Lovable is your personal AI software engineer.
You describe an idea and then you get a fully working product.
The reason is to enable those who have had like such a hard time finding people who are good at creating software.
That's been their absolute bottleneck and let them take their ideas and their beams into reality.
You guys hit 4 million ARR in the first four weeks.
You hit 10 million ARR in the first two months with just 15 people.
You're the fastest growing startup in all of Europe.
How did you decide on lovable as the name?
It's so sweet. The best word for a great product is that it's lovable.
A lot of jargon that I like to use to emphasize what we should be striving for is building a minimum lovable product and then building a lovable product and then building an absolutely lovable product.
So I took that jargon with me in the company name.
People wonder just what jobs will be more important, what skills will be less important.
Doing a bit of everything, being a generalist, is I think much more important than it used to be.
If I'm putting together a product team today, I would really obsess about getting as many skill sets as possible for each person I hire.
What have you done that has allowed you to grow this fast with so few people?
People love the product.
That's the driver of the growth.
Today, my guest is Anton OCK.
Anton is co -founder and CEO of Lovable, which is essentially an AI engineer that takes an English prompt and codes a product for you in minutes.
You can then talk to it, iterate on the product, and then launch it to the world.
It's one of the fastest growing products in history, the fastest growing startup in Europe ever.
And as Anton describes, their goal for lovable is for it to be the last piece of software that anybody has to write because it'll be able to create all future products for us they launched just a few months ago in the first four weeks hit 4 million arr in the first two months crossed 10 million arr all with just 15 people absurd in our conversation we covered a lot of ground including a live demo of lovable how their team operates how they hire what does most enable their team to scale this quickly with so few people pro tips for using lovable how it all started how he recommends you build product
teams going forward with tools like this existing what skills will matter more and less going forward plus how to think about lovable versus competitors and so much more if you're trying to wrap your head around how product building will change with the rise of ai tools this episode is a must watch if you enjoy this podcast don't forget to subscribe and follow it in your favorite podcasting app or YouTube.
Also, if you become a yearly subscriber of my newsletter, you now get a year free of perplexity and notion and superhuman and linear and granola.
Check it out at Lenny's newsletter .com.
With that, I bring you Anton OCK.
This episode is brought to you by Cinch, the customer communications cloud.
Here's the thing about digital customer communications.
Whether you're sending marketing campaigns, verification codes, or account alerts, you need them to reach users reliably.
That's where Cinch comes in.
Over 150 ,000 businesses, including 8 of the top 10 largest tech companies globally, use Cinch's API to build messaging, email, and calling into their products.
And there's something big happening in messaging that product teams need to know about.
Rich Communication Services, or RCS.
Think of RCS as SMS 2 .0.
Instead of getting text from a random number, your users will see your verified company name and logo without needing to download anything new.
It's a more secure and branded experience.
Plus, you get features like interactive carousels and suggested replies.
And here's why this matters.
U .S. carriers are starting to adopt RCS.
Cinch is already helping major brands send RCS messages around the world.
And they're helping Lenny's podcast listeners get registered first before the rush hits the U .S.
market. Learn more it gets started at cinch .com slash Lenny.
That's S -I -N -C -H dot com slash Lenny.
This episode is brought to you by Persona, the adaptable identity platform that helps businesses fight fraud, meet compliance requirements, and build trust.
While you're listening to this right now, how do you know that you're really listening to me, Lenny?
These days, it's easier than ever for fraudsters to steal PII, faces, and identities.
That's where persona comes in persona helps leading companies like linkedin etsy and twilio securely verify individuals and businesses across the world what sets persona apart is its configurability every company has different needs depending on its industry use cases risk tolerance and user demographics that's why persona offers flexible building blocks that allow you to build tailored collection and verification flows that maximize conversion while minimizing risk plus persona's Orchestration tools automate your identity process so that you can fight rapidly shifting fraud and meet new waves
of regulation. Whether you're a startup or an enterprise business, Persona has a plan for you.
Learn more at withpersona .com slash Lenny.
Again, that's withpersona .com slash Lenny.
Anton, thank you so much for being here.
Welcome to the podcast.
It's a pleasure to talk to you, Lanny.
Great to be here. I don't know how you have time to do this podcast.
Your life must be insane these days with the pace at which you guys are scaling, just how much is changing in AI every day.
So I just extra appreciate you making time for this.
I think you said it's 10 .30 your time is when we're doing this.
I'm a bit tired, yes.
Mostly from the crazy pace of everything, but yes.
This is going to be an invigorating conversation.
You're not going to be able to sleep.
I'm sure, I'm sure.
Okay, so for folks that are maybe a little bit familiar with Lovable or not at all familiar, what is Lovable?
What's the simplest way to understand it?
I'd say Lovable is your personal AI software engineer.
You describe an idea, and then you get a fully working product from the AI.
And what this means is that entrepreneurs actually, today, they turn their ideas into real businesses.
um we have a lot of designers and product managers that create the first version of their product ideas to show to their teams and some of them become founders because of like their the empowerment from this um but also developers themselves they actually writing code or creating products much faster and um i mean the reason it's pretty obvious for me so I'll spell it out, but I'll spell it out.
The reason why we're doing Lovable is that I don't know about your mom, but my mom doesn't write code.
And almost all my friends from throughout my life reached out for help.
Like Anton, I need to build something.
How do I find a great software engineer?
And we're building for this 99 % of the population who don't write code.
um currently if you're technically inclined you get much further but over time naturally the way to build software is by just talking to an ai that's how we sit I love the way that you guys describe it and uh you didn't mention it but I think it's like building the last piece of software ever how do you how do you how do you phrase that yeah we say we say we're building the last piece of software the last piece of software okay we're gonna do a live demo but first of all can you just share some stats on the scale of this business at this point because it's quite absurd yeah so we launched lovable
less than three months ago and now we have 300 000 monthly active users and 30 of those 30 000 of those are actually paying and it's growing on this at the same rates like just almost only through organic word of mouth okay and i'll share a couple stats in terms of revenue just so folks know this and we'll have this in the intro too i think you guys hit 4 million arr in the first four weeks you had 10 million arr in the first two months with just 15 people you're the fastest growing startup in all of europe and you guys had to rewrite your entire code base recently and you couldn't ship any new features
for a while is that right that's that's right yeah people were saying like oh you're shipping so fast and we were all quite frustrated because we wrote our service in this kind of scripting language and then as we started scaling we just know we have to throw everything away and rewrite it in a more performant way.
Okay before we get to the demo last question you shared there's some companies that have started based on Lovable.
I didn't even know that so what are some examples of companies slash businesses that have launched off of Lovable now are actually companies?
I mentioned designers using lovable and and one of our early users harry he he started shipping real web apps to his clients instead of just shipping designs and then he went on to say okay wait i'm going to start an ai startup and and his he his company he like launched on product hunt and everything and making money is just like lets anyone upload their photo libraries and then it's cat legs the ais parses and categorizes it and if you go to launched .lovable .app like this is an app built with lovable which is a kind of a product product hunt version where you can see a lot of businesses or small
sass and featured there okay cool so we're going to come back to some of this stuff but let's get into a let's get into demo i i rarely do demos on this podcast but i'm finding that i think it's really important for people to see these products in action because in a large part This is the future of product building.
And a lot of people hear about, oh, yeah, AI is coming.
And I don't think a lot of people actually see what the latest tools are capable of.
And so I love showing these sorts of things on this podcast.
So Lenny, I was thinking, did you ever consider making a copy and build your own Airbnb?
I haven't. Let's go on.
How about you do that?
Let's do it. Let's do it.
Okay. So we're going to make our own Airbnb.
Okay. So I just put in the first prompt for an Airbnb clone.
Okay. And what's the prompt just for folks that aren't watching?
Two words, Airbnb clones.
That's the problem.
I like start simple.
And then what you get is that the AI says, okay, I'm going to go through what does a beautiful Airbnb clone look like?
And it goes through a bit of design decisions.
And then I'll zoom out to see more of it.
we have this just UI that is, I mean, it has all the nice things you would expect from an Airbnb clone where you see different categories and you can see two listings from Airbnb with login buttons and everything.
So far, it doesn't have the functionality of Airbnb, it just has the UI.
I would now ask for an improvement on some of the functionality.
Like if I'm switching category, I want to see different listings, let's say.
But if you have any thoughts on what we should build next, let me know.
Okay, and so you had this preloaded, so you didn't see how long it would take, but how long would this normally take for it to just write all this code and have it for you?
The first prompt takes 30 seconds.
30 seconds, okay. And it's like a very good copy of Airbnb.
Yeah, I love that you don't have to show the design, you just tell it Airbnb and it was...
Okay, so your question is, what would I want to add to my own version of Airbnb?
be what i've always wanted to explore buying the place that i look at just like is this for sale so what if we see what that would feel like if you're just like a way to buy buy a listing okay so let's let's how about we add i mean prompting is important here so let's be specific but we would ask um creating an add a button on the listing which says purchase this this airbnb home is that perfect is add a bottle and i'll be even more specific it will pop up a modal um to purchase the listing perfect and i love so i think something as you're typing i'm just going to share thoughts as you're doing
this so the site that you asked this ai engineer to build like it's actually a functioning website you can browse around it's not just a design the say obviously there's no like actual listings here like there's no actual houses here say you were trying to like actually build airbnb and you wanted to start adding like actual homes that plug into this how does that sort of step work so as you say this is just kind of the mock -up ui but it's also also interactive if i want to add login and add listing management then we will connect something called the backend so where data is stored where users
log information is stored and i can show you how to do that first let's just try out where we got with this short prompt of adding the purchase listing and it it didn't do exactly what i wanted i said add a button or i didn't say what a button should say here but it says book now And if I click book now, I get a booking confirmation.
So the AI was like, okay, it didn't really, it was probably surprised by you wanting to buy the listing since it's Airbnb, right?
So it still says book the listing, but it shows a pretty modal where I can click confirm and pay.
And then it says booking confirmed.
I'll just say real quick, I love that this is actually a really good example of why being a good product manager is important.
And a lot of wasted time happens when you're not clear about the problem you're trying to solve and why you're trying to solve it and all that kind of stuff.
So it's really cool that this is a use case where you have to be really good at explaining what it is you want.
And it's interesting.
You don't have to tell this AI why.
Humans want to understand why is this important.
Mostly you need to be very clear about what it is you're doing.
And I love that's a really strong PM skill.
PM is really good at that.
So we have to. Hey, explaining exactly what you expect and what you're not getting is even more important with AI than with humans.
But so I go into hooking up more of the actual functionality.
But first, I'll actually show you something like how was the fastest way to change what went wrong.
It's created buttons that say book now and I want them to say buy now.
And what I could do is to select this item and say, change it to buy now.
But what we just released is that you can actually edit this.
This is a fully functioning product, but you can edit it visually.
Like you do in Squarespace and Wix and so on.
So I'll just change the text to buy now.
And then it instantly changes.
It actually changes deep down in the code base, but it's very fast to do that.
So I think people listening to this and seeing this, if you're not aware, like this is the cutting edge of tools like this.
No other tool out there lets you generate code from an AI engineer and then actually just like change a small element of it of every other tool that I'm aware of.
You have to like ask the agent, do this for me, and then you hope that it does the right thing.
So this is a huge deal, which you just showed.
Now it says buy now.
And that's something you just launched.
Correct, we just launched this a few days ago.
But I want to go into building the full functionality.
But what it looks like is that you connect an open source backend as a service, and that's called Superbase.
And I have this instance to connect to that's completely empty, just like one click to set that up.
And now it's connected to the backend.
it's just like automatically generating and explaining um generating some code and explaining what i can do next and what i would do now is say let's let's add log in let's say let's add log in and where is it actually hosted on the back end everything general yeah so everything can be one click deployed and then it's running it's hosted by like a cloud vendor which is hosting i think a huge chunk of the internet it's called cloudflare um and the backend is hosted by the also a good cloud provider which is called superbase amazing okay uh let's wrap up the demo that was unless there's anything
else was there anything else really important you wanted to show no i mean okay i just explained what you what i would do next i would say okay let's add login um let's make the listings editable by the users so users can upload listings and then this is going to take a bit more time but with patience and good prompting skills you're going to get to full working Airbnb.
That was a really good piece to add so basically like this is getting to a place where it actually is not so different from actual Airbnb people can log in they can add their home you can add internal tools to add listings for your say sales team ops team basically it just will allow you to built a marketplace uh that looks a lot like airbnb um amazing okay thank you for the demo i think for a lot of people they're like yeah i've seen this kind of stuff for most people like holy shit it's unreal what like it's almost like we're taking for granted now you can ask uh an app to build you a whole
website and that costs probably like a few pennies it took like five minutes versus like it would have been tens of thousands and like weeks and weeks and months even build just a prototype i mean these tools as we see here they're already very good like it looks really good as well um but mainly i would say they're getting better very very fast and i'd say like one of the bigger bottlenecks is now they're not integrated into the current way that you have your existing products and so on but it is getting better so fast so fast i think the best thing for people who are interested in this or like
interested in just being a part of the future economies get your hands very dirty with these tools because being in the top 10 in using them is going to be to absolutely set you apart in the coming months and years so let's let me follow that thread so say you are magically able to sit next to everybody that is using lovable for the first time and you could just whisper a tip in their year to be successful with lovable what would that tip be it takes a lot to master using tools like lovable and being very curious and patient and i we have something called chat mode where you can just ask you know
like to understand like how does this work like is i'm not getting what i'm what i want here um am i missing something what should i do is is is the best way to be productive is also one of the best ways to just learn about how software engineering works which is you don't have to write the code anymore but it is useful to understand how software or how building products works so that so i think that's the patience and curiosity is is super useful the second part that we spoke about is that being i would if i would sit next to you I would probably say like, hey, you're not being super clear here.
Like, for example, don't say it doesn't work.
Just explain exactly what you're expecting and which parts are working and which parts are not working.
And that's a lot of that's something that a lot of people don't do naturally.
I love that. Like when you have an engineer you're working with, that is a very expensive mistake to miscommunicate something, to just forget about a feature, to forget about a requirement.
in here it's you do that and then like 30 seconds later you're like oh okay sorry that was wrong and then you could just try again that's true it might it might be more costly with humans uh okay and the first step so the first step is chat mode so you could just so your advice is chat with the what do you call it do you call it an agent you got what's like the the term for the thing that you were talking with uh yeah lovable is an agent just lovable yeah okay so you're Talking about lovable.
By the way, where did you, how did you decide on lovable is the name?
It's so sweet. I think it's all about building, I mean, a great product.
That's what I want more people to be able to do.
And the best word for a great product is that it's lovable.
A lot of jargon that I like to use to like emphasize what we should be striving for is building a minimum lovable product and then building a lovable product and building an absolutely lovable product.
So I took that jargon with me in the company name.
That is great. Absolute lovable product, ALP is the new MVP. Okay.
So we talked about this, the scale you guys have hit at this point.
I imagine it's far beyond 10 million ARR.
Do you share that at this point or are you keeping that private?
We don't anchor on the numbers, but I could probably do a 2x tweet about this quite soon, yes.
Okay. So it's far beyond 10 million an error at this point uh it's uh one of the fastest growing startups in history the fastest going startup in europe i want to zoom us back to the beginning what is the origin story about how did it all begin what was the journey to today i i think i was not impressed by like what people were doing with the large language models when after especially after i i was using them way back but when chat gpt came out they were starting to get really good at taking a human instruction and spitting out code and then um people in my team i was the cto at a yc startup
they felt like oh nanton you're exaggerating this is not going to change anything in the coming years so i wanted to prove a point and i created a open source tool called gpt engineer where you write something like create a snake game and then it spits out a lot of code a little different files and then opens the snake game and then i tweeted a video about that and um gpt engineer is to date the most popular uh open source tool to showcase the ability for large language models to create applications and it's like 50 something 50 something thousand github stars and like dozen of academic references
and i know that i'll just add that it like github shut you down because i thought it was some kind of attack the like how many stars you're getting how many people are using it right yeah so that was that that came later that that's with lovable lovable um earlier was always creating new projects on github when someone used lovable and it was we asked them is it fine like how what's the limits here they said oh there are no limits but once we started creating 15 000 projects per day.
So there were a lot of usage.
Then some engineer, when it was on call, maybe they woke up in the night and they saw their servers were taking too much load because of us.
So then they shut that down completely and we got this email that said, oh, you broke some kind of rules and we didn't know what was going on.
That's similar to a story I heard when ChatGPT was originally being trained.
Microsoft servers or blocked it because they thought it was some crawler and it was just actually like the very first version of chatGPT being trained on data.
Anyway, keep going.
So I built this tool called GPT Engineer and I was thinking about, I mean, we're seeing the biggest change humanity will ever see, I think.
Like before you had manual labor being taken over by machines, but now it's actually cognitive labor being being done better than humans by machines and what's the best way to have some kind of positive impact here it's not to make engineers more productive which is there's a lot of companies using it to make engineers more productive microsoft you build co -pilot and so on but it is to enable those who have like such a hard time finding people who are good at creating software.
That's been their absolute bottleneck and let them take their ideas and their dreams into reality.
So enabling more entrepreneurship and innovation by building the AI software engineer for anyone.
And then I grabbed a previous colleague of mine who has also been a founder, Fabian.
And I said, we should build something like GPT Engineer, but it has to be for the people who don't write code.
and that's the story.
Okay, and then that became lovable.
There was like the shift from open source into a product that anyone can use, but also pay for.
Makes sense. Okay, so from that point, I saw a stat that you started making a million dollars in ARR per week once you launched lovable.
Is that true? Yeah, so we launched, so we actually called the first version of the product like GPT Engineer App.
And that was very different in some ways.
And we launched that under a wait list and said, like, oh, yeah, we have this wait list and we got a lot of feedback and iterated.
And finally, when we thought the product was really good, we said, OK, now we have a lovable product.
And it was mainly on the AI that we did a lot of improvements.
Once we launched that, that was 21st of November.
So that's almost three months ago.
We just hit like one million ARR in a week and then it kept growing at that pace.
it's still growing at even faster than that pace.
Faster than one million ARR per week.
Holy shit. That sounds like product market fit to me.
You said that you did a lot of work on the back end.
I saw you tweet about this, that you guys figured out some kind of unlock on scalability, like a new scaling law that allowed you to build something like this.
What can you talk about there that kind of on the technical element allowed you to build something new and the successful?
There are many scaling laws, I would say, when you build AI systems.
And this one in particular is about when you put in more work, the product reliably gets better and better.
And what you've seen generally when you have AI building something is that it can get stuck in some place.
It starts, it's super good in the beginning and then it gets stuck.
what we did was to painstakingly identify places where it goes stuck and and there there's a different approaches but address like different ways how we do it but address the places where it gets like tuned entire system quantitatively and having a very fast feedback loop to improve it in the areas where it got stuck the most important areas it still does get stuck sometimes But that's the scaling law and we're still early in that scaling law I would say.
And so when you talk about things getting stuck it's like the AI agent just saying like I don't know what to do from this point or like they introduce some kind of bug.
Is that an example of getting stuck?
It introduces some kind of bug and then it's not smart enough to figure out how to get out of that bug.
I see and this is a common problem people have with tools like this is they like get to a certain point and then it's like well I don't know what to do I'm not an engineer.
Like here's a bug it's running into or the infrastructure is built the wrong way and so it sounds like one of the paths to solving that is what you're describing is you make the AI smarter to get to avoid more and more of these places they get stuck another is people just learning how to get AI unstuck uh there's something when we had omjad on the podcast from replate he said that this is like the main skill that he thinks people need to learn is how to unstuck ai when it runs into a problem uh just thoughts there i don't know anything along those lines come up as they say that i mean this is something that uh
is a problem today and it the frontier of where this is a problem is very rapidly like receding back so So what we did was we identified the most important areas.
So specifically adding login, creating data persistence, adding payment with Stripe.
Those are the things that we made sure it doesn't get stuck on, for example.
And the places where it gets stuck today is currently something where you can use being very good at understanding and getting unstuck.
But in the future, it won't be so important.
this experience is just going to not get stuck.
And I know you're not talking super in depth about this because this is one of your unfair advantages, this kind of stuff you figured out, so I'm not going to push too far.
I know you want not everyone to do exactly the same stuff.
So I want to zoom back to the pace of growth that you guys have seen.
One of the big stories, everyone's always looking at you guys, 15 people, 10 million ARR in two months.
That's absurd. I don't know if it's ever been done in history.
if so it's maybe a couple other ai startups recently how have you been able to do this what have you done that has allowed you to grow this fast with so few people i'd like to take credit of like having done everything end to end in the product um but what but we're building on top of taking on the oil here which we have discovered oil which is are the foundation models right um and then what you what you've done is that we're obsessed about what's the right way to present this to a user what's the interface for the human to get as much out of this as possible packaging together i i showed you
in the demo that you how you can add authentication and making this work seamlessly together as a whole that's what we've done and then people love the product that's what that's the driver of the growth uh the for getting awareness we've mainly been posting what we've shipped on social media that's that's how people know about us so building in public is is how people usually describe that so it's like uh i think it's like you guys have the advantage of the demos are just like holy shit you can do that and then you guys share the numbers that you guys are growing out so it's innately interesting
and shareable but i imagine most people have something interesting to share i guess is there anything that you think you did that other companies maybe haven't done that make the product so lovable that i mean the team is everything in building a great product so i i just give uh they give a big shout out to to team that has written the code i had written the code recent much of the code recently i would say um and the i mean you you want people who can ship really fast and have good taste for what is simple, what's the right abstractions.
And I think that's what we've done differently and have this obsession for just making it better and better and better.
This episode is brought to you by the Fundrise Flagship Fund.
Full disclosure, real estate investing is boring.
Prediction markets are exciting, meme coins are a thrill ride.
Even the stock market can swing wildly on a headline.
Hello, DeepSeek. But with real estate investing, there's no drama or adrenaline or excuses to refresh your portfolio every few minutes.
Just bland and boring stuff like diversification and dividends.
So you won't be surprised to learn that the Fundrise Flagship Real Estate Fund is a complete snooze fest.
The fund holds $1 .1 billion worth of institutional caliber real estate managed by a team of pros focused on steadily growing your net worth for decades to come.
See, boring. That's the point.
You can start investing in minutes and with as little as $10 by visiting fundrise .com slash Lenny.
Carefully consider the investment objectives, risks, charges, and expenses of the Fundrise Flagship Fund before investing.
Find this information and more in the fund's prospectus at fundrise .com slash flagship.
This is a paid ad. Okay, I want to come back to the team because I know you have a lot of thoughts there.
In terms of writing code, how much do you guys actually use AI to write the code that is building lovable how does that work on your team we have set up lovable so that we can change lovable with itself we have done that um since there is a lot of like hyper specific things um in terms of running a separate like we spin up as dedicated computer for each user it doesn't do everything Lovable doesn't do everything so we use the tools that are for developers not for the 99 % most of the time everyone uses AI all the time it's also a great course for experimentation and other tools like Cursor and stuff
like that I think Cursor is the one that almost everyone uses in the team yeah okay cool we did a survey recently on tools that my listeners and readers use in cursor like 17 % of all people that read my newsletter use cursor already which is absurd and you guys were in there too okay so kind of along these lines there's obviously other competitors and companies in this space so everyone's always wondering you bold replit cursor is a different kind of thing what's the simplest way to understand maybe how lovable might be different from say bolt and replit which i think are probably the closest
the packaging for non -technical people is what we what we aim for and and i showed you in the demo that you can edit the text like you can release to change the colors and so on instantly with without having to go into like code editor and without having to wait this about 30 seconds for the ai to to the full change so that's the big way that we think about packaging it and then for you know making sure that this can be used as productively as possible in a larger team something that's different from I think all the other tools is that it is synchronized with GitHub and that means that you can use
Cursor if you're or the people in your team that want to be more low level they can use cursor and while the people who don't want to mess and set up their local file system and commit to github and so on they can use lavables not getting stuck is i i think the most important thing for people and that's why we came we kind of entered the space late we haven't done the same type of marketing there's as many others and we're still um from the people that i talk to Ranked as the one that works most reliably.
I love it. Okay. So this point about how you can just use Lovable to build a lot of it for you and then get into Cursor to edit and tweak is a really big point.
And you're saying other companies aren't as good at that.
I don't know if any other does that.
I don't even let you do that.
Amazing. Okay. And then what's kind of like the vision for Lovable?
Like, what's the end state of this is this everybody can build anything they want sort of thing what's the simplest way to understand where you're going in the next i don't know five ten years i mean i have to say so we're building the last piece of software and it is inherently very hard to predict how the world looks like in five years these days it's very hard and but the last piece of software how i see that is that it's almost instant to go from what you want to change in a product or what you what product you want to build to having it fully working end -to -end integrated with any of your
existing systems or integrated with the kind of very powerful third -party providers already today you can just ask add a chat with open ai and then you get a chat with open ai in your in your product but um that's like just work working perfectly is the something that's coming in the coming two years I would say um and then after that there is a lot of things in building a product that is not just engineering side right and I think um an AI can be very useful in aggregating and understanding your users so like if you uh if you use the analytics tools you know that there's something quite common which is to
see how users have interactive with the product.
AIs can do that on an absolutely massive scale and propose changes to a human to say like, oh yeah, that sounds like a good change to make it a bit more intuitive.
And it can also automatically run, spin out A -B tests so that you can see with data are these improvements to the product.
So I think that's on the horizon as well quite soon.
Like what's interesting about this in one way is people wonder just what jobs will be more important what skills will be less important let me share a thought i have and then i want to get your take and see where you go with this it feels like what is getting more valuable is being good at figuring out what to build and then knowing if the thing you had built is correct and good and ready so it's like discovery ideation uh idea part of the step of launching a product and then it's like taste and and craft just like is this the thing is this going to solve people's problems because the building
now is being done more and more and it's interesting it used to be the reverse engineering was the hardest most valuable skill and now it's like figuring out what to build you could sit there and you could just tell it what to build and a lot of people get to your screen i'm sure and they're like i don't know what to build i don't know what people want and it's like that's the thing now so i just reactions to that and thoughts on what skills will matter more and less i mean if you're if you want to if you're a founder or you want to build something.
I totally agree that figuring out what are pain points and seeing like there are pro often currently solutions to every, some kind of solution to everything.
What is the, and how can you make this 10x better somehow like figuring that out is super important.
When you have an existing product, then I think taste and like refine it, tasting what is, what is good is even more of the important part the tech like the engineer skill set is still going to be important because that that helps you understand what are the constraints so what you can build and i just think a lot of software engineers are probably a bit scared now like okay am i out of a job and what's going to happen but they should see themselves as the people who translate the the problems that are by a human probably to technical solutions.
And but they do have to subtract themselves up a few steps, not just like looking up in their tech stack, like, oh, I can just do the front end changes.
Engineers or technical people are very good at understanding what are the constraints technically, and they should see themselves as that translators.
Is there like a, like, is it almost like you want to be, learn the eng manager skill of overseeing engineers versus like the actual engineering skill, or do you think it's still going to be really important to learn how to code and be really good at that?
I mean, doing a bit of everything, being in general, is I think much more important than it used to be.
And if I'm putting together a product team today, I would re -obsess about getting as much of as many skill sets as possible for each person I hire.
They should know how architecting a system works.
preferably they should know design they should know they should have product taste they should know how to talk to users i think everyone should be able to know should know a bit of all of that preferably easier said than done it's hard to find people that know all these things so let's segue to hiring and how you hire how many people do you have at this point is that something you share yeah now we're at 18 18 okay wow so i love that you it sounded like you're about to say oh we have a hundred people now no 18 okay so you went from 15 to 18 uh okay great so what do you look for when you're hiring
people the way i saw you describe it on twitter is you look for cracked engineers the best crack team in europe things like i guess just specifically what are you looking for when you're hiring i think the most important thing is that people care a lot and they're not just like oh i'm here for a job i'm here for being as a passenger on this journey but everyone should really care about the product the users and care a ton about the team how the team works together and that you're always contributing to making the team work more productively together and that like care or preferably obsession gets
you very long ways and then you do often want to have like absolute absolute superpower in some dimension to be able to understand and do as many things as possible like have this generalist brain that quickly learns any skill but be super super good in one dimension and that's for us that's mostly cramming as much out of AI out of the large language models and understanding the entire parameter space of what you can change to make our product perform better.
So how do you actually test for these things?
You know, like some of these things describe everything everyone's looking for, like they care about the user, they want to collaborate well.
Just like when you're, cause like you have 18 people building in a company that's growing more than a million AR every week.
Like that's an absurd scale.
And the people you've found are clearly world class.
and i think a lot of people are going to like want to hire the type of people you're hiring so when you're actually interviewing how do you suss out some of these things like their ai cramming skills their team building collaboration what do you actually do i ask people what they've done before and they these people that i'm describing they have often done something where they care a lot about what they've done before uh and dig into details about the technical things that they did and then um i mean we do the normal thing of giving showing a very hard problem that is a bit unorthodox that someone
hasn't seen before preferably and see how they think through the thinking reason through that i then something that i think is more uncommon is that we do i pretty much always have people join the work simulation for at least a day of an awful week.
Awesome. Okay, so work trial, that's awesome.
So basically they work with the team for at least a day.
You said like sometimes a week.
Yeah. And I love this point you made about they cared deeply about something they previously worked on and you look for just like obsession with the thing that they built last or something they worked on.
Like what percentage are engineers of these 18?
18 so 12 at least right code in at least part times 12 out 18 okay cool uh you're when we were setting up you're like oh our engineer's creating content now i think that's a cool example of of how people do a lot of different things yeah uh also okay so i have your job posting that you shared once of like your actual job description i'm going to read a few lines from it.
It's very inspired by Shackleton, right?
Would you agree? Cool.
I love it. By the way, did you write this or did you have AI write this job description where you like create an engineering job description?
In fact, let me read it to you.
I don't even know. You may not know what I'm referring to.
I'll read a few lines here.
Long hours, high pace.
Candidates must thrive under a high urgency under AGI timelines approaching.
Difficult mission ahead.
Honor and recognition in case of success those seeking comfortable work need not apply and then there's a few other things collaboration with other exceptional minds purpose larger than any normal engineering role generous share in the venture success amazing thank you thoughts yeah so I did get some help with the formatting of this but then I was mostly me doing the exact tracing of the different sentences so good and I love that you know to some people it's gonna be like holy shit i'm not signing up for this but to a lot of people the people you want is like yes this is exactly what i want to
be doing great amazing okay cool so so it feels like one of the elements of hiring here is uh create a really good filter to be clear about just how intense this is so that the people that want that are the ones drawn to you okay and then you're also you're in sweden uh fastest growing startup in europe ever thoughts on building in europe sweden versus the us slash san francisco yeah so this this ambition level that you you're talking about in the job ad is more uncommon in sweden and i think that is the like the biggest unlock that someone like me who sees that this is the like the time in human history
when you have the most impact per a worked hour and that's why we have to be super ambitious like just up the ambition level and then we can maybe retire and have ai take care of most things in society um that and and inspiring people to be this ambitious in the place where the average ambition is lower but the talent the raw talent is much more available is is a great recipe i think that's a great recipe so then and that's what i think it's some kind of advantage there and it's a bit of a double -edged sword but but it's some kind of advantage so i'm hearing is like there's there's incredible
people in europe they're just not uh they're harder to find and what i'm hearing is like the key is how do you suss them out and get them to want to talk to you yeah most people in europe they haven't thought that oh do going on an extremely ambitious mission is what i want to do so that's figuring out who those are is is a big part of it awesome okay i want to talk about prioritization i imagine all these things that i just shared about just like how ambitious this mission is how much you're doing the last piece of software you must have a bazillion things that people ask you to build that you
want to build what's your approach to deciding what to purge as an actually build i just top line i think identifying what is the biggest bottleneck what's the biggest product problem and iterating or fast on saying okay this is the biggest problem let's really resolve that the problem and then picking the next one um and not overthinking not like dreaming out the long roadmap that's my my default there's a very very simple algorithm um understanding what is the mostly the biggest problem is not always a simple simple problem i think yeah so we spend time as one should on talking to users list
reading up on what people are writing we have a feature board or people do a lot of requests, as you say.
And then when we pick one of the problems, we're quite engineering -led.
For a product like ours, it's hard to have product managers that are not engineers say, oh, this is what we should do now because the right solution to the problem might be entangled in things that are like technical details or they might be entangled in technical details of like, okay, yes, this is the biggest problem, but we should have this larger technical initiative that's going to solve all of these problems.
So it's quite engineering -led compared to many other product companies.
As it should. I'd be worried if you guys had a product manager at this point.
That would make no sense right now.
I imagine the answer is it's chaos and there's no actual defined process.
But just like what does it look like generally?
Like what's kind of the cadence you guys operate on?
How do you take an idea to like build it, spec it, launch it?
What does that look like if you have something?
If you look back like three months, we mainly said, OK, let's do this weekly planning.
We have like a fig jam board where we have all the main problems and then we have kind rank them which are which else do we focus to one of our focus on next or this week um and then we have a demo where we say okay this are this the things we shipped this week so to get everyone on the same page and we do have a bit more of a roll up now and where we say like here are we're going to make so sure you can support custom domains next they're going to add a collaboration uh after that and and like the biggest problem now or the biggest initiative now that solves the biggest problem is making the system
more agentic um and that has a bit of a longer roadmap but we still do the cadence of weekly planning these are the next the things we're focusing on this week it's mostly there's a good word for this that you i would want your help with but polish like we're fixing the bugs and polish this week and that was the planning on monday that was actually this week was uh polish polish week i love that uh how far is this roadmap that you are now having i mean it's uh clear over the coming month but it stretches out three months and then but within within one month it's probably going to look a bit different
okay and then what are the tools you use just for folks that want to understand like the latest tools so you said fig jam what else is in that stack of tools i mean we do so many things in our company in linear because it's just us an amazing product so we do talent application tracking in linear oh and after going through and and this thing a lot of the other made custom made tools for that uh linear and then fig jam so simple uh how soon until one of your engineers is an agent engineer an ai engineer do you think do you have a sense i'd love to dig into what what does that question actually
mean um i think we've been talking about like oh ai that would require um or something playing chess that's that's ai like if you even a computer can play chess that's ai and now that's like oh no that's a chess program and we're always shifting this forward and forward um i think anything that a human doesn't do is just a smart computer system right so like what isn't when is something when is a software engineer and agent i think it's always going to be just we're building and lovable is just an interface that humans interact with create the software that they want.
And then how we solve that?
Is that going to be an agent under some definition?
Yeah, sure, I think so.
But that's less important to me.
Okay, I like that. Let me ask this.
You guys are moving super fast scaling like crazy.
You described a little bit about your process, weekly planning, big jamboard of ideas and now there's a roadmap that you're kind of thinking out in the future.
Is there anything else that you found helps you move this fast that gives you a lot of leverage over the small team you have to ship quickly and move fast that you haven't already mentioned we we work from the office most of the time i think it's pretty nice then you can say like hey i think we're thinking wrong about this thing or like shouldn't we actually do this other thing and especially i think lunch like eating lunch together is a pretty productive hour where you're cross -pollinating i mean people are constantly thinking subconsciously as well about the how to solve the different problems
and which the most important ones are and then being in office um has this like focus or most of the time you should be focused but you also have this like high bandwidth where everyone has a bit unstructured communication i love that uh the answer to uh the ceo of a company that's the one of the most advanced ai tools in the world is one of your answers to how to move fast is like lunch together i love that that's so human and so it makes all the sense in the world but i love that that's still a part of this yeah okay you talked about this kind of on the same thread you talked about if you were to
start in a team like a new product team today say you were head of product somewhere or head of rpm vp of product somewhere building a new product team scaling a product team what would you do going forward that's different from what people have done in the past in terms of who you're hiring how you're structuring them that kind of thing just like what do you think people should be thinking as they build product teams going forward knowing tools like lovable exist and all the other stuff that's going on i mean everyone should be excited about using ai I think that's a pretty big one.
And then the team working really well together is like the launch.
You have to like sit down and solve problems together.
You should, at the bottleneck for most products these days, it's not going to be as much on the engineering, but having good taste, good intuition about your users.
and um that i mean engineers and everyone preferably in the team should have that like willingness at least to want to go through that motion and listen to the users um and truly understand what they care about what's kind of like the background of most of the engineers and people you've hired are they like is there anything like in common are they just like super uh impressive humans generally like you know champions of programming contests stuff like that i don't know like what are some attributes of the folks you've hired so far i think raw cognitive capability is the strongest like diamond
the strongest correlate of being at lovable but there there is this startup mindset that i i think is also very strong and being a bit more being much more interested in moving very fast and iterating fast and having like a lot of structure a lot of process and thinking about the business as a whole more than thinking about my specific profession my specific craft that i'm see myself like wanting to dig in into only amazing okay so smart like very smart entrepreneurial acts like an owner yeah doesn't just uh isn't just like this isn't just a job but they feel like they actually have agency okay this is great there's
something you said kind of along these lines that uh i think is important that one of the things that gets you excited about what you're building is giving people superpowers and especially people that don't add a code basically 99 of people is there anything along those lines that you think is important to share?
It's very clear to most people who have been engineers or been founders that there's so many that have failed in their endeavors because they didn't have someone that know how to solve the technical parts.
And now that we're close to having people know that this tools exist and they solve everything, it's going to be an explosion of entrepreneurship and better software product.
We're not going to settle for all the annoying, bad technology that we use today.
And everyone who has an idea is going to say like, okay, I'm going to build this thing and show you that this is the best version of the product or what our company should be doing instead of having long meetings or, like, writing up documents.
So it's going to be empowering across a lot of different professions and places in the world.
What's next for Lovable?
What's kind of, like, the next few things they might launch as this episode comes out?
I mentioned this agentic behavior.
And when I say agentic, what it means is that you give more freedom to the system to decide what happens next.
It might want to write a test, run those tests, and say, like oh the test failed let's fix those so so that's um one of the big unlocks for getting further faster and on then there's some more like obvious things that you want to do to go all the way to easily go all the way to making money with lovable and that's that's like how do you set up so that it's hosted on your specific domain how do you collaborate the seamlessly with your team and making that that easier so that are just obvious things um and something we're thinking about is to help just founders succeed after they built their first
version and like how do they get more users how do they get feedback how do they get the word out if they build something useful i was just gonna say that that's that's exactly where my mind went is like everyone's gonna be building all these things no one's ever gonna get any traction with these tools because no one knows how to find users get anyone to basically go to market and growth is like a whole different skill so that is so cool that you're thinking about that how do we run some paid ads for you how do we think about a ceo how do we think about word of mouth reality referrals that is very
cool okay yeah we already have some playbooks that we that we help the people building with how do you do those things that you can find up on a blog oh interestingly this makes me want to buy some meta stock because all these apps that everyone's building they're going to be running paid ads on facebook and google oh my god what a good business those other guys got uh i want to come back to you said that you can work on your existing code base this is actually a big question for a lot of people they see all these tools they're all like amazing for prototypes and concepting you talked about how
you can actually do this within your existing code base use lovable let me correct you there you cannot use it on any existing code base um we kind of have a research preview of importing your code base but what you can do is if you start in lovable then you can have engineers editing it how in whatever tool they want to use for editing it okay cool that's great clarification so i guess just for people because a lot of like most listeners who are not building something brand new they're working within an existing product so you're saying that that is coming you can use lovable in the future in some
form with your existing app and product.
Great. Wow, that's huge.
Okay, because that's basically the most people.
So that's going to be a big deal.
Okay, final question.
We have a segment on this podcast called Failure Corner.
Okay. Where most people come to this podcast, they share all these stories of success and everything's going great and here's all the things always winning.
You guys, this is a good example.
It's just up and to the right, the fastest growing product ever.
what's an example on something totally failed in the course of your career and what did you learn from that?
I'm a bit hard -pressed to find something that totally failed, but I think there's a bit of a product lesson where I was the first employee at an AI startup here in Stockholm called Sunalabs.
And the premise was just, okay, so humans learn in different ways.
If you personalize that you get two standard deviations more uh effective learning so there uh there is a lot of products like education software that helps you learn um that is not personalized and we could be we were building an api to personalize learning and the i mean the ai and sold.
It was pretty good.
But the thing that we were doing in the end was to say like, okay, here's this product.
Someone has to build a product or some way to learn, be it like English and think Duolingo.
And then the people that have that product have to use this advanced AI APIs to start making it personalized.
And it was a very hard like retrofitting like oh you have to switch out the engine and put in this ai and it's it's well the big learning here is in that it didn't work very well for the company i mean the company wasn't super successful in this the big learning is that you have to start with like how is this product working end -to -end and then add ai or think where should we add ai so that was a big learning for me that um you you really want to see like how the what is the big picture of the user what's the big picture of how this how do you think the user experience should be and then add something
with ai uh to solve specific problems and now some of the labs is doing great but it's not on top of that product specifically that's i think it's a lot of people hear this and they're like of course but i think it's so hard to actually remember this point when you're have some cool tech and you're like holy shit everyone needs to try this they're gonna love it and then you don't realize like no one actually cares if it's not solving a problem for them you know there's like a lot of novelty products that like everyone want to use for a little bit and then like forget and it's not I don't actually need
this often and so I like what this makes me think about is there's all these product lessons for what is likely to help your product be successful and an app like a tool like Lovable can help you do this.
Because if someone is building something, you can guide them.
Okay, what's the problem you're solving for somebody?
How many people have this problem?
How much does this matter to them?
Maybe we should add like the Lenny mode.
She activates in Lovable.
It activates like this product coach.
Like when she asks infinite questions, you're like, no, wait, hold on.
Why are you doing this?
What? Let's take a step back.
Crazy. Everyone's going to be like, get out of my way.
Yeah, exactly. What's your, yeah, what's your experiment plan?
That's actually, I think there's actually a big opportunity there to save people because, you know, there's like a play around with this thing and then there's like, okay, but really is this anything people actually want?
I love it. Can we call it Lenny mode?
Is that a fine with you?
100%. Awesome. Let's do it.
I'll license you, no cost.
Sure. Okay. Okay. We made a deal here.
let's do it okay uh Anton is there anything else that you wanted to share anything you want to leave listeners with uh before I let you go and go to sleep I think again the the world is changing quickly and it's very fun you should see this like have fun in all of this change um and the best thing you can do for your current profession or if you want to have a new job is to be in the top one percent in knowing how to use the ai tools so go out there use use lovable use other ai tools and become um make sure to understand or try to understand as much as possible in how to use them productively
um that that's that's something i i tell all my friends generally and i like love the audience to know as well okay well i gotta make try to make this even more specific for people uh how do you know if you're in the top one percent like what's like a heuristic almost of like slash how do you get there is it just use it a hundred times a day what else what can you recommend yeah i i think if you spend a full week on trying to reach an outcome right the best way to learn is like i want to do this thing and then i want to use ai to do that thing uh and you spend a full week you're you're in the top
one percent in the global population if you have friends that you surround yourself with friends who have this obsession or they also care a lot about this then you'd be quickly in the top 0 .1%.
So what I'm hearing is like find a problem that can be solved.
Like find a problem, a pain point for yourself or someone and then end to end like fully solve that problem.
Spend a week getting from idea to like a thing that somebody's actually using.
And you're in the top 1%.
I think at the top 1 % by just spending a full week and making like asking ai if you don't understand so making sure that you can understand yeah like that's the thing people forget you just ask like like you would would you ask the chat feature of lovable in this case or would you go to cloud or chat gpt to ask for advice i mean my recommendation here if your product is to use lovable to build software and learn that ai tool if you're and then you should use chat mode and chat mode i have to add is something you activate in your user profile it's not launched like in the main product so it's in labs
but if you add that flag then you can use chat mode if you're if you want to learn some other ai tool then you should i mean ask that tool or ask claude about how how that topic that domain works okay amazing uh where can people find you where they can where do they find lovable and how can listeners be useful to you?
Lovable posts, updates, and memes on lovable underscore dev on Twitter.
We post things on LinkedIn as well and there are a lot of things coming out and changing in how we build software.
So you can follow lovable underscore dev and you can follow me at Anton Osika at Twitter.
I'd love more feedback on where people see this is a huge change for them and we there are a lot of people posting about that on twitter but there's a we have a discord where you can share like oh this is how i use lovable and it was super useful to me and and uh feedback dot lovable dot dev you can give you can ask for uh new features there's a lot of people asking and outboting what features you want next so and that's super useful that's the most important thing for us we just want to solve people's problems.
Amazing. Anton, you're doing incredible work.
What a journey. I'm excited to have you back someday when we see more chapters of this journey.
I have a lot more to learn.
As do we all. That's why people listen to this podcast.
Anton, thank you so much for being here.
Thank you so much, Lenny.
Bye, everyone. Thank you so much for listening.
If you found this valuable, you can subscribe to the show on Apple Podcasts, Spotify, or your favorite podcast app.
Also, please consider giving us a rating or leaving a review as that really helps other listeners find the podcast you can find all past episodes or learn more about the show at lenny's podcast .com see you in the next episode
