Anton Osika, Co-Founder and CEO @ Lovable: Hitting 85% Day 30 Retention - Better than ChatGPT

0:00 we have month one retention that's 0:01 better than chat gpt's month one 0:03 retention Anton Osa he is the co-founder 0:06 of lovable. Dev fastest growing company 0:08 in Europe what a title growth starts 0:11 ramping up after we launch growing 1 0:12 million AR per week at some point and 0:14 that just keeps accelerating the most 0:16 important thing is talent and culture 0:19 and there's more raw available talent in 0:22 Europe when a competitor has a lot of 0:24 money do you have to raise too ready to 0:27 go 0:28 [Music] 0:39 Anton fastest growing company in Europe 0:41 what a title but first thank you so much 0:43 for joining me today man thank you Harry 0:45 it's it's always fun to talk to you that 0:48 is very very kind of you it's the 0:49 British accent but I want to start 0:51 actually pre- lovable and I spoke to a 0:53 couple of your investors in your first 0:55 company depict and so I just want to 0:58 start there what are your big takeaways 1:01 from depict that shaped how you think 1:03 about lovable we scaled super super fast 1:07 at depict as well and and we like moving 1:10 just very fast Scrappy we did that we 1:13 nailed that really well I think we did 1:15 really well on this high potential 1:18 Talent as well quite Junior with high 1:20 potential Talent you can see that from 1:22 all the uh companies coming out from the 1:24 picted the the pict Mafia is is 1:27 absolutely real um but what I think we 1:30 what works in the beginning is to say 1:32 yes onto a lot of opportunities and try 1:34 out what works once you be become more 1:36 people and you have to follow up and 1:39 maintain everything that you start you 1:41 you have to be much more focused so we 1:44 said yes to too many things at at the 1:46 pick and uh we didn't like take this one 1:49 thing that we could do 10 times better 1:51 than anyone else uh and then at as the 1:55 economics uh macro wise turned worse we 2:00 we didn't continue the scaling 2:02 trajectory that we were on initially I 2:04 was just reading an article actually 2:05 with Paul book heite the founder of 2:07 Gmail and he said you need three great 2:10 features in a product and it's really 2:12 that simple and make them really really 2:15 great do you agree with that product 2:17 Simplicity towards feature depth yeah I 2:21 I think just on a produ product level 2:23 you should say no to as many things as 2:25 possible and make it more of like an 2:28 apple feeling the things you do you do 2:30 them the purpose okay and so we have hey 2:33 don't say uh yes to everything is there 2:36 anything else which you did or didn't do 2:38 that really shaped how you think about 2:40 the early days of lovable no I think 2:43 moving fast um and I I would say Talent 2:48 is the most important thing and culture 2:50 how you work together every day like how 2:52 people interact and collaborate quickly 2:55 uh those are the two most important 2:56 thing for for almost any company okay 2:59 let's just unpack talent because it's 3:01 been a I spoke to Frederick at creandum 3:03 before this and he said we had to chat 3:05 about this too you favor Talent overe 3:08 experience which sounds kind of obvious 3:11 respectfully you're going to go for like 3:13 an like an experienced untalented 3:16 person but how how do you think about 3:18 your hiring lessons around experience 3:21 versus Talent um so I think experience 3:24 can be a negative thing in some cases 3:27 you often want people who are super 3:29 ambitious they have a lot to prove and 3:32 they are more open-minded towards how 3:35 how you should work together in a team H 3:37 so um I mean for many RS I I think uh 3:41 junior Talent is is first of all super 3:43 easy to to get them into the company 3:45 they don't they're not uh already 3:48 committed to some like many their 3:50 projects the best people that you can 3:52 hire at a young age they would go on and 3:54 become Founders and then you can't hire 3:55 them anymore right so so that's why 3:58 junior uh people are often quite good 4:01 will you hire them if they haven't done 4:03 what you're hiring them to do before the 4:05 benefits often of hiring someone 4:07 experiences well I can see they've been 4:08 at X company for four years they can do 4:11 that at my company will you hire someone 4:13 who hasn't done what you asking them to 4:14 do before in most cases yes uh but I 4:18 mean if it's engineering you have to 4:20 know software engineering of course um 4:23 and you for some roles you definitely 4:27 want in that domain someone with a lot 4:30 of experience like that who can coach 4:33 and who can uh tell the more Junior 4:35 people what Great Looks Like These are 4:37 going to sound like strange questions 4:38 but you mentioned the word ambition 4:40 there did you always know that you would 4:42 be successful when you were younger 4:44 building did you always think I will be 4:47 successful in something that I do I no I 4:50 don't think so I I I I was always 4:52 frustrated with how people around me 4:54 didn't understand things as quickly as 4:57 me it felt like and then after at some 4:59 point I felt like well sometimes it's me 5:02 um being too naive I like that's 5:05 something I learned over time but uh 5:07 many of the like truths about what's 5:09 going to happen in the future have I 5:11 have a very good track track record on 5:13 and I think that's my this one of the 5:15 superpowers that have made me have made 5:17 me successful so I I felt I had that 5:19 superpower I didn't know it would 5:20 translate into building something 5:22 successfully how did you first make 5:25 money what was your first 5:26 entrepreneurial thing Anton ooh I think 5:30 so I always uh nerded out with kind of 5:34 computers and setting up our um like 5:38 land parties when I was a kid and so on 5:40 and so I noticed that all neighbors and 5:45 like friends or families they had the 5:46 comp computers and they had some issue 5:48 and then they called they wanted my help 5:51 and many of them wanted to pay me 5:52 afterwards so that became a bit of a 5:54 side hustle when I was I was a young 5:56 teenager did you do gaming when you were 5:58 younger yeah 6:01 yeah you're like the I'm so sorry to be 6:03 rude but you're like the easiest like 6:05 archetype because you fit like all the 6:08 characteristics of like successful 6:09 Founders that I have on the show it's 6:11 like number one made money early number 6:13 two exceling gaming uh both very very 6:16 clear um archetypes I want to move to 6:19 lovable so GPT engineer starts as a side 6:23 project where does the idea come from 6:26 we've left to pick G uh gbt engineer 6:28 starts as a side project where does the 6:30 idea come from um this was the spring 6:32 after chat GPT came out and I've been 6:35 playing with the the precursors of that 6:37 as well and I from already like a year 6:40 before then I felt there's an there's a 6:42 massive wave coming from scaling up 6:44 these models with more data and the p is 6:47 not set up for for currently not set up 6:49 for leveraging that after like I was 6:53 actually traveling with my now wife as 6:57 her engagement trip and when you're 6:59 traveling I get extra creative and there 7:03 um I don't think there was anyone who 7:05 was talking about like AI agents at the 7:07 time but I uh during those like sitting 7:10 on an airplane I was I started writing a 7:12 lot of like okay you can you should 7:14 hook them up and make uh you basically 7:17 put the large language model in a for 7:19 Loop and then you can have it to do a 7:21 lot of agentic things and then uh when 7:24 I'm back in Sweden I'm like okay where 7:27 do I apply this obviously on software 7:30 engineering and I'd been talking to 7:33 people about this and I felt no one was 7:34 really sufficiently imaginative to what 7:38 I was thinking about and I had to like 7:40 prove a point in that this is already 7:42 now with the current first versions of 7:44 chbt apis you can build an agent that 7:47 writes code and then I put together that 7:50 there were two I drank a lot of coffee 7:51 and then I just crammed away and um I 7:55 got the first version that really 7:57 impressed people you write create a 7:58 snake game and then you get a running 8:00 snake game on your computer yeah so that 8:03 that was the the first version of what 8:05 what we're building now how long did it 8:07 take you to build V1 with that 8:08 caffeinated session I think it was um 8:13 like one main weekend and then a bit of 8:15 Polish over with a few hours here and 8:17 there on two two weekends after that 8:20 what are the biggest lessons or advice 8:22 from building many different v1s for 8:24 many Founders that are listening it 8:26 depends what you're building for 8:30 um for most like firsttime Founders I 8:34 would really focus on the user and the 8:38 user problem and say like how can I get 8:41 one person to love what I'm building in 8:44 in this V1 uh that that's my general 8:46 advice I didn't do that I just put out a 8:48 video on Twitter and it got dozens of 8:50 academic references and millions of 8:53 people using it so take me to that so we 8:56 have this weekend we released the 8:58 product what happens then dude yeah so 9:01 it wasn't clear to me that I would build 9:03 a business on this at all I thought it 9:04 was fun that there was like and this was 9:06 an open source project so I I started 9:08 nourishing like a community that went on 9:10 to work on this open source project 9:13 while I I talk went to my co-founder and 9:16 said like Okay so this thing is 9:17 absolutely huge um I've been thinking of 9:20 doing something else to be honest and 9:22 they here is like pretty like this is a 9:23 bit of a wake up call for me that it's 9:25 it might be time to find a good 9:26 replacement for me as a CTO at at picked 9:30 that's what happened next and then then 9:31 a few months passed and so a few months 9:33 pass the community continues to grow 9:36 what happens then no but I I I figure 9:38 out a good replacement for me I decide I 9:41 need I want to have a great co-founder 9:43 that I can have as like my partnering 9:45 crime here and and there was a guy who 9:47 is the most super efficient zero fluff 9:52 and and engineer and 9:54 entrepreneur so he had sold a company 9:56 previously that I wanted to work with 9:57 and I I I biked to apartment and said 10:00 hey let's take a walk and uh plan the 10:02 future so he I got him him on board and 10:05 then we started building and created the 10:08 first version of what became loveable 10:10 lovable okay so we're building the first 10:13 version of lovable you've got your 10:14 co-founder talk to me about that time 10:17 when did we release and how did the 10:19 official release go with lovable as a 10:21 product in company so the the launch of 10:24 lovable like the product was one year 10:26 after we started building and um in the 10:28 meantime time we launched like great 10:31 listed preview versions called GPT 10:33 engineer app that was I mean that that's 10:36 the getting user feedback cycle and um 10:41 building building up I guess some some 10:43 excitement about what we were work 10:44 working on and employer branding as well 10:47 and the the first versions I I think 10:49 they were very good as they were good 10:51 they were not very good and we had some 10:53 people really liking it but the like the 10:57 Thea moments didn't click for 11:00 sufficiently many people like okay this 11:02 is how I get real value from this when 11:04 we when we went on to iterate over the 11:06 coming year we packaged together all of 11:09 these things so that you can ask today 11:13 lovable I want to build the basically a 11:15 SAS business and then people have built 11:19 their entire SAS companies and get made 11:20 made money by just prompting our AI I 11:24 just want I just want to break down a 11:25 couple of different things that you said 11:26 there you mentioned the weight list do 11:28 you have any big lessons or advice on 11:30 how to do a weight list strategy well I 11:33 think the weight lists are useful in 11:35 that you can control exactly how many 11:38 people you want to get on board and take 11:40 user interview interviews with with so I 11:42 think that's the like just get 11:43 sufficiently many people on the weight 11:45 list find a good way to qualify who you 11:48 want to talk to like you have you 11:50 probably have a different hypothesis on 11:51 who should get who you should talk to 11:53 who you're going to sell into who you're 11:55 who has the most value from your product 11:57 and then uh qualify only those and talk 11:59 to them when you do the user interviews 12:01 and user feedback sessions what are your 12:04 big lessons or piece of advice on how to 12:06 do them well what questions are good 12:07 what questions are bad any lessons uh so 12:11 there there I think for us there are two 12:13 different type of user to use we we had 12:15 this this type where we just see them 12:17 use the product and see like where do 12:19 they do they understand the product and 12:20 so that's more of a you user experience 12:22 interiew um and the other one is uh to 12:26 understand like if they have just tried 12:28 the product a bit we ask them okay so 12:30 you tried the product a bit what's why 12:33 are you even interested in this and ask 12:34 them like what problems they're facing 12:37 in their business try to identify what's 12:40 the biggest pain point that you that 12:43 they're actually looking to solve it 12:44 might be oh I want to get more customers 12:46 and I think I can get more customers if 12:49 I can show to my customers that I can 12:51 get the first version out with AI more 12:53 more quickly H so those those type of 12:56 questions how does it change the 12:58 structure of teams when you look at 13:00 teams today you have obviously software 13:02 Engineers you have designers you have 13:04 developers you have front end you have 13:06 backend how does it change the structure 13:09 of teams themselves Harry if you want to 13:11 create a personal website for you it's 13:14 super productive you can do it you don't 13:16 have a team it's just you you just 13:17 create it with AI if you want to ship 13:21 the first version of yourselves and 13:22 start making money it's all you it's not 13:24 a team like a team just slows you down 13:25 maybe you get some input from a designer 13:27 like he how does this how do you think 13:28 it's just look and and so on uh but once 13:31 you have existing 13:33 software with users and then you want to 13:37 iterate and change that software 13:39 AI 13:41 might absolutely bring 13:44 down kind of mess up your entire 13:46 codebase I might mess up your code base 13:49 so then you would want to work with a 13:51 software engineer that knows how to 13:54 bring have quality and consistently keep 13:58 quality in the product you mentioned 14:00 like people took a bit of time to find 14:02 that aha moment how important is the 14:04 time to aha moment I think it's very 14:07 important and it's funny because you 14:09 asked me this question and I I think we 14:11 are very bad at making the time to a 14:15 moment super short I think we could like 14:17 double our conversion rates if we become 14:19 better at uh like speed to a h moment 14:24 but you I can say something we're doing 14:26 let me take credit for something when 14:28 you come to love 14:29 you just see a prompt box it's very 14:32 inviting instead of getting to a landing 14:33 page you you see a prompt box and then 14:35 for the ones that do enter a prompt box 14:39 you get a quite a quick a of the first 14:42 aham and there has to be many aha 14:45 moments in lovable it's quite a complex 14:48 a software engineer is a very complex 14:49 feature but uh that that we are doing 14:51 well and that's what I I would recommend 14:54 just give the user something interactive 14:56 with instant reward you mentioned that 14:58 the prompt box itself I've had many 15:00 guests on the show say before that 15:01 almost the biggest sin that chat GPT did 15:05 was make chat the default UI for a 15:08 future of AI do you think that bluntly 15:11 chat and the prompting that we have 15:13 today in in lovable and many other 15:15 products is the right default UI for an 15:18 age of AI I think yes prompting can you 15:22 can do almost anything with prompting 15:24 and explaining your thoughts in in 15:25 written form it's um er it's also easy 15:29 to implement an itate on but but it's 15:31 going to get more advanced over time 15:33 with not just prompting I love the way 15:35 you took a deep breath there it's like 15:38 this is a weighty 15:39 question I'm thinking about it a lot I 15:41 mean you're building the interface for 15:43 creating software and no one knows what 15:45 that interface is going to look like but 15:48 the prompt thing yes I think it remains 15:51 but I'm kind of going chronologically 15:53 through the story because it's an 15:54 amazing story um you rejected YC at some 15:59 point yeah talk to me why did you reject 16:03 YC we felt that at uh best YC would be a 16:09 lot of dilution and some acceleration 16:12 but a lot of dilution and some 16:13 acceleration and at worst it would be a 16:15 distraction H to go to SF and go through 16:19 all of these like fun things that happen 16:21 when you go to YC um and we we just took 16:24 some funding instead and uh built focus 16:28 on Talent the seed round when does the 16:30 seed round come does that come post 16:33 launch or pre-launch so that comes 16:36 before we launch the first the first 16:40 waight list of our product even okay how 16:42 did that seed round go so before you've 16:44 launched the weight list of the product 16:46 how did that round go how did it come to 16:49 be yeah so I I I have this advice that 16:53 I've always gone by which is to work 16:55 with investors that you like and I I had 16:58 some people that I know from before that 17:00 I just think are amazing people that I 17:02 want to have by my side and if things go 17:05 sour if things go well uh and I spoke to 17:09 a few uh I kept it very very brief I got 17:11 like a preemptive offer and I said yeah 17:13 let's let's do it I love it how big was 17:15 this round so we took three million and 17:18 then we added on um I got the advice to 17:21 say like yeah just to get a lot of cash 17:23 because you never know what happens in 17:24 the markets so we we raised a quite a 17:27 large pre uh with an up to uh almost $8 17:31 million would you advise Founders to 17:33 raise quite large preed rounds if they 17:35 can if the money is on the table would 17:37 you say take it depends on how you want 17:39 to operate like if you like talking to 17:42 investors which at the time I was like 17:43 no I just want to build the technology 17:45 then then the answer sure take ra a big 17:48 pred so you can you get time to figure 17:51 things out um if 17:54 you like talking to investors which I um 17:58 I think you should 17:59 if you think it's an interesting to just 18:02 talk to a lot of interesting people who 18:04 care about understanding the market then 18:06 I would go more race more iteratively 18:09 smaller smaller runs we're seeing a lot 18:11 of Founders today be immensely dilution 18:15 sensitive from day one in a way that 18:17 they haven't been before like 10% is 18:19 like the max they're willing to give up 18:21 on a round how did you think about 18:23 dilution sensitivity uh here again like 18:26 I have a very wise person who was like 18:28 no dilution does does matter so much 18:29 it's all about the size of the pie and 18:33 I I think that was I was both affected 18:37 by that advice as so as my like no 18:40 minimized ution this is this is my life 18:42 life's work so that's mainly how I think 18:45 about it now so we now raise this round 18:47 we have the wait list we're skipping a 18:48 little bit take me to go live day when 18:52 when does that happen and how does go 18:54 live go we go live with lavall the 21st 18:58 of November last year 21st of November 19:01 last year that was only four months 19:03 ago yes four months ago isn't it nuts 19:06 how much like life changes in four 19:08 months it's true it can change really 19:10 fast yes yeah so November 19:13 2024 you launch from day one is it just 19:17 nuts How does it go um so I mean we had 19:20 users paying users on like earli version 19:24 that was called something else um and 19:27 then our launch 19:29 I think it wasn't one of these like wow 19:31 launches I I think we could have gotten 19:36 like 10 times more press on the launch 19:38 100% um but people start noticing like w 19:41 this is really good and and we continue 19:43 to quickly improve things in the 19:45 products chip things very rapidly um so 19:48 it ramps like growth starts ramping up 19:50 after we launch and we we like wow we're 19:53 growing 1 million AR per week at some 19:55 point and and that just keeps going 19:57 accelerating through that was in 19:59 December and that it just keeps ACC 20:01 accelerating 20:03 we it was a bit on the technical side it 20:05 was a bit frustrating because we have we 20:08 have some a lot of scaling issues that 20:11 we run into and the team is like okay we 20:13 can continue to uh patch this but they 20:17 say like we let's rewrite everything 20:20 while we're seeing this explosive growth 20:22 and there's so 20:23 many quick fixes that we want to do on 20:26 the product side so you roted ASAP and 20:28 stabilized it then yeah so it took a bit 20:32 more than eight weeks I mean it's not 20:34 completely done to date but that we took 20:38 eight weeks and then now we're shipping 20:41 faster can I ask you you mentioned that 20:43 like a millionaire r a week how much are 20:46 you growing now a week two million year 20:49 per week two million a week now 20:57 yeah like 20:59 it's so funny for me dude CU like when 21:01 you live in a world Adventure you don't 21:03 you don't kind of get this I don't think 21:04 and I don't mean that rudely to you but 21:06 like your companies go from like one to 21:09 four in a year and that's like a 21:13 four a in a year amazing that would be 21:16 good and then you're like growing 2 21:18 million a in a week does that sink in 21:20 like do you know how nuts that is no I 21:23 guess no I I mean um I a lot of things 21:27 last few weeks have just blunted me and 21:29 I'm just I'm just focused on all the 21:31 things that we have to fix and like 21:33 improve and that's all I think about 21:34 what do you think are the most common 21:36 ways or methods that company's 21:39 development process slows down for 21:42 Founders listening what should they 21:43 watch out for that often happens what 21:47 makes product development slower is 21:49 usually that you have a complex product 21:51 and you have a lot of requirements on 21:54 your product so you you said earlier har 21:57 that you have three three things that 21:59 are good in your product I think that's 22:01 uh generally A very wise thing to do 22:04 Simplicity leads to product Direction 22:06 knowing where you're going and what 22:08 you're working on when you look back 22:10 since the start of lovable where from a 22:13 product perspective did you invest and 22:15 spend time where was the benefit of 22:18 hindsight you shouldn't have done I 22:21 think at Lev ball we we thought a lot 22:23 about the community and like Community 22:25 features inside of the product and the 22:29 that I think that could have made a lot 22:31 of sense if we had seen slower growth 22:34 but now that you just like growth is not 22:37 something we not a concern you don't 22:38 need Community features to power growth 22:41 so that I think that was pretty much 22:43 waste waste of effort do you agree with 22:45 the sentiment build it and they will 22:47 come the growth has been amazing from 22:50 day one it feels like since gbt engineer 22:52 even people have just kind of come for 22:54 the product because it's been amazing 22:56 and different do you believe and they 22:58 will come actually does stand true today 23:01 if you have a very strong vision of 23:05 how that where there is untapped 23:07 potential if you really know that deep 23:09 down and you have track record of 23:12 showing that then build it and they will 23:14 come is going to work um uh and and you 23:18 also have the runway your personal like 23:21 energy Runway and so on to really make 23:24 it work and they make them come but for 23:28 in most most cases yeah that is too 23:30 risky it's too risky to just build it 23:32 and they will come and it's you can 23:34 build it and make them come or try to 23:37 make them come at the same time and 23:38 that's much lower lower risk you said 23:41 there about energy and it reminded me of 23:43 a statement that Nick revolute said to 23:46 me in a show when he was talking about 23:47 his investing and he said the most 23:49 successful Founders that he invests in 23:51 are between 30 and 35 they don't have 23:55 the naivity of the incredibly young 23:57 founders but they don't have the bluntly 24:00 tiredness in some ways or that you know 24:03 you have more energy when you're younger 24:05 of older Founders how do you think about 24:08 that given your age today I think energy 24:11 is uh super important and uh I think the 24:15 navity is is a benefit often but uh I 24:20 mean I I I made mistake a lot of 24:23 mistakes was the first time manager as 24:25 well that uh dep picked pretty much what 24:27 was the biggest mistake he made thinking 24:30 that we should change the culture and 24:31 become a more of a scale up something 24:34 slow moving or like more management 24:36 layers when we became 40 people that's 24:39 that was the biggest mistake I did why 24:41 did you think 24:42 that that was I mean my 24:46 co-founder 24:47 um and other senior people we spoke to 24:50 said like oh now you have to hire 24:52 Executives and so on and that uh was um 24:57 that was a bad idea that okay and so did 24:59 you start hiring Executives pretty much 25:01 yeah or yeah we did and then they don't 25:05 work uh one person I hired didn't work 25:08 out uh which slowed me down and they 25:11 kind of set us back a lot and so when 25:12 you think about that what would your 25:14 advice be to other Founders don't 25:16 believe the you can scale way 25:18 longer without exag many Founders hire 25:22 more mercenary style maybe like skilled 25:25 people but just like they are running in 25:27 their Lane 25:29 um I hire generalists and I like try to 25:32 empower them as much as possible and if 25:35 you have a lot of General super super 25:36 smart generalist that you're um doing a 25:40 lot of smart like new initiatives and so 25:42 on adding Executives on top of them um 25:47 is high risk and like reward is 25:52 questionable it's at some point of 25:54 course it makes sense does culture break 25:56 at any point when you're scaling user 25:59 base so fast you're scaling Revenue so 26:01 fast I mean usually I think it does yes 26:04 or it changes it changes it evolves um 26:07 for this is something I'm very mindful 26:10 about and I why I'm so scared of adding 26:13 too many heads what are you worried 26:15 about the most important thing for 26:19 everyone at our company I which I talk 26:21 about is uh to role model how much you 26:25 care about the product the users the 26:28 entire the team how how well the team 26:30 works and role model and Mak sure other 26:33 people care it as much as you and feel 26:36 that that comes from a feeling of of 26:38 ownership of the culture and the team if 26:40 you're a lot of people that gets diluted 26:43 typically 26:44 so um yeah that that's that's what 26:47 breaks or that's what's harder okay so 26:50 we have the team scaling we have user 26:52 base we have Revenue 26:54 scaling we're making actually a lot of 26:56 money at this point why a series a I 26:59 just had some scheduled checkins with 27:02 investors and there was this one guy and 27:07 his team that had like obsessed about 27:10 lavable for the last few months and they 27:13 they're hearing fedck he's here in 27:15 Stockholm at kandum and uh he G he just 27:19 gives gives an amazing impression on me 27:22 and I'm like okay I could I can wait 27:25 because I know it's very clear to me 27:27 that this is just the beginning and I 27:28 can I can could raise it later but we 27:32 can accelerate by adding an investor 27:35 that is a partner in on in helping to 27:39 find more amazing people in um just as a 27:43 sounding board Fredick helped grow 27:45 Spotify from from nothing um and uh I we 27:51 decided to raise a small round okay so 27:53 you decided to raise a small run I have 27:54 to ask this you have many well-funded 27:56 competitors in the US I think it was 27:59 Neil Murray that said you put like we're 28:02 Devon that actually works from like very 28:04 early on yeah which I thought was bold 28:08 um when a competitor has a lot of money 28:11 do you have to raise too I don't think 28:13 so no I mean you can just we could 28:15 bootstrap like you don't you can 28:17 bootstrap most things so you never have 28:20 to raise but they can outs spend you on 28:23 Talent on customers on marketing no yeah 28:27 I mean if you're 28:29 I I'm not afraid of any of those if 28:31 being out spend on any of those I think 28:34 it's the only thing that matters is 28:36 execution so if you can outperform your 28:39 own execution then I'm scared where 28:41 could you improve your execution today 28:43 we spoke about speed of shipping amazing 28:46 culture's great the talent's phenomenal 28:49 where would you say execution wise you 28:51 could 28:52 improve I think you can always improve 28:55 on like how the decision Loop on how 28:58 fast you take decisions and communicate 29:01 those decisions so that everyone is 29:03 really on the same page that's uh how 29:05 how you do that how do you make them 29:07 today how could it be improved I think 29:10 it's about doing fewer I think we could 29:12 do fewer things as well at lovable like 29:15 there's a lot of uh people with great 29:18 ideas that each each idea in isation is 29:20 great but you can only have so many you 29:22 should only execute on so many at the 29:24 same time we mentioned the team and the 29:26 culture one and we mentioned there 29:28 you've been pretty Ardent around 29:30 building in Europe keeping the team in 29:31 Europe being a European company a lot of 29:35 people say to me by staying in Europe 29:38 you are deliberately not doing what is 29:40 best for your career if you were in the 29:42 valley you would be more successful what 29:45 is your response to them the most 29:47 important thing is talent and culture 29:50 and it's more raw there's more raw 29:55 available talent in Europe uh the 29:58 culture isn't on the like the US has 30:01 more culture that fits succeeding as a 30:04 startup I have to say per default so 30:08 what what specifically do you think it 30:09 is about that culture so it's about then 30:12 theault of thinking big and being super 30:15 ambitious and being very committed to 30:18 making things work that that in Europe 30:21 people are more like about uh living a 30:24 balanced uh life and uh we're we're kind 30:29 of taught that we 30:31 should um in in Sweden we we talk 30:35 about the law of jante which is that you 30:38 should be better than others okay yes 30:42 this so This concerns me so how do you 30:44 respond to people who say you would be 30:46 more successful if you were in the 30:47 valley we would be S so if you're a 30:49 Founder I think you you 30:53 have more free entropy of like talents 30:58 to use and channel into some into 31:00 something successful but there are of 31:01 course many benefits from being in the 31:03 valley I think it's a bit of playing on 31:05 the on hard mode here from Europe and I 31:08 get excited about playing on hard mode 31:10 and showing that you can create a 31:12 category defining company uh which is 31:15 lovable in our case from Europe are you 31:18 positive about Europe moving forwards we 31:20 are in a doom loop around Europe which 31:22 we both disagree with and hate are you 31:24 positive and what guides your thinking 31:27 yeah I'm uh I'm super positive I think 31:29 there is a like an strong Underdog 31:31 mentality among all of us Founders now 31:34 which is usually a a winning concept to 31:37 be to feel like the underdog and having 31:39 wanting to prove others wrong I also 31:41 think there's incredible superpowers in 31:43 using the Arbitrage pricing of 31:45 incredible engineers in Europe and 31:47 selling into the US you can build huge 31:51 companies selling into the US from 31:53 Europe doesn't mean you can't sell into 31:55 the US just because you're a European 31:56 company true no it's a very Global 31:59 Market dude um I've got some questions 32:02 which are a little bit spicier but I 32:04 have to ask them okay so a spicier one 32:08 is if one were to criticize lovable they 32:11 would say amazing God look at the 32:14 revenue growth look at the user growth 32:16 but is this like AI um Sugar Sugar 32:19 Revenue in other words it's not 32:21 sustainable and it will churn very 32:23 quickly how do you respond to it's not 32:27 really real stainable revenue and it's 32:29 not sticky we have month one retention 32:33 that's better than chat gpt's month one 32:35 retention on paying customers and so 32:38 it's about 32:40 85% and that um that is just going up 32:44 like they of course there's a lot of 32:46 things to be desired when you working 32:49 with an AI system there's a fraction of 32:51 people some people that come in and 32:53 they're they just like flip up their 32:55 credit card because they want to try 32:57 they they want to learn which is 32:59 rational thing to do and then they know 33:01 that they are going to stop using it 33:03 almost they almost know so we have a 33:05 fraction of users who are like those 33:07 people wow 85% month one yeah wow that's 33:14 great because you're also too early to 33:16 have month six what could you do today 33:19 to increase retention most significantly 33:22 the easiest thing we can do is to 33:26 give more more of the important aha 33:29 moments to ensure that every all of our 33:31 users get more of the important aha 33:33 moments on how to use the product I I I 33:35 don't I don't really respectfully agree 33:38 or get you because like you give prompts 33:41 Under The Prompt box so you say like 33:43 build me a sass app or build me a dog 33:45 website or whatever the prompt is dude 33:47 all you need to do is click it and you 33:49 see the code being written and then you 33:52 see it being created like the aha moment 33:55 is pretty obvious no so there are many 33:58 aha moments for or like education 34:01 moments to get the most value get the 34:04 full value out of using it and there the 34:06 most important of those are with regards 34:09 to when the when you feel like you're 34:10 getting stuck and the AI doesn't 34:12 understand you and there's a there are 34:15 many things in how to get around around 34:19 that that you can pick up as user it is 34:22 how you about how you prompt it's about 34:24 how to understand what doesn't work and 34:26 explain or explain clearly what what the 34:29 what the problems you're seeing what 34:31 what you find could be the problem when 34:33 you're building a more complex feature 34:36 and it's about that you can actually 34:37 onboard an engineer to do small changes 34:39 to the code base like those are things 34:42 our users should know and not everyone 34:44 knows them what's your norstar matric 34:46 today if you have one metric on the TV 34:48 for the whole team to focus on what is 34:51 it um it's the number of uh users that 34:57 go all the way to getting like users on 35:00 what they built getting getting 35:02 something hosted with users on what they 35:03 built that's what we focus on what's the 35:06 amount today so we have almost 40,000 35:10 paying users and then that's the proxy 35:12 for that wow 40,000 PES as okay do 35:17 you care about the time it takes to go 35:19 from start to website created or not we 35:22 care about I I have to say Harry like 35:25 the these things about the onboarding 35:27 and how much can improve um is not 35:29 something that I we have had focus on 35:34 like we were just making the core AI 35:36 Parts better better better that's what 35:37 we focus on okay so you say about making 35:40 the core AI Parts better better better 35:42 yeah amazing people also say ah rappers 35:46 on top of other people's models why are 35:49 they wrong with that assumption I mean 35:51 it's very easy to create a cool demo 35:53 with just a rapper it's super easy to 35:55 make a cool demo with just a rapper the 35:56 hard part is to get to or get close to 36:00 100% you get what you're asking for and 36:05 there's a lot of small 36:07 details in making that work there's a 36:09 lot of small details and there's a chain 36:11 of at large language model API calls and 36:13 other algorithms that run and that chain 36:16 is like you can continue to optimize 36:18 that for years without reaching 36:19 Perfection whose models do you sit on 36:22 top of today we we use all of open AI 36:25 models Google Gemini and the main 36:29 Workhorse is anthropics Claud model for 36:32 writing the code I was going to have 36:34 this in a quick F but I have to ask it 36:35 now you've got anthropic at 60 billion 36:40 you've got open AI at 300 or you've got 36:42 grock at 50 which do you buy and which 36:45 do you sell okay so um I 36:50 think I would I I care about the best 36:53 talent here and I feel like uh Elon is 36:57 very good at tent so I would buy I would 36:59 buy Gro I think they're also very 37:02 ruthless in like finding business 37:03 opportunities we we'll see about that 37:05 but I feel like they could be good at 37:07 that even though anthropic is my 37:08 favorite I love the the culture and I 37:10 love the like the leadership there um 37:13 and I will short open AI because they're 37:15 they're like have been very good in the 37:18 scrappy face but they they haven't 37:20 proven that they can um we have a clear 37:24 product Direction and focus now over the 37:27 last years 37:28 so I would push sorry I'm it's a Sunday 37:30 we're allowed to be a little bit more 37:32 casual I would push back on you and go 37:34 the two biggest things that matter in 37:35 this next wave is brand yeah recognition 37:38 on the consumer side and it's consumer 37:41 facing frontend products and when you 37:43 look at Brand everyone's mother knows 37:45 chat GPT they don't even know open AI 37:48 but they know chat GPT and then on the 37:50 consumer product side respectfully 37:52 they're way ahead of anyone 37:54 else that's true am I wrong I just I 37:57 just think we're still early in the days 38:00 of AI and if you look at the Enterprise 38:02 Revenue like anthropic is almost caught 38:05 up to open AI from nothing from being 38:09 absolutely dominated and I don't know 38:12 what Gro is going to do here but I 38:15 expect them to be having a like open a I 38:17 lost all their best talent to to 38:20 anthropic I think Gro might be able to 38:23 pull something off here I spoke to a 38:24 very wise friend of mine before this and 38:26 they said that the biggest conc that I 38:27 have is that open source or Mega corpse 38:31 of the world with massive distribution 38:33 advantages come in and win the market 38:36 how do you think about that and is that 38:38 a concern the big Mega Corps they move 38:41 very slowly in many domains so they're 38:44 never going to in many years they're not 38:46 going to have the best product on the 38:48 market there are distribution advantages 38:50 for some of this Mega Corps and that's 38:52 like the marketing and distribution is 38:54 is what I'm more concerned about um but 38:57 overall it's going to be like a growing 38:59 Market where you can easily as a startup 39:04 be both best position for some parts of 39:06 the market one thing that's going to be 39:08 a real shift I imagine that you have to 39:11 face is the shift from plg and kind of 39:14 prosumer to Enterprise how do you think 39:17 about that shift and is that just not a 39:19 concern given the speed of Revenue ramp 39:21 how do you think about that um yeah I 39:24 think if we would do Enterprise I would 39:26 want to do that really well so we're 39:27 holding off with doing Enterprise for 39:29 now we're holding off and U our goal you 39:33 may asked about the north Zone north 39:35 star metric what we want to have is we 39:38 want to be the best place for for 39:40 Builders to create products and get the 39:44 hund get get a million of the most 39:46 talented builders on lovable and if we 39:48 succeed in that that's a great segue 39:51 into many other areas including 39:53 Enterprise when you look at 39:55 Shopify what do you learn from them 39:58 blazing a trial in the way they have 40:00 done as efficiently as they have done i' 40:02 I'd I'd love to hear from you har what's 40:04 the what do what are you most impressed 40:07 by with Shopify honestly I would say 40:09 it's their narrative around everyone 40:12 being an entrepreneur I think Harley's a 40:14 phenomenal communicator and Storyteller 40:16 at bringing very real businesses to life 40:19 and how Shopify changed their processes 40:22 their sales and really made an impact 40:26 and I I love to and I think he's 40:28 obviously a brilliant technologist but I 40:30 think without Harley's storytelling 40:33 narrative entrepreneurial Vibes to life 40:38 it would be a different company so I 40:39 would learn from that a lot in terms of 40:41 how to tell your customers 40:43 stories yeah to not totally I I think 40:47 Shopify um has been able to execute on 40:50 many things and just created a good 40:53 package I they've iterated fast that's 40:55 my impression they've iterated fast on 40:57 serving all the needs for for building 40:59 your e-commerce and I think this 41:01 velocity is was one the most important 41:04 thing for building the best product when 41:05 we look forward dude what would you most 41:08 like to build in the product but for 41:11 whatever reason you cannot why what 41:13 would that reason be I think I can build 41:15 anything into the product really like 41:18 team would block you um that PRI 41:20 prioritization not the right time okay 41:23 so something that we have to say no to 41:25 right now yeah so I'd love to build in a 41:29 way for Founders to get everything you 41:33 get out the Y combinator into lavable so 41:36 that that includes like some marketing 41:38 infrastructure why comor has that uh all 41:41 the entire Playbook and going above that 41:44 and even saying like oh here's your 41:47 Incorporated Delaware C comp company 41:49 your stripe setup you just have to focus 41:52 on the product and talking to users and 41:54 creating content for marketing when you 41:56 look at partnership ships today apis 41:58 that we have today into the different 42:00 providers that's that's quite a real and 42:02 possible product yes yes it's a it's on 42:06 the road map wow yeah good okay well YC 42:11 are going to love you after this show 42:14 uh um final one when we look forward 42:18 dude what concerns you most if you had 42:20 to choose was it be regulatory 42:22 challenges competition or hype Cycles I 42:25 think if someone competitors is super 42:28 super good at marketing uh that would 42:30 concern me how important is 42:33 brand I think brand is mainly the 42:37 outcome of your product there's other 42:40 factors too and then so I mean product 42:45 is the important thing and then product 42:48 good product plus awareness just creates 42:49 brand that's that's I think was 42:51 sufficient so it's more of a downstream 42:54 effect of something else then it is 42:56 important when you're doing 2 million a 42:58 week in in additional ARA you're going 43:01 to have more and more investors want to 43:03 give you 43:04 money how do you think about that and 43:09 what what makes it worth it to take 43:12 versus this is a distraction get out of 43:14 my face it's currently it's a 43:16 distraction I think it might make it 43:20 worth it it will make it worth it when 43:22 we know know how to exactly how we want 43:23 to spend the money or it's a partner we 43:26 really want to work with and um it's 43:30 like terms where we we would never 43:33 regret saying yes to them what if done 43:36 would be a massive needle mover for the 43:38 company one or two more of the perfect 43:42 technical product 43:45 hires dude I could talk to you all day I 43:47 want to move into a quick fire around so 43:49 I say a short statement you give me your 43:50 thought so yeah what do you believe that 43:53 most around you disbelieve I think we 43:56 have models today that are really smart 43:58 I think that's where people don't agree 44:01 with me they're they're smarter than 44:03 humans but they don't have memory in the 44:05 same extent as we do they don't have 44:08 context how fast will we ramp context in 44:11 memory given scaling today it's the ramp 44:15 what you need to ramp is to decide what 44:19 like the the process for storing memory 44:22 I think is what you have to ramp and it 44:23 has to store all of these things like my 44:25 conversation with Harry has to be stored 44:27 somehow into the system and all like my 44:30 childhood almost has to be stored as 44:31 well I don't know it's going to take 44:33 years you can buy and hold one public 44:36 stock for the next 10 years which one do 44:38 you buy and hold um some Talent play 44:43 would and what comes to mind is Tesla 44:45 and they're also interestingly 44:47 positioned outside of uh software wow 44:50 okay what was the most important trait 44:52 in a Founder that no one talks about U 44:55 okay I think great judgment people 44:58 probably talk about but but I think um I 45:01 like people who hire like see potential 45:03 in people it's like okay this this 45:05 person can become amazing and and the 45:08 Judgment in picking Talent like that 45:09 what's your biggest weakness as CEO 45:11 today um I'm not as good as multitasking 45:15 as I would like to 45:17 be if lovable failed tomorrow what would 45:21 be the reason like investors WR a 45:23 premortem which is like a reason why 45:25 something doesn't work what is that um 45:28 we lose momentum and excitement and 45:32 that's like that's what fuels us today 45:35 right all the excitement and the 45:37 momentum what have you changed your mind 45:39 on in the last 12 months you don't need 45:42 to be attached to One Foundation model 45:44 provider they're all going to be amazing 45:46 you don't need to there's not going to 45:47 be one winner here are we seeing 45:49 Foundation models be completely 45:51 commoditized 45:53 yes we won't see specializations like we 45:56 do today I know Claude is better for 45:57 codee and 45:59 Engineering do you think that will 46:01 really uh equalize yeah I love the sweds 46:05 one word 46:07 answers um what's your favorite failure 46:11 um I I think so when I was 17 I failed 46:15 my exams at school Anton I failed them 46:18 badly yeah like D D's across the board 46:21 and I I looked at myself and I was like 46:23 I'm either going to work behind a bar or 46:26 I can make something of my life mhm and 46:29 I chose to work harder than ever and 46:31 make something of my life 46:33 um that is my favorite failure I think 46:37 my favorite 46:39 failure um I 46:42 probably that I didn't 46:45 um the pict didn't work out so I could 46:47 do more things what concerns you most in 46:49 the world today I think that leadership 46:53 in this in the world today I I love that 46:55 whole leadership to be true idealists 46:58 and not like function of circumstances 47:01 and the people to pleas um and be much 47:05 more open-minded to people who disagree 47:07 with them and super intelligent as well 47:11 but leadership in the world today uh are 47:14 usually like a bit corrupted and uh 47:18 narrow-minded and that's what concerns 47:20 me do you think they would say you're 47:22 naive I uh like I want to be naive to to 47:27 the challenges of leadership to the 47:29 challenges of having a you know a voter 47:31 population you sometimes have to be 47:33 certain ways I think we used to have 47:35 like better presidents and leaderships 47:37 in throughout the last century so no I 47:40 don't think I'm too n but I'm definitely 47:42 always a bit I totally agree with you I 47:44 think the state of leadership today is 47:45 woeful um who penultimate one what's 47:48 your biggest short in the public markets 47:51 I not so read well read on the public 47:53 market it would be a SAS company it 47:55 would be like a SAS company that has a 47:57 per seat 47:59 pricing where th those their 48:02 ICP um is going to be replaced by AAS 48:07 that just replaces the employees so the 48:09 the number of SE goes down what's the 48:11 future pricing in an AI world yeah I 48:14 think it depends on how defensible your 48:16 you like your business is if you're 48:19 selling to like Enterprises that never 48:21 change SAS software you should try to 48:24 make some value based pricing I don't 48:26 know how you how you get the good value 48:28 based pricing but I would do something 48:30 like that otherwise I think you should 48:32 should don't innovate so much too much 48:33 on pricing do what do whatever works 48:34 today I love the facial expressions uh 48:37 my my final one I like optimism and I 48:40 think we need more of it in the world 48:42 I'm incredibly excited for the next 48:44 decade because I think we'll find cures 48:46 to my mother's got multiple curosis 48:47 which is a horrible disease I think 48:49 we'll find cures to diseases that we 48:51 really didn't think possible before what 48:54 excites you most about the next decade 48:57 that we have coming I think it's that AI 49:00 is hopefully going to make us humans 49:03 understand each other better and be 49:06 better at doing win-win playing win win 49:09 and that would be awesome I'm excited 49:10 about that so like you our leadership 49:12 actually being um enhanced by super 49:15 intelligent AI Anton listen dude I've so 49:18 enjoyed doing this thank you so much for 49:20 joining me and thank you for inspiring a 49:22 generation of European entrepreneurs 49:24 that they can build unbelievable 49:26 businesses is in Europe thank you har it 49:28 was it was super fun