Why the Best Founders Are Neither American nor European

Meanwhile, a decade ago, you built a drone company. That one got acquired by an American company, and that American drone company then blitscaled with American investors. They grew the company quickly past what the business could actually sustain and then walked away. Now, most people would get disillusioned by venture capitalists and just take a job after that. But you start another company, an even more volatile space than drones. What did you learn from the drone experience that fundamentally changed the type of founder you are today?
The drone venture was like I don't know eight years long. Yeah, about seven to eight years long. And I I went through all the phases. You start a business with a completely new market back in 2012. [snorts] No one was using drones at all. We were like crazy guys trying to take pictures from the sky with like toys, literally toys, super expensive toys that were crashing like 30% of the time. I'm not

kidding. Uh and of course 10 years before the all the you know the war in Ukraine and so on. So very very new for the regulation and so on. So we we saw the market growing up which took a while and it's it's interesting. The pari is interesting. Then we raised some money. Then we sold the business. Then we joined the acquire. Then we went bankrupt. Then we managed to buy back the assets to keep the business going.
Yeah.
And basically we sold it a second time. So yeah, lots of ups and downs, but that it's actually a fun journey. It's not just about one tough moment when you when you get laid off uh in one hour, you know, when you go out of business. So lot lots of lessons in that in that journey.
And how did that change you as a founder? Are you now allergic to blitz scaling?
I am not. Um I know that you can have very very low moments when you think

everything is just done and you just got to go find a job and just wait a few weeks to a few months and you're back in business again. Um, that's one thing I've learned because when we when we sold the business, I could have raised money and I eventually sold it probably a bit too soon. Should have kept it. Um, when we went bankrupt, who would have thought that we would be in business a few weeks later, even the customers were starting to leave and we managed to keep them and so on. So, I think, yeah, resilience, that's the sorry, the term I was looking for. I think it taught me about resilience. Um
that that doesn't even sound like resilience.
I think resilience is automatic that comes to founder that almost sounds like optimism. Even though things look really bleak, you have a fundamental belief that even

when the house is burning and half burned, something good can happen and this turns around. So never give up in that in that sense is there's resil res. Yeah,
but optimism in the sense that you believe in a better tomorrow.
Resilience is the polite term for business books.
I think the real term is kind of a bit of craziness or you know foolishness.
Naivity is the naivity. Yeah.
Stubborn stubbornness. Um but yeah, no and and we we've been through lots of ups and downs with the AI venture as well. Um, first scenario was not an AI company at the beginning. It was a metaverse company, of course. Yeah. But it was not a crypto company though. Um, we we skipped that part. Um, we initially had a product to scan your I would say objects with a phone. So you would go around an object, anything, a chair, a mug, nothing. and

then get a 3D model from the phone from the phone which was um back then pretty revolutionary to get a 3D model by just taking pictures from your phone and I thought it was the right it was one very interesting way to facilitate the creation of 3D contents at you know at a bigger scale and it turned out it was it was a pretty terrible idea. It was way too long, way too complicated. The the output were pretty bad in terms of quality. So that we tried, you know, we had some, you know, early testers and customers. We tried for about a year, a year and a half and we realized it wasn't the right product. It wasn't the right market and we had to change. That was late. That was mid22 and we had to pivot pretty hard with a few months of uh cash left and yeah, like a few
[laughter]
you know um yeah we had to also it's sad but we had to let go half of the team

and and work really hard and that was pre-chreai I mean pre preach GPT at least of course open
GPT4 at least
GPT 3.5 which was the first oh 3.1 maybe I forgot one of them um but we were using that we were using GPT before CH GPT stable vision 1.4 4 and 1.5 and midjourney v3 that was like the prehistory stone age of AI and I I thought it would be super cool to do a business in that space and that's where we pivoted and we got super lucky because then Chad GPT came out 23 and then 25 24 I mean we we all know the the story of the uh the AI wave uh we we got early we got in early we got in at the right time and that's another lesson for founders is timing matters more than all your smart plans about a product and a market and a need and so on. And timing is everything

in sickness and in health in soft launch and in live ops through every meta shift and every platform policy change. That's the kind of loyalty vibes built between your game and your players. AI powered targeting, a [music] real metal layer rorowass optimization that actually compounds. That's what makes Vibes the fastest growing loyalty platform in mobile games. Till churn do us a part. Vibes vybs.co.
Yeah. Well, you have had a great timing. I mean, drones a little bit too early.
Way too
metaverse on time but kind of late. But you still were able to raise and now AI just when the wave was picking up there was no wave. There was a there's a feeling of a wave and you jumped in on the AI.
Yeah. Yeah. And um by the way the the there was a a discrepancy between

the feeling the sentiment from us founders and from investors versus the users specifically in gaming. The users in gaming, they took years, two years more than us founders and investors to realize that this was a big thing to look at. And for the first two years after we raised for the AI uh product for the first two years, I mean it was a ruthless um a ruthless time evangelizing, doing demos, but not really convincing any big customers. And things change in 25 and things completely change in 26. Like 26 is another story.
How much of these pivots, especially as we're talking about these, you know, metaverses and AIS and crypto and drones. How much of it is driven by the investors and how much is it driven by founders? Because ultimately as a

founder, as you said, you were running out of money. Your money doesn't come from customers in the beginning. It comes from the investors. So it's driven by cash first and foremost usually um no matter where the cash comes from. Um in our case it was very much driven by the founders and only the founders because we could have said okay sorry guys it was a preede round a small round uh bad idea. You know people are grown-ups the the investors they they would have understood that we would have just fold the business. Um and I I know a lot of funders that would just have faulted the business. No drama. By the way, the AI product was not too far from the metavos product. It was about content creation, just not via scanning, but via AI models and printing and and and and stuff. So, the the DNA was kind of similar, if you see what I mean. The tech, the underlying tech was different. Um, it was very much driven by by founders and I would say investors

stepped in to help really like by the way, their friends, it's their money, they're good professionals. Um,
but they're LG's money.
Yes. Okay. Well, they put Okay. Yeah, they have some skin in the game though. You can't say. You can say no. Yes. Um, no, they they stepped in, but they they had they weren't in they weren't at least at that time in their position. They weren't in the trenches like we were. And I think a pivot Yeah. If it's not driven by the founder, you can just forget about it. If if the if the investor is asking for a pivot and mandating you know forcing the founder to run that pivot given a pivot is very hard um you know psychologically in terms of workload and so on it won't work.
I've seen investors venture capitalists push really hard for pivot to whatever is hot right now. And by push pushing

hard, I mean threatening, promising like the brighter future. You do this, you'll be rich. You don't do this, you will die. And and just arguing with the founders, pushing them to make the pivot. If the founder is has not been receptive, like saying, you know, up to fire half of your team, the everyone who's not on board with whatever is hype, let them go. Just just go ahead. This is this is the the direction. And then that same direction ends up in a disaster in about a [laughter] year in because whatever was hype is is you know already too late if you're jumping on a hype train. Is that your experience?
No. I mean for the AI pivot that wasn't our experience at all and and we had some mini pivot by the way but after the seed the seed round that we did in 20 late 22 uh on like which you know should we should we bet on video or images? Should we bet on gaming on marketing? Should we whatever um should we raise a lot? Should we stay profitable? Should we you know these are like mini pivots

kind of and it was always very much founder le and in good in good condition with the the investors.
What's what's a mini pivot? What's a mini pivot?
It's like
I've never heard of that.
I'm sure your UI team [music] is good. The question is whether the data feeding their decisions is keeping up with them. Pathmatics from Sensor Tower will give [music] them what your internal dashboards can't. Historical spend across every major platform. Competitor creative rollouts, share of voice, trends, and seasonal patterns you'd never spot otherwise. It's the difference between playing with fog of war and having the whole map revealed. Pathmatics from Sensor [music] Tower because you want your UA team to be OP. like we are typically okay one example we are pretty strong in gaming I felt that gaming was the best market for AI visual AI for scenario

because why gaming is big okay hundreds of billions of dollars
gaming is content heavy and most of the money in gaming has to go towards content
so that's a big pie right
marketing sec content second
that's okay yeah but marketing needs content video content, UA content, and so on. Yeah, it's still content. Okay.
And so if you if you're a consultant like a Mckenzie BCG guy, Bane, you're like, "Yeah, let's go gaming. It's the best market for for AI products for your stuff."
Mhm.
It turned out it was the most demanding. It was the most um not reluctant, but the most um the hardest to convince, the most conservative actually. And yeah, and it's completely uh unexpected that an industry that has innovated for uh rendering, GPUs, you know, multi multiplayer and so on, it's so innovative. Um it's still very conservative when it comes to new tools

sometimes. So a mini pivot would say would be like okay let's forget gaming or just keep gaming on the side and let's focus on marketing e-commerce social social media and so on. You you keep the same tech you slightly change the course of your uh of your business.
Got it. So many positioning it's it's not a pivot
positioning position. Yeah. It's not kind of Yeah. Yeah.
Everybody listening to this is probably not surprised that you're European. So, I wanna I want to ask
I cannot despite what eight years in California you're you're you're holding the fort on the French accent 100% despite my US passport. No way.
It's it's okay. I've had a I had a French co-founder. I'm very fluent in French glitch. What do you call it? And so, there's this pattern with European founders. You move to the US, you America, you adopt the the American startup culture

because there's sort of like an implicit message that European ambition is somehow insufficient. Yet you are French. You've raised from American investors. You're selling to American studios. You built your first company in France that was acquired by an American company. So, do you feel like you're expected to perform a version of yourself that fits what Silicon Valley expects? And has there been a moment where you thought that maybe the performance is costing you more than the access is worth?
Okay, it's a tough tough questions. You have to perform in Silicon Valley a certain version of yourself and if you're if you're a Silicon Valley native, that's just your version, right? The fact is there are very few founders that are Silicon Valley natives. Actually there are a few um but most founders come from all over the US and all over the world including Europe

including Middle East you know Asia and so on and it take years to find out who that new version of yourself is. not weeks, not months. And the big mistake that European founders do, which I did, by the way, back in 2016 to 14, 14, whatever, 14, 15, land in San Francisco, you know, find some trade show to, you know, find an excuse to book a flight, find a trade show, go to an incubator and believe that, yeah, you're going to raise quickly there because there is so much money, because your business is looking good. And actually it doesn't work like this. You've got to leave. You've got to really be in the the ecosystem for many years to understand what the new version of yourselves will be. And yes, so there is there is another version.

Is it is it the bad one or good one? I think it's just another one, you know, and it's um and is it costly? I don't think so. I mean I I could have listen I would have never done the pivot that we did back in 22 in Europe ever. Never ever. And I found some investors. First our investors supported but the new investors that were shown that were that were presented with a new business plan, a new plan were crazy enough to say, "Yeah, these guys actually messed up for one year, but they're crazy enough to show me something that looks cool. I'm going to fund them." um you you don't find that in Europe. Um so is it costly? I don't think so. Is it does it require hard work? Yes. The the myth um according to which you you learn in Silicon Valley and you make a lot of money doing your startup your software startup and so on. It's completely uh it's for the books and for the you know blog post. Uh most founders, yeah, the

fables, most founders work,
I don't know, 80 80, you know, 80 90 hours a week in the with ups and downs, of course, but it's and in AI, it's no AI companies, you know, AI is supposed to make us work less, right? Supposed to take take off some work and and workload and so on. Most AI most AI
Yeah. Well, most founders work like twice more than they they would work in another company, you know.
Yeah. We were just starting this podcast. You were talking to me about how today is looking good because you spent Sunday all Sunday working while your family of four kids was uh was with the uh with the in-laws. So,
yeah.
Yeah. That's the uh the life of the founder.
Trick for a good Monday. Work on Sunday, you know. No Monday emails.
Yeah. Start early. Uh, what was the version of you, the French Emmanuel who moved to Silicon Valley like? How did

that change you as a founder? Saturi, it's the Japanese Zen Buddhist term for sudden enlightenment. That flash where everything clicks. We're told that Heroic Labs team had theirs when they realized that most studios run live ops like a horoscope. Vague, kind of generic, same advice for everyone. So they built Sattorii, real time experimentation, player level personalization, events that actually respond to player behavior. Say Sanonara to your old live op stack and Konichiwa to Sattorii from Heroic Labs. I mean, I [clears throat] hope I hope it's not going to be um too cliche to say that the ambition actually in Silicon Valley is way beyond what I've imagined. And I I can [laughter] can give you an anecdote that could be cliche, but it's a real anecdote. I'm getting ready to pitch

one of the top investors on Sand Hill Road. Back then, pre-COVID, you used to go on Sand Hill Road office by office and you would pitch investors in person. Um, and I get prepared and I meet I meet another pretty experienced founder, French American founder in Palo Alto and I and I show him my my deck and uh the plan and he's looking at the deck and he's quickly saying, "Well, okay, great story. You should be fine. There's an issue with your financials. Like, what's the issue?" and he says add a zero. I'm like add a zero what? Like you just add a zero everywhere and on the in the Uber between PaloAlto Starbucks and Sanhill and I was really pitching one of one of the top guys like think top guys and I was adding a zero in my deck and I pitched the the whole slides and it was like yep that sounds like a good plan. I had actually had a second meeting with the same investor and he

probably would have lost with my initial plan that was 10x smaller. They know in Americans they know that your plans might be completely I don't know you you may exaggerate but they appreciate the fact that you're willing to aim for the moon. You know what? You see what I mean? Like they actually not aiming for the moon would be a sign would be a bad sign and um would be would be something that would led them to not investigate further.
Yeah. So if you approach this in a in a typical logical matter, here's the target addressable market. Here's how we would take it and like not even accounting that the market could actually 10x or 100x and you're taking a percentage out of that. So, so they want you to not only take a size of the market, they want you to grow the market as itself and own it after that.
I just love the the whole what I found

in Silicon Valley. I I loved every every part of it except healthcare cost. Um but the yeah the ambition, the fact that people are always very very much looking for the the innovation, anything new, it's I don't know, it's nowhere to be found in the world. Not that I dislike what I what I have in France, what I have in Europe. Um, all my family is, you know, very much European. Um, and there are some great um there is a great culture as well that could be very beneficial for startups. But it's I think the mix is the best. It's like when you when you have that ambition and forwardlooking mindset that is very much us but you you know how to be grounded sometimes which I feel European are better at. I think you're invincible. I I feel okay. I just I just just made the US a very American um very American

remark here saying.
Yeah,
of course. Uh the best.
Um you said about something about ambition and how it's important to be ambitious, how to it's important to 10x even your wildest pitch. Uh, do you agree with the sentiment that European founders lack ambition compared to the counterparts of other continents? Or is it just that that in America you have access to such a ridiculous capital that it allows you to be reckless? 7 years. That's not a sponsorship. That's a marriage. And that's exactly how long we've been partnering with Appsfire. [music] And the reason it has lasted is simple. Every growth team I talk to [music] on this podcast, every operator who comes and breaks down what is actually working for them, [music] eventually the conversation lands on measurement. And in gaming, we refer to reliable measurement with a single word, [music] absier. Just think about it. Maloco, Meta, Google, and Unity all back

this platform too, which [music] means the people competing hardest for your budget all agree where the single source of truth lies.
[music]
Uh not a simple question either, but it's it's it's all about your um referential your everything's relative. Um I don't think European founders are not ambitious. They just grow in a market in an ecosystem that they may not that they that may not convey. There could be some very ambitious stories and and so if you grow in the US then you will learn that sky is the limit and if you grow in Europe you may learn that no you have to build um
the regulation is the limit that's what you that's what you [laughter] learn um yeah you say it didn't [laughter] so it's not I think I think Europe it's not that European are not ambitious I think they are in their own environment

and you can find some very aggressive and very ambitious founders in Europe as well that don't need to go to the US all the time.
Would you recommend to European founders to go to the US?
100%. So much man. So much even I mean even
that answers everything. I mean [laughter] it's not even a question to me is but even like four five years you know you don't have to spend your whole life you don't have to you know give up on everything and sure you'll miss a lot of the good food the culture the lifestyle yes uh you'll have some sacrif there will be some sacrifices um but um but there are stuff you learn in the US you cannot learn by living just in Europe now of course Europe is not just one block there are like different countries you know uh different cultures within that beautiful continent and and I I'm pretty sure some countries maybe do better at startups

than others. Um but even within communities and within cities and so on, you'll find some pockets of uh I don't know ambition. Um but yes I mean at least try one my my became my friend my mentor my advisor Eric once said when I left he said you know worst case you may fail you may not make much money but you learn so much and that's that's very true you'll learn so much more moving there than if you stay and trust that you'll make it you'll make it just uh in your home turf. So yes, do move in the US even with the current uh ecosystem right now. Just do just do it.
Yeah. Yeah, I 100% agree. I I was there. It it is a it's a it's a magnitude difference and um everybody around you is ambitious or borderline crazy uh with

what they're trying to do or sometimes just crazy. And it is quite the different environment where the craziness is not um shunned upon but the lack of craziness is.
Yeah. And failure is celebrated.
Yeah. It it truly is um depend
it's not stigmatic like you can you can run wei works to the ground burn $40 billion and then raise a new round have an Apple TV show made out of you that yes failure is not is not a stigma
it's not a stigma and if you can prove that you learn from it then then it's going to be a good very good point a very good um strength typ typical European would would consider failure a stigma So yes, do move to the US, find a way um and you can start with just some business active business trips and at some point when you find the right place, the right visa, the right business partner, the right investor.

Yeah. Uh something that we talked about before is uh that scenario is doing well and you are on your way to profitability. uh which for an AI company I mean that that doesn't profitability is not in the vocabulary of an AI company. Let's just be honest. So I I'm I was left wondering is this an identity choice based on your experience with blit scaling? So kind of burning money super fast to to grow your revenues even at the risk of of the company. Uh or is this is this um like you not giving power to others to to kill what you've built? Um neither I would say. Um so first yes I think next month is when we when we hit the when we hit the profitability checkbox. So I'm waiting for this I'm waiting for the numbers to be uh on the you know on paper and then maybe do a

little blog post or something.
Um
ARR with a Y with a with a with a Y column as well. Remember that. um my investor one of my investors was saying looking at numbers he was saying yeah you you are swimming against the current uh trying to be to be profitable as an AI company um because everybody's like spending like there is no tomorrow the margins are very very low if like super bad
um there are two things first we massively improved how much we can produce with a small team. So we are only a 15 people right now. We may grow a little bit especially with you know things going well, new contract and so on. We may grow to like maybe 20, 25, maybe a bit more but reasonably uh at least for now until I raise more money which is still being discussed. Um but AI made us so productive that we can deliver a lot with a small team and then be productive
by working by working six [snorts] to

seven days over the week. Yeah, I know. I know.
For for for 18 hours.
If you if you want, we can say default alive versus profitable. That's a more uh you know, more true statement. Um and then no, I think I think it's just that I didn't want to overspend too fast, too much, too soon, which is one of the mistakes. Yes, we did back in the days with the drones. We we felt the market would grow faster. We felt uh it was the time to have the biggest brand without a product which is a mistake. Uh and yes and eventually we we the investors have to pull the plug. Um right now it's not going to happen with scenario for sure. It's no not not going to happen. Um whether we need whether we need to chase profitability at profitability at all cost is still a

real business question because if you're the only one to to do that while everybody's spending like crazy you may not you may miss a business opportunity. You may miss something but at least I know the the business is healthy. Yeah. Um which which brings me to a point where I mean if you want to talk about it but it's um that whole GPU the price of GPUs the price of the models the tokens um there's a massive imbalance right now and the token price will raise there's no question uh tokens have been massively massively subsidized by VC money and and public money and so on um and it won't stay this way so I'm I'm not sure exactly how it will end up but it's I'm seeing models getting better but getting more expensive as well and I think budget will increase
more expensive in how many tokens they spend. If Apple owns your checkout, Apple owns

your player, their receipt, their refund flow, their win back email, leaving you as a tenant in your own economy. Appch charge hands you the keys to the castle you've built a worldclass web store [music] and a direct line to your highest spenders the platforms have been quietly keeping from you [music] for a decade. The 30% you save is the boring part. Owning the players relationship is the actual unlock app charge provides. uh basic basically the models are getting better meaning you don't have to try many times over and over again to get what you want whether it's code image video 3D whatever okay the models are getting better but each inference is getting more costly um and it's what you see with GP GP image 2 or Gemini nanobana uh Google openai amazing models honestly versus

the the models we had the stability you know SDXL and so on like Gemini and GPT image 2 are amazing. But a good image, a good GPT2 image is easily 4050 cents, while a Flux or SDXL image would cost you 5 cents or maybe less. 1 cent, but you know, you get there with one prompt if you if you do if you if you do it with
if you're good at prompting.
Well, if you're good, it's not just prompting. It's good at, you know, building your workflow. So you usually it's never one prompt. It's more like a series of prompt. Uh but if you if you know what you're doing, if you're in control, yes, you get there very you can get there without back and forth and you're not you're not playing uh you know you're not with a slot machine. You're not at a the AI casino uh which we've which we've uh which we've been through for two or three years.
I'm still in AI casino 100%. when it comes to generating images or

whatnot. Um, isn't it good for AI company for an LLM that it is a casino? You end up doing more prompts. More prompts is more tokens. More tokens is bigger subscription. That's my pitch to the market actually to stop prompting like you're in the casino to be in control. Um, be very mindful of your token spend. spend some time learning about AI, which models and it takes it takes a while, you know, it takes a few weeks to a few months to develop the AI muscle and then you know where to go very fast, very efficiently. you save time and you save compute. And I think the market still especially in gaming still missed a lot of the market missed this. And when when you said at GDC when you met a gaming executive or a CEO that said no AI is not working at least not for mundane task like support or QA and so on like he said AI is not working for content.

No, it is. But it it requires still today quite a lot of education on boarding and workflow research. So you've got to do the work ahead to be to be efficient down the line. It's not an easy sale.
Yeah. I I just don't believe that that happens until the prices go up. There's no incentive.
Um what do you mean? Well, in order to be efficient with your usage of AI, the prices have to go up because now it's just really easy to just pull the the slot casino uh handle.
But once once you have only few polls a day, then you become really good at writing the prompt and almost treating it as a junior employee.
So my my feeling is the the budget won't matter as much as the time saved, you know. It's not about the budget. Spending a few hundreds to a few

thousand per months per person won't matter versus I did something in three days that I was doing in three weeks or three months, which which I'm seeing right now, by the way, which I'm seeing I'm seeing studios doing in days what they used to take weeks. What used to take weeks or months. Um, and when you're at that scale, you don't really care about your token spend too much. You still look at the report of course but you care about that efficiency productivity gain in terms of time spent time time saved sorry time saved
I understand that you're you're talking more from the company perspective I was talking more from the individual perspective I think those are a little bit different but kind of the same as like when you run out of the uh the the tokens it's uh
it it does have an impact on your workflow. Yeah, because I never run a token [laughter] actually.
Of course not. You're running on unlimited tokens and be like these tokens are are too subsidized while

blasting
ocean of token. Yeah.
So maybe I should live the real life of a of a real consumer one day.
Well, just pick an pick your third LLM that that you have only a premium subscription to, not the unlimited one, and then start using that in your in your base work. And you'll notice how quickly you run out of tokens, especially with image generation. Uh yeah that's true. I mean it and it's still an issue that we are facing in terms of for the product. It's how to let people consume as much as possible while on the other side making sure we we don't lose money by giving them unlimited access that will that they will waste in uh use useless generation. Um we'll we'll find a way.
Yeah. Well talking about money and talking about profitability. One thing that I wanted to ask you is how do you find the right type of an investor? Because essentially you're saying we're capital efficient but we also raise venture money which is an equation that is kind of guaranteed to create major

tensions in the board meetings.
No tension. You know my investors you can check be surprised if they say that it's being tense right now. People are pretty happy. um we are getting prepared to raise more as the market may require us to grow more aggressively but if we do so we have the healthy unit economics uh to grow for the past two years we've proven that we can grow a healthy business and I feel that it's going to make our series A way more um convincing than an average AI series Okay, at least that's my that's my approach here. It's not like trust me the market is going to grow and I I'll find my pricing and so on. It's no I found my market. I found my pricing. I have a margin. Give me 30 million and I'll like 10x where I am today using pretty much

the same recipe. Um yeah, European grounds and American ambition again. [laughter] Why not? Um
well, you're building an AI and that kind of means that every conversation you have is is contaminated by a lot of hype. Um investors, they project the uh their AI thesis onto you. The clients by because they are afraid of missing out. uh they're afraid that their competitors are using your tools or similar tools. So I'm wondering how much of scenario's current traction, which is very impressive, is real demand versus AI FOMO.
I'd say none. Okay, it's bold. Uh but I'd say none. 70 plus% of the revenue is with enterprise or mid-market plus customers. They don't buy. Sure, the investors and the

executive they do mandate their team to use AI, I get it. But it usually requires a pretty long process of validation. And the sales cycle is very long. We're talking about four, six, nine months of back and forth, a demo, a P, a pilot, a PC, like very long legal, IT, uh, finance negotiations and so on that you cannot mix this with FOMO and hype. there are too many guard rails in the in the process. Um and then when they start they usually start pretty small um a [snorts] few few dozen people start testing but the studio has like thousands of of employees and we see them all of them 5x 10x over you know 12 18 months because the product works. If the product wouldn't work then just the team of of 20 would keep using it. That's it. they it wouldn't scale it wouldn't spread. So I don't think I

don't think our in our case I don't think we we are a hype machine. Um dynamics may be different in the consumer space and and we see that on Twitter I'm a lot on Twitter and socials. We see consumer going from one model all the way to the other model you know Sora to VO to Cance to Nanobana to GPT and back and so on. And that is a different uh space and it's not all creators by the way. It's a category of completely AI first, AI crazy creators that are switching non-stop and chasing the hype. It's not the same uh audience. It's not the same crowd for the enterprise users that don't have time. They just need to produce. They they've been hired for a job. The job has to be done. And it's not about tweeting and bragging about that latest model you have tested.
Yeah. or somebody else has tested and you see the results. This is 4.2 and the other one was 4.5. We need to switch to

4.5 in the review that somebody made.
Yeah.
Um,
so no, I think
you mentioned that the game industry is
Yeah, good.
You mentioned that the game industry was very conservative. So, how are you able to not how you able to how do you sell this to to the game industry? Um the the best entry point for us is not the top executive. It's not the CEO or CFO that's looking for efficiency gains and so on. It's the user. It's the art director the technical artist um you know lead artist and so on. Whether it's 2D or 3D and doesn't matter. Um, and I do succeed so far because myself and the rest of the team, we know our product really well. It is not just a rapper. Hey dude, like look at this cool website where you have a lot of model. Trust me, it's nice. No,

we know the use cases. We know like the workflows. We know how to how to prompt. And every time I speak to a game studio, even though I'm not a game artist, I'm not I've never trained I've never been trained in, you know, Photoshop, Illustrator, Blender, and so on. Despite that, I know how to use AI for their use case. And I think it's very important um when selling to the gaming market to know that it's not it's you you know exactly how to use it for them. I don't I don't think many AI companies in in the visual space in the you know visual content space uh do that. Yeah, I I understand why marketing teams would be all about this. I mean for them the uh the amount of content is vital for for for finding those creatives that that that work for the the shots on goal. But you're talking a lot to game teams and there's the perception that

once you start using these tools there's going to be less positions especially on the art side. So how are you able to convince the art directors to start using something that may replace them or their teams? So I'm I'm telling you what I'm seeing. None of the customers that we have and I'm talking about I don't know 20 25 large studios from what I know none of them has laid off or reduced their team after adopting AI. They've upskilled their team and they've tried to just do more with the same crew. And I I feel that it's um I feel that it's by it's just the only and the right way to approach it. If you hire people, it's because they're talented. No matter the tool they use, right? They should be using any tool for their skill, for their creativity. So I I don't and I'm

talking about really big names and they they don't they don't really use the team. Now they may not hire as much as they could have hired. Yes. Um, and for I've got plenty of stories. I've got one story that's interesting. A small game, an old game that was declining, the studio was able to keep it alive just because AI brought back healthier unit economics for the game. So, the game kept running with a rather small team, I don't know, 10 20 people plus. Otherwise the game would have been sunset and everybody would have to would have had to move to another game or another company. So it's you know it's it's not just bad and AI would replace our jobs and so on. AI will change balance. Yes, for sure. U but it's not just about everyone be everyone being laid off.
Yeah, I understand. I I mean the uh like every AI tools company would say

that we're not replacing employees, we're empowering them to do more, better, and faster. But what do you think happens inside 5 years? Like what happens to creative workforce? And again, I'm not trying to be doomcroller because when you look at the data, the unemployment rate is not increasing. Maybe that happens in five years, but it's not correct now. At least based on economic data.
I mean, we are a perfect example of a company heavily using AI for literally everything. Support, marketing, sales, code, everything. Um, I should be three times or, you know, two times or three times the size to produce what we produce without AI. So, we're able to be 2x or 3x 10x more efficient thanks to AI. So yeah, I I know I know that myself. On the creative side, um the talent is not about knowing which

button does what in Blender or Photoshop. It's a total misconception. The the talent is where is how to know, you know, what's the tone, what's the expression, what's the motion, what's the movement, what what what are the colors and so on. I mean, no AI models knows this yet. They all do pretty pictures. Can it be art? Can it be a game? Can it be a movie? You know, um that is a that is a talent that uh that that the LM have not learned at all. And and that's why I'm not worried about, you know, creative good creators won't be replaced by AI for a while because it's all about their intent, not about their knowledge of the, you know, of the software. I don't believe in magic and I've seen investors being bamboozled so many times before with with whether it's

does something in Unreal Engine and it looks magnificent and they show you that this is a team of five that did this but it's a demo and as you said like there's different things between demo and production quality. Uh there's a vast vast vast difference. one is mere u a facade and the other one is actually a building. So um yeah what I'm seeing is a lot of facads.
We'll see a category of creators that have never been trained or used to they've never really used in depths the traditional creation software Blender, Unity, Unreal and so on. they've never used them but they they came straight to into AI either from the start or they I don't know they they moved you know changed career and so on and they have great ideas and great vision and AI allows them to express their vision and their work um while they could not have done it in the

traditional way without spending years learning the tools. So is it good or bad? I think it's good. It basically opens opportunities for other types of creators including UGC you know you know individual
but that has already happened most of the engineers using Lua and Roblox they they came out through without going through the traditional route of game development.
How many what's the percentage of creators on Roblox versus players? Like 5%
1% it's between a few percent.
So maybe you're you're Yeah, but that changes over time. Over time, all the platforms move to the same place where the amount of creators is a tiny faction. Whether we're talking about Instagram or Facebook or LinkedIn or Roblox, it starts with many and then it ends with only a small percentage.
Yeah, but the AI may change that fundamentally.
I don't believe that.
That that's something we'll see in a few years.

Yeah, sure. How how difficult it is to take a photo on your phone?
Yeah. Why still why why only 5% are creators on Instagram?
That that was my pitch back in 22. By the way, I think it's up to the content creation companies such as game studios to facilitate the access of the tools to everyone so everyone can find how to create more within a game within um platform and so on and they they know about it you know I mean Instagram is Biden is providing that. Let me give you another example. LinkedIn writing is not that simple. Everybody knows how to write, but but writing is kind of difficult. Well, AI has made it easier and it's an utter shitow. It's unusable at the moment with slop on slop. [laughter] So that's another scenario where where again what happens is professional writers and those who are interesting you know people whether like

I don't know I have a Steven Bartlett's book here on on the uh on on the table he's a creator on a on a multiple different platforms his content is becoming more and more engaged because there's so much slop being produced by democratizing the ability to create and so it favors the ones who are professional Yeah, I was reading a theory according to which slop didn't wait for AI to exist.
Of course not.
And a lot of platforms had a lot of slop for a while even premini and and so on. Uh so there is still a place for high high quality content whether it's made by AI or not.
Exactly. And and know my point is just the the more democratization of content creation you put in, the better it is for those who are actually really good at content creation. And of course, there will be the ones who are born out of it. Just like on YouTube, uh some of

the uh the creators become massive just because they are really investing a lot of time in content creation. Everybody can take a video. There's only one Mr. beast and that Mr. wasn't raised in in you know one day he invested a lot in the new tools and the new platform and continues to do that continues to hone in his ability to create stories like it has happened with everything so it's not the easiness to create it's still the taste that you need to have in order to be good at content creation not the tools
yeah 100% yeah 100% that's that that's why I am not yet concerned about all creators creative you know, losing their jobs, but they they they've got the studio itself, the brands. They have to keep up with developing their their lore, their taste, and so on. That matters so much. And and finding the people that that will help developing better taste, better lore. So, I think it's a I think it's a plus,

but I'm not Yeah, I I would I would agree that it's not obvious yet. Yeah. What's for sure is the the wave is here. the cat is out of the bag and and the the few very few studios I I I don't I don't feel there is any studio left by the way that's ignoring or refusing. They're all getting into it even though they sometimes
you'd be surprised. [laughter] You just don't work with them.
I do work with them. I do work with them. I I know. I know. I think it's I don't understand how somebody can just not even test at a certain scale and
because they read the news because they read the news and they're be they're scared out of their mind and they think that by not engaging with the new tools that's going to
save them.
Yeah.
Um Yeah. That's nonsense.
Or by trying the tools and saying it

doesn't work. But I've had the question so many times. The the it doesn't work is the worst.
Yeah.
It doesn't work the way you tested the day you tested with the model you tested. And every time there is I've got a challenge, a creative challenge like can you do that with AI? I'm like I certainly can but it's going to take us a few days to a few weeks. We'll get at 50% at first and then we'll see whether we can get to 60, 70, 80 and at some point whether we plateau and we'll never get to where we want or if we keep grinding to where we want but the magic doesn't happen in one you know one day one prompt and um it's too easy to say yeah I've tried that model it doesn't work um the the brightest case is 3D 3D is a very is very less mature than 2D or 3D makes messy meshes. It's like a polygon soup with the texture is is bad,

the UVs are bad and so on. But when you mix a human manual workflow, traditional workflow, Blender and so on, and a and an AI workflow for some of it, then you can go much faster, but it requires some back and forth and it takes a few some time to figure out figure out. If you stay on the AI side and you're like, can you do what I was doing in Blender with a few clicks? No, you can't. Well, it's It doesn't work. You're going to wait. Yeah. You're going to wait six months and you're going to lose a lot of knowledge, a lot of expertise.
Yeah. It's like me using Adobe. I could say it doesn't work [laughter]
after after one.
Now, Adobe actually now Adobe is in is in clone. Kobe has MCP. Adobe has you know agents.
I correct. Notion doesn't work. I know it works for some. For me, it doesn't.
Notion has a great AI. One of the greatest. Uh, I haven't used it in a while. I hate it. I goddamn hate notion. [laughter]
The the the the UI. You may you may I

see
not the UI, just the inability to work with it.
Like I'm unable to create that's a different podcast. I'm unable to create a sheet. And if I can't create a sheet immediately, then I'm not the problem. Not the problem.
Um, my deep belief, and that's beyond scenario, beyond beyond image, video, 3D, whatever. My deep belief is you wait a few years and every interaction we will have with any software will have an LLM layer. So what matters the the the LLM will know better than you how to interact with the software whether it's Excel, Unity, Adobe, whatever, notion, whatever. But you'll have to know what to ask the LLM and how to and how to get stuck, how to avoid getting stuck into what the LLM wants wants to do because it's been trained in some ways to uh to generate something something by all means, you know,
to guess to guess the words that you're trying to put out. Yes, that would be

great. I I know how to use Excel, but it's ridiculous that I have to come up with various formulas when I exactly know what I want to achieve. if I can use an LLM for that, ask for what I want and it gets me that that's that's great. That's a fantastic use case. So, um anyways, I wanted to move from
from talking about AI specific elements to just to close the uh the episode talking about you and about the uh about the uh the sacrifice you make as a founder because this is not your first rodeo. You've done this before and I feel like you went into this engagement with your eyes open. So, what did you tell yourself that would be different this time around? Uh, when it comes to being to the cost of being a founder, I've I I we have avoided a few mistakes and I am making plenty more.

Um, on the few mistakes I've avoided, I went I I think I went straight to the market that I that that the right one, the the US market, the Silicon Valley investors and so on. That's a mistake I made in my first company and I think we are a different company because of the setup. Okay. Um, second, my co-founder was my my co-founder at the previous business. And so we we've gone through fire and flames and we've known each other for a while and we work so much more easily now because of that past experience. Okay, so it's not like the same thing. Let's do it again. Let's go back in the trenches again and again. Okay. Um, I I was not expecting AI to be so ruthless. and so hard in terms of the constant news. The models keep rolling every week non-stop and you need to keep up and

keep up and keep up. AI never sleeps and and so does you. So do you. I I I feel that I despite the sacrifice sacrifices, I'm a better founder because of the because of that first experience and I'm doing more and it's more fun than the first business. So, it's kind of more work, more knowledge, uh probably more mistakes and more fun. Is that just a better product market fit would probably be the uh the reason why it's so much more fun.
Possibly we had a good market product market fit. It just was too early.
Yeah. So
just was too early.
That's it's a good hypothesis. But now you actually have a good product market fit. If you're talking about being profitable before series A in an AI company, that's a great product market fit. I know that the clients you're working with, they're gigantic. They're

the best studios in the uh in the world and they all adopted and they all said this is the best one.
No question there's product market fit but in AI your product market fit can change pretty fast.
How how difficult it is to hire for an AI company in Europe because I know the salaries in the US and they're ludicrous.
Yeah. Well, all of our engineers are in mostly in Europe. Um, and it's not too hard and it's not too expensive. You just
I think you shouldn't be telling this.
Well, you've Yeah, you just got you've got to manage them the way you manage Europeans. They don't work the same way as Americans.
No, but they do.
They're religious about their time off, about the work life balance
for our engineers. uh without trying to generalize too much, they they really care. It's like craft. I see like

European developers consider their code like craft and possibly the US ones as well. Um but with with the with these days, it's like about like ship ship ship, you know, make vibe code as much as you can. AI will figure it out. Um and what I'm seeing from our side, especially my CTO is based in Paris. uh they they care a lot about the the code base and their and their you know their their craft and it's why by the way they were pretty reluctant at the beginning to put code in the loop or codeex or whatever LM it took us a while to convince everyone at the company to use AI even for code and I vividly remember discussions about engineers not willing to you know code code with AI Then it was it was interesting. I was like, "Are you crazy? You've got to do it." And uh it it took us a few months.

They were not convinced at the beginning. So, no, not not too hard. It's not too hard to hire. We're going to need to hire more in the US on the on the sales and marketing and customer support side. Um and having teams on both sides of the of the Atlantic is difficult. I've I've lived that for 12 years now. you know 10 10 12 years it's difficult culture wise time zone wise and so on so
yeah compensation wise
expectation wise execution wise Americans expect you to be at a hyper speed or they jump off the train about in about 9 months
yeah um yeah I I think people will accept that we may not grow as fast as the hyperscalers when they see the the community economics and the the healthy business we're growing because the the we're not chasing the consumers the consumers that are

churning it's it's a general trend consumers in AI are churning a lot and are not paying much so the margin are terrible and the churn rate is horrible and it's no secret okay so numbers go up yes but behind the scene it's difficult our customer profile is
but much bigger companies very loyal zero churn. Pretty much zero churn.
Yeah.
Um I know and and um yeah. So
yeah, that's that's why I honestly wanted to talk to you. I I I knew I first asked about the the references. There's too many AI companies wanting to hop in on a podcast and tell their their FOMO story and there's very little behind their ARRS. I mean, and then there's companies like yours that uh that don't talk a lot on the podcast and you have a very tangible AIR ARR uh when talking to your large uh clients. So, uh so yeah, kudos to you on building I'll keep it this way.

I'll keep it this way. Uh despite the sacrifice sacrifice and despite the the the Sunday doing emails.
Yeah. because it's fun by the way, you know, and I'm I'm I'm having a lot of fun and it's a I think it's a very deep change that I was not expecting to live. I was expecting some tech to be better and better, but I think this wave is not like the other ones. And so, yeah, it's, you know, it's cool to to live through it.
Yeah. I mean, again, this is a European conversation. When you're doing uh emails on Sunday, it's seen as as a failure state. But it's not. If you look at from the other lens, it's seen as success rate because your business is doing so well and because you care so much about your business that you're willing and eager to put in the work on Sunday. Uh I'm I it's it's just it it's just a different lens and you can view things from from different perspective.
Yep.

So last question. You've been through the the hype cycle before. You know how this ends for most most startups. You've actually been to the ending before and you're now in the uh in the new beginning. Uh you know the personal cost. We've talked about it multiple times. Uh you've seen the investors lose their interest after first you know super funding and blit scaling your company. You've seen markets shift. Uh you've shifted your company pivoted multiple times. You've seen technologies commoditize and the effect of that on on a on a startup. Now, knowing all of this, what is the actual thing that gets you out of bed today? And I'm not looking for a mission statement. I'm looking for something more real, something that uh that other founders can relate with. Um, I think one one success, even the tiny success, the customer you win, the the the round you close and so on, um, gives you so much dopamine versus all the pain that you've had for the previous weeks

or months. Then you just want to do it again. I think that's the real it's just it's chemical you know like the the success after some hard time hard time selling hard time pitching hard time uh hiring and so on getting the success gives you so much pleasure that that you just want to start again and then when you and it compounds because usually you start with the small successes and then the bigger ones and then then you feel good actually despite Yes. despite what you have to go through. Um, yeah, that that's on the on the personal side and I think when you when you look more like on the rational side. Um, being in the arena when things happen is rewarding versus looking from the outside. That's just rewarding. My kids, you know, in 10, 20 years tell me what did you do like when you work? Yeah, I

was at during the air wave and it was crazy. Let me tell you how we how we managed to uh you know navigate as a small company growing in the that space. It's it's rewarding.
I I get it. You're not hooked on the sale or the sales process. You're hooked on the the outcome of what you've built and sort of getting the customer's approval of what you have been building is exactly what they want and what the market wants.
Yeah. the and the the fact that people are eventually saying, "Oh, he's right or you're right after hearing you're a fool for like five years." That's that's powerful, you know, like yeah, you you've um you stick to your uh vision and eventually you make it no matter the patient, no matter the product, that matters. So, I think that's that's may that's probably what most a lot of founders experience without even realizing it as well.
Proving others wrong. Yeah, maybe that's something to do with your childhood.

[laughter]
I've talked to enough founders that everybody tends to have that sort of a chip on their shoulder. So, uh the need to prove that they were right all along.
That's a [clears throat] powerful motivator. Um Emmanuel, I think what you're building No, I don't think I know what you're building is very valuable. I know this by talking to your clients. uh and it's very refreshing to talk to a founder in the AI space that is incredibly rational but ambitious at the same time. So it's uh it's been an utmost pleasure having this conversation with you.
Likewise. Thanks Mishka.
Thank you everybody for listening and scenarios link in the description. If you want to make Manuel happy become a client they retain. So
just just ping me and and give me a challenge. You can I'll I'll take it up.
All right. Thank you everybody and thank you Emanuel.

Bye-bye.
You did it. You made it to the end of the episode. As a fan of the show, it would help us out if you'd subscribe and leave us a review on the podcast service of your choice. Those stars and likes, while a small action on your part, mean a lot to us and help to grow the show. More importantly, are you a member of the Deconstructor of Fun Slack group? If you have 5 years or more of gamesindustry experience, go to deconstructorofund.com/slack and apply [music] to join. Join the games industry best professional community filled with peers always willing to lend a hand or subscribe to our newsletter to get all the latest insights [music] from the deconstructor of fun content creators. Thanks for listening.