From Zero to One: Anton's Story

0:00You need to be more like who are the

0:01users? How are they interacting? When

0:02are they interacting? What should the UX

0:04be like? I felt I could do it better

0:05myself. Extremely unafraid to just say

0:07this is [ __ ] Why are we doing it like

0:09this? Many companies that's the most

0:11important thing. In the future, humans

0:12are not going to look at code. They're

0:14not going to write code. We should

0:15obviously not be building for those

0:16developers. We should be building for

0:18everyone else. We kind of had product

0:20market fit and we kind of lost it. Yeah.

0:21>> At each point in the journey, we could

0:24have done a lot like some of things that

0:26would have made it into a extremely

0:27successful story like which is a bit

0:29sad. The most important things you can

0:31do to improve a product is to

0:43>> let's go.

0:44>> Let's go. Okay. So, welcome Anton. It

0:47was very clear for us that you would be

0:49our first guest for Silicon Valhava. You

0:52truly embody the entire Silicon Valha

0:55movement and everything beyond that. So,

0:58I really want to start off asking how

1:01are you doing? Like, how's your mental

1:03state?

1:03>> I'm pretty good. We had a super intense

1:06week and then yesterday we had Iris

1:08lovable summer party and we like

1:10everyone went swimming together

1:12afterwards. It was a bit crazy like

1:14early stage vibes and then now I woke up

1:16and went to this bearish boot camp. So,

1:19I'm I'm a bit tired coming into the into

1:21this podcast session here. Well,

1:24everyone's pulling you in every

1:27direction right now. If you have to

1:29guess, how many unread messages do you

1:32have throughout all social media

1:35channels?

1:36>> Okay. Um, I think there's 150 WhatsApp

1:40and 100 the iMessage and then the the

1:44other one that's okay. LinkedIn is like

1:46thousands. I think in emails emails of

1:49maybe 10,000 but a few hundred that I

1:53should get back to I guess.

1:54>> Oh wow.

1:55>> Last few weeks I there's been a lot of

1:57urgent important things. So I it's been

2:00building up.

2:01>> But you're using AI tool to manage your

2:02inbox.

2:03>> Yeah. You said shortwave is good and now

2:04I'm using short wave.

2:05>> Exactly. You're using shortwave. Yeah.

2:06It was Alexander from EQT who mentioned

2:08it first. So a big prop to Alexander.

2:10>> It saves me like a small part of the

2:13time.

2:15Thank you. But yeah, I think it's a very

2:18interesting topic how you should manage

2:19your inboxing cadence and what

2:21expectations you should set in terms of

2:22how people can reach you. Generally, my

2:24expectations are like, well, the only

2:26reliable way to talk to me is to find me

2:29and shake me physically or to call me

2:31and that's might work.

2:33>> Yeah. No, I can uh I can agree on that.

2:36And I always tell people who reach out

2:38to me because they want to reach out to

2:39you and always say guys like you need to

2:42find Anton. That's the best way. just

2:43find him.

2:44>> But you know talking about finding Anton

2:47uh you know we want to drill in a bit

2:50into your childhood and who you actually

2:52are and what shaped you growing up. So

2:54you mentioned already you know that you

2:57know you your father has an interesting

2:59story also and uh you know let's talk a

3:02bit about your upbringing your parents

3:04your father and how that shaped you.

3:06>> Okay. Yeah. My mom and dad they know met

3:12when they were teenagers and then uh

3:14they had they got crap pregnant when my

3:17dad was 19. My mom was 20.

3:19>> 19.

3:20>> Yeah, I was 19 and we lived in like a

3:22super nice house from was my

3:24grandparents out in Brahma. We didn't

3:25have any money. I think my my dad worked

3:28like side hustle things.

3:30>> Um and I went to a school that's called

3:34like a boulder school.

3:35>> Yeah. which is a lot of freedom and much

3:38less like oh you have to do all this

3:39homework and like you get um filled up

3:43with things to do. So it kind of

3:45triggers more creativity among among

3:47kids. I think for some kids I think it's

3:49pretty bad because you should actually

3:51learn learn discipline

3:53>> but I was super introverted when I get

3:57from I think kindergarten was more like

3:59I felt at home but in in in school super

4:02introverted

4:03>> um and was like walking around thinking

4:05about and how things work and was

4:07obsessed about how things work and

4:10technology. And then I think when you

4:12got into like 12th grade, I started to

4:14be more um like, oh, you can just do

4:18things. You know the quote that's the

4:20tech Twitter

4:22>> nice just do it,

4:23>> but like you can just do things and then

4:25see what happens. And and some pretty

4:27good teacher because there were very

4:29many weird teachers in that school uh

4:31had this session where like everyone's

4:33supposed to share something they learned

4:34in life. Yeah. And I think when it came

4:36to me, I was a bit surprised by my

4:37answer, which was like it's fun to

4:38optimize things and like think about how

4:40to optimize things, which is a big big

4:42part of of how I like to do things as

4:44well. But when you actually just go out

4:47and do things in the world,

4:49that's when you learn things and when

4:51things happen. And I've had a lot of

4:52that attitude since then. I think a lot

4:55of bias to action. Yeah. And some of the

4:58things that that always always excited

4:59me outside of um

5:04like technology is to build things. And

5:08um my dad with my brother uh he he said

5:12like oh you know we could try to build

5:15in Swedish it's called a lad bill.

5:17>> Yeah. Wooden car that you build yourself

5:19that you can go downhill with.

5:21>> Yeah. So I think in the neighborhood

5:22where we grew up there was like a lad

5:24bills competition and then uh there was

5:27this thing like oh you know you could

5:28put a big battery and put in an engine

5:30on it and that was like one of the

5:31things that triggered me to create

5:33things and cool

5:34>> this thing of that you can create build

5:37things and create things has always been

5:38a big driver in what what I want to do

5:41and I think many people find that in

5:44themselves when they find a good outlet

5:46a good to tool. So like I I think

5:49lovable is like the biggest expression

5:51of that in my life in making people be

5:55able to create. But I did some like a

5:57lot of stupid things building. I like to

6:00mix weird things together. So I found

6:03this recipe book with some friends where

6:05you could create bombs. So you like mix

6:08sulfur and a lot of things. And we uh

6:11realized that oh you you can't like

6:13either you have to buy vodka from Poland

6:15from some weird guy or you can just mix

6:18sugar and apple juice and with wine.

6:21>> Wow.

6:22>> So we did lots of that like dozens of

6:24liters of wine.

6:26>> But it seems like you know you're

6:28engineer by heart. Uh but you are quite

6:31creative and you've been quite quite

6:33creative since you know younger ages.

6:36>> I I think yes. they have too many ideas

6:40and like

6:41>> hard time to like really fully execute

6:43on them I think because there's too many

6:46good ideas and then the next one comes

6:48in and so on.

6:49>> Yeah.

6:49>> Uh so that's been like learning how to

6:52be able to take make something like all

6:55the way which I think is the hard part.

6:56>> Yeah.

6:57>> Uh has been what's actually been driving

6:59my success.

7:00>> Yeah. And that's you know what DK has

7:02said also you know ideas are easy

7:04execution is hard and it's a cliche but

7:07>> but you know looking at your household

7:08would you say that you know you come

7:10from a more entrepreneurial academic or

7:13what type of upbringing if question so

7:16from my dad's side he he grew up with

7:17like nine siblings super poor as well u

7:20and but he I think he's like engineer

7:22like pretty his dad was an engineer so

7:24on and had a like rough he had a rough

7:27childhood and then my

7:29>> mom's side was also

7:31math and

7:33>> doctor's backgrounds but we grew up with

7:35like we had to get we were we were on

7:37social welfare like that's all we had

7:39four kids right I had three siblings

7:41>> you were four in total

7:42>> yes yes and I was like one of my angst

7:45when I was a kid was that I had like

7:47others had like this nice clothes and I

7:49always had like maybe some old clothes

7:50for my siblings

7:51>> when I became a teenager I stopped

7:53caring a lot about [ __ ] and had some

7:56really good friends that were also like

7:58ner just nerdy like me but very outwards

8:01going.

8:01>> Yeah.

8:01>> And those were also many of some of them

8:03were the ones I went to

8:05>> um university studying physics.

8:07>> That's really cool.

8:08>> And that that's Yeah. I think that the

8:10group of friends really shaped me.

8:11>> Who was fifth grade? You mentioned that

8:13you were introverted but you know

8:15>> fifth grade. Yeah. I think was

8:16>> were you men in the classroom or

8:18>> No, I think it was like so some at some

8:21point I think that fifth grade people

8:23called me Einstein. I was still like

8:24there pretty quiet but then there was

8:27the one there was an interesting

8:28technical problem I always had answer to

8:30it something like that

8:31>> you already mentioned a bit university

8:33and and you studied engineering physics

8:35at KTH right uh was becoming an engineer

8:40always a given to you or did you

8:42actually consider you know studying

8:44something else or even studying abroad

8:46and not in Sweden

8:47>> you asked me who this vterusa is like

8:49he's actually my uncle and he was he

8:51took it quite seriously to me like anon

8:52you you should you see him you have a

8:54lot of potential, you should think

8:55seriously about what you're going to do

8:57>> and you should study hard. Um, and

9:00whileas my dad was more like this doing

9:02a lot of hustle things like building

9:05this car. Um, and now he has bought a

9:08big farm and like he works completely

9:10all the time on different a lot of

9:12different initiatives but not as like

9:14long-term view as as um

9:17my uncle who told me like you should

9:19maybe you want to be an engineer like

9:21this is a very good path. Um and I I

9:24remember going out and seeing like okay

9:26computers is what I'm

9:28>> been doing a lot to code when I was 12

9:31after seeing the matrix the movie like I

9:33I want to know this.

9:35>> Uh but someone told me that uh engineer

9:38physics is like it's Sweden the most

9:40>> general education where you can do

9:42anything afterwards and you um can feel

9:46this hunger for understanding things

9:48like how do things work. So, so that was

9:51uh being able to do that, going into

9:52research afterwards is what I made me

9:54decide that.

9:56>> Looking back at your childhood,

9:57>> would you say that you know you have any

10:00chips on your shoulder because you also

10:02have a brother?

10:03>> Yeah.

10:03>> Who is or older or younger?

10:06>> He's younger.

10:07>> Younger. Where are you out of the four?

10:08You are

10:09>> 30.

10:10>> Your brother is also an entrepreneur.

10:12>> Did you have a lot of competition at

10:14home between the siblings? you know what

10:17how was it going up and and you know

10:20yeah your mother is an entrepreneur

10:21>> so I think like being super quantitative

10:24about this I think it's a low

10:26probability that there's like something

10:27deep freudian about my childhood here

10:31I just like have this drive to do to be

10:34creative and make things uh like happen

10:38in the real world

10:39>> and I think that drive is what's like

10:41why I work super hard

10:44>> I guess they could be the chance there

10:46could be something there in that uh I

10:49was insecure as a kid and I like needed

10:51then when I could sort of prove myself

10:53that has been a big driver for me yeah I

10:55but I don't think so

10:57>> maybe I'll do psycho analysis at some

10:58point I'll be able to give you good

10:59answers

11:00>> but it's still clear you know that you

11:02were creative you were hustling around

11:04but you also have a very mathematical

11:06mind engineering mind also it's a pretty

11:09rare combination

11:11>> so today sauna is one of the most

11:14well-known own AI startups in Stockholm,

11:17but you know that wasn't always the

11:19case. So I would love to go back to your

11:22time at Sauna and you know going from

11:25CERN to Afield to joining Sauna as a

11:29founding engineer in 2017 if that's

11:31correct. How did that happen? How did

11:34you how did you meet Joe?

11:35>> Let's go back before 2014. I think there

11:38started to be a lot of

11:39>> um development in deep learning and

11:42there was there was

11:43>> after moment in 2012.

11:45>> Yeah. And now I think quite many people

11:47read like papers preprints on archive on

11:50what like how things work. But uh in

11:532014 I I figured that oh you can

11:55actually read this preprint and it's

11:56super cool like it's this m of math and

11:58so on but you can understand how these

12:00learning systems that are like human

12:01brains work. So I started to read those

12:04and talk about them thanks to other

12:07friends from physics that uh we like

12:09wanted to meet up at Starbucks once per

12:12week uh and do I think it was something

12:15like uh shield coding or something and

12:18then we changed it to Stockholm AI

12:21>> and that was that was around 2014 2015.

12:24Um and then I think I did it. Um there

12:28was then there were some big events in

12:30stock. Um that maybe already do you

12:34remember who organized the first ones?

12:35>> K I know natural lun was

12:38>> yeah we were

12:39>> we organized at also but that was later

12:41on. Yeah, I'm trying to remember this

12:43one Dutch guy who who organized.

12:45Anyways, there were some big big events

12:47as well and uh on one of them I met uh

12:49Anna

12:50>> who who is

12:52>> has this background in marketing and was

12:57Joel's co-founder.

12:59>> So Anna was Joel's co-founder. I met her

13:01at one of the stock bigger stock events.

13:04>> Um and

13:05>> when was this like 2016 17?

13:07>> Yeah. So this was 2016 I think. Yeah.

13:09Yeah. And then me and some other stuff

13:11spoke about how do we make stock amazing

13:13and then she was like Anthony you should

13:14come and meet Joel for lunch whatever

13:17and then yeah Joel was very interesting

13:21guy at the time.

13:22>> How old was Joel back then?

13:24>> He wasn't even 20. I think he was 19 and

13:28he

13:29>> we were same class Joel and I

13:31>> No way. You're that young, am I?

13:34>> I'm too young.

13:37>> And Joel had a lot of ambition and uh uh

13:43I love loved his like energy and and he

13:46had uh he was he understood that you

13:48could put together a lot of smart people

13:50and build something amazing.

13:51>> Yeah.

13:52>> Uh so I was like okay I mean and then we

13:55worked together on like okay what are we

13:58going to do here? we're going to use AI

13:59for learning faster and there was a lot

14:02of like super academic really good ideas

14:04that could have worked.

14:06>> Um and then there were more like

14:08engineering approaches of like just

14:11optimizing some an API that helps other

14:15developers build a learning product.

14:16Yeah.

14:16>> And we kind of pivoted around there. We

14:18got in some really smart people into the

14:21team. Yeah. Yeah.

14:22>> And that's like getting those really

14:23really smart people in the team also. So

14:25what ultimately led to Sana becoming

14:27super successful she is today right and

14:29Joe was like extremely high standards

14:32and in like around culture and around

14:36what type of individuals you want on the

14:38team

14:40>> and yeah it was a super fun journey.

14:42There was there was a lot of wasted time

14:44on doing crazy AI [ __ ] that didn't lead

14:46to anything but we won like big machine

14:49learning competitions and so on because

14:51team was very good.

14:52>> Did Oscar text work at the same time?

14:54>> Yeah, we also over overlapped. Yeah.

14:55>> Yeah,

14:55>> it was great. And I think for the

14:57audience to put things into context like

14:58Oscar is one of the actually researcher

15:01at Google brain who was quite seinal in

15:05the creation of the transformers

15:07architecture

15:08>> uh and and self attention also

15:11>> uh you know speaking about the level you

15:13also had back then at sana.

15:15>> Yeah but I I mean he's super impressive.

15:16I mean there was other people I'm even

15:19more impressed on by I think and is

15:21amazing.

15:23>> He's amazing. Yeah. If um Joel could

15:25describe the 2017 Anton, what would be

15:28>> 2017? I think he would say I was like

15:32bold and uh like understood things

15:36faster than anyone else. I was

15:39overconfident and sometimes difficult to

15:43work with

15:44>> to be honest.

15:45>> I guess like there was not as much

15:46maturity as there is today.

15:48>> And you were mentioning that you pivoted

15:50a lot. What what is your biggest lesson

15:54from sauna going zero to one especially

15:57around pivot? Yeah, I mean one of the

16:00thing was like before I joined Jo was

16:02like, "Oh, we're going to build this app

16:03where we can learn." And I think we

16:04should have done that, but Joel was just

16:06like, "No, no, no. We have to build an

16:08API and build an uh like focus on the

16:13pure AI and then um someone else can

16:16build products around that."

16:17>> And I think uh it was a big learning

16:20that that's sounds cool and could work

16:23like but it's very abstract. Um and

16:28the yeah the learning is basically

16:31the the most important things you can do

16:34to improve a learning experience or like

16:35a generally a product is to

16:38>> um be very very close to how looking at

16:41how a user is using the product and

16:43taking this super high level big bet

16:45decisions on how it should work.

16:46>> Yeah.

16:47>> Not like optimizing in some sequence of

16:49learning materials. M

16:51>> um and that's uh that's why just an API

16:55doesn't make as much sense. You need to

16:56be more like hey wait so who are the

16:58users? How are they interacting? When

16:59are they interacting?

17:01>> What should the UX be like?

17:02>> Um and I think that was like the the

17:06biggest realization we should have

17:09should have thought about much earlier.

17:11>> Um there's also some big realization on

17:14like how you do

17:16um go to market

17:18>> I think from the time. So Son was

17:21obviously like very salesled and did

17:22that well like that's why it's

17:24successful today. Yeah,

17:25>> I think it was a pretty good approach

17:28like make maybe

17:30um selling fewer products at the same

17:32time because we had like the API for for

17:36learning then we had the voice product

17:38and we had like an analytics product

17:40>> and

17:40>> yeah the voice product that was spin out

17:42that was sold too

17:43>> but that was also really cool because it

17:45was successful like that voice product

17:46was state-of-the-art in the world

17:48>> um where you could get on your

17:50pronunciation and there are lots of

17:51successful apps doing doing only just

17:53that right

17:53>> yeah and there's now a startup actually

17:55that took that technology.

17:56>> Yeah, they acquired that technology.

17:57Acquired that. Yeah. And there was the

17:59the part of the first part was also

18:01acquired like the

18:02>> AI to learn faster.

18:04>> Exactly.

18:04>> Right. Was that the reason why you left

18:07SA or what was the actual reason why you

18:09left and how was your relationship with

18:11Joel when you left?

18:13>> Yeah, I think it was uh good. We me and

18:15Joel didn't work so closely together at

18:18SA. I was a bit frustrated about like

18:20how we had uh

18:24pivoted so many times. I think that

18:25naturally like that's generally what

18:26happens when you people pivoted a lot of

18:28times. Yeah.

18:28>> And then

18:30>> uh and I felt I could do it better

18:31myself.

18:32>> So I was excited to work with two

18:35wonderful people who uh were more

18:37similar to myself in a few different

18:39dimensions.

18:40>> Yeah. So that was all and lonely at the

18:43>> coming to that are you the kind of

18:45person who works best under someone or

18:48actually you have to be in control and

18:50make the last call.

18:52>> Um it's a good good question. I think um

18:57generally if I'm like very open I think

19:03I don't add a lot of structure on how I

19:07work naturally

19:09and and now I find ways to force that

19:11structure around on me when I'm like I'm

19:13leading everything.

19:14>> Um it is good for me to have someone

19:16that tells me like

19:18let's let's keep it simpler. But the

19:22reason why because I have like been

19:24thinking should I be CEO here at Lovable

19:27>> or should I find someone that does like

19:28over the CEO partner thing and like I

19:31run I do what I do best and the thing is

19:34if I'm if I have like have the last day

19:36I have a very very easy time to see

19:39what's wrong how it should be better and

19:41be extremely unafraid to just say this

19:43is [ __ ] like or say it in a nice way but

19:46say like why are we doing it like this?

19:48Why aren't like this is isn't this

19:50obviously better? And then make things

19:52um make us stop doing stupid things.

19:54Make people work faster. Make the right

19:56people to talk to each other.

19:57>> Um and that

20:00uh some in many companies that's the

20:02most important thing in many stage

20:04that's the most important thing. I've

20:05been find trying to look at people who

20:07like

20:07>> is can I delegate these two things to

20:09someone else? No, I cannot delegate.

20:11>> I don't trust anyone to delegate.

20:13>> I'm looking forward to find them. Of

20:15course,

20:16>> getting that kind of exposure being in a

20:19early stage startup, is that something

20:21you would recommend to all of the people

20:24out there who are graduating university,

20:27>> thinking about what's next, not going to

20:31perhaps one of the large investment

20:33banks or consulting firms like everyone

20:35tells you to do? Would you say that

20:37going to a early stage startup is the

20:39best path to go today if you want to go

20:42the entrepreneurial year route or would

20:44you say maybe perhaps something

20:46completely else? Yeah, I think if you're

20:50early in your career, if you get a

20:53chance to work at an early stage

20:54startup, which is difficult, an early

20:56stage startup with good people, like

20:58with good people,

20:59>> uh then you should definitely take that

21:00chance. And then

21:02>> I mean maybe you should try other things

21:05as well like the more breadth you have

21:07the like the better you can understand

21:10the early stage startup. If you also do

21:12some like you go into MBB and then you

21:15and you also do start like oh right this

21:16is things are completely different then

21:18can also um sell into an enterprise

21:20company better if if you seen how that

21:22works. So many experiences is always

21:25better.

21:25>> Yeah. And when you're young you have you

21:27know much higher risk tolerance in

21:28general and I always tell younger people

21:31also that you have nothing to lose.

21:34>> Exactly.

21:34>> But like you said I think the most

21:35important thing is that you choose the

21:36right one with the right people.

21:38>> Yeah. Because if you choose the wrong

21:39one naturally that would be a very poor

21:41experience that might actually guide you

21:44wrongly because of that one data point

21:46that you got from that experience.

21:49>> You know talking about early stage

21:52startups you already mentioned Oliver

21:54and Leon. I remember in late 2019 when I

21:57met Leon and Oliver at uh Espresso House

22:00at Sve and I had been tracking them for

22:02a few months since they actually started

22:04registered the picked and and I was

22:06impressed by what they were doing uh

22:09especially in terms of the

22:10recommendation you know systems and

22:11utilizing deep learning and met them. I

22:13was still like thinking what the heck am

22:15I doing Oliver here Oliver like the

22:17other Oliver was I think 16 is 19 I was

22:20talking about doing my first angel

22:22investment in them didn't end up doing

22:25it for some stupid reason I think I told

22:27them that yeah we need a 1 million post

22:28money valuation and they wanted

22:30something else and they were smart

22:31enough to decline my offer but I still

22:34remember I was very impressed by the

22:35hustle but still they were young uh and

22:40UA when you joined there U was also

22:43young. Yeah.

22:44>> Oliver was even younger. Tell us a bit

22:46about the picked how you ended up there

22:48because you joined a few months after

22:50they started company. uh what was the

22:52story behind that and and you joining

22:54the picked not as the CEO but as the CTO

22:57and co-founder

22:58>> and was it obvious that would you would

22:59be a co-founder when you joined the

23:01picked

23:02>> on this young part I I when I was 20 I

23:05was also like well this this obvious

23:06that young people would do a much

23:08because I was into university I knew all

23:10the smartest people in university like

23:11obviously they would do a much better

23:12job than all these old people in all

23:14these companies so I've been thinking

23:16about that for for years right I mean

23:18there is a lot of things like my first

23:20meeting with uh Lun and uh Oliver that

23:24was serious like they were they didn't

23:27communicate clearly where they were and

23:29like where we're meeting up they were 20

23:30minutes late because they were somewhere

23:32else like it was obvious that they were

23:34like lacked so they lacked so much

23:36professionalism and maturity so I was

23:38very I was a bit hesitant but yeah this

23:40the story was just I knew loness a bit

23:44from something called effective altruism

23:46yeah

23:47>> which is like my my um what's been very

23:50I cared a lot about in my life since I

23:52was like a young teenager is like why

23:53are we humans um destroying the planet

23:58and being so bad at coordinating and

24:00collaborating um and like most many of

24:05your like super focused on environment u

24:08and then effective alum is a bit more

24:10broadly but like how do you actually how

24:13how do you apply being analytical to

24:14solving these problems rather than just

24:16like oh give money away to someone that

24:18say that we should they say koala bears

24:22like it's a bit more like okay how do we

24:24actually solve the root causes here uh

24:26and a lot of smart people so mental is

24:30there so I had a lot of like empathy for

24:33him and Oliver had the same they like

24:35they both shared that and they were

24:37obviously smart I think yeah Oliver also

24:41did a good job at like grabbing coffee

24:44with me and asking me a lot of very very

24:45technical questions about recommendation

24:47systems I was like okay this guy is

24:48super smart his [ __ ]

24:50>> Um, and then I was I had had some like

24:55deep thinking about should I join or not

24:57and um I did and I think it was um

25:03yeah they I mean there was it was very

25:04early. They had been hustling a lot but

25:06they had some things that really worked

25:08um and they continued to work. We went

25:10through Y Combinator and then uh we just

25:13kept growing faster and faster. We

25:14reached like 1 million R

25:16>> a year after we got admitted to YC which

25:19is um like I think legendary Swedish

25:24standards.

25:24>> Was it a criteria to join as a co

25:27co-founder?

25:28>> Yeah.

25:28>> Yeah.

25:29>> Yeah. I mean I these were like two young

25:32kids right and I was like okay

25:35we' already discussed a bit you know the

25:36CEO role and and you also said that

25:38you've been pondering a bit now at

25:40lovable when are lovable should I be co

25:42or not and you started as the CTO and

25:44co-founder depicts how did you assess

25:47Oliver who was 17 if he could be the CEO

25:52and your CEO co-founder or did you even

25:54assess it at all uh or did you more

25:56assess his technical skills and you felt

25:59that he's technically strong. Let's go.

26:01>> Oliver was like so extremely obsessed

26:04and had so extremely strong opinions

26:07about everything building a startup. Um

26:11so like him being so him being good at

26:14CEO came from him being extremely

26:16obsessed and like studying all the YC

26:19material again and again.

26:21>> Uh like reading a lot of books. I think

26:22he read the Elon Musk book like five

26:24times like again and again.

26:27Uh and so he then he got very very

26:30strong opinions about everything from

26:32there. So he kind of

26:33>> he was he was a very good CEO for the

26:38early years for his like his age and

26:40like of course very many things are

26:43difficult if you haven't done it before

26:44>> of course. Um, and then

26:47as we

26:50grew as a company, I think Oliver and

26:55Oliver is just like being so uh strong

26:59opinions and just like no, we're going

27:00to do it like this.

27:02>> Doesn't

27:03um you need to kind of adapt your

27:05leadership style and because you have to

27:07listen more to people. Yeah. And then

27:09his like strengths like couldn't those

27:13are very strong strengths that he had

27:14they couldn't show and he um I guess he

27:18it was disheartening to for him that he

27:20was like oh all of this listening

27:21listening listening and feeling a bit

27:24>> um like you can't um just do things

27:27can't do things.

27:28>> Yeah.

27:29>> That sapped my energy as well. The p

27:32scaled too fast and we kind of had

27:34product market fit and we kind of lost

27:36it.

27:36>> Yeah. Um and I mean at least clear

27:39product market fit and then um like

27:43you're a lot of people and and then

27:44everyone has more like strong opinions

27:45about many things because you're

27:48>> everyone is like uncertain about is this

27:50not really working like should we do

27:53this everyone has strong opinions then

27:54it's consumes a lot of energy and Oliver

27:58had been working like seven days a week

28:00since he was 14 so

28:02>> yeah exactly that was the story

28:04>> very intense intense founder Oliver

28:07Uh you already mentioned Y combinator

28:09and that Oliver you know read all the YC

28:12material which actually I would say was

28:13quite rare in 2019 2020 in Stockholm.

28:16Now it's more you know common I'd say

28:20but you guys you went into YC in 2020

28:23the summer 20 batch that was the first

28:25remote batch also due to co

28:27>> but you were still like one of the first

28:31Swedish YC companies not many before you

28:33had been there like who was the driving

28:34factor for you guys to apply to YC

28:37>> so I think all applied before I joined

28:40as well and driving factor was just that

28:44this is the best place to learn like

28:45it's as simple as that. Yeah. What was

28:47your, you know, feeling when you heard

28:49about it

28:50>> that you would join YC? Were you

28:52positive or hesitant or neutral about

28:54>> I was I think it was a positive. I I

28:56didn't know I guess my exposure to YC

28:59had been through like hacker news and

29:01reading some like reading uh posts from

29:04Sam and Paul Graham.

29:06>> Yeah.

29:06>> Um and all like all very obvious being

29:09obsessed and talking about it.

29:11>> Yeah.

29:11>> So that and I was excited.

29:13>> Yeah. And look for a good reason. you

29:15look at you know YC companies if you

29:17actually know about those open AI was

29:19essentially incubated at

29:20>> at YC but how would you say YC has

29:23formed you as a

29:25>> as a founder and CEO

29:27>> I mean if you look at the effect the YC

29:30has um it's mainly around about

29:32surrounding yourself with great people

29:35>> uh and then it's about like time

29:37pressure but it and it's like a school I

29:41guess like you you have some good uh

29:44some there's some good group partners

29:46that can coach you. Yeah.

29:47>> But they aren't you can only do so much

29:50in a few sessions and it's three months,

29:52right? It's like nothing compared to

29:54studying at university. So I mean it's

29:56not a it's not a game changer in your

29:58life. But yeah, the great people in the

30:01um network and some of the like how fast

30:06you can move like just like where you

30:09can actually move much much much faster

30:11is what you can take with what you

30:12mainly what you take with you which is

30:14very valuable

30:14>> and actually YC has a very specific

30:17culture and and you know just looking at

30:19the companies coming out from YC a

30:21successful culture looking at depict and

30:23the culture you had to depict you know

30:25and I I spoke with Oliver also about

30:28this uh a a few weeks ago. Uh still

30:30quite amazing to see what type of

30:33founder factory depict has become

30:35national lovable. Uh you know you were

30:37the CTO co-founder depict uh now

30:39Vessence that was at YC with Henrik and

30:41Ludvig tandem with Oliver uh and uh even

30:46Max from Lora was at at depict as an

30:49intern.

30:50>> Why?

30:51>> There might be one more. I I don't know

30:53if I can say anything but we'll see.

30:55>> Okay. We need we need to we need to

30:56talk.

30:58I actually I actually think I know now

31:00that you mentioned it.

31:02>> Uh but we need to talk.

31:04>> But why why why do you think the picked

31:07has become such a founder factory? Now

31:10>> the most important criterion for um when

31:16coming into YC is like you are obsessed

31:19about building startups and you are

31:22really want to get into YC.

31:24>> Yeah. Um, and you are smart and like

31:27everyone that the picked had all those

31:29attributes and that's a big part of we

31:31selected for like people really like

31:33were very obsessed about startups.

31:36>> Yeah. Yeah.

31:37>> Simple as that maybe.

31:38>> Yeah. Well, it's I guess

31:40>> but it's you know and you know it as a

31:42founder also and when I've been building

31:44companies uh you know with our joint

31:46friends also culture is not something

31:48you put on post post-it notes and so

31:51it's it is what you do daily and you

31:53know looking at the picked I remember

31:55when you're building it was still like

31:56it was intense like you were crazy

31:58obsessed

32:00>> like no doubt and I guess that was like

32:02the number one factor both in recruiting

32:04people and how you work the depicts or

32:06was there something else besides there's

32:09obsession like technical strength or

32:11>> yeah technical strength was very very

32:13important and that's I think is um

32:16behind

32:18like the why people were came have

32:20gotten into and started successful

32:22companies

32:23>> the lot of focus on like smarts and um

32:29less focus on seniority

32:32>> uh which I think is

32:34one of the factor the bit didn't go as

32:36well as it could um But also one of the

32:39factors why there's so many successful

32:40companies coming out of it.

32:41>> Yeah.

32:42>> So what would you do different now

32:44looking back building the picked and and

32:46then we will naturally move into lovable

32:48but

32:48>> yeah a billion things but um when we

32:51were growing fast coming out of IC I sat

32:53down with Oliver like Oliver is great

32:55like we're growing but this is um

33:00we're doing too much YC mantra do things

33:03that don't scale. We should take we

33:04should take a step back and be like

33:06analytical here in terms of what how do

33:09we create the most value for these

33:11customers like how do we create the most

33:12value for them and pick something smart

33:14and not just be like as much brute force

33:16going for putting forward

33:18I don't know if I was right on that uh

33:22but I would do it differently I would do

33:24it like that or like Oliver said like no

33:26no noon

33:28the most important thing in a startup is

33:30momentum and we have momentum we just

33:32have to keep pushing and So everything

33:34else we got figured out, just pushed

33:35forward more momentum. And since then,

33:38uh, I kind of changed my mind. I was

33:40like, "Oh, Alder is right. Wow, this it

33:42actually worked." Like we uh reached

33:44like $1 million AR. We we raised a huge

33:46series A.

33:47>> Yeah. With Tiger and None EQ and but

33:51doing something different like it's more

33:52maybe not.

33:53>> I'm not emotional here. I'm just

33:56like I can't speak about this.

34:00So Oliver thesis there kind of had a lot

34:03of points for it but but after like

34:06already at our series A um I think we

34:10had been um like how do you say this we

34:14had a great great engineers that could

34:16solve any problems for customers and

34:18instead of like just making things super

34:21scalable um and or like making things

34:24reasonably scalable.

34:25>> Yeah. Um so I think if I would have done

34:29again I would have been more analytical

34:30about what um was the how do we offer

34:33the most value.

34:34>> Yeah you did a lot of things well given

34:37you know looking at the founders coming

34:38out from the pics. So I think

34:40>> for sure

34:41>> it's quite fair that you know

34:43>> but but there's a billion things like so

34:45at each point in the journey it we could

34:48have done a lot like some of things and

34:50which would have made it into a

34:51extremely successful story like which is

34:53a bit sad like there was at each point

34:54in time there was so many things we

34:56could have done to make it extremely

34:58successful story.

34:58>> Yeah.

34:59>> We should have just been even more like

35:02no no no we've kind of lost product

35:04market fit here. That means we are mega

35:07scrappy again. Exactly. And like if you

35:09don't want to be mega scrappy right now

35:12like this is

35:13>> uh how we're going to operate completely

35:15differently we should have yeah the

35:18biggest part the biggest frustration

35:20throughout the journey was that

35:22>> uh we promised a bit much to customers.

35:25>> Yeah.

35:25>> Uh and when they were always changing

35:27things on their technology side it was a

35:28lot of like work from for us.

35:30>> Yeah. More like unscalable consulting.

35:33So there was a I mean there was like

35:34maintenance job to be done with some

35:36customers and then what we should have

35:38done is is to be like look we promised

35:40this we cannot do that uh like you uh

35:45we're doing something else you know and

35:46be more like pull off uh band-aids that

35:50hurts.

35:50>> Exactly. But that's that's always tough

35:53for founders especially saying no to

35:54customers but at the same time no is the

35:56most powerful word for founders whether

36:00it's you know customers recruiting or or

36:03you know whatever you want to say.

36:04>> Yeah actually let me just say two two

36:06things that should I would emphasize

36:08it's the being more analytical about

36:10where we create the most value and then

36:12going much more founder mode. I mean we

36:13are not very very afraid to go founder

36:15mode in the end. I I think if we had it

36:18would have worked out much better.

36:20>> But it's tough when you have PMF and you

36:22lose it. At the same time, PMF is not

36:26something you get and you lock it. It's

36:29a living organism all the time. And I

36:30guess you got to experience that at

36:32Depictand.

36:33>> Depict did a lot of great things for our

36:35customers. Uh and this is what you

36:37should always be focused on. Um we did

36:40too many things at the same time like

36:43too many new customers at the same time.

36:45If we had focused on the ones where we

36:46did the most great things and built that

36:49into a stronger product and strength, we

36:51would have been super super successful.

36:53>> Yeah. Yeah.

36:54>> Now let's talk about lovable. So

36:57>> fastest growing startup, unicorn status

37:02under two years.

37:04But let's go back before any of this was

37:08given. So I want to go into that number

37:11one reason for startup failures is due

37:14to co-founder conflicts. What made you

37:17trust Fabian and why why choose Fabian

37:20as your co-founder?

37:21>> Yeah, Fabian is very very easy to trust

37:25if you talk to him.

37:27>> End of discussion.

37:28>> End of discussion. That's it.

37:30>> He's a very high signal to noise. Mhm.

37:33>> I mean, he doesn't talk as much as many

37:36others, but it's like when he does, it's

37:37often very high signal and um it's

37:40always very high signal and he is very

37:44very reliable in like getting things

37:47done and um

37:51doing the hard things. So, I just want

37:55really wanted him and his like calm

37:57structure on on everything. He also

38:00uh can just build anything build

38:03anything and end.

38:04>> But you Fabian worked at Depict also.

38:06>> Yeah. So he worked there with me and he

38:09was like he was junior but he was like

38:10Anton I I'd love to take lead on this

38:12team. He told me I was like oh really?

38:14Um I didn't expect that. I'm like but

38:16sure you'll do it. And then he did a

38:18good job. I hired someone who was like

38:21super senior engineering manager over

38:23him and that was mistake. I should not

38:25have that.

38:27>> That's a huge mistake.

38:28>> Yeah. Yeah. And he because he was a

38:30previous founder and so on before

38:32Venezuelan

38:39the CTO.

38:41>> Yeah.

38:42>> Yeah.

38:42>> Given that you had worked at Depict also

38:45together, that's maybe also why it was

38:48very natural how you

38:49>> work together. Yeah.

38:50>> The roles uh you knew each other. How

38:52long had you known each other? You and

38:54Fabian?

38:55>> Two years.

38:56>> Yeah.

38:56>> Almost. essentially got to know each

38:58other when Fabian started at Depict

39:00>> one and a half years. Yeah.

39:01>> Yeah. Yeah.

39:02>> During your time there, your open source

39:05project GPT Engineer became of sort of

39:08like a accidental success if that's

39:12>> if I'm not uh if I'm not mistaken. Do

39:15you think that lovable would have

39:18existed without that success?

39:20>> No, I without something else because of

39:23premises in AI 100%. I would ask

39:25something else. So I I I really pushed

39:28Oliver like because uh three years ago I

39:30was like I was it was obvious to me that

39:32>> Oliver had picked to be clear.

39:34>> Yes.

39:36>> It was very obvious to me that we're

39:38we're on this point in time where we're

39:41um at an inflection point. There's going

39:43to be so much things you can do if you

39:45train good AI models. Uh we have a lot

39:48of smart people here. Oliver, why don't

39:50we do something like this? Like there's

39:51so many different examples of it. He

39:54told me um I don't want to do a hard a

39:58hard pivot like it's too much work. Um,

40:01and I made me start thinking about okay,

40:05what what should I be doing here? Like

40:06uh I don't want to uh abandon the picked

40:11uh which I did you could say.

40:13>> Um

40:15and then when this GPT engineer

40:18success came and which was like very fun

40:20because there was this huge community of

40:22people in the on the open source world

40:24that wanted to be part of it. I felt

40:26like okay but this is a point in time

40:29where

40:31I should think about what am I actually

40:33doing now and then it triggered me like

40:34okay I need to think about what I'm

40:35doing should I do this GP engineer thing

40:37first I was like nah probably not um but

40:41what became very clear to me was that um

40:45people were building tools for

40:47developers like GBT engineer and that's

40:51great like that's fun is a kind of

40:53interesting problem to solve But the

40:56absolute majority of the world, they

40:58just are so frustrated by that they

41:00cannot code and they don't know any

41:02great engineers that can help them. 99%

41:06and um in the future humans are not

41:09going to look at code, they're not going

41:11to write code. So we should just we

41:13should obviously not be building for

41:15those developers. We should be building

41:16for everyone else and completely new way

41:19of uh creating software products

41:22>> and being able to do it. so that they

41:25become more lovable because you can move

41:27faster, you can do much more, you can

41:28iterate much more faster and the people

41:31closest to the problems

41:33>> 99% of the best ideas

41:35>> can be uh come reality thanks to anyone

41:38being able to just tell an AI to be

41:40their software engineer.

41:41>> Yeah. Mhm.

41:42>> So then that like that that vision was

41:46just so obvious to me and in terms of

41:49how you would like win if this becomes

41:52competitive which I obviously understood

41:54it would become become competitive

41:56>> is to

41:59be smart in terms of how what you allow

42:02the AI to do. Mhm.

42:03>> So you take all your hardearned

42:06learnings as like building product like

42:08CTO type of experience and you put that

42:12as guardrails into the entire system.

42:14>> Yeah.

42:15>> You don't start training your own models

42:16from scratch. That's going to be this

42:18super competitive race where um you're

42:22going to just spend a [ __ ] ton of money

42:23on compute. Uh but you build a really

42:26good application. focus a lot on the

42:28user experience

42:29>> and then and that this is is what we set

42:32out to build in Fabian.

42:34>> But just to grill you a little bit there

42:36like

42:37>> when GPT engineer exploded that was

42:40summer 2023.

42:42>> Yeah. Spring.

42:43>> Spring. Yeah. Yeah. Now your vision

42:45seems very obvious like where we are

42:46right now and and with the explosion of

42:49web coding as a category even. Did you

42:52consider at some point two years ago

42:54that should we maybe go more like the

42:56path of cursor where we build something

42:58for software engineers rather than for

43:0199% who can't code and cursor has

43:04naturally also also done well

43:06>> uh but did did you did you really think

43:09about that path also

43:10>> yeah think about all the different

43:11directions you could go um there like it

43:15was very tempting because you get a lot

43:17of data from the get- go cursor gets a

43:20lot of data Yeah.

43:21>> Um but yeah, this I decided this this

43:24part like two years ago, the building

43:26for the non technical people

43:28>> and I guess I always question everything

43:30a lot. So there was many parts through

43:32through the journey where I I questioned

43:33like how should we actually make it

43:35should we sequence it so that we do it a

43:37bit more for developers and get a

43:39stronger product market fit in a smaller

43:40segment. U but I'm happy we just kept

43:42pushing until it got so good that it

43:45completely blew up. What were your first

43:48six months like at Lable? You know,

43:50mistakes or failures, early wins?

43:54>> I think we So, so that was uh like

43:59autumn, end of um 2024.

44:03And

44:03>> no, 2023.

44:04>> 2023, sorry.

44:06>> 2023. And then, uh I was uh to be honest

44:12a bit burnt out. So he was like, "Okay,

44:15I'm super excited about this vision."

44:17Um, and then but it was we had a very

44:21small team. I think we should have at

44:24the time

44:26spent I should have hired a few more

44:29great people. We could have just

44:30accelerated much much further. Um and

44:32then there was um of course like

44:36uncertainty because doing being a good

44:39product and being good at distributing

44:41it uh is

44:44um it's um magic formula. It's magic

44:48formula. So we had a lot of different

44:50thoughts on how we would do that. Um but

44:55what we did was we launched on product

44:57hunt like with an alpha version that was

44:59closed and so we could onboard people

45:01and see what happened.

45:02>> Um

45:02>> it was still called GBT engineer.

45:04>> It was called GPT engineer app. Yeah. So

45:05it was like a different product and then

45:07we got feedback on it and we kept

45:09iterating. Um I think we could have with

45:13like really good marketing um and

45:20like a more yeah a more specific use

45:22case we could have already then have

45:25started making a [ __ ] ton of money I

45:26think if we had executed faster. Yeah,

45:29>> some less lessons learned from me was

45:31that if I had like had my been my best

45:33self and had full energy would have

45:34moved faster from assembling a great

45:36team which we have done today

45:38>> and been orders of magnitude better

45:40position that we are now.

45:42>> And in the very beginning of lovable a

45:45lot of the shiny

45:48VC logos rejected to invest. But

45:51speaking of the, you know, of the very

45:54shiny logos and so forth and fundraising

45:56is heavy on founders, did that give you

46:00a lot of motivation of giving getting

46:02all of those nos or how did that affect

46:04you?

46:04>> Yeah, I didn't get so many nos actually.

46:06I had read this book which is like keep

46:08it [ __ ] simple. Just work with

46:10someone you like. Um,

46:12>> and

46:15like don't waste your time. Get look at

46:17what outcome you want to get. I think um

46:21there was a billion ways I could have

46:22raised from like this any shiny logo if

46:24I had the sequence better and I I just

46:26put together a good run a tight

46:28fundraising process 100 100%. So I I I

46:31guess I could have I messed up I guess

46:33you could say

46:34>> it's a good idea

46:35>> and um and but it was I wish I didn't

46:37read that book

46:39because if you're like if you have my

46:43background and you don't race from a

46:45shiny logo it's like a super bad sign.

46:47>> Yeah. But it's it's a good it's fun

46:48learning that like it it the book was

46:50true. It's like I didn't it didn't

46:51matter in the end like we we we got to

46:54where we are without any shiny logos in

46:56the first round.

46:56>> And speaking of that on fundraising what

46:59would be you know both at depict and

47:01lovable you raised a lot of money and so

47:04forth. What would be your what would be

47:06your recommendation to all of the

47:08aspiring founders out there who about to

47:11go fundraising and how to go about it?

47:13>> Yeah that's a good question. I I think

47:15default case just be super scrappy and

47:17find product market fit. You can do it

47:18without any investors and then when you

47:20have product market fit like you have

47:22product market fit maybe don't you don't

47:24need to raise

47:25>> uh or you can then start building

47:27relationships and reject investors until

47:30you have so good terms they like an

47:33obvious yes. Uh that's the default. If

47:35you are if if there's something you have

47:38that will make it um easy to

47:44get like good terms and get someone you

47:47really want to that that's going to

47:48actually be very supportive

47:50>> then I would u see if you can take the

47:53chance and but do it by

47:56thinking of investors as a potential

47:58partner in the future like you're making

48:01friends here

48:02>> and

48:03>> uh good investor

48:06If you show that you are have a lot of

48:07potential, they are going to want to

48:10help you and be on to have you on their

48:11good side

48:13>> uh even if you're not racing and even if

48:15you're saying like actually I'm not

48:16racing in the coming six months but

48:18>> um I'm uh yeah this is my philosophy

48:22about the company and why I'm I'm giving

48:25away equity to others

48:27>> under the assumption that I want this to

48:29become a hundred billion dollar company

48:31right so

48:32>> and you would get it at an insane

48:33discount Right? So if you think there is

48:36something already before product market

48:37fit that will make you a very attractive

48:40investment then I will have those

48:41conversations

48:42>> and then uh see

48:45who which people you were like naturally

48:47will be good friends with regardless and

48:49we or and or if it's just someone that

48:51is so good that you won't have them on

48:53your side and they they want to give you

48:55a good offer

48:56>> and then when you feel like who you know

48:59who those are h you go you go out and uh

49:02pitch M

49:04>> how did you approach valuation early on?

49:07Are you more like you know it's going to

49:08be an auction I don't decide valuation

49:10at the investors or like with lovable

49:13were you clear that it's going to be on

49:16this valuation at minimum if I go and

49:18race now again like what I decided to do

49:21was to work with with people I really

49:23liked which I did which is I'm very

49:24happy about it turned out super super

49:26well.

49:26>> Yeah. But if I would give general

49:29advice, I would say

49:32uh get like if you're building a hundred

49:35billion dollar company, get low dilution

49:37in the

49:38>> Yeah.

49:38>> first round, but only raise uh raise a

49:41little bit. That's what I would do.

49:43>> Yeah.

49:43>> And there's also ways to be creative if

49:45you're in a super good um

49:47>> position.

49:47>> Super good position. Yeah. If we talk

49:49about timing,

49:51if you're being brutally honest

49:53and let's say you would have started

49:56lovable one year earlier or one year

49:59later or even two.

50:01>> Yeah.

50:02>> Do you think it would have been the same

50:03success today?

50:05>> No, I think it would been difficult. I

50:07mean, there would have been something

50:08else, some other unique part you could

50:10have built. But

50:12>> what I decided like a year ago, I guess,

50:15was like, okay, the goal here is to get

50:17as much mind share as possible for this

50:19category that doesn't exist yet, but

50:21this new category

50:23>> um and build

50:26uh the full vision, which is we're not

50:28even close to now, but any company, any

50:32company can ship all their software and

50:34anyone a nontechnical person in the

50:36company can can be part of shipping that

50:37software just by talking to an AI. And

50:40it's more reliable. is more secure. Uh

50:42you can ask AI anything about your users

50:45and uh just like you go to the analytics

50:47departments. It takes two weeks but now

50:48we get it instantly like that full

50:50vision we're not doing close to but

50:51that's what we decided I decided like

50:53mind share and building towards the full

50:55vision um and getting that full that

50:58mind share that we have today of like uh

51:02building an app is best is just done

51:05using lovable uh would have been more

51:08difficult if you start now

51:09>> when you were building lovable

51:10especially late 2023 did you inherently

51:15just trust that the model performance

51:16will get there Yeah,

51:18>> because I guess like you know Sonet 3.5

51:21in the summer of 2024 that was the first

51:24model actually that performed so well

51:27that you guys also could build lovable

51:30as we know it today and actually have a

51:33lot of other features integrations and

51:34your own things that you built around

51:36it.

51:37>> But uh you know would you say one that

51:40you trusted all the time that the models

51:41will get there and two what was the

51:43biggest break? Yeah, I thought the

51:44models would get f better faster much

51:47faster than I thought I thought it would

51:49get better much faster than they did.

51:51>> Um, and we've done a lot of smart [ __ ]

51:54on the AI side.

51:55>> Yeah.

51:55>> But that's not as valuable over time

51:56because AI is just getting better by

51:58time.

51:58>> Exactly.

52:00>> But what when things get really tough

52:02for you and things get tough during

52:05startup journeys, what is it that

52:07actually drives you? What's what's your

52:09purpose that drives you forward and not

52:10just quitting? Hm.

52:13>> Money.

52:15>> I know it's not money with you.

52:18>> I think it's two things that it's the

52:23um what we can do in terms of shaping

52:27this world by building a

52:30community and like a generational

52:31product

52:33actually living that footprint and

52:34making sure we do things responsibly on

52:36the way there. And then it's

52:40uh seeing like me contributing to the

52:44team succeeding.

52:45>> Mhm.

52:46>> And they've done a lot of hard work and

52:49continue to do so. And I I get a lot of

52:52energy from seeing a team work really

52:53well together and succeed together.

52:55>> Yeah.

52:56>> And it's my duty to like make sure they

52:59succeed. M and in that very you know

53:02beginning of starting a company do you

53:04think that momentum is the you know the

53:09most important thing or did you even

53:11think about modes in the beginning or

53:13how do you

53:14>> yeah I thought that I thought about

53:16modes most of the time that was a waste

53:20of time

53:22>> it's more about moving fast and having

53:24momentum and adapting fast

53:26>> actually I'm I'm very curious about this

53:28one so Sam Elman woman always tells that

53:31no one talks about the early days

53:33because they're so embarrassing.

53:35>> Oh yes. This is why this podcast is so

53:37embarrassing.

53:38>> Exactly.

53:41>> What is like one thing that you haven't

53:43told you know out there um public that's

53:48been embarrassing in the early days of

53:50lovable?

53:54I think I think it's embarrassing that

53:56um we spent too much time and too too

54:01often said like oh we should do this

54:03thing differently like this or do this

54:05thing differently like this

54:06>> rather than

54:09>> saying like okay we can re-evaluate

54:11every three months unless there's

54:13something huge that happens. Uh so I

54:15mean you should be pivoting fast in the

54:17beginning but I caused too much like oh

54:20uh guys let's do this instead.

54:23>> Fair enough.

54:23>> But you had a few launches also with GP

54:26engineer or when lovable was still a GP

54:28engineer app but yeah maybe they're not

54:32something you're embarrassed about

54:33because that's just part of the journey.

54:35>> Yeah I think that was that was like we

54:37should have just done those done

54:38launches faster and

54:39>> yeah exactly

54:40>> it faster. that the

54:42>> Are you worried that all of the big AI

54:45players would at some point crush you?

54:47Like given they're both partners and

54:49competitors,

54:51how do you perceive that?

54:52>> Yeah, I think every company should think

54:54like, oh wait, if someone creates AGI,

54:57uh why don't they just let that AGI make

54:59a cheaper copy of myself? Like I think

55:01every company could think that, but um

55:03the world is more complex than that. And

55:07we're going to have AGI, they're going

55:08to have AGI. Um, so it's all about

55:11execution and

55:13building trust with the humans that are

55:16going to pay for your product and

55:17building a by creating a lot of value

55:19for them.

55:20>> So if we can create a lot lot of value

55:21for our users then we're building a lot

55:24of strong

55:24>> defend like brand which is a mo and then

55:29>> um that's what we should be doing.

55:30Simple as that. Keep it simple.

55:32>> Would you say you know looking at the

55:33big AI labs that you're working with who

55:35are also now creating their own lovable

55:38competitors? Yeah.

55:39>> Would you say that you're a paranoid

55:41optimist as a founder, you know, looking

55:44at them?

55:44>> No, I'm just optimist.

55:45>> You're just not paranoid at all.

55:48>> Yeah. I mean, I think every time someone

55:50launches something that's completing is

55:51there's like a wait a second, what is

55:53this is going toffect is good, which is

55:55healthy, but like I you could look at

55:58Google, they have some kind of competing

56:00product. No one knows what is what is

56:02that? Like

56:03>> I tried it at some point. I was like,

56:05no, no one liked it. And

56:06>> yeah. So, let's talk about your social

56:09media like founder presence. There's

56:11another Stockholm AI founder, not saying

56:14any names, who said that there was one

56:17Anton before he went to the US and

56:21then another Anton when he came back

56:23>> and he was fine-tuned

56:25>> and he was fine-tuned

56:26>> in the US. Yeah.

56:28>> Talk about that like what what happened

56:31over there.

56:32>> How did you become so good on social

56:34media? No, I just became a bit more

56:36based, I guess.

56:37>> Yeah, you didn't post that much. Even

56:40like going back just a few years, the

56:42early days of lovable.

56:44>> I don't think I post too much now

56:45either. I am too busy. But I um I do

56:49make sure to be like, okay, let's what's

56:51the shortest thing that will resonate

56:53with someone if I post it and then I

56:55post it.

56:56>> But you have a very specific style which

56:58resonates well.

56:59>> Okay.

56:59>> Like kudos to you.

57:00>> Thank you. Uh I know maybe as an

57:02engineer it's hard to talk about being a

57:04social media profile but

57:06>> no just write something that resonates

57:08with people and keep charging.

57:10>> Speaking of that, what would your advice

57:13be to upand cominging founders out there

57:16or actually founders that might be you

57:18know introverts not comfortable with

57:20social media and so forth and building

57:23in public?

57:24>> Yeah.

57:24>> What would be your advice to them? Look,

57:26if you are obsessed with something,

57:28you're going to become really good at

57:29it. And if you're if you're smart, then

57:33you're going to get good at anything

57:35fast. Um, take sales. Like people who

57:39are obsessed about sales and do it a

57:41lot. And think about what they should

57:44have done differently next time. They're

57:46very like thoughtful about it, they

57:47become extremely good at it. I think um

57:50Oliver at the Picto also a good example

57:52of this where he's like has a super

57:54weird style and he seems a bit weird

57:56when you talk to him but he's actually

57:57very good at states.

57:58>> Yeah.

57:58>> Yeah.

57:59>> Yeah.

57:59>> And you should just use that fact and

58:01become good at the things you're not

58:02good at as a founder

58:04>> because if you do then you can hire so

58:07much better people and you can manage so

58:10much better um than if you just um

58:15avoid things you're not good at.

58:16regarding the distribution if there's

58:18and you know advice to founders out

58:21there if you think back what is your

58:26number one advice for distribution

58:29>> yeah it depends a lot on your product

58:30right I think uh sales is the most

58:32common form of distribution and you need

58:35to be good at sales regardless if you're

58:37if you're a founder

58:39>> sales as in building relationships

58:42>> um where people start trusting you

58:45>> by giving them value

58:47>> because now you're also you know going

58:49more sales levable. I'm not doing that

58:53and I should I should be I'm doing too

58:55many other things but I should be um and

58:58but because there's so many like

59:00extremely large companies

59:02>> like uh Spotify

59:05claim that are using love pay for it

59:08lovable um and uh when distribution on

59:12your question I think becoming good at

59:15sales is probably

59:17very very valuable you should you should

59:20default to that you can also So create

59:23entertaining content on the internet.

59:25That's

59:26>> the second one. Uh

59:29for hundreds of years like good

59:30interesting content has been the way to

59:32do marketing. And

59:33>> exactly

59:34>> um and if you're doing sales at a large

59:35scale you also want to do marketing and

59:38uh if you are good at understanding what

59:40content is people will resonate with um

59:44it's going to help you a lot. You

59:46mentioned that especially when you were

59:47younger you were introvert and uh

59:50personally I feel actually that the best

59:51content comes from opinionated engineers

59:54who when you meet them might feel like

59:56introverts but when they write yeah

59:58>> it's very opinionated

1:00:00>> and naturally the hardest thing is to

1:00:03push these types of individuals to start

1:00:05doing it.

1:00:06>> Yeah. But what's your recommendation?

1:00:08Especially like take a young KTH student

1:00:11now who is very technical, very strong,

1:00:13you know, how do you push through that

1:00:15chasm to just write about your thoughts

1:00:18for example because it's so valuable for

1:00:20so many but many don't do it like

1:00:23looking at your background also what

1:00:25pushed you. I say like now I'm doing a

1:00:27saying paradox here like do many become

1:00:29all become good at all these things but

1:00:32it starts with taking one of them and

1:00:35being like absolutely obsessed about

1:00:37being good at that thing. Yeah.

1:00:38>> So I and I think on writing uh your

1:00:43thoughts it's hard to break through the

1:00:46noise on that today unfortunately. So I

1:00:49think the first move where you can like

1:00:51be obsessed and get quick feedback which

1:00:53is

1:00:53>> absolutely crucial to become good at

1:00:55something get quick feedback. It's more

1:00:57about um

1:01:00yeah doing being on TikTok or something.

1:01:02I don't know what's the best one. I I

1:01:04would do like an analysis on where your

1:01:07where your uh skill set is like where

1:01:09you have something unique to offer and

1:01:10then

1:01:11>> be be on a platform where you got a lot

1:01:13a lot of quick feedback to become really

1:01:14really obsessed about that and good good

1:01:16at that

1:01:17>> because you were still with loveable you

1:01:19know also the early days you know you

1:01:22everything seems like an accident to

1:01:23some extent but you've been still quite

1:01:25strategic also in what channels and how

1:01:27you push and that has been successful uh

1:01:31or am I

1:01:33And how did you approach it? Like say a

1:01:35year ago were you thinking about okay we

1:01:37have this great product?

1:01:38>> Yeah.

1:01:39>> How can we be smart about distribution

1:01:42and do things in a scalable way and find

1:01:45the right channels?

1:01:46>> Yeah. Before we started it was clear to

1:01:49me that um I mean if you look at GP

1:01:52engineer we came got this insane

1:01:55distribution right. Yeah. partly because

1:01:56it was open source and free but

1:01:58>> um that was just because there's this

1:02:00appetite for a cool AI products and

1:02:04>> especially on Twitter when goes viral

1:02:06there.

1:02:06>> Yeah. So it was obvious to me that oh

1:02:09the this seems like a good way to solve

1:02:11the distribution part and I can spend

1:02:13most most of my most of my time on the

1:02:15product.

1:02:15>> Smart. So naturally, Anthony, you're

1:02:18very pro- Europe uh in your communic

1:02:20talking about communication and writing

1:02:22also and you're often talking about, you

1:02:25know, we need to show that we can build

1:02:28not only a hundred billion dollar or

1:02:30euro company from Europe, but a trillion

1:02:32dollar company from Europe. And actually

1:02:34given where we are right now, we're a

1:02:35bit at an inflection point in Europe.

1:02:37There's a lot of forces right now

1:02:38pushing us. And uh it's still a bit up

1:02:42in the air if we can actually push

1:02:44through in Europe now to make sure we

1:02:45can seize this moment. But you know

1:02:48challenge you a bit. So Lovable is a

1:02:50Delaware CRP.

1:02:52What's the story there? You know pro

1:02:56Europe but Lovable is still registered

1:02:58in the US. What's the background?

1:03:00>> Yeah look you can definitely build

1:03:03worldclass companies and product like

1:03:06category leading companies from Europe.

1:03:08It's all about execution and the team

1:03:10you put together. I'm

1:03:13rational. I need to do what's best for

1:03:15the business. Yeah. I

1:03:17>> It's easier to do many things if you're

1:03:19incorporated in the US while still

1:03:20having 90% of your people here.

1:03:22>> Agree.

1:03:23>> If um our bureaucracy continues to be

1:03:26bad and like there is some reason why

1:03:28it's too bad for us to stay here, we

1:03:30would like I would move the company I

1:03:32would move the company from Europe to to

1:03:33somewhere else.

1:03:34>> Yeah. Yeah.

1:03:34>> It's just how how it is. And that's why

1:03:36I I think uh and that's actually the

1:03:38truth why which politicians have has to

1:03:40think about in terms of making European

1:03:43countries as successful as possible. I

1:03:45think that's I think like looking at the

1:03:47companies I backed during the past year

1:03:50>> over half are overse corps y

1:03:53>> uh building in Stockholm also but it's

1:03:55like I think you hit the nail on the

1:03:57head Anton what I was tell the founders

1:03:59is it's not your work to be a politician

1:04:02your job is building a generational

1:04:04company something big and you do it

1:04:06wherever you feel is fit for purpose to

1:04:10do it and then it's other individuals

1:04:12and naturally founders can drive that

1:04:13also but you need to build a company and

1:04:15I think it's very refreshing hearing you

1:04:17saying that at the end of the day you

1:04:19know your biggest

1:04:22you know responsibility is lovable and

1:04:25you build it where it makes sense at the

1:04:27end of the day but uh can we build a

1:04:30trillion dollar company from Europe

1:04:32which feels very elusive right now

1:04:34looking at the comp tech companies that

1:04:36have come out from Europe

1:04:38>> I want to ask can we build a$10 trillion

1:04:40company from Europe that I don't know

1:04:42>> a trillion dollar of course

1:04:43>> yeah and uh

1:04:45looking at building you know from Europe

1:04:47and looking at where lovable is right

1:04:49now and where lovable was like when do

1:04:52you feel is the right time to start

1:04:54scaling abroad? Uh I know you guys you

1:04:57mentioned in the Swedish newspaper also

1:04:59that you feel bad recruiting talent from

1:05:02your friends here in Stockholm that

1:05:03you're going out to London and US also

1:05:06but when when you know when is the right

1:05:08time when do you feel when was the right

1:05:09time for you? There's um a lot of people

1:05:11who have like okay what what's the right

1:05:13answer and they try to brand their

1:05:15startup by like finding someone who

1:05:16knows what the right answer is to

1:05:17everything. And I I think

1:05:19>> generally most of the time you want to

1:05:22be uh analytical and look at where

1:05:25you're uh where you're at. Write out the

1:05:28pros and the cons of different options

1:05:30and pick the best that one is that's

1:05:32best for you. Um get a lot of like ask

1:05:35people for a lot of data and opinions

1:05:37but and take what's best for you. So it

1:05:38it depends and

1:05:41um I want to make an example out of

1:05:43that. You should always figure it out

1:05:45yourself.

1:05:46>> Yeah.

1:05:46>> I can say that there's a huge value both

1:05:49person on a personal level and for the

1:05:51company to be in one place in person.

1:05:53>> Yeah.

1:05:53>> Do you think that in general like Nordic

1:05:56startup founders go abroad too late?

1:06:00>> Yeah. So on the commercial side you

1:06:01should probably most of the time think

1:06:04about the US market first

1:06:05>> instead of going out in Europe. Yeah.

1:06:08Yeah. Unfortunately. Yeah.

1:06:10>> That's the raw truth.

1:06:11>> Should you just build for the US from

1:06:13day zero like for the because US is the

1:06:16number one market in technology with

1:06:19maybe the most I wouldn't say demanding

1:06:22customers but most savvy when looking at

1:06:24the companies coming out from there and

1:06:26like you know linear has open AI and

1:06:28cursor and perplexity all these top AI

1:06:30companies as customers also. But what's

1:06:32your take on that like building from day

1:06:34zero for the US market even though

1:06:36you're in Europe? Yeah, that's what I

1:06:38would also say. Try to articulate. I

1:06:40think you um get the most magical

1:06:44companies that are first led by early

1:06:48adopters pro self doing self-s serve.

1:06:50>> Yeah.

1:06:51>> And that doesn't that does not need you

1:06:54to be in the US at all. And then you

1:06:57layer on depending on how the complexity

1:06:59of the product you lay layer on sales

1:07:02part of that on that.

1:07:04>> Are you guys using linear by the way?

1:07:06Yeah, of course. Of course. Good. Good.

1:07:08I think it's just an example of this

1:07:10like thinking about modes.

1:07:12You think car didn't think about modes.

1:07:14He was just like this product is a [ __ ]

1:07:16is still not nearly as good as it can

1:07:18be. And just kept kept going at it.

1:07:20>> Yeah.

1:07:21>> Until perfection

1:07:22>> and incredibly product design focused

1:07:24also, which I love about the Nordics.

1:07:26>> Yeah.

1:07:26>> Which you feel in lovable, you feel it

1:07:28in linear like that's still that Nordic

1:07:30flare

1:07:30>> in terms of UX design, integrations,

1:07:32everything.

1:07:33>> Silicon Valhalla.

1:07:34>> Silicon Valhalla. Exactly. Thank Thank

1:07:36you. That's a great compliment that

1:07:37we're similar to linear in lovable.

1:07:39>> I think so.

1:07:39>> I mean that's very appreciated. I think

1:07:42we have some some way to go on some some

1:07:44because we're because we're moving so

1:07:46fast.

1:07:46>> Yeah. No, but it's still like a great

1:07:48user experience.

1:07:49>> There is

1:07:51>> great momentum now in Stockholm. Uh I

1:07:53would say Nordics, but especially

1:07:54Stockholm now within AI

1:07:56>> and many have asked me Oliver, what's

1:07:58the reason now for the momentum here in

1:08:00in Stockholm especially. And naturally

1:08:02everything is relative like compared to

1:08:04San Francisco. It's still quite small,

1:08:06but I still feel that we have a good

1:08:08shot now of making Stockholm the at

1:08:12least a application capital of Europe,

1:08:14the main hub.

1:08:15>> Let's do it.

1:08:15>> Exactly. And that's what we're doing.

1:08:17That's why we're doing this podcast

1:08:18also. And uh I think we will be it in

1:08:21within 12 to 18 months it will be

1:08:23obvious that Stockholm is the main one.

1:08:25But asking you on you've been you know

1:08:27I've been involved in the AI system in

1:08:29Stockholm maybe for a little less than

1:08:3210 years about 10 years. uh but you've

1:08:34been a long time here and we already

1:08:36mentioned Stockholm AI

1:08:38you know that was the largest AI

1:08:39community in Nordics back in the days uh

1:08:41started by among others Lars who is now

1:08:43a deep mind KTH all these founder

1:08:46factories but what do you think is the

1:08:48reason why Stockholm is so strong and we

1:08:50have this momentum now uh here

1:08:53>> there's this recursive effect which is

1:08:57that when there is a great founder that

1:09:00is a good role model it creates other

1:09:02great founders or great role models.

1:09:04>> Yeah.

1:09:04>> And that's the main main factor. The

1:09:07other factor is it that in London a lot

1:09:10of smart people go into finance because

1:09:11you make so much money.

1:09:13>> Uh or they going to some big tech

1:09:15company that has a big office there.

1:09:17>> Yeah.

1:09:17>> Not as much in Stockholm.

1:09:19>> Yeah.

1:09:19>> And given that you've been that

1:09:21inspiration and still are for many

1:09:23people out there to start companies and

1:09:25we want to encourage people to start

1:09:27companies. But on the other side of the

1:09:30coin, what is the one mistake or you

1:09:34know any advice that you wish no other

1:09:37founder will go through? Um the one

1:09:42advice for other founders

1:09:46um surround yourself with great people

1:09:49and be very very direct about like what

1:09:54way of working do you want to have and

1:09:56how do you and who what type of people

1:09:58do you want to hire?

1:09:59>> Yeah.

1:09:59>> Yeah. Team is the most important.

1:10:01>> Yeah. It's all about people.

1:10:02>> Mhm.

1:10:03>> Even in even if we're closing getting

1:10:06close to AGI, but

1:10:08>> it's even more important with people.

1:10:10>> Exactly.

1:10:12But talking about the fact that we're

1:10:13all just people and human beings uh even

1:10:17you son you're just a human being at the

1:10:19end of the day and and uh you know we we

1:10:22all have self-doubt and you know I think

1:10:26everybody especially I guess the more

1:10:28successful you become you start dealing

1:10:30with imposter syndrome every now and

1:10:31then like you know am I really this good

1:10:34and I've talked with you know decacorn

1:10:37founders who just deal with it all the

1:10:39time uh But how do you manage, you know,

1:10:42self-doubt or imposter syndrome?

1:10:45>> Or do you have it? Maybe you don't have

1:10:46it. I don't know.

1:10:48>> No, I don't have so much.

1:10:50>> That's good. That's good.

1:10:51>> On the high.

1:10:52>> Yeah. Yeah. But I think that's great.

1:10:54That that probably lets you execute more

1:10:56freely uh and execute faster if you

1:10:58don't.

1:10:59>> I think the reason I don't have it is

1:11:01that it's just like I know how bad I am

1:11:03and

1:11:04I don't take it's not so much anxiety.

1:11:08Well, just you saying that means that

1:11:11you maybe have, you know, every now and

1:11:12then some.

1:11:13>> I guess I've gone through more with

1:11:14that. Yeah.

1:11:15>> Yeah.

1:11:15>> But I mean, it's all about knowing where

1:11:19um you have your superpowers and how to

1:11:22build the best teams depending on like

1:11:25how the works and how you work

1:11:27personally.

1:11:28>> And if if you do that, then you're going

1:11:30to succeed.

1:11:31>> So, if you weren't an entrepreneur, what

1:11:34would you be?

1:11:35>> Yoga teacher.

1:11:36yoga teacher.

1:11:37>> That's why my joke at least that would

1:11:39be very healthy.

1:11:41>> I like it. Agi by the end of 2025. Yes

1:11:45or no?

1:11:46>> No. We unless we already have it now.

1:11:49>> That's what I

1:11:50>> Where's Lovable in five years?

1:11:53>> Lovable is like the default interface

1:11:55for humans to AI and the and the digital

1:11:57world generally.

1:11:58>> And Anton, does he have kids in five

1:12:00years?

1:12:01>> Yes.

1:12:03Wow. I get a cup now.

1:12:05>> I don't know the answer.

1:12:08>> So, uh, the most significant AI moment

1:12:12in your life.

1:12:13>> It's the alpha zero paper. Yeah.

1:12:18>> Some reinforcement learning that the one

1:12:21of the Dota games.

1:12:22>> By the way, I recommend everyone

1:12:24watching. It's still on Netflix. Diggo

1:12:26movie from 2016. It's still it's almost

1:12:2810 years old.

1:12:30>> Still mindbending. Move 37 especially

1:12:33>> favorite AI tool can't say lovable

1:12:36>> such a

1:12:38is white combinator worth it

1:12:41>> yes

1:12:42>> any role models or mentors that shaped

1:12:44your path

1:12:45>> so I have I have a mentor now is Alex

1:12:48he's shaped me a lot

1:12:50>> is now the best time ever to build a

1:12:52company

1:12:53>> I hope so I I think so yeah

1:12:56>> momentum or moes when going zero to one

1:13:00>> momentum

1:13:00>> thanks Son, thank you.

1:13:01>> Thanks, Anon. This was amazing.

1:13:03>> Thank you for doing this and uh share

1:13:06letting people share their stories about

1:13:08how hard it is to start a company.

1:13:10>> And uh you're naturally the first one,

1:13:12but I think I just came up with this

1:13:14idea and we might do this after every

1:13:17episode, but who do you think we should

1:13:19invite to this podcast?

1:13:21>> Can you meet Oliver? I'd love to watch

1:13:22that episode.

1:13:24>> Absolutely. Oliver at home. Yeah. Depict

1:13:27CEO, former depict CEO and co-founder.

1:13:30>> Exactly.

1:13:30>> Yeah. see the other side of the coin

1:13:32says about you.

1:13:33>> And then we get Joel also.

1:13:36>> Thanks.

1:13:36>> Thanks so much.

1:13:37>> Thank you, Anon.