Never wrote a line of code, now a $6.6B unicorn : the vibe coder - Lazar Jovanovic [Lovable]

0:00I think what you did to VC with bootstrapping is what vibe coding is going to do to traditional software development. The only people that should be afraid is people that watch me and you today and do nothing about it. Then you should be afraid. You have to put your ego aside. It is always the human fault. If you're waiting for a job description, you're not optimizing for the right thing. Today, your competition is zero. I don't think anybody's going to need to write code by the end of 2026 already.

0:25Lazar Jovanovic:Today on Billions, I'm sitting down with Lazar Jovanovic. Lazare is the world's first official professional Vive coding engineer at Lovable. A startup that hit 100 million dollars in ARR in just 8 months. Faster than any software company in history. Here's what's wild. Lazare has never written a single line of code manually. Not one. Yet, he ships production apps daily faster than senior engineers. From Lovable merch store to internal tools used by a company now valued at 6.6 billion. His weapon, AI, clarity, and of course, taste. Lazar, thanks a lot for being here. Thanks a lot for having me, man.

1:04It's a pleasure.

1:05Lazar Jovanovic:I mentioned also taste because you're really good at design and you've done some amazing things when it comes to design, I feel, because, you know, with AI and AI generated, it was more or less all the same, but we'll get back to this later. As my first question, I want to start, you know, you were mentioning that non-technical people, they can also have like an advantage over engineers because ignorance is kind of like a superpower. So how exactly does that play out, you know, day to day? Yeah, 100%. So I will start with a quote on this one. It comes from Warren Mason and it's about Kurt Cobain.

1:47Everybody hopefully knows who Kurt Cobain is, former Nirvana lead singer. He said about Kurt that he has this neyvet about not knowing musing theory at all and what goes together, which is great because he doesn't then have to make the right moves. He just has to make it different. And I think that's exactly how I look at this now, which is people that don't know the rules make up the rules as they go. And sometimes go against the rules, which is great because if we are all honest with each other, we don't know what the capabilities of the models powering these apps that are AI app builders are even capable of.

2:30I don't think that even creators of these models know. One of the proofs of that is that Claude Code was used to Vibe Code Claude Cowork in just seven days. in theory that would be impossible like it would be impossible in theory to even think that something of such a high production value that is now shipped to people as an enterprise-grade product is vibe coded by the tool itself um i think that people at open ai and i i may misquote them but like uh from my understanding the 5.3 the latest open ai model was largely written by the previous model. So again, if models can write models, why can't I do whatever comes to my mind, right?

3:16Or at least try. I think that's the key here. I meet with a lot of engineers and I don't want to get, because a lot of people, when I talk about this in public, they think that I'm saying that being an engineer is a disadvantage. No, absolutely not. Like if you're an engineer, that's a 10X engineer that is experienced and accepts these tools as the reality of what's today and what's coming tomorrow, you're unstoppable. I can't compare with you because you have all the layers of experience. What I'm saying is that for us that are non-engineers compared to a traditional engineer that just simply doesn't want to accept the reality of the moment we live in, we have this huge advantage of just like immersing ourself really quickly and leapfrogging in the sense of like, if you just know a little bit about software architecture, That doesn't mean you need to know how to write code.

4:06So I don't write code. You're right. Your intro was 100 % correct. But that doesn't mean I don't know how startups are built. That doesn't mean I don't know how software is built. You have to have some level and some degree of knowledge. But I think the big advantage is really not going into the nuances and just try, right? Just try. Prompt it. See what happens. And I think the way that plays out for people today in tools like Clovable or any other tool is really opening up capabilities for them to unlock their creativity without having to wait for resources or permission to do so. And that's pretty exciting to me today.

4:43Lazar Jovanovic:Yeah, definitely. And something that a lot of people have been talking about is the fact that you might be able to create software, you might be able to create apps, you might be able to use Lovable to create your own application, whether it's on your phone, on your computer or whatever. But what's hard about software is maintaining the software and making sure that as it evolves, things don't break. Like, I mean, we've seen it at Lemlist, you know, eventually all software company goes through a phase where you have technical debts and then it slows you down a lot and you have to rebuild kind of from scratch and go.

5:22Lazar Jovanovic:So what's kind of your vision with all these AI tools, you know, where something like can eventually break and where it's kind of hard to maintain the code long term? I think as the models get better, this will no longer be the bottleneck. Right now, it is a little bit. And for you to scale, at least from my point of view and looking at projects that I was able to scale, you can't just vibe code your way to that. You need to have a system. You need to have a framework that I'm trying to get people to adopt. Like, in a way, you're in this position where you need to start treating these tools not just as AI coding tools, but also almost as like technical co-founders and working alongside them as a companion rather than you bossing them around what to do.

6:13What that means in practice is this. Whenever you're trying to develop a feature that will require future maintenance, you need to properly document it, both forwards and backwards. That basically means that before you build something, you need to be very clear and very specific how that is supposed to look and work, document it. And usually I do that in the form of PRDs, which are for non-technical people. That means a product requirement document. Basically, just a breakdown of what we're building. Then when I'm done building it or where the tool is done building it, I try to back up the knowledge of defining the source of truth of how was it built?

6:58What are the relationships and dependencies with this new feature, new menu or new option? And like simply by documenting, you're already steps ahead of the curve. And even if something breaks, you could then point your agent in a direction where it's supposed to look and that way make the job of maintaining software easier. Now, is this a perfect workflow? Probably not. Like, does it help being technical when things start to break? Yes, most definitely. But I just feel that the underlying models are key. their context memory windows which means like the amount of workload that they can take in a single request are constantly expanding um i mean just two days ago we had the big uh sonnet 4.6 launch from entropic massive model amazing biggest context window ever best reasoning for coding ever i just the way i feel is like it's a problem but a problem that will be easier to solve than some other problems that vibe quarters are experiencing and the two bigger problems actually in my book are like why are you even building the thing that you're building like what's the idea behind it what's the reason behind it which is like a lot of us that are like just shipping a lot need to start asking ourselves should we ship this in the first place who is going to even need this that's number one then number two if we know that people are going to need this.

8:27How are we going to get that in front of their eyes? So distribution, again, same problem as we always had. I have a good friend, she always says it, like if coding was the problem, every developer would be a millionaire. But because they're not, that's just another proof that coding was never the issue. It was always something else. And I think we just made it easier for us to get to a point of solving the biggest problem in the book, which is attention, you know and getting your product in front of people's eyeballs and that's why like always try to shift away from making these discussions too technical because that's not the problem or if it's a problem today i i'm in that tribe of thinking that six to 12 months from now it

9:13Lazar Jovanovic:definitely won't be anymore and yeah i 100 agree with what you're saying and things are moving like so fast that is just every problem that people are raising now they're not gonna be a problem and to your point you know of uh developing like by trade i'm like an engineer so when i started i was doing like c as a language which then you had like different layers and javascript and so on and so forth back in the day you had assembler so obviously you know it's just like layers that are getting less and less abstract and now people with just language can essentially like get their app which but you still need the logic as uh as you mentioned and and something that um that i was curious about lovable is your your strategy and your target audience like do you want people who are using lovable to essentially like build their business and make revenue out of it or is the main target let's say like i don't know just a kind of consumers who want to create like cool apps for their friends and just like have fun or like what's kind of the the strategy and the people you're targeting the most i'd say if i'm honest like we just want to empower anyone to build anything they want um if you just want to make a game for your son because they're yelling in the car you can do that in lovable if you want to build a business as you pointed out that can make you some money you can do that in lovable as a solo founder if you are an enterprise and And you're looking for better ways to prototype internally so that you can reduce the time frame and shorten the cycles from an idea to a production grade software.

10:53Again, you can use Lovable for that. Like we're amazing at that use case. Another use case that I think we're starting to really become good at is internal tooling. And I'm serious level enterprise grade internal tooling for like engineering and non-engineering teams. So I think we are on this mission and trajectory to empower the 99 % who don't know how to code. I think that's kind of the overarching mission of like, if coding was your bottleneck, we will solve for it. But I also feel that that doesn't mean that engineers shouldn't or couldn't use Lovable. And we have a lot of engineers who are saying that, hey, maybe I won't use it today for something production grade, mostly because the not because the tool isn't capable, but because Lovable is kind of constrained a little bit on the tech stack.

11:46You mentioned a couple of different languages. So Lovable uses React framework as its primary framework. And I understand that in some companies, that's just not that it's not applicable to whatever stack they're using. But that doesn't mean they can't just come in and vibe code a good looking user interface connected to their database and do some data visualization. Maybe something like, you know, maybe you're not building features on Lovable, but you can build a feature adoption tool on Lovable. You plug it in, pull in all of your A-B tests, for example, add an AI model on top, and then you can see how are people commenting about your features or how are they adopting them, for example.

12:23So I think the mission, from my point of view at least, is clear. This is the future and the present. I think that, and this will resonate with you too, probably. I think that vibe coding will do to software generation, to what bootstrapping did for VC. So, you know, I know your story and your background, right? And like 10 years ago, you were probably considered the craziest guy ever getting into one of most competitive spaces, email, right? Against well-funded companies with$1 ,000 in your pocket and an idea. You had no prior experience doing something like that. you had no prior success, no prior personal brand.

13:07Like you were just like, hey, I'm just gonna do this my way. And not just that, I'm gonna go and not do anything corporate, the way I hire, the way my people that I hire behave. I will never forget one post, someone from, I think it was your customer support lead. It was like five, six years ago, but it still stayed in my mind. She was talking about how, you know, if people attack you as a support manager, as a brand, you can just tell them to piss off. You don't need that type of a customer. I was like, wow, like six, seven years ago, that was not the standard. Now everybody's doing that. Now everybody's protecting the culture, protecting the brand.

13:49Everybody wants to be a solo bootstrapped founder. But when you did it, everybody thought you were crazy. And I think what you did to VC with bootstrapping is what Vibecoating is going to do to traditional software development. Everything that I'm saying is going to become normal. The only difference is this. It took you a while to become the smart guy in the room. It took a minute, right? You were crazy for a long time. I think with VibeCoding, that is just going to get shrunk completely. So let's say, I don't know, I think it took people to understand the power of being a Bootstrap solo founder for about five or six years.

14:27With you, Peter Levels, Mark clue all of the people making it mainstream with vibe coding i think we're talking in terms of maybe two years you know like everything is just going to happen faster because the models are getting developed faster because everybody's attention is is now moving much faster like if i if personally if you ask me in 2023 how long is it going to take i would tell you that it's going to take at least 10 years to get to where we are right now and it's like just early 2026 i I just think everything is shifting in such a fast pace that is super hard to keep up. But that's also for anybody listening and thinking, well, does it make sense for me to then jump in on this?

15:08I'd say 100 % yes, because that is actually an advantage. Because everything moves so fast, everything that I know today and talk about today is almost irrelevant six months from now. Like I will have to relearn my job and reapply for my job in six months again with another skill set. So if you jump in today and let's say you watch my, I'm doing this interesting series called 28 Days of Lovable where like I'm recording videos every day. You just watch that and you start building. Me and you are almost at the same level, really. So you're not that behind, right? So I think that was a really good project, actually.

15:43Lazar Jovanovic:Like congrats on that. It's very awesome. I saw it like on LinkedIn and I think it was really cool. I actually shared it with our CMO because I think the way you're like showing and, you know, having people do it with you and within like this contest. And it was really, really cool. It felt like a kind of, you know, Christmas Eve, like when you have like the calendar with like all the chocolates and every day it's something new. And yeah, so very cool. Yeah. Yeah. No, and I just want to like put a full circle on the subject of like where we're at. Again, I don't know if you agree or disagree. You're longer in this than I am.

16:22But again, that's just my feeling. Bootstrapping really disrupted certain things. And even if it's not in monetary terms, it just disrupted the way the world works. I think the key word here is permission. I'm not saying that either founding a company was ever gatekept or learning how to build software was. But if we're honest with each other here, there was a little bit of a gate there. There's definitely gatekeeping. Yeah, yeah, definitely. You want to build a company?

17:00Lazar Jovanovic:Yeah. Yeah, go ahead. No, to your point, I think whenever you wanted to build a company and the reason why most people wanted to go through the VC route, which was my case, because you were pro-boostrapping, etc. like the the reality is i tried to fundraise when i got started and i got rejected so many times so it's like of course it got easier like once we reached like 10 million in arr and then sold like 20 of the company for 30 million and then we scared like to 40 50 million in unrecring revenue everything gets easier but the reality is like when i got started no one would talk publicly about what numbers, how they would acquire customers, how they would do things.

17:46Lazar Jovanovic:So everyone would say, you need to go VC because that's what the press talk about. That's what you read on TechCrunch at what. And eventually for me, I got lucky because it started working. So I'm like, okay, at least I can share the stuff I do. And maybe some of it was luck. Maybe some of it was just finding new ways to do things. And I feel what you're doing with Lovable is definitely the same. you know it's like building software feels gatekept because even if you look at you know like I know so many people in France who build websites for businesses and for a long time you know when you had to actually be a developer and know like HTML, CSS and people had to work to create a website these guys like the developers they were spending like time they need to you know like check where the website was hosted sometimes they were issue with hosting platforms so they had to to spend a bit of time but eventually you know there was that phase where a lot of people didn't know that you could go on whatever website buy a template edit it yourself and then that's it you know you don't have like maintenance on anything and these persons were actually making good money making people believe that every time they need to make a change it's actually super complex and you need to be a crazy developer etc and this for me is like the worst gatekeeping ever and i feel like with vibe coding it's exactly the same like lots of people still believe that you need an army of developers to build things well actually as you say so well i think your your motto is like demo don't memo which which i think is great and very cool and And you can build like a lot of things by yourself.

19:31Yeah, 100%. You saw, you summarized it. You hit the nail on the head there. Yeah, it's not that people, nobody's forbidding you to do it, but it takes a while, you know, to, I feel, and this is another kind of a touchy subject that people have polarizing opinions of. I get asked, should I learn how to code, right? And my answer is a definite no. And people always like that are developers, they say, no, that's the wrong way. People should know how to code. The way I see it is, it's going to take me five years, maybe 10 years of learning how to code just to be at the level where AI models of today are.

20:11I really don't think that five or 10 years from now, anybody's going to write code. And if I'm honest, I don't think anybody's going to need to write code by the end of 2026 already. It's going to be a thing of your own choice. If you want to write code, great. Eventually, as I said this before, writing code will be like calligraphy. It will be a form of art. It will be something that if you want to fine print your code, do it by all means. Nobody's going to prohibit you to do so. It's just not going to be needed or efficient. So for me, that layer of permission has gone away. And now there's probably different levels of gatekeeping that we need to solve for,

20:48Lazar Jovanovic:which is, again, a lot of people feel like this is a doom and gloom moment, which is AI is coming for all of us, especially in tech. Everybody who's sitting behind a computer, they're going to be out of a job a year or two from now. And I feel completely the opposite. I see this as a massive opportunity unlock. I also think that as we progress further down this path of everything fake, for the lack of a better term, right? Fake videos, fake images, fake this, fake that. We are number one going to start to value again in-person events, in-person meetings, in-person deals and everything done human to human.

21:27And I think same way taste is the currency, the key currency will become trust, right? There was a big polarizing moment. I don't know if you saw that on LinkedIn between two authors of very viral posts. One of them was something big is happening and the other one was called something messy is happening. It was like a clash of two worlds where one guy is saying, you know, if AI didn't knock on your door yet, you know, you better be ready. It's coming for you, right? It was a bit of a, you know, dystopian reality of like, oh my God, like we're all done. And it's all doom. And the other one was a little bit of an opposite of that, which is acknowledging that AI could be coming for people's jobs, but that it's not actually like that and that people don't hire people only because they're efficient.

22:22The person tried to compare, to use an analogy of law, like you can teach AI everything about Anglo-Saxon law, for example, and you can represent you in court, but AI will never be able to read the room, the jury, to see what they respond to. they will never know what argument the judge appeals to because an attorney works with that judge on dozens of cases before. So it has all of that context, right? So there are certain layers in which I think this may seem like a COVID moment where like it's a major cut off, but I don't think it is because with COVID, we went into lockdown everybody on the same day.

23:05We were mandated to stay at home right so the shift was for everybody all at once yeah yeah now it's like people still get to decide whether they want to adopt this and you wouldn't believe how many people are actually still on the sidelines i live in a very small bubble intentionally i only expose myself to builders i i'm on discord of lovable i'm on linkedin only following builders so to me it seems like oh the whole world is building then i go outside uh and play volleyball with my friends and nobody knows the name of my company. Nobody knows what my company does. Nobody knows what by-code coding means.

23:40And most of them don't care. Like I showed them and they're like, you know, they're not builders in their hearts. And it's fine. You know, things will shift. But I think what I'm trying to say is, you know, things are definitely progressing fast. Maybe you don't need to learn how to code to take advantage of all of this, but you also shouldn't be afraid to make a step. you're not far behind you're actually still super early and don't worry if you the the only people that should be afraid is people that watch me and you today and do nothing about it then you should be afraid like if you listen to me and follow me and you still don't feel like you should do anything about it then my friend it's if anything does happen to you it's it's kind of your fault it's not the fault of ai but if you just take a couple of steps really just minor ones vibe code whatever or not even vibe code just start using ai in your day-to-day life a little bit more you will not just find it interesting and fun but you will be prepared for what's coming yeah i think

24:42Lazar Jovanovic:it's quite insane like i'm i'm actually like um launching a new business and um it's it's really funny to see how things have changed because you know i've been like a ceo of 150 plus people etc for like for for years and now you know i'm started like uh alone you know like i mean with my uh my actually partner and and girlfriend and you know we have to do all the things from the start like build a website like uh do slides marketing and then i'm like i remember how much time you know it took back then in 2018 and i'm like with the eyes this is like so freaking fast and you can do like so much more and it's just uh it's just insane and it's like uh it's a real pleasure to be able to build.

25:27Lazar Jovanovic:And I was quite curious to get your thoughts on, you know, like we're talking about Saspocalypse. Like you've seen probably it on Twitter, like stock market going down for all software companies, etc. Especially after Anthropik announcement of Cloud Co-Work, all these things. On your side, like lovable, like the clear positioning has been like the last piece of software. So it's like anyone can build whatever they want. But if people build things, it's still a software with an interface. If we like kind of hear like what's happening on the stock market, etc. With AI agents doing everything and essentially their view is like you will never need an interface ever again apart from a chatbot.

26:17Lazar Jovanovic:Like what's kind of your view about it? Is it something you're talking about internally at the company? Like what's the long term vision? I mean, it ties back to what I was just saying. I think anything that is as extreme as like you're only going to need X is probably not going to be entirely true. People still live in spreadsheets. People still live on pen and paper. You know, you go to a doctor here in the U.S. You don't even get there's no electronics like you just write with your hand, like everything. Right. Patient intake forms will, I don't know when they're going to go online even. So it's like, you know, I feel, you know, maybe it's a vision of the future.

27:01And yes, I feel like that synergy between an interface and a chatbot is probably a closer reality, which is the software is a little bit more personalized to your liking and your experience. I think that's the big benefit here that when you design software and you build software in the past, because of the maintenance that you mentioned earlier, you kind of have to strategically make certain tradeoffs and say, look, we're going to have to go in this path. Whereas here, like I can just build features on top of features in seconds, in minutes. So like this is the way I build, for example, one of the internal software that we use at Lowable that I built for our GTM team.

27:42I built it as an interface originally because I needed that so that we can interact and learn which things can be automated, which things pull more or less inefficiencies in day-to-day workloads. And wherever we could, we introduced a chatbot layer that you can just dictate into it to make decisions for you. But it's based off of that interface. So I think today, at least I can say, I see it as not one or the other, but both. You can just marry the two and have the best experience. I don't even know, like there are people, for instance, on Lovables Discord that I talked to about this stuff. Like there are a lot of people thinking, oh, we're designing for phones is stupid.

28:26We're not going to use phones in a year or two. We're going to have a wearable. We're going to have glasses. Like we should design interfaces for that. So like everybody has an opinion about this. What's interesting again is the question is what do I do about it for the viewer? and again that's why i say don't be afraid look at how many opportunities there are there are so many potential ways that this can develop just pick your lane pick any lane and you will likely succeed because i think some people will decide they don't need monitors or phones they're going to use wearables and glasses other people are going to go the other way they're going to keep the traditional setup there's maybe going to be i'm seeing mentions of like uh that open ai wanted to launch a Super Bowl ad for like a physical device of some kind.

29:17But then last minute, they backed out of it. So there's this whole mystery around what's that going to be about? So I don't know. They did like a partnership.

29:25Lazar Jovanovic:They bought like Johnny Hive, no? Was it like the company? Yes, exactly. And then I saw something elliptical, almost like metal. And I don't know, maybe it was even fake. Maybe it was a fake video. Like that's the problem today. You don't know what's real. So what I think is that the answer is really, if I'm honest, nobody knows what's going to happen. We internally at Lovable have this, you know, kind of a great forward looking culture. And everybody in the company is a very early adopter. Lovable is not number one by accident. You mentioned all the numbers. Getting those numbers can only be done if you have the best talent in the world.