Zuck pay NFL-style contracts—mega sums—for the best people. How do you think about and analyze that, and how difficult will it be to get the best talent moving forward?

For me, it's actually more difficult than for Zuck to know which engineers are going to really thrive, push the culture forward, and push the ways of working in the products forward.

For Zuck, there are these 10 people who know everything about how to train foundation models, and he's paying more for that knowledge than for these people. The talent itself is so good. It's probably pretty good as well. So it's very different.

Do you think you don't need the same caliber of engineering talent if you're working in the application layer?

It's just very different. I think one of those people Zuck is hiring wouldn't perform as well as the engineers on my team doing what we're doing. It's a very different type of talent.

If I knew who the perfect engineers to hire were, I could maybe step up our compensation bands to get exactly those people. But I don't know who the best people are. I need to figure out: Are these really, really good people to work with? Are they moldable? Are they going to work well together on this team?

And then give them top-of-market compensation rates for that.

You've built an incredible team, also, of less obvious talent in the early days, and then you hired amazing rock stars like Elena Verer. When you look at your hiring process and your talent-assessment process, is there anything that's nonobvious?

For us, I look for people who have either extreme trauma or extreme masochism.

Yes.

And being serious, I think not enough people are opinionated. You said brand is important. Great brands are opinionated. People love them or hate them. Lovable—good example. What's nonobvious about hiring or talent assessment for you?

I like to think a lot about slope. If I talk to someone and learn a lot of things from talking to them, and I notice that the conversation is very dynamic and exciting, that usually feels like a very good indicator that they're going to adapt to the organization and that their slope will be very high.

Otherwise, I think there are good ways to understand how they performed in the past. If I could be there with a video camera when they worked in the past, that would give me a lot of signals. So that's usually what I spend a lot of time on when I'm talking to new candidates.

When you think about slope, it's noticeable with you. We haven't known each other for a huge amount of time, but when I compare when I first met you to when I met you today, it is still very different in terms of your leadership. Where have you not progressed in the way you would still like to?

I think I still operate Lovable in a very scrappy, startup-y way, even though we're at the later growth stage right now. Adding a bit more structure in a few key areas is somewhere I'm looking to progress, or to start being an excellent operator.

3. Founder Mode vs. Structure: Can Chaos Scale?

And the joys of this show—and of being friends—is that we can have a discussion, not just a one-back-and-forth interview. Do you think you actually need that? We've had founder mode so propagated and praised, and being close to the metal—Jensen having 52 direct reports. I would say structure and that middle layer are where slowness and apathy come.

True. Yeah.

Do you think you need that?

That's a good question. I think I'm always going to operate with most of my impact coming from founder mode. But given that there are so many things thrown at me and coming in from all the different directions, I do need a protective layer that introduces a lot of order in how we prioritize all these incoming things.

That comes down to a well-running organization. For a well-running organization, you need a very organized manager somewhere at the top. I'm not planning to be that manager myself, but to surround myself with great leaders who do more of the organization.

Do you have a protective layer today? Because I get probably 25 intro requests for you a week, and I probably make one a month. Do you have someone who does filter?

Yeah, I do. It's a wonderful, chaotic protective layer that works together as a close team. I don't really have a name for it. It's just the people working closely with me.

I'm not sure who you've interacted with on the team, but the team is made up of previous-founder-type generalists who work closely with me. I just work in terms of quick feedback: “This is not what we should be doing. This is what we should be doing.” It works okay now. I think we can do even better.

So you said talent was number one and brand was number two. If we think about a great brand, what does a great brand mean to you?

I think a super-concrete example is the Apple ecosystem, where they obsess about details, maybe too much, so they move slowly. But that's what builds up trust and a very strong brand.

That's what we're aiming for as well in every interaction. Every time we update the product, how do we make sure we roll it out so that we really understand the users and their reactions to all the things we're changing very rapidly in the product and the company?

4. The Brutal Truth About Defensibility in AI Startups

There are a couple of questions which everyone has, where they will throw them as a critique at Lovable or at anyone in the space. I think one is protection—defensibility.

When you think about defensibility today, is brand the core element of defensibility, or is there something that people do not see?

I think you need to build a product if you want to be maximally defensive, where, if you are on this product and the platform that product is on, you don't want to leave because you have so much value that you've created on the platform that you're getting automatically every day.

I think that's what Lovable is becoming: this product-building platform where Lovable today is your technical co-founder. We want it to be your co-founder in general, handling all the admin and setting up your finance operations. If you're on a platform like that, you probably don't want to leave.

Would you say to all founders building an AI startup from day one: Don't worry about defensibility. It comes over time?

Yes. I have a friend who has this fun analogy for an AI startup, which is that AI startups are like chickens shot out of a cannon up in the sky. If you start getting traction, then it's all about flapping fast as a chicken, because there are new chickens shot out of cannons every day. If you keep flapping faster than the other chickens, then you're going to do great.

I think that's a good first level of analysis for how you should operate.

I'm just going to say, for any vegans who are listening, no chickens were shot out of cannons. That is the most extremely Swedish way of doing it. Reid Hoffman says it's about running off the cliff or whatever with a power glider and just flapping. That works, too.

5. What Does Make the Marriage Successful?

Yeah, I think that's my recommendation: Just execute fast and grow faster. Then, when you're starting to get up there, you can start maybe thinking a bit about defensibility.

6. Unit Economics: Are AI Companies Doomed to Bleed Cash?

Totally get you. That's the one criticism. Another is, when you look at these businesses—and a lot of people are criticizing this with your Replits, your Bolts, your Lovables—they're not actually very good businesses in terms of unit economics, and so much is passed through.

Bluntly, if I give you a dollar, how much is passed straight through to Anthropic and OpenAI?

I can't give you the exact numbers, but if you look at the paid usage, it's the majority. It's not everything.

Okay. How does that change over time?

As our business develops, we're looking to get most of our revenue once you, as a user, are saying, “I love this platform. I'm never leaving.” But today, in the beginning, you're paying to build, pretty much.

Over time, we just want to create so much value that you stay on the subscription, and a small part of the cost goes to their AI compute.

Will you be able to make money through not optimizing models? What I mean by that is, in the future, you may not need the very best, very latest model to do the simple About Me website. And so you can route users.

7. Quick-Fire Round

Yeah. I think as all applications develop, the AI is going to be adapted to those applications. For most things, it's super simple to do. It's like you're driving a car and you don't think about what you're doing. When you're in a new situation driving a car, then your brain really goes on fire.

We're not there. We're not close to being there yet. I think for us it's too early to optimize for that because the AI, every month, is doing new, completely different things. We just want to be able to iterate really fast on what the AI is able to do and not optimize the models for what it's doing.

That's really interesting. So you build for what tomorrow's model can do, not what we have today.

Yeah, to quite a large extent, yes. Generally, when I think about models, there are models that are very thoughtful and deep-thinking, and now we put as much of our work on those models. In the future, it's going to be a mix. When it's obvious what you should do, then it doesn't cost any money. It's super fast.

But when it's a new situation, which building a software product often results in, then it has to think much more.

One other area where you can see real margin expansion is also in token selling. When you think about how you price tokens, given prosumers and consumers don't fundamentally often know the price of tokens, you can actually have quite a considerable markup on token usage. Do you think that is a place of real elasticity to gain margin or not?

Yeah. So we looked at the numbers. This was a few months ago, but we looked at how much revenue is flowing through the AI from Lovable applications.

Okay.

More than $10 million in AR was flowing through the AI. All of that revenue needs the user to go through this somewhat complex process of setting up the connection to the model providers. So that's something we're just simplifying. We're looking to simplify it, so stay tuned for how we enable more simplicity, first of all, for our users with that. And then, if we can reduce the underlying cost, maybe we can take a margin there as well.

How do you think about mental plasticity to delay margin optimization?

Mental plasticity—what?

The willingness to wait for margins to come. What I mean by that is, if you look at Deliveroo, its margins were terrible in the early days, and over time they get better and better as you have more and more people use it and more density, more orders in small areas. You've got to be patient, so to speak. Same with OpenAI, same with Lovable.

How long does one wait before thinking about margin optimization?

I have these 2 conflicting perspectives on it. One is I speak to Nick, who built Revolut, and he just tells me, “Anton, you need to compute the payback time, and then you need to do super-hard performance optimization on acquiring new users.” You, of course, need to have good payback times in terms of profits per user, which makes sense.

If you can do small changes in margins, it actually affects a lot how fast you can grow. But the other perspective, which I index a bit more on right now, is you just want to have as much mind share and as many users who just love the brand as possible right now, and then you can think about that later.

So exactly how I trade those 2 perspectives off is—I mean, you need to look at the weights in my neural network—but it's some combination of the 2.

It's so funny. I think it's absolutely number 2, and then it's number 3, which is Nick's incredible performance optimization: really understanding funnel metrics from CAC to LTV and how you drive efficiency through the channels.

8. The Security Bombshells No One Talks About

Do you know what's ironic? You kind of then go back after that stage to an art, which is where Nik is now: we've done that so well, we have to sponsor race cars because brand again becomes the most important thing. It's so interesting, that kind of bell curve where you see it go up and then come down again. It's kind of like that with brand, where you have it at both ends of the spectrum.

Yeah. I think if you can do everything at once, the company benefits a lot, but you usually can't. You should be focused.

What would you most like to do now that you're not doing or can't do?

I would like to rethink how applications are built. What is the best way to build an application? Right now, what Lovable does is take all the best practices from decades of how great software products were built, but that's not how the future is going to look.

All software applications are going to have some type of AI. They're going to have extremely seamless payment and checkout flows. That's something I'd love for us to spend time on figuring out and making possible for our users—not to just have a superhuman AI engineer, but to have an AI engineer that builds the future of applications.

When we look at the model providers—we mentioned margin optimization there—you also have to have that mental plasticity. We saw OpenAI kind of suggest or proffer Lovable-style competitors. To what extent do you feel OpenAI and Anthropic will come after Lovable in the way that Claude Code comes after Cursor?

I think in the long term it just comes down to execution of a team. Many people are going to offer what we're offering today. We just need to offer much more when that time comes and give a better user experience, give a better value proposition to our customers.

Who do you worry more about: OpenAI doing it or Anthropic doing it?

9. GPT-5: Game-Changer or Overhyped Disappointment?

I think what we're betting on is to be the gateway for humans, right, and be the best user experience for AI. So far, OpenAI is doing that better than Anthropic. I see them as a more serious competitor in 12 months.

How did you analyze GPT-5, and when you look at performance post-launch, are you more or less bullish on OpenAI?

We looked a lot at how—sorry—we looked at a lot of how GPT-5 would impact our users before we decided, “Okay, let's put this into the product.” We looked at how long it took to get responses. We looked at our qualitative and quantitative evals, and then we just vibe-checked it in many different ways. What we concluded was that it's oftentimes too ambitious for our users.

That's why we decided, “Hey, this is very smart, so let's give it to all our users and see what they tell us in terms of what's good and what's bad.” What we found was that for the use cases when you have to solve a really, really hard problem, it's great.

In terms of whether OpenAI is doing a great job, I think this was a really smart, obvious choice for them to say, “We have these 5 different models that you have to select in ChatGPT. Let's just bring it down into 1 model, GPT-5.” They definitely should have done that, but it comes with a lot of trade-offs. So far, I'd say they executed pretty well on it. The model is still too ambitious.

There's also a question of when you set the bar at AGI, and then you get model optimization and model routing, which is really essentially what it is. Capability-wise, it hasn't been a step-function improvement over what we had before.

No, I don't think the biggest part of GPT-5 that's disappointing is that now they have to optimize all these different things into 1 model. Before, it was different models, and they had to do it really fast. So it's inevitably going to fall short in some dimensions. I mean, it just is a disappointment that you can't improve in all the directions at the same time.

How do you use OpenAI versus Anthropic within Lovable today?

We have this very complex agentic chain where we pass the user's response and the application information through many different models. We take really fast and small ones, and then for code writing we usually use Anthropic. Right now, you can say, “I want to use GPT-5,” and that's better when you're solving a really hard debugging problem.

Super. And you've seen it be better than Anthropic when it comes to a hard debugging problem?

Yeah.

What do models not do today that would be a step-function change in what Lovable can do?

Something I'm super excited about is that the AI has more context about who it's talking to and how it should be answering, to guide it through our specific application. Solving that problem is something that we have to do.

We have to do it both with how we build this agentic chain and, over time, by building an absolutely world-class team and paying $100 million to get the people who train the models. That's on the horizon for us: to get it to be hyperpersonalized for you specifically.

10. How Lovable Hit $100M ARR in Just 7 Months?

When you think about hyperpersonalization for you specifically and the users that you have, you recently announced $100 million—an amazing milestone to hit in 7 months. For years, dude, it was like $0 to $10 million in 2 years was the gold standard. That was fucking nuts.

That's what I was brought up on, which makes me feel really old. It's amazing to see. My question to you is: when you look at the revenue breakdown of the $100 million, just guesstimate—what is the split between hobbyists, pro devs, and kind of normal people? How does it fit between the different segments?

You're right. People do everything with Lovable. They come with their idea to build a software business and product, and then there are a lot of people in large companies that use it as, “Okay, now I can show what I actually think we should build in the business.” Then they build a working product, and then they can decide, “Are we going to give this to our engineering team?” and they actually implement it.

Then there's everyone else who builds their personal website or their small business website in a few minutes. 80% of people are in the first category. They're building really complex applications. In terms of revenue, 80%.

Wow.

Yeah. But the second segment is actually growing very fast because enterprises are slower to wake up. You might have seen this product leader from Google who says, “We're never again writing a document about a product. We have to use Lovable or something to build out a fully working demo.” So that use case is also growing very fast.

In terms of the third use case, a lot of people have been burned trying to build nice websites in no-code website builders—Wix, Squarespace, and so on. If you can just always do everything in Lovable, with a UX that I think is more sophisticated while moving fast, that's also growing. But I think the first 2 are the ones who are really game changers.

Okay, so let's go back. 80%, sorry, is actually building complex apps?

And then 10% is enterprise and 10% is hobbyists. Something like that, yeah.

Is that what you want it to be?

We want to build for the new generation of AI-native founders that build maybe 1-person unicorns soon. The funny thing is that those people also have jobs, maybe in large, successful companies, and they want to help their friends and family build simple websites. So I think this is a good split.

Is that an optimal market to go after if you're thinking, “God, I sound like such a VC,” but value extraction—which is an AI founder building a mega-business on Lovable—great, you, 1, have to have a lot of mechanisms for value extraction, be it payment solutions or you name it. But if they're single-seat and it's just tough to get true value extraction from that, is it not much better to be a hobbyist for everyone—for mom and pop to build the about-me website—where it's 7 billion people?

Our mission is to enable a lot of people who have the opportunity to build businesses, but have been held back by not being able to write code and not having access to capital to hire engineers. It's obvious to start with the people who are going to build businesses, and then it naturally trickles down to everyone else as a function of that. Those, I think, are the best people to start building for, and where you can extract value, I think less about that. I think about our mission.

We should have a Lovable holiday fund, which is, like, every year—

We pay the most talented people within large enterprises for a 1-week holiday.

Then they build their business.

And then they just use Lovable for the week, and after the week—

They quit their job.

I think it'd be the funnest thing ever.

Sounds good.

Yeah. But I can expand a bit on the business thinking there as well. The future, I think, is that many of the largest businesses haven't been created yet. Now, with AI, you can move much faster, get much closer to your customers, drive prices down, and we want to be the enabling tool for that movement.

I think, over time, that's going to result in a lot of revenue. But if we can do all of these use cases that you're asking about at the same time, what our use-case percentage of adoption or revenue is going to converge towards is just: what's the spend from enterprises? What's the endgame spend from enterprises? What's the endgame spend for consumers on tools like this if we continue to dominate this completely new category?

Obviously, I'm super grateful to you for taking my money, but the number-one reason why I would invest in Lovable is because the market TAM, or the TAM expansion, is actually incomprehensible. It's very much like Uber: you could never have foreseen the market expansion that would take place. Very much like Lovable, saying “website builders” is an existing market is completely the wrong analogy to understand how big Lovable could be moving forward, and that's a common thing in the best venture investments ever made.

Yeah, it's really interesting. I also just want to say on the enterprise use case, if I was the CEO or the CTO of a large enterprise company, I wouldn't think in terms of, “How can we make our engineers more productive?” I would think in terms of, “How can we get the most information about what we should build as quickly as possible into new products or into existing products?”

That requires everyone in the company to be able to work in 1 place, to change and edit their products, and propose new changes to them. It's hard for us, or for anyone in that matter, to build a product that does that tomorrow. It's better for us to start with the founders who are building it from the ground up and then move the same experience into the enterprise.

It also allows for this incredible democratization of ideas within companies. I interviewed the CPO at Duolingo for 20VC, and he actually said that 2 designers—not traditionally people who come up with and build products from day 1—created chess in Duolingo. They did it with—I can't remember what the tool was. I hope it was Lovable. Please say it was Lovable. That was actually their first iteration of it, which I thought was an amazing instantiation of this.

So I totally get you there. Does that mean, then, that we lose the design process and the brainstorming process? Do we skip that and go straight to prototyping?

Yeah. Look, to date, what you've done is that you've taken an idea, and then you've gone through many, many, many steps until it's a fast-growing product. All of that I call the product life cycle. One part of it is writing the code, which is where AI has now made it much faster.

There are many steps after that. There are steps before that, which are to mock it up, validate it internally, and validate it with your users. What we've done so far is take all the first steps until, “This is validated. This is what we need to ship,” and even have external users on it, and put that into a few minutes or a few hours of building.

That's where we've seen the most maturity in our product. The steps that come after are something we have to build out as fast as possible so that you don't need a product, design, and engineering organization. It's 1 all-in-one tool where anyone with the best ideas can spend the most time.

The steps after or the steps before? What I mean by the steps before is, at the other end of the spectrum, Figma doing Figma Make. They're like, “Hey, we'll always have the design process, and then we'll move with you into the second phase of that product life cycle, into the build or prototype phase—from design to prototype.”

To what extent do you worry about entry from that earlier standpoint?

Entry from Figma?

Yeah, entry from Figma, given the fact that they own the design part of the phase and can then move into prototyping.

I think humans are sometimes too obsessed with small details being perfect, which makes you move much slower. The way of doing it right now, with the design—where 1 person does all the design very slowly and in great detail—is going to be replaced by AI doing it.

You talk much more high-level, and you talk about your design philosophy, and then the AI does the implementation of the design. Then you, as a human, go out and get all the context from the other people—“Do you think this looks good?”—and give it all as feedback into the AI.

Then there's a very seamless, opinionated way of taking it all the way to a product, with all the marketing and growth functions built in, with AI behind it, as well as all the tooling you need to evolve a high-quality software product, which comes with testing, quality assurance, and so on. That's a new way of doing it.

Thinking you should be doing design with Make and the Figma tool—for me, my thesis is that for most people, it will slow you down too much. Does that make sense?

It does. What happens to Figma then?

I think for some pixel-perfect things, it's going to be amazing to continue to use Figma. I don't know how the distribution will look in terms of doing it in a more opinionated way—which is what our platform is becoming—versus how many companies want to continue to do it like you do it now in a tool like Figma.

Do you think you're opinionated enough?

I think that our tool enables a lot of flexibility at the cost of some velocity in our product development. But I think it's a pretty good sweet spot. You can build with Lovable, and then any engineer can come in, edit, and take over if they want to.

So, yes, we would be moving faster if we were even more opinionated about how things should be done.

What are you not opinionated on that you would like to be in a dream world? In a dream world, we would be even more opinionated about how you build an application, and we would know what the future of building applications looks like with AI being such a core part of it.

I don't think it's possible because of how AI works, and because the best UX with AI for the products that are built with Lovable changes so rapidly. At some point in the future, I'd love to be there.

When we can be more opinionated, you get the right level of detailed adjustments on how the AI works for your product, and how the backend flows, workflows, and automations work for your product. Right now, we support a lot of different things, so you need to be really good at prompting.

Do you think we will prompt in 5 years' time?

Yes, I think so. But maybe it will evolve in terms of how you do it. Hyper-personalization takes care of a lot of the detailed prompting that we have to do today.

What does that mean?

Prompting is basically providing context to an AI about what your goals are and how you want it to do something. It's like when you have great employees: they know everything about how you want things to work. You just have to say, “Let's go to Stockholm and do a hackathon,” and then it magically becomes what you want it to be.

It pretty much does. Honestly, I sent a picture to my mother beforehand, and she's like, “You had nothing to do with that.” I'm like, “No, I did not.” And she's like, “I know.”

Right? You can't just tell ChatGPT, “Let's go to Stockholm for a hackathon.” It's going to come up with something different from what you had in mind.

You can either prompt it in great detail, or you can make sure it knows how you think. That's what we'll be evolving toward.

We said the word “opinionated,” and we spoke about it with regard to the company. I love your social media presence because you're also opinionated in your social media presence, and I think it's respectful, but also just quite blunt. You've been opinionated in how you talk about competition, specifically Replit and companies like it. How do you think about whether or not to engage in an opinionated stance against the competition?

I don't think so much about competition. The only thing that matters is that we make our product the best product and continue to deliver on our value promises to our customers.

There was a competitor that found a lot of apps that had been poorly made, and they said, “Oh, this is a vulnerability.” I spoke to a lot of security professionals, and that wasn't really how you would normally announce a vulnerability. So I went in and bashed them as a result of that. I think that was a very reactive thing, and it's something I'd be happy to share with that competitor in person, face to face.

It was interesting because Jason Lemkin, who's a friend of mine—I don't know if you saw it, but he was using Replit. I can't remember exactly what happened, but they basically had a massive security breach, or they deleted his entire database, or something bad happened, and it was code red for them. The takeaway for him was that security on all of them is just nowhere near where it needs to be.

It's wrong for Replit to bash Lovable. It's wrong for Lovable to bash Replit. All of you guys suck at security. Is that true?

Yes. Let me say it a different way. First of all, we talk about security company-wide every week. Every day, I hear something about security because we take it so seriously. There are so many different fronts on which to make it much more secure than if a human did the application development, and that's why it's so important for us to be the best in the world at security.

So you're saying it's more secure than humans?

Not yet. I said this at some point: If you take your average developer, who normally works in a large team where they have a lot of support, and that human goes out and builds an application, they are going to create software that has security holes, on average.

It reminds me of self-driving, where, for the world's best driver, I'm sure you are better than self-driving. But for the majority of people—and especially for the majority who are tired—

Yes.

Humans have the potential to be hungover, high, malfunctioning in some way, and actually wildly dangerous, whereas self-driving is much, much better.

Very much so, yeah. I would say I'm very proud of what the team has done so far with security, but there's more to come.

When we look at the many different competitors we've spoken about in different ways, from your Figmas to your Replits, if you move forward 3 years, what does the space look like then?

Again, I focus on what our product does and how we serve our customers best. I don't really predict that. If we get the majority of the profit share in this market, that's amazing. If it's spread out across different companies, that's also fine, as long as we build a product that lasts for generations. I do that by building the best product for our customers.

And this is why brand is so important for you.

That's how I think about it.

Do you mind if developers go to Lovable, get 60% of the code from there, and then fine-tune it? How do you feel about that?

From the get-go, 2 years ago, I decided I was going to build Lovable for a world where humans don't write code anymore. We're quickly moving there—very, very quickly.

Today, I don't mind at all. It should be flexible. Some humans have their own way of doing things, and I think it's good to have an ecosystem where you can use many different tools on a product over time.

I think it will converge toward the very opinionated platforms we're building toward, and everyone's just going to look at the cost-benefit of doing it with some other tool as well. Only using Lovable is going to be the obvious, highest-velocity, highest-quality choice. That's the future.

Today, does AI make 1x engineers 10x, or does it make the 10x engineers 100x?

It does both. It really does both.

For junior engineers, or 1x engineers, they often are bad at something. If AI can bridge that gap, it takes them from zero to one and enables them to do something they weren't able to do before. That's like a 10x: they go to 10x or even more. If that's the case, then it's more valuable for them.

If it's a 10x engineer working on something where you need many years of experience to work on a system that a new junior engineer completely doesn't understand, then the 1x engineer is useless. No AI is going to increase their velocity, whereas the 10x engineer is going to maybe become a 100x engineer.

How will the size of engineering teams change in the next 5 years?

I think that, for the best companies, engineers will really act as this translation layer. They will need to be thinking more in terms of product, being a product manager. There's a higher elasticity in such engineers.

You might see many companies saying, “Oh, more engineers—we can do even more, even faster.” They're out talking to the customers and changing things with AI super fast.

But the skills required to be a good engineer then change with time.

Yes, definitely. Being a generalist becomes more and more important with everything, with AI, so that you can understand how things come together as a larger whole. Then you use AI for the deep expertise that you don't need as much in the future.

11. Should Anyone Still Study Computer Science?

When you think about the skills required, there's a lot of people asking today, “Should I bother studying computer science if we're going to see Lovable be the last software that we ever need?” How would you advise your little brother if he were questioning whether to study computer science at university?

I think university is not the best place to learn. It doesn't matter what you're studying. You should be out there and really understand how the world works in terms of how work translates to value creation, and you don't learn that at university.

University is a way to train your brain to learn new things and meet a lot of interesting people.

Would you encourage your children to go to university?

This is now almost 20 years in the future, so it's hard to say something about 20 years from now. I think it's a great experience to have had in life, so why not? But it depends on what outcome you want to reach. If you want to have a job where you make the most money, no, they shouldn't go to university. I think the opportunity cost of those years is very high.

True. Yeah, in the UK in particular, we just get very drunk for 3 years.

That’s generally how it is. We generally study generalist subjects like geography and history, and honestly, it is a little bit of a waste of time, in which case you can utilize those years so much better, given your stamina, your energy, and the plasticity of your brain at that age.

Which is why I highly advocate against it.

I agree. If you just do a very, very specialized job for those years, maybe you’ll become less of a generalist, so there’s a trade-off there. Of course, at university you’re exposed to many different concepts, which can be useful.

We mentioned earlier AI and enterprise. When you look at the biggest enterprises today, they’re not able, for data, for permissioning, and for security, to use AI. Are we going to see the biggest shift in incumbent power in the next 10 years?

So you mean if they’re not enabled, there’s someone else that comes in and is enabled? I think you see this in banking, for example, where a bank is a software company, right? It’s all about software systems, and the old banks are moving much slower.

I think, yes, there are going to be some companies that are built from the ground up for AI to change their systems. Anyone who’s exposed to customers, understands the legal requirements, and so on, can move much, much faster in creating a good customer experience. So, yes, I imagine there are also some benefits of having been around for a long time in the enterprise, in banking. There’s a certain element of trust and so on.

I don’t know how large the shift is going to be across the different segments of the enterprise market, but many companies will get disrupted by cheaper, much, much better alternatives.

It’s interesting you said there about trust. How loyal do you think Lovable users and customers are? Do you think there’s a high propensity to switch and an ease to switch, or do you think people are fundamentally loyal?

It’s 50-50. Some people are just super, super loyal to a brand. Many people, if you do something that hurts your brand, will switch, and they’re just out there looking to maximize some kind of cost-versus-capabilities objective.

You can think about both of those groups simultaneously. If you have the best product with the best value, you’re going to get everyone.

We spoke about people being threatened by large incumbents. What question should large CEOs and business leaders be asking today about the future of AI, their companies, and how they use it that they’re not asking, do you think?

I think one of the biggest bottlenecks for these companies is going to be some kind of change management for the humans in the organization. I think they should be asking, “How have similar companies to ours done change management very, very rapidly?” They should get that conversation into the leadership room and then maybe across the entire organization, to start studying those examples of where change management has been very successful.

Then they should look specifically at which AI tools they should be using. Should they be adopting something like building their product on top of Lovable, 100%, because then everyone can collaborate? Or should they hire some new type of people who come in and upskill everyone?

12. Work-Life Balance Is Dead: Inside Anton’s 10x Culture

You said about speed there. You said about talent earlier.

Bluntly, dude, I get really fed up with everyone saying that Europeans are about espresso and taking the summer, and that it’s August and July so we’re not going to work. I advocate for a very aggressive work culture, which, you know, is 996.

How do you feel about the importance of unwavering hard work over balance in the desire to win?

Over a 10-year period, I would advocate for some balance, but over a 2-year period, if you really care about something, then you should make sure that you get your exercise and sleep in really, really well, and maybe have something that relaxes you. Then just work your ass off. That’s what you should be doing.

Do you agree, then, with Scott from Cognition, who clearly said to all Windsurf employees, after hiring them or buying the company, “It’s 6 days a week”? It’s unwaveringly relentless, and if you don’t want to sign up for that, you can leave.

I would say, in how we think about it, you are here to have 10x impact over other people at other companies. If you don’t have 10x impact, to do that, you do it by being very talented, being good at your job, and being very focused.

For some people, you need to put in a shit ton of hours, but not for everyone. I would prefer to speak about whether I’m seeing the impact. Am I seeing that if you told me you were leaving tomorrow, I would be like, “No, you are such an important part of this company. You have to stay”? That’s how I push performance and impact.

Do you do the keeper test?

Yeah, we do the keeper test.

Has it made you change how you construct teams?

The keeper test?

Yeah.

Yeah. I think it always makes it clear to people that I need to figure out how I can have more impact. That’s one part of it.

Then I think in terms of, what does this organization look like? Is it optimally set up to succeed right now? Culture is such an important part. If you’re just throwing people around too much, it hurts the culture and the ways of working.

But doing this business exercise, saying, “Is this organization set up perfectly to win?” definitely shapes how I build the organization.

You mentioned Nick earlier at Revolut. He gave me the best answer I think I’ve ever heard on culture. You know, I’ve done 3,000 shows. When culture comes up, it’s like first-principles thinking. I’m like, “Fuck it. We’ll edit this bit out.” Always.

But he said the best thing ever. He said, “I don’t think about culture. I think about winning.” The single biggest determinant of human happiness is growth and development. When you’re winning, you’re most optimally positioned to grow and develop.

And so, if I create the conditions to win, you will grow and develop. Supporting that, the other thing that people like to do is accumulate wealth as well as development.

And you will accumulate that by winning because of your share-price increase. That’s a good quote.

It’s a really good way to think about it. If we win, everyone will be happy. There are very few places where they’re losing every day, day in, day out, and blissfully happy. It doesn’t happen.

What’s not great about your culture today, if you could change it?

There’s a certain personality type that takes a lot of initiative. They’re very, very excited about new ideas and doing novel things. As your company matures, you like that; it’s still an important ingredient. But your first priority needs to be making what you have high quality, continuing to keep it high quality, and improving the quality across everything you’re doing.

I want us to be even more like this: let’s improve the quality. Let’s improve how we do things. Move slow so that we can move really, really fast.

And you want to be more thoughtful around where you spend time and where you don’t.

Yeah. There’s this cowboy-versus-farmer analogy, where a farmer is optimizing things for the long term. I think we can do a bit more of that, optimizing things for the long term. But we always strike the balance of doing.

Do you think you’re in that phase of company-building yet? I actually prefer optimizing for the short term. I don’t know if you’ve spent much time in China or with Chinese development teams, but they are unbelievable in their psychology around building. They optimize for the short term incessantly and then just sticky-tape the shit.

If you have super-clear product-market fit, you have a brand to defend. You cannot sticky-tape everything. You can do that in sprints to move and innovate, but you really need to focus on whether these pieces are put well together.

Spend a lot of time moving things around in the organization and in your product so that it has high quality, maintains high quality, and you can build faster upon that foundation.

13. Why It’s Better to Build in Europe

Can you imagine if Apple were like, “Ah, fuck, we deleted your cloud. Sorry”?

Back to what I said earlier about Europe and summer and you not moving fast enough: you said before that it is better to build in Europe, and I don’t want this to be an advert for Europe. Why do you think it’s better to build in Europe?

I think there are many good things about Europe. There are also things that are better in the US, for example. I mostly think about this: I want to prove that you can build a generational product, a generational company, and a generational team from Europe, and part of it is on hard mode.

What parts are on hard mode?

Hard mode is that the network isn’t as great in terms of how many individuals and companies have worked on, and have context for, all the different stages of building an amazing multinational company.

Completely agree. There are no Elena Verers in Europe.

Yeah, very few. Maybe we’ll get there soon. But I think access to capital, and people who will quickly give you a lot of distribution and help you with distribution and brand, are also part of hard mode.

Do you think access to capital is a genuine problem?

I think there's so much money in Europe that, actually, it's not a problem for us. As I said, it's not a bottleneck for us.

No, it's not.

And you're going to start to see very soon—I'm sure you're probably already seeing it—but Lovable spinouts, where anyone who leaves Lovable will get a term sheet straight away.

100%. True. Why is it better to build in Europe?

Yeah. We're the best talent.

Okay. So you said the hard thing there is, like, number 1—

What was number 1 again? Sorry.

Wasn't there a network and individuals that have done it before?

Yeah.

That's one thing. Is that it?

I think it's easier to get distribution and to be on the center of the world stage in San Francisco and New York. But we've been able to pull that off from Stockholm, which is good proof that you can do it from here. And then, in the US—

Why do you think you've been able to do it? I have my theory on why I think you've been successful.

I think it's about storytelling and sharing everything we're doing at the company, and empowering other people who are using Lovable to tell their stories. That's how we've been breaking through. I think we understand that you should be building in public and sharing what you're doing. I think transparency is everything.

People like to follow people, and you've combined the two very well. You're incredibly transparent around your ARR growth. It's easy to be when it's as good as it is, but you're incredibly transparent in a way that most people aren't. Then it's led by you and your voice, and the two combinations of the cult of personality—you and Anton, you being Anton, not you in the third person—and then you and growth are what really drive the success in that way. So that's kind of what's harder. What's better? Why should everyone build their company in Europe?

We are the biggest talent magnet in Stockholm right now, which is amazing. You can't—it's much, much more difficult to be that in San Francisco or New York. So we can really pick up all the underutilized talent and 10x their performance by being in a 10x better culture, ways of working, and with amazing colleagues. Being able to be that top 1, I think, is the biggest one.

There's a culture of humility and low ego, and working really, really well together as a team, that I think is stronger in Europe. This way of thinking in terms of efficiency and doing much, much more with less is also stronger here.

You have inherently higher churn in the Valley. When you have a bad day, OpenAI offers you a bigger package, and it's like, “Ah, I'll leave and go to OpenAI.” That prevents the compounding of knowledge within teams, which I think is so valuable.

Yep. Totally agree with you there.

Would Lovable be less successful if it were in the Valley?

I honestly don't know. I think it would be very successful regardless.

Did you ever think about moving?

Yes. When I was about to start the company, everyone was, of course, telling me I should go to San Francisco. But we just kept building, and we found some great people in Stockholm, so we kept building it from here. I'm happy with how it turned out.

Dude, it's been amazing. That was a decision you made, and it worked out very well. What did you do in the Lovable journey that, with the benefit of hindsight and some experience, you wish you hadn't done?

When we started, we had this idea. The vision was very clear; the sequencing was not so clear in terms of what we should be doing. We had this open-source community that was excited about a tool I made a few months before we started the company: GPT Engineer. I think we should have scrapped that completely and 100% focused on what the future looks like in terms of opening your browser and just building your product there, which is Lovable.

Why should you have scrapped that?

You should be very, very focused on doing one thing—

Was it not crucial for customer development and customer feedback?

No, I don't think so. There was, of course, a plan for how to incorporate it and get more value together. Open source can be very useful for many businesses, but in terms of maximal focus, it was just a bad idea to do 2 things that were a bit too tangentially related.

That's definitely one thing. We talk a lot about doing 1 thing, finding the bottleneck for the company, and solving for that bottleneck. That's the best way to move really, really fast.

What is the bottleneck that you'll be discussing at the board meeting tomorrow?

I think the bottleneck for our long-term future is identifying the technical product engineers who will take the product to its next phase and innovate on many, many fronts at the same time. That's not the long term.

If you think about the product today, it's giving our AI more capabilities that really make it a polished user experience and give it more of those capabilities, so that you can build out your full company and grow your business on top of Lovable. Then I think a bottleneck is how we serve this extreme amount of enterprise customers, learn and pull from them, at the same time as we focus properly on founders building on Lovable.

Will Lovable have an enterprise sales team?

Yes.

And become an enterprise company? You will not become an enterprise company, but you will have this enterprise sales team?

Are you nervous about it?

I'm not so nervous about it. It's just about talking to and understanding your customers, and making sure they have the tools to get value from the product. That is how I see that part.

There are, of course, many enterprises with top-down enterprise sales teams that hustle to wine and dine CEOs. That's not what we're going to do.

What is the hardest role to hire for?

I think hiring engineering leaders is very difficult because it's so hard to predict how their past performance will translate to our organization.

Have you made mistakes on hiring?

Yes, I've made some mistakes.

What did you do that you wish you hadn't done when hiring?

I wish I had been in the details when I delegated too much, and I wish I had been more proactive about asking: Does this person want to reach the outcomes? Are they excited and inherently motivated by the outcomes that this role is going to be most important for? Where is it going to be most important for that role to reach those outcomes?

I think something that's really interesting when you talk about leaders is your co-founder Ashley, who was in the video for the OpenAI release.

Yeah.

What struck me about that is, respectfully, it was one of the first times I've seen him front and center, not you. How do you think about exposure between the 2 of you, given you are very much the face of Lovable?

I'd love for Fabian to have more exposure, but I also want him to be focused on building the product. That's what he's focused on. It's probably much easier for people to relate to Lovable if they see 1 person and keep seeing that 1 person. To date, that is me. I think it will continue to be me.

What do you think is the biggest secret to a successful co-founding pair scaling at the speed of Lovable?

I think the most important thing is just the horsepower and adaptability of the founders. If those are maxed out or high, you must be able to work together. If you have sufficiently low ego, it's going to work.

But if you really want to work extremely well together, I'll take an example, which is Fabian and me. He's not very big on doing some weird new way of doing things. He's just like, “Simplify as much as possible.” He's quite introverted and quiet until he's really shaped an opinion about what's most important. I'm on the polar opposite side of the spectrum, saying, “We should use this new crazy thing.” That polarity is actually very productive for both of us.

When you think about that, and then you think about the fact that you're married and very happily married, what makes a successful marriage so successful?

Fabian and I can talk about anything. That's extremely productive, and we can talk about turning every stone and challenging each other around anything. We have a lot of humility. I think that's very valuable and very important, and the same is true in my marriage.

I think humility.

Humility. Yes, a lot of humility. I talk about my faults a lot in my marriage, for example.

Does success make marriage harder or easier?

If you have 0 hours to spend time with your partner, it makes it more difficult.

I also look at 90% of relationships and see that they often struggle, and a lot of arguments are based on money, which is an inevitable thing that's very hard, especially as the cost of living goes up. For a lot of people, that then doesn't become a problem.

True.

Yeah. I don't think your marriage has changed much.

Not so much now.

Very humble Swedes, aren't you?

Yes. I haven't changed my lifestyle since Lovable was successful. Maybe I think less about monetary decisions, but no, my lifestyle is pretty much the same. Do you not?

What does a Lovable product look like at the end of 2026? I mean, it’s your perfect co-founder that you go to with your idea from the idea stage, but also all the way up to growing your business once you have customers, and taking care of what Elena is doing: optimizing the product for growth, optimizing the product, and optimizing your communication with your customers, whether it’s through email or different marketing channels.

Yeah, it’s one obvious way to do the entire product lifecycle.

So, you eat the whole stack, then. You do everything from email marketing to SMS marketing and everything in between.

Yes. Obviously, this is also what an enterprise wants to build its products on. But in the interim, they’re using it for individuals—people in teams building out ideas that enterprise companies should be doing—and they’re doing that very, very productively.

Is benchmarking for models and evaluations bullshit? I had Edwin Chen from Surge AI on the show. You know Surge AI; it’s like the Scale AI competitor, but it’s actually phenomenally successful. It’s never raised a dollar, and it’s $1.2 billion in revenue.

It’s unbelievable.

He was like, “The benchmark evaluation is bullshit.”

I mean, they turn more and more useless over time. There’s something called Goodhart’s Law: when you start optimizing for a number, that number stops being a good measure for success, even if it was previously a great measure of success. Obviously, that happens with all benchmarks over time, in some sense.

What metric within Lovable means less over time?

It means less if we start optimizing for it, right? One example, I guess, is how many people click the thumbs-up button on messages. We can say fun jokes or whatever that, for some reason, just trigger people to be more likely to click the button. Just asking the human, “Click the button if you want to do it,” means we’re hacking the metric. That’s just one example.

Dude, we’re going to do a quickfire round. I’m going to hit you with some incredibly unfair questions, and you can give me your thoughts, okay? What wildly held belief about AI do you think is just very wrong?

I think AI is much better than humans, and most people don’t agree.

Do you not think they do now?

I think most people don’t agree, and the reason is that it’s often very, very stupid. It’s very stupid. But if you give it all the context, or you build a purposeful system for what it’s stupid at, it’s smarter than humans.

Do you think we will see a plateauing, or do you think we will see a continuous exponential progression curve?

I think we’ll see plateauing on the things that we care about, which is a lot of nuance and being good at all the different things at once in the same model. There’s probably—

People are looking at GPT-5 now and saying, “Ah, we’re hitting a stage where improvements are actually much more incremental.”

I think what we’ve seen so far is sigmoid curves across many different dimensions at the same time. We’re going to see a plateauing. There are some sigmoid curves where I think we’re still in this exponential phase of the curve. Those could be things like science and engineering, and bioengineering, where AI is just going to continue to exponentially become extremely powerful and generate a lot of new medicines and new ways of treating health.

Grock, Anthropic, OpenAI. You can invest in OpenAI at $380 billion, Anthropic at $180 billion, and Grok at—I think it’s $100 billion. Which one do you invest in, and which one do you short?

I’d invest in Grok, and I would short—what were the numbers again?

$380 billion and $180 billion.

Okay, $100 billion. I would probably short Anthropic because—no, I would short OpenAI, let’s say.

Why would you buy Grok and short OpenAI?

I think it’s more the slope of the Grok team. They’re doing something that I respect a lot, which is hiring missionaries for the data-curation part, and they call it AI tutoring. I think the morale is much, much better in that team than in both of the other teams. The morale is super high. OpenAI has gone through all this mess, right? Grok has good morale as well, and they’re growing faster on the enterprise side, from what I’m hearing.

Do you think OpenAI wins the consumer market, in terms of being the next-generation Google, and Anthropic wins the developer and enterprise markets?

No, I think it’s going to be unknown. There’s going to be something else happening that we don’t know about.

Do you think there will be a leading model that has not been created yet?

Yes. From China.

Do you worry about China?

I think Chinese companies are not as good at really understanding their users, so I’m not very worried. I do think there’s a 50-50 chance they’ll have the best model. We’ll be using a Chinese model at some point, and that makes me a bit concerned because I—

Chinese models at Lovable?

If we would.

Yeah.

And I would have to look into the details and see what’s bad about that. Do we give them data we don’t want to give them? But we just want to do what’s best for our customers. If that means going for a Chinese model and there are no negatives, then yes.

I completely agree. The multitude of models coming out of China is terrifying. Every week there are four new ones, and they’re all as good as the last one. The speed of distillation is just fucking insane. Are the models of the future open or closed? Which model wins?

I think the best ones will always be closed. But if you want maximum flexibility and some kind of open ecosystem around them, it might be that open ones are the ones most people choose.

You can have dinner with anyone, dead or alive. Who do you have dinner with, and what do you ask them?

I think I would have dinner with Newton because he was religious and super smart. I’d talk about what he was like in his age and why he was religious. He invented so many different things, and he’s a bit of a role model. He’s dead, so I can’t meet him unless I see him.

No, sorry, I can’t help with that one. There’s no intro there that would work. That’s amazing. What AI company doesn’t get enough attention? Like I said, Surge AI for me is one. It’s like Scale AI, but fundamentally a much better business. Barely anyone knows it, and it’s ridiculous. Which company does no one pay attention to that everyone should pay attention to?

I think the browser companies are interesting. There’s Strawberry, there’s Comet, there’s Dia, and there’s Perplexity. I’m very excited to see what happens to these companies.

What do you think happens to Perplexity? So, they want to create the phone, I think, and I think that’s a good bet. You think it’s a good bet? Would you invest in them at $18 billion?

It depends on what options I have.

That’s amazing. Your laugh there just kind of said it all.

That’s very funny.

Who’s been the single most instrumental person for Lovable who is not in the company?

It was atlena. He ran sales and like CEO at Muro, and he was on Dropbox at at N segment. I get a lot of input and help from him.

What’s been the biggest?

I think just talking about how I think about things. He’s my coach, and I talk about how I think about things. Then he asks me questions and tells me, “Anton, you have to step up in this. There needs to be a bit more structure in this area.”

I love him. I had him on 20 Sales, and he was fantastic. What have you changed your mind on most? Penultimate one. I thought that, bluntly, you’d see the commoditization of model performance and models, and that value accrual would be very difficult and there’d be a race to the bottom. I think that’s clearly very wrong and very stupid of me to have ever thought that, and there’ll be very valuable model providers. What did you believe that you now think, “That was wrong”?

In the context of Lovable, I thought we should be building an agent before the models were ready for it, because the models were starting to get optimized for an agentic system. What I realized is that, no, no, no—you need to have a product that as many people as possible are using today, so that you can optimize not necessarily the AI, but the entire user experience for those users. That’s your data flywheel that you want to use.

Do you worry about job displacement at scale over a 10-year time period?

I worry about us humans globally not even understanding what we want to achieve on this planet. If there’s a lot of rapid change, with white-collar workers being out of a job, and we humans get super worried, concerned, and scared, then all hell is going to break loose. That’s what I’m worried about.

But if we’re a bit more thoughtful and say, “Okay, if there would be an insane amount of job displacement, this is what we think we should do and this is what we want to achieve. This is how we make sure people can have some made-up job in the interim,” then we would 100% solve that.

8 out of the top 10 paying jobs today did not exist 15 years ago. I always think that’s an interesting stat.

And we always overestimate job displacement with new technologies.

Yeah, interesting one. I think we're going to have maybe a shift away from some very glamorous jobs, which people will get depressed by. Similarly to how being an artist was so cool, but clearly you can't make any money as an artist. I think we're going to see that again now for a lot of knowledge work, and that's going to be funny.

My kids? Brain surgeons? Brain surgeons. No AI is going to come for brain surgeons for years. Come on.

Maybe. Have you seen the robots, though? The surgeon robots? They're pretty good.

I'm going to be honest. You're not having people be like, "I'm going to choose the robot version of that."

Not for a while.

It's a good job. Is there anything else that concerns you with AI when you look forward?

I think, again, as humans, we're very, very good at competing. In many cases, that's amazing. That's how you get all the best companies. That's how you get great technology. But in some cases, we're competing and then we go to war with each other, or we start preparing for wars.

I think if we can be better at thinking big-picture across superpowers, that would prevent the scenario where you have an AI that can kill all the people in another nation in an instant, and that being triggered without us actually wanting that to happen.

So, yeah, I'm concerned that us being so competitive in a world where things happen much faster is going to lead to some unexpected results that no one really wants.

Which competitor do you most respect?

I think OpenAI is pretty good at building products. I think there are other foundation model labs that will do even better at building products, and those are the ones we should think about for the future. But mainly, focus on just: What do our users want? How do we make a better product?

But you must look across your Figmas, your Bolts, your Replits. Who do you respect?

I respect Figma.

Figma.

Yeah.

Why?

Because they're good at listening to their users and building a good product. If they can translate that to the full product life cycle, I think they're a very formidable competitor.

Everything goes to plan. We hit all of our numbers and everything works. If that is the case, where then is Lovable in 2035?

We're the mostly used interface for humans to AI, and that's a very huge market.

Dude, it's so much better doing it in person. I've so enjoyed this. Thank you so much for agreeing to do it in person, and I've loved it, man.

It was fun.