If AI does everything, what am I good for?

Julien Klein, Jonathan Bélisle (Paracosm) and Vincent Lamanna (Crewdle) answer, no filter, the question every entrepreneur is asking themselves in silence in 2026.

Spotify Apple Podcasts Key points 01What do you sell when AI does everything?Two skills are missing from every résumé in 2026. Vincent and Jonathan name them. The Grumpy IT archetype that doesn’t exist anymore. Why selling leadership leads nowhere. And what you can sell instead.

80% of entrepreneurs are behind. Most don’t know it. The others can’t admit it. Three CEOs talk about their own shame around AI. No filter.

No theory. Concrete moves. The Dream/Learn framework that Crewdle is rolling out. The first step when you’ve never really used AI. And why velocity without direction leads nowhere.

The AI Sprint, cohort 01 now open. One month to see the invisible opportunities already right in front of you. Without burning down what you’ve built.

We talk about what AI really takes. About what it will never touch. About the silent shame that locks up 80% of entrepreneurs. And above all, about what you can do starting Monday morning to reposition yourself without burning down what you’ve built.

Three entrepreneurs. Three perspectives. One single question: if AI does everything, what are you good for?

Late at night. In the shower. When you open LinkedIn. The thought goes through your head. Nobody dares to say it out loud.

In this third episode of Blind Spot, Julien Klein (monExpansion), Jonathan Bélisle (Paracosm) and Vincent Lamanna (Crewdle) say it. And the three of them answer it.

AXIS 1 | What do you sell when AI does everything?Two skills are missing from every résumé in 2026. The Grumpy IT archetype that doesn’t exist anymore. Why selling leadership leads nowhere, and what you can sell instead.

AXIS 2 | The silent shame.80% of entrepreneurs are behind. Most don’t know it. The others can’t admit it. Three CEOs talk about their own shame around AI, no filter.

AXIS 3 | Monday morning, where do we start?No theory. Concrete moves. The Dream/Learn framework that Crewdle is rolling out. The first step when you’ve never really used AI. And why velocity without direction leads nowhere.

The answer to “what am I good for?” isn’t in your job title. It’s in what AI will never be able to do in your place, and in the way you own it.

The trap doesn’t get solved by staring at it. It gets solved by making a first move. Two options.

Assess your 6 core needs Identify your form of the trap Personalized verdict + concrete next steps Access to the interactive tool Take the diagnostic → THE BOOK The Talent Trap 208 pages. The complete mechanism, the exercises, the stories of Hong, Sofia and Emma. The 10 signals + the 6 locks Digital vs Organic explained Your expansion profile The 90-day prototype We weren’t able to confirm your sign-up. Your book is on its way to your inbox. Send me the book “If I could do this, you can do what this book asks of you. This book was born from a conversation I had right before someone passed away. It’s written so that you can have your rebirth.”

A shoot. A channel with 25K subscribers. A promotion agency. The good talent comes to you.

Full transcript If AI does everything, what am I good for? Blind Spot · Episode 03 · Guests: Jonathan Bélisle · Vincent Lamanna Read the full transcript Lately, there’s a thought that goes through the head of almost every experienced entrepreneur, maybe you too. It goes through late at night, in the shower, or when you open a LinkedIn post. Nobody really dares to say it out loud. So we’re going to say it here.

If AI does everything, what am I good for? Today, with Jonathan and Vincent, we take the bull by the horns to answer it. And I’m warning you, the answer isn’t in your résumé. Welcome to Blind Spot 3.

Before we start, a quick word: if what we’re going to say today speaks to you, know that I’m opening the first cohort of the AI Sprint soon. One month, four group calls with coaching so you can see the invisible opportunities you already have right in front of you, whether you’re an employee, a freelancer or an entrepreneur. The idea is for you to find meaning again in what you do, to gain independence and, above all, to get your energy back. And all of that without burning down what you’ve already built.

The waitlist is open, the link is in the description. Now, let’s go. Nice. So, listen, welcome to this third edition of Blind Spot.

Today, uh, we’re going to cover a bit of the angle. Okay, if AI does everything, what am I good for? And we’re going to look at this a bit from your client’s perspective, meaning the people you advise, not you as an entrepreneur, but we’re going to try to put ourselves in the client’s shoes to try to bring them some value, to guide them a little through this swamp of, you know, everything that’s popping up everywhere. Tons of new opportunities but you don’t know where to start. A bit overwhelmed, or, I don’t know how to translate that, swamped by everything that’s coming.

So we’re going to go point by point. Uh, the first axis I’m proposing is: what do you sell when AI does everything? Because with Crewdle and Paracosm, correct me if I’m wrong, but you help clients realign themselves with the new tools, the new processes, the new possibilities that AI brings. But, well, the first question would be a little bit: what is the talent that you could sell for more but that doesn’t show up on a résumé, in relation to these new aspects, these new possibilities?

It’s funny because the conversation Vincent and I were having this morning was exactly about how, once Vincent and I start interacting with a client and the client feels the change that’s about to happen, they let us into their life more. There are people who’ll call this change management, but Vincent and I, what we called it, we split it in two, and Vincent, I’ll let you start, but we split it into the capacity to imagine and dream growing, and the capacity to learn and unlearn and challenge them. That is, you really have to help our clients both imagine better, think better, clarify their vision, but also help them realize that the capacities they had before, in the old world, also have to evolve. We’ve kind of, we’ve kind of set off the trigger of the nuclear fusion that makes the rocket take off, but at the same time, we go into the cockpit and give instructions to each of the astronauts, telling them, you, while we’re taking off, you’ve got this to read; you, you need to learn to meditate; you, you need to read such-and-such book; and you, well, you need to get to work, you know, your skills are too low. Here are some actions to take. So the challenge Vincent and I have is that three years ago, we lived through what all our clients are living through.

Yeah. For me, my role was to do interfaces, and every time I told my clients, can we think about who the users are? Can we think about why we’re doing these things? Can we think about the app-interface model?

Since the interface is basically free now because execution costs zero, I finally have the time to give my clients reflective work, research, things I couldn’t sell before but that I now have time to do, while what I used to sell no longer has any value in monetary terms. I’ll stop there. I’ll stop there because I want, I want Vincent to synthesize differently what we talked about this morning. Go ahead, Vincent.

Yeah. I think, you know, to come back to that part, what’s not visible on the résumé, in fact you, we’ve already, we’ve already touched on the question in our other, in our other episodes, but AI is an amplifier, and then, separate from that, the notion of raw skills, you know, that shows up on the résumé. In my opinion there are two things that are evaluated less, less today in our, in our hiring processes. Uh, the first is really the capacity to communicate. Uh, that’s something that’s becoming hugely important.

Uh, that’s, that’s something people aren’t used to, in the sense that often, you know, we take for granted that, well, there’s a manager on the team. The manager is better at communicating, but the people working on the team, that’s not necessarily their strength. Whereas now we have, we have a paradigm shift where in fact communication becomes a key skill for every member of the team. You have to, you have to develop it. That’s the first thing.

And the second thing is a bit like Jonathan was saying, you know, uh, we don’t have excuses anymore. Meaning that if you’re not hardworking, if you’re not competent, it’s going to show extremely quickly because the, the, the goal of this tool is to accelerate us and to move from a model where, for example, you charged hours to a model where you charge a result. Uh, that, I think, is the big number a lot of people have a hard time, let’s say, letting go of. They were so used, for years, to valuing that model.

And now it’s a completely different model, but one that’s much more in line with the final result, which is a good thing, but we weren’t used to that model for, for decades. Yeah. Yeah. Because you could, like you say, you could hide a little bit.

The first example that comes to mind is the grumpy IT, the archetype of the guy behind his computer; you ask him something, he tells you to get lost a bit, uh, because he knows the passwords, he knows the processes that you don’t. Uh, yes, I’m talking from real experience, and, uh, I kind of feel like that’s not really possible anymore either, because there’s a service that has to come with it now, a kind of customer care that’s way more important than before, when the highlight was the technical side. The guy knew how to fix that problem, and you had to go beg him, kiss his boots so he’d fix your problem, and he could almost afford to be a bit, you know, off-putting. I, I saw that a lot in videos, uh, of guys explaining this about devs, that devs had a bit of an exaggerated aura precisely because they were doing things nobody else understood.

So they took advantage of it a little, and that, that seems to be ending a bit. So I get the impression that, you know, to simplify it as much as possible for, uh, for Mrs. Smith, let’s call her that, who’s watching the show and knows nothing about, uh, or Mr. Smith, right, we won’t, it’s, uh, really the, the aspect of, uh, uh, support that’s much more important. You talk about communication, you know. Here, I’ll come back to the example of my wife.

She manages the invoices for Communauto. Uh, but not only does she manage them, she also goes and checks: do we still need to manage paying this invoice? Is it still relevant for the business? You see, that’s an accountant thing.

Usually, they go like, that’s not my job, I do what I have to do. But she has this proactive side that makes her, whatever tool comes along with it, I find that this customer care, this care for work well done, the, the proactive side, I think that’s what’s going to be valued the most now. I don’t know what you guys think. For me, when we were talking with Vincent about integrating it, one of the advantages is that philosophically, when you become proactive, that is, when you discover the potential of AI in work or in our lives or in the way you apply better questioning, uh, when you realize you didn’t know everything, that you could have taken an easier path than the one where you stubbornly waste a ton of time, anyway, proactivity, before you, proactivity, you become vulnerable.

There’s a comfort zone you transcend that was your trade, the one you knew. And three years ago, Vincent didn’t believe AI was going to be in his company. I even think, Vincent, you told me you didn’t want AI in your company. Like a lot of people in development who also resist adopting agentic AI in an aggressive way.

They’re very cautious. Okay. As for me, I kind of didn’t have a choice because my trade was already seen as, not a nuisance, but a form of intellectualism. Oh, UX, we think about strategies, why do we have time to think, Jonathan?

We have a week, you know, a week to think about a project that’s going to cost a million, but we’re going to put all the money into development. That 30% that used to go into development is now put into strategic reflection, which lets a project be much deeper, first in a better understanding of the market the product is coming into, a better understanding of the user experience, of mental models, not just the personas, I can give examples, you know, of saying, hey, Jonathan wants to propose a system to analyze users’ mental models, oh, that takes a long time to do. Not anymore. Now it’s exactly, it’s no longer longer now, it won’t cost you more to do it. So let’s do things right. Let’s take the time to do things right, oddly, let’s take the time to do things right.

It’ll go fast. There are, these are things that, when I look at, say, the notion of, to go fast you need a direction, and velocity is only if you have speed plus a direction, otherwise you crash into something. Okay, so an example, we start, we say you have to have a mindset, a beginner mindset, you know nothing. It’s very hard to do because everyone identified, like Vincent said, with what they do, we identified, that’s it, we identified with the doing. Okay, that’s great, I mean, me, I was a workaholic when I was 30. Vincent, if you’re not a workaholic, I don’t know what you are.

But you know, we worked a lot, okay, to become what we are. We put tens of thousands of hours into becoming what we are. We reassure ourselves too. To reassure ourselves, but you have to reach a deep vulnerability, a bit like you, you do coaching, Julien, you understand?

You have to reach a vulnerability about what you don’t know. You’re a, AI is going to reveal them. And if your mindset is to say, perfect, I’m allowed to be vulnerable right now, but it’s like there’s a coaching portion in our project with clients, it’s not just a solution that searches, it’s also psychological support. Because when the collapse of your comfort zone comes, when you’ve built your little cubes, your little safe spaces disappear, you need other people around you who’ve also made the same bet.

Vincent and I, it’s going well, and it’s going well with you too, Julien, because we made the same bet. Okay, but people on the same thing, that’s when things really start. In any case, my point about potentiality is always there, but you still have to be with people who make the same bet as you. If you’ve got folks heading toward the open water and others heading back to shore, you’re all going to capsize.

You know, if I understand correctly, the talent you pay more for that doesn’t show up on a résumé is actually being ready to question yourself. Yeah, in any case, for me that would be the first thing. The first mindset is beginner’s mindset, you know nothing. You’re going to have to accept that the direction, even if it’s not always clear, good one.

Velocity is the equivalent of going fast in one direction, not changing direction every time. If you’re afraid and your whole team is afraid, the boat is going to capsize. You know, that’s actually what Vincent and I observe. It doesn’t mean everyone takes the same direction, but Vincent and I, we have the same direction.

We say AI first, uh, sovereignty for our, for our clients, so giving them a technical capacity and a philosophical capacity, both at the same time. So you, Vincent, you’ve already pivoted with your business, so as an entrepreneur, for you it’s almost natural and logical to adapt to what the market needs while respecting your ethics and your ambitions. But your own clients, maybe they don’t have that entrepreneur mindset, or how, how do you, how do you juggle that? What?

What we see in the market right now is that know-how, let’s say, or more traditional execution is worth less and less, in the sense that, uh, to build a web app, to build a website, to do, people are willing to pay less and less. On the other hand, the opposite, when you get into the more support portion, and sometimes, sometimes it’s even just listening to people talk to us, uh, people are willing to pay much more, uh, for that part. Uh, so there’s enormous value, I think, in the market right now in human interaction, more than, more than interaction with the machine. So true, that.

Uh, that, I think, is the point you have to capitalize on, but you also have to understand the message. Meaning that, as Jonathan was saying, it’s still a big change that we impose, or that’s imposed on us. In fact, it’s not some other party imposing it. The market is just there now.

The market is there now. AI is here, and, uh, and we have to adapt. But, but that change takes a huge amount of effort. As Jonathan was saying, as much on the psychological level, you know. Uh, I remember when the first tools came out, uh, you, we asked ourselves the question, you know, do we do something with AI?

And I remember, I was the one who said no, our business isn’t, it’s not about that, we’ll never do that, you know, and, well, three years later we’re a platform that offers only that, you know, but still based on something that’s the foundation of what we’d done, that is, we’d developed a decentralized communication technology, that was, that was really our, our strength. And in an agentic system, where it works well is when you have a good communication system between your agents, your different systems. It’s a pretty natural progression. Yeah. But that’s really it, it’s the support, the human interaction.

Like I say, it’s not so much about guiding them as it is about listening to them. People need to talk about their experience, about how they’re living through it, so that sometimes it’s not even about chiming in or asking them tons of questions, it’s really just about listening to them and taking the time to let them decant too. It’s so much that. Uh, just before we started recording, I was giving you the example of the fact that I joined an incubator.

When I left my salaried job, I had the corporate reflexes. So I told myself, hey, I’m going to sell leadership because I manage litigation, that’s my expertise. I’m very good at it. I read the time for Netflix, Scanline.

It’ll go well. End result, it’s not at all what the market needed. I lost $4,000 on ads. So now, I signed up, uh, first dipping a toe into the thing at 1,500 euros for a group incubator.

And there, I watched, for, there’s one live a week, and in that live I watched, uh, Flavie Prévôt actually, Solo Nation, already very good, she does great podcasts and all. And in fact, what made me want to pay for the premium, which is 6,500 euros, is that she looked at what you’d done and told you, from my experience, here’s what I would do. And in fact, that has enormous value for me because, well, she’s at the next stage from where I want to get to. You see, like you guys are at the next stage from what your clients want to reach.

It’s not that I can’t do and set up sales pages, all that stuff, you know. By the way, that, that can be a trap, you know, just diving into making landing pages forever. I made some. It didn’t work.

What I need is for someone who knows better, uh, or who’s already walked the path I’ve walked, to look at me from the third-person perspective and tell me, “Listen, based on where you’re good, based on where your strengths are,” because I know I have something important or valuable, I have value to bring to people, I know it, I’ve tested it one-on-one, but I can’t manage to convey that value one-to-many, you see, like a product. And then, to have that support, like you say, from this person who’s already walked the path and tells me, hey, uh, push more on this, drop that, it’s useless. That, that has value, you see, I, I’m ready to put down 6,500 right now even though I’m losing money every month. But for me, it’s important because that, that, that has enormous value.

But it’s interesting, the timing. One of the things that changed after 27 years of practice is that when AI arrived in my life, I realized I was going to harvest my, my clients. That is, I was going to prepare them to work with me instead of immediately working with them and losing my momentum. Meaning most clients aren’t ready to work with me.

That’s it, that’s what I realized in my life. I’m a wild card. I’m kind of an unidentified thing in myself, you understand? I’m a creature, you know.

And my clients have to get used to the fact that they know they have to work with me but they’re not ready. They’re not ready with, they don’t have the questions. It’s just that they met someone like a shaman. But what kind of question do you ask a shaman about the future?

About what they should do? What practice they should adopt? What diet? Okay.

Yeah, because three years, I nurtured relationships in my life for ten years with one person. That person in particular is probably going to become a client, but I hope, there’s no rush, because it’s a conversation, and he tells me it’s coming, it’s coming. I’m getting ready, I see it. And you build trust because you leave room.

You’re not in sales mode. There’s nothing to sell with a personal revolution. It happens in your head when you’re ready. When you’re ready, you know.

And then when they’re ready, those clients, we catch them on the fly when they’re ready, and I’d tell you that we develop capacities in our clients. So what Vincent has set up in his infrastructure, which is super interesting, is that I bring my philosophical design ideas, all that, where we clarify the client’s vision, their intentions, but we also identify, not just what they want to do, but what they can’t do. Yeah, that’s it. Exactly. On a really wild path.

Vincent and I, you know, Julien, we do this podcast, it’s kind of like declaring our learnings and trying to be open enough about the knowledge we want to transfer. But you know, when clients realize they’re ready, it doesn’t mean they’re, uh, capable. You know, you can want to go camping, but you can never get a fire started when it rains. That’s, that’s what support is, really, it’s that you filter between what they should do and what they have to let go of and what they maybe have to learn, because if you want to get there, that’s, that’s cool.

The direction is clear, unfortunately, you have no speed. Your speed is very slow. Uh, it’s pouring buckets to get there. You’re going to need boots.

You don’t have boots. And it’s, it’s a bit of a classic example I give, but you have to, you got, preparedness is more important than luck. Not luck. AI doesn’t give a damn about luck. Yeah, that’s clear.

That’s the annoying part, actually, because there were folks who thought you could be lazy with, no. Exactly. Lazy, it’s going to be hell in your life. If you’re a lazy person, it’s going to be, it’s going to be really hell.

It’s going to be a kick in the butt to get you back to the edge of your chair. That’s, that’s it for me. AI, but exactly, it’s annoying, but it can annoy people, for example, that constraint. Yeah. When you’ve been sitting comfortably in your chair, cozy, for 20 years, of course it’s annoying. Precisely about that support aspect, it also requires that you, when I say you I mean Crewdle, you also have to up your game in terms of communication, of what you do to support clients, because, you see, I’m doing the exercise for my positioning, and I’ll admit it’s very hard to communicate something very simple in a powerful way. How do you do that?

It’s a challenge for everyone. I’m just looking at where our model is at now. We’re going to give a little preview, but we’re coming out with a new platform probably early next week. And then, well, that comes and changes our whole business model again.

And then we’re going to focus more on providing tools and support versus offering a turnkey service where we did everything for the client. We’ll keep doing some of it, but, let’s say, I think there’s more value in supporting and showing people how to do it than in doing it for them and hiding all of, all of this revolution, so that the client just sees the final result and thinks it’s easy, you know. Yeah. You want, you want, at best, instead of done-for-you, you do done-with-you, and it’s like they learn to fish rather than you giving them the fish every time, you know.

Exactly. And you know, what I tell myself is that, at the level of the message, the message actually evolves every week. Meaning that every week, the more we talk with people, the more we realize the worries. The worries also evolve pretty quickly. The, the, it’s still an industry that moves fast.

Mm-hmm. Uh, you know, what was possible to do a year ago versus what’s possible to do today versus what’s probably going to be possible to do starting, I’d say, in one to two weeks, because GPT-6 is coming, probably in the next, the next 6 days. Oh, whoa! Okay, that’s going to be another new step, because GPT-6 is probably the same kind of capability as Mythos, which was, which was announced over at Anthropic.

So we’re going to enter a new, new range of models that are even more powerful, that are more, that have even greater reasoning capabilities, that probably have more speed too. So, so we’re in a thing that’s constantly evolving and we have to adapt our message, the way we talk about it, according to that. For sure, having visited the west coast, the valley, a few weeks ago versus what’s happening here, you know, over there it moves much more than here, uh, yeah, you get off the plane, fresh meat, you arrive at the airport, the doors open, and all you see is billboards talking about artificial intelligence everywhere, you drive on the highway and it’s nothing but billboards about artificial intelligence, you get here and we have basically none, we don’t have that, that same picture. So we have a market reality too.

Uh, but we realize that it’s still a subject that’s pretty well covered in, in the media, in, there isn’t a week where we don’t have at least a few newspaper articles on the subject. Uh, so, you know, what people think of it, the experience people have with it also evolves, uh, over time. So, uh, so it’s a message, it’s a message that evolves. It’s, it’s pretty difficult, actually, to have a clear message today.

And again, it also depends on the audience you’re addressing. Yeah, that’s clear. And that makes the perfect, the perfect transition to, uh, the second point, which is, uh, well, the silent shame, a bit, of entrepreneurs or business owners who think they’re completely left behind, who, who are behind, uh, who think they’re behind, and sometimes rightly so too, because, because they’re not putting in any energy. So of course, at a certain point, uh, the more behind you fall, the harder it becomes to catch up. I’d say, though, that if you’re well supported, you can also quickly filter out everything that’s noise and tests that are pretty useless, you know, like, uh, for everyone who jumped on Open Claw, which now, in fact, you can do directly in Claude with what just came out this week, more or less.

Uh, if you had a personal example with AI releases where at some point you said, this is, this is new, it’s heavy, I feel a bit intimidated, I don’t know if I’m going to get into this. Does that happen to you? Uh, me, I miss the train every morning. I miss it on purpose.

I miss the first one, I take the second. But no, it’s almost a joke I’m making there. But, uh, it’s a good question, actually, it makes me think, so now I’m thinking live. Did I feel ashamed about not wanting to get on board with something I felt was inevitable?

Uh, it’s because me, I’m afraid of rejection, okay, in my life. I was rejected my whole life because I was kind of, like I told you, a wild card poet. It’s hard for me to show it. I, I just, I just want to put that back in balance.

You get praised very strongly by a certain category and rejected very strongly by others precisely because you have, yeah. No, but I’m saying it to, to give my example, but you’re right, it’s, I can be rejected by everyone, because there, whereas I mean, I lived rejection badly. And in 2014 when I did my first machine learning and AI projects with WIA, I got called a poet, that it would never happen, and I was dreaming in color that you could just say something and it would appear. Me, I saw it maybe as an intuition, okay, as something that was mathematically inevitable, but let’s be honest, it was abstract, it was, I was communicating with very colorful language, so not very business.

Okay. The biggest challenge I have with the shame of not being understood is language. It’s that for people to follow you, you still have to adapt your language. Okay.

Yeah, that was super hard for me, because since I’d understood things for a long time, I was too early, it pissed me off to dumb it down. It pissed me off, excuse the word, but it almost aggravated me, or it aggravated me, to have to dumb myself down intellectually. At the same time, that was just it. You have to bring it back to a 5-year-old, because, an example Vincent gives, when we look at our clients doing prompts, or just the basic conversational intelligence it takes to work with AI, it’s like having a really good conversation with someone.

When I talk with Vincent, Vincent will never tell me design, it’s more complicated than that. The word design means nothing. Design of what, man? Business design, graphic design, system design.

Yeah, you already have another level of conversation. The fact that people think that when you talk to AI you can be lazy, that’s false. In fact, you have to be as clear as if you’re talking to a good friend you want to describe your dream to, well, be precise. Use the right words, be refined, be aggressive, but at least be clear. And we realize that clients think AI is going to solve that language part, whereas in fact you have to develop it.

You have to, you have to, it, and that’s something that, well, it forced me to work on that, because precisely I was a bit up in my own bubble, you know, I, yeah, you were up on your cloud with, I refused to come down, but even Vincent told me, you know, you know, come back with, intellectual, it’s, it’s intellectual, ontology, don’t use that, you know, ontology, that’s for us internally; not long ago, Vincent, we had words in our own interface, we’d say everyone should understand this. Our CFO would say, what does that mean? Well, it’s the structure. Well, you know, write structure instead of writing the other word, because people understand structure, and the rest they don’t get, who cares. And even if you know it, but the word that comes before structure is, up there, you know, it’s systems, who the hell cares. Yeah.

Nobody. That’s, that’s my, my frustration that I had to get through, my shame, was that I had to accept that people didn’t understand me and that I had to adapt my language for everyone. In fact, I shouldn’t have stayed up on my cloud. In any case, for me that was my, that was my big learning.

Very good. And, uh, the example I also got from Flavie, precisely, my, my new coach, is, uh, that on her YouTube channel, she doesn’t even talk about scale. The word scale is too complicated. She talks about growth or some more general-public word for her YouTube channel, you see.

So totally aligned with that. Go, you, are you ashamed? Me, it’s more from a technological standpoint. Uh, and I have two, two big examples, since we made the shift to AI.

The, the first is that when we got into AI, uh, for us the question of sovereignty and data confidentiality was a very important subject. So we stubbornly insisted, for maybe a little too long, on using open-source and local models, and we were very limited, in fact, in our, in our capabilities. Mm, uh, mainly because the hardware, you know, and I’m talking about this, it’s a bit more than two years ago now, but the hardware two years ago wasn’t, uh, wasn’t there yet for, uh, for that kind of model. Uh, now with Gemma, maybe it’ll be a little more, uh, more interesting.

Exactly. Exactly. But, uh, despite the fact that, uh, you know, I’m a technology guy, uh, accepting, let’s say, technological changes still takes me a little time. And the best example of that is when Open Claude came out at the end of January.

Uh, my first reflex was to say, we’ll never use that. It’s too dangerous. Yeah, yeah, that, that wasn’t a bad reflex. We’ll never use that.

It’s too dangerous. And then, uh, finally after a few weeks, I started playing with it, and finally we decided to, to make it into a product, uh, but where precisely we handle the security part, we handle the, we, in fact we give the tool but in a way that makes it secure for the user. We control the access points, we control the environment the agentic system runs in. Uh, there’s always a little lag between when a new technology like that comes out, which is a little confronting, and the moment I adopt it, uh, there’s always a little lag, but something happens.

It’s interesting that Vincent says technically, because at the technical level, one of the things that struck me the most is that, say, I work with Vincent, Lovable does a partnership with Claude last month, a month later Claude copies Lovable entirely, and Vincent sees Lovable, because me, I work with Lovable, with Replit, Cursor, all the vibing tools possible, and Vincent says, how do you put that into production? Well, I can’t, and everyone has to redo my code, it’s just prototypes done quickly, but I lose the production deployment. He said, “You don’t want to keep the production deployment? What do I do?” A month later, Vincent gives me my own Lovable, you understand? And here we’re not talking about a year or two or three years of product development. Here it’s Vincent with his little team redoing software that normally takes two, three years to build, that, you know, I don’t want to say it’s finished, it’s never finished, but it’s impossible.

What’s going on here? That’s something I also had to accept, because me, I was always an imaginative guy, I’d get told I dream in color, but here I’m not dreaming in color. It’s the technical guy who sends me an email. You know what we talked about yesterday, it’s live, it, it can’t be.

And I have to accept it, because even my own rhythm isn’t right anymore, because me, I’m still in the middle of thinking, and then there I’m thinking about stuff, and then Vincent says, you can use it now, stop talking, come work with us. So there’s a quantum leap between, is the tool I’m going to adopt better than the one I can build? Because before, I bought tools, I wasn’t the one building the tools, I adopted them, I became an expert, a user expert. You say with, plus, what I’m thinking, it becomes the tool, Vincent creates tools based on our conversations.

Whoa! Yeah, that didn’t exist before. It, it wasn’t this fast in any case. It’s, thinking about your technical side, it even changes the conversation with Vincent, because even Vincent is there, okay well now, you have to put it to the test and we improve it.

The notion of feedback is more, uh, it’s more, on one side, uh, for Vincent much more pleasant to receive feedback to move his software forward. Before, it was almost a nuisance. It was like, it was seen as bugs. It was seen as a, oh, well, we did a release and now, you know, we’re going back into, we’re going to do a, we’re going to spend a month on review.

No, no, now it’s, it’s like every day, it’s what’s not working, we fix it. What’s not working, we fix it. But, but that, you’re so right. Listen, I remember in the old world, okay?

I’m referring to the period before, like the old world, but, uh, we wanted so badly not to get any feedback on our work, it was incredible, because it meant tons of work. Exactly. Exactly. And then, we could spend months refining, refining. Sometimes it even required reworking the architecture.

That was dreadful, because, and then it meant we’d made a mistake, and it was looked down on because it was going to cost a lot of money to fix. Whereas today, like Jonathan often said, me, that’s what I’d tell Jonathan, tell us what’s not working. Tell us what, tell us what you don’t like. Tomorrow you’ll be able to try a new experience.

Yeah, that, that changes everything. Uh, well, for my part, the shame, it was, uh, it’s a slightly different angle, uh, it was refusing to be supported, precisely. You see, this can speak to your clients, because in fact, uh, I went with a coach, uh, beforehand when I was still in Europe, who cost me 15,000 a pop and who was, who was really not suited to me. Uh, I take responsibility, but I also think he knew very well that he was going to take my money and that it was going to go nowhere given my situation.

So I was kind of vaccinated against, against, against support, you see. I was like, because basically he gave me hope that I’d be able to make a living from my talents and become independent. That’s what I want to sell too today, you see, with monExpansion. Except that in reality, uh, he didn’t, he didn’t deliver.

Uh, and on the other hand, then, I saw, precisely, uh, I don’t mind plugging her because she’s really, she’s really top-notch. Flavie, Flavie Prévôt with Solo Nation. I was like, yeah, honestly, what she does is cool. And, am I going to have the same experience? And in fact, uh, it was by following what she was doing, by seeing how much she, she made alliances with other podcasters, how much she integrated herself into the ecosystem, that reassured me extremely.

I tell myself, it can’t be that she’s not good if people partner with her to that extent, you see. So the fact that she went with others reassured me enormously. And then afterward, I did the test with the, the entry-level offer. I went, “Okay, it’s good.” Well, she knows exactly what she’s talking about.

On top of that, she’s a former marketing director for Sixt car rental. Uh, the other guy before, he was a guy who’d been a waiter before, so he didn’t really have corporate experience, you see. She, she was a bit more grounded in the real world, and she went from salaried work to entrepreneurship, which is exactly my process. Uh, so then I went, “Okay, it’s good, go for it, I’m dropping my barriers.” But, but you see, it, I could have gone much faster, uh, but, well, that, it was a, it’s a mental hesitation, and, and there you go, that’s how it is.

It was a bit of my little shame, but it’s, it’s interesting because I have clients maybe who, who are going to listen to the podcast that I worked with, where the support starts well at the beginning. There are expectations that get set. Me, I was a super optimistic guy. Be really careful about that with AI. People fall in love with a potential, but if you don’t have the, again, you feel ready but you don’t have the capacity, you’re the one with the capacity, as a coach, but the clients don’t have the capacity yet when it comes to the real work, yeah, exactly, of doing it themselves, they realize they’re not super good yet at, like me, the questioning of the machine and generating with a prompt, it takes them 500 prompts, they’re like, it’s not as easy as Jonathan said, it seems like I paid a lot relative to the results I’m getting.

Yeah. Okay. Uh, me, I think the shame also comes from the fact that at the start of the conversations, the expectations have to be in line with what I call the maturity of the organization. That’s why Vincent and I are redoing our approach, because beyond the technical realization, the learning, the learning experience becomes essential to, you know.

That’s mainly what happens, in fact, uh, as a result, uh, because you learn that if you’re clear in your intentions, you’ll reach a better result than if you execute it yourself. It’s completely reverse-engineering from how we learned to build our trades. Uh, and that, if you don’t learn it, honestly, you’re going to suffer psychologically, you know. There’s, no, no, but there’s psychosis that’s already observed, there’s emotional resistance.

All that, I see it every day. There are people who’ll call it change management, but I just don’t want to deny the fact that you do have to do change management sometimes. It’s just that the speed AI moves at, what it takes is adaptation. It’s, it’s curiosity and adaptation.

It’s not even, there’s, I can’t manage your anxieties. You’re going to have to learn to regulate your nervous system. First off, doing AI is like telling people, learn to regulate your nervous system. Yeah, that’s true.

And then, that’s your door too, what reassured me enormously about getting into the premium version is that she did a qualification call with me. So of course, it’s the routine, you don’t, step by step, but she wanted to make sure, in fact, that she could, uh, be the right coach for me. That’s exactly what the other one hadn’t done. The other one, he just wanted his 15,000 bucks, and after that it was settled, you see.

Whereas she was like, “Wait, yeah, I’m going to take your money, but first I want to make sure I can deliver.” And that, it, it’s true that I do this now. I, I do it now because of the experience I had with clients who maybe are listening to me right now, and it’s a better, a better outcome, in fact, what I do is, sometimes you take on a client and you think the client is ready. In fact, they’re ready because they’re with you. Okay.

And then you give an impression of confidence, but in fact they don’t yet have the confidence themselves. Really important not to drop a client, to say give me your 15,000 or your 20,000, you do a residency and everything’s fine, but as soon as you leave, the confidence disappears in them, and then they resent you, but they resent you, they see it’s not working, if you’re not there, you know, and that, that’s what Vincent and I want to fix, it’s, it’s all well and good, the technical capacity, but philosophically you also have to be transformed by the end so that the capacities you were missing, you know they’ve been mapped, you see where you are and where you want to be, and the gaps; if you don’t give that to your client, that’s where there are frustrations that I find pretty striking. Vincent, what are the prerequisites for working with Crewdle? Because I imagine that’s the, the best way to approach a client too. Yeah.

Actually, I’d say there aren’t, there aren’t that many prerequisites, uh, you know, we really work across all sorts of industries. Uh, no, but in the sense of, relative to what Jonathan was saying, there’s one point, it’s, do you, do you want to do AI, meaning, do you have the will, and not, uh, you know, there’s a difference between, say, uh, oh, I’d like to do AI, you know, I’d like to do AI, and I really want to do AI and I understand what it implies to head down that road, you know. Uh, that, that’s going to require changes, it’s going to require questioning, it’s going to require a whole journey, and if the client isn’t ready to get into that, it’s, it’s not even worth it, you know. And me, me, I see it, you know, in the sense that, uh, when we make proposals, uh, you, we’re still in a position that, that, I wouldn’t say excellent, but not catastrophic either. Which means often when we send proposals, I don’t even follow up with the clients afterward.

Meaning it becomes a bit of a way to filter. Oh, it’s, if you really want to work with us, you’re the one who’s going to come back, you know. Uh, and the people who get the proposal and we never hear from them again, well, they’re probably people who weren’t ready to, or who didn’t want to. Exactly.

They didn’t want to. So those people, we don’t chase them, we don’t chase them, we let them go, you know. And then sometimes some come back, some will come back 6 months later, a year later. Then they’re ready. Then they’re ready.

They’ve, they’ve moved along, they’ve lived, they’ve lost clients, they’ve, you know. No, no, but, and by the way, that’s one of the reasons why we ourselves adopted AI at that point, we were in survival mode for our own company. In fact, we were going to die. We were sure, in fact, that we were going to die, that we wouldn’t survive where, where we were. So we didn’t have a choice but to adopt AI, and in fact it’s AI that kind of saved us.

Uh, and today, well, we made a business out of it, but I think the clients who right now are, let’s say, too comfortable. Yeah, the will won’t be there to make that change. Yeah, yeah, yeah. But coming back to comfortable.

Car Magic is a methodology we’re going to bring, that I’ve developed over some twenty years. I’m bringing it inside the Crewdle agents. The Crewdle agents are therefore going to have tons, they already have tons of methodologies we’ve given them, like skills. But this is going to be unique to Crewdle.

I’m not going to put that on Claude. You know, it won’t be something I give out publicly. If you come into Crewdle, you’re going to have a kind of Jonathan AI agent that’s in design-critique mode, you know, that says, “Do you know why you’re thinking about this? Do you know why you’re asking me this?” These are philosophical things that you never take the time to ask yourself.

Our system, it’s in philosophical mode at the start. The, for, I think, Vincent, we’re going to call it, I don’t know, do we want to reveal our, our big, our big projects yet, but you know, we’re going to call it Dream, the Dream function. Yeah. Oh yeah, in fact the idea, the idea of the ecosystem, because in fact that’s what we’re going to launch in, in a few days, it’s, it’s an idea of having an ecosystem rather than a product.

And then, depending on the stage the company is at, it’ll use some products more than others, and it’ll evolve within that. So we cover, and all our products, we try to make a single syllable that’s an, that’s an action. Uh, so the first action is chat. Okay. Chat, that’s really the entry point.

You’ve never played with AI, start with that, ask questions. So we’re really in a very, very simple interaction mode, not too confronting. And then, well, we evolve through that. And then, like Jonathan was saying, we want to introduce his methodology into our platform through two strands.

The first strand, which is the Dream strand, which is basically going to help us structure our brainstorm and find a direction, because as Jonathan was saying, velocity is a combination of a direction with speed. So the tools we have for building bring the speed, but now if you don’t have the right direction, it’s not much use. Like my example of doing leadership coaching, which got me nowhere. Exactly. Exactly.

So we’re going to have the, the, the Dream strand. But the second strand, which, in fact we thought about this morning while discussing a case, uh, uh, let’s say a slightly more complicated one, but we realized it would probably need a Learn strand. Okay. A strand where we support, we develop the client’s abilities to use AI, whether it’s their capacity to communicate, their capacity to think, even technical capacities, because, you know, uh, like Jonathan was saying, the choice of words is still important, you know. Yes, we can simplify, on our side in our UI, the words we use, but at some point, if you don’t talk about, I don’t know, ontology or design tokens or, with AI, well, your project is going to be pretty limited, you know.

Uh, you won’t, you won’t make an incredible PRD, for example, if you don’t know the terms. At some point, you have to, you have to take the time to sit down and understand the terms. Same thing for an architect who wants to draw up the plans for a house. Yeah, there are terms they have to learn, you know.

And then AI can help, but the human has to have the knowledge of those terms to be able to dialogue and exchange with AI, or, very good. By the way, I’ll do a little plug for a VES event, so the Visual Effects Society, which I’m now part of at the board level. It’ll be the 28th, uh, sorry, the 29th, uh, of April at NAD at 6 PM. Uh, there’s the technical director of Rodeo FX, Peter Nobs, who’s going to do a, a little talk on the origins of machine learning.

So, uh, to make the link with what you’re saying. I think it’s going to be interesting to come back a bit to the fundamentals. He, he studied that in his student head when he was still in Munich, and, well, he has the know-how of how you do it in production with Rodeo, which runs really well today. So that was a little aside.

And in fact, Vincent, what you just said is a perfect transition for, for the last part. It’s okay, it’s cool, but where do we start Monday morning? Let’s imagine that me, I’ve never used AI, I’m a business owner, whatever the field. Where do I start, Monday?

You start, V, you go for it. Listen, me, I have my little idea about this, having seen it. You know, like I was saying, the first step, uh, don’t start thinking about building complex systems or doing autonomous agentic stuff or, uh, no, go very, very simple. Open a chat interface and start using it to accelerate your information research.

Okay. Yeah. The intelligence, because that’s the first area where you’re going to be confronted with hallucinations, and you have to develop the reflexes to detect those hallucinations and to learn how to dialogue with the machine to explain to it that it, that it hallucinated, because it’s fine to take an AI, we can tell it, listen, you lied to me about X, or, can you give me the sources, or sometimes even, me, I saw it recently, uh, someone using it next to me, and I knew the answer to the question that, that the person asked, and, uh, over time, sometimes the information evolves.

So the answer AI had given was based on, uh, its knowledge base on which it had been trained. So it was an answer that dated from two years ago. Yeah, I challenged Claude with the Chemtrail thing, it didn’t like that, but you know, at a certain point it was like, well, there’s, there’s a court ruling where the name Chemtrail was named, so it was faced with the fact, and you’re like, okay. So that, that’s, for me, in my opinion, really the, the first step.

It’s the first step. It’s where we start to get familiar with AI. How is what you just described with your AI OS, operating system, that you’re going to build? Is that also a bit to support people in this progression, if I understand correctly?

Exactly. Exactly. In fact, we have ultra-simple tools. In fact, it’s funny, because us, we started with the most complicated tool in the end in our chain.

Yeah. Which is our studio that lets you build agentic workflows, fully automated pages. Okay. But, but then we, we, we deconstructed from there to bring it back to the simplest thing, which is the chat interface where we communicate in a pretty simple way.

So it’s a back-and-forth, right. There isn’t, there isn’t too much of a team process. But it’s interesting what Vincent says. I have to jump on this, because me, I was segmenting it into thirds.

You know, you had conversational intelligence that you learn to develop, meaning asking questions, realizing you’re not that good in your inquiry, that philosophically you have to develop conversational capacity if you want AI to go well. And then detecting, like Vincent says, the lies or the hallucinations, however we see them, which is a probabilistic, non-deterministic system. So that, that generates a prediction, the answer. There, you learn it’s not really intelligence the way it was sold to you, it’s something else. And you get familiar with that.

Then we connect that quickly, on our side, with our systems in document intelligence. You put data in there, you realize AI can also read data. There, it becomes an introduction to the notion of an agent. It can read documents, it can synthesize them, it can see them, it can browse.

So normal actions a human would have done, where now it’s no longer just a conversation. There’s like an external action that happens. There’s information that’s going to be searched and that comes back. You can do supported research.

The notion of correlation appears in, in the flow. And before we introduce the automated systems, it takes, it can take a year. Yeah. Yeah.

Yeah. Yeah. Before a client is with an autonomous agent, there’s a mental preparation, because recently we even had people internally at Crewdle who were introduced to autonomous agents, and it’s a personal revolution. It’s like, okay, okay, because there’s a fear that gets transcended, after that it’s like, once you’ve lived through your first, your midnight swim. Yeah, you realize there was something you were afraid of, that you’d passed, and once you’ve touched it, it’s very hard to go back to your position, I’d say, of resistance. You’re like, okay, what can I do now?

What question can I ask myself now that I know that’s possible? Mm, it’s a bit like preparing people, you have to, you have to prepare people to do survival, you know. It’s like, learn to make a fire, learn to make a fire with wet rocks, learn to make a fire underneath a, a shower, learn to stay in the forest without expectations. It’s step-by-step survival kit, you know.

Yeah. Well, it’s, it’s a bit, if we take the analogy of just hiring an intern and giving them more and more responsibility, and accepting that you also have to delegate. You, just that, in the real world without using an agent, is already complicated. So, you imagine, on top of that, a machine that obviously doesn’t know all the contexts, that you have to explain everything to, because in fact it’s a bit stupid.

If you don’t explain to it, it’s maybe even more stupid than an intern. So, but once you’ve explained it well to it, it’s far less stupid than an intern. So there, that’s clearly, it brings yet another dimension. Uh, well, listen, very cool.

We’re getting there, we’re pretty much, well. Just to, uh, add to that, the big danger too, it’s important that people realize this, is, uh, because you bring up a good point about context, and context management is in fact a hugely important thing in AI. Uh, and the danger, in fact, is that AI has a tendency to want to please the human. Oh yeah.

Yeah. And if there isn’t enough context, it’ll invent the context. Yeah. It’ll make you persevere in your error.

Exactly. Exactly. And, and it can go far with that, you know. And I was reading the other day, there’s someone who, who, who wrote an article about this, but what it said is that whatever your point of view, if you want AI to justify it for you, AI is capable of justifying your point of view, a bit like the internet.

But from there, no, again, I have to harp on this just because we talk about it nonstop, but learning is transversal across all the features of our ecosystem, because you can learn at every stage, you can have training, books to read, coaching sessions, whatever. Learning is really transversal. But say, Dream, it’s also about saying, “Okay, I dreamed about this, am I ready to make a product?” Us, we’re going to say no. A nice idea, it’s, there isn’t even a concept.

I won’t write the concepts for you. We’re going to, I’m going to ask you questions first. There are living platforms that’ll take your idea and make you a landing page, straight pipe, in one minute. Whoa!

I made a landing page yesterday. But, but you don’t have the, behind you. Tactically, that’s not the right approach. We’re going to take a more patient approach.

We’re going to say, we’re going to ask you a few questions to refine your vision. We’re going to clarify your intention. And after that, the system is going to say concept, and then there. Yeah.

Otherwise, you’ll, we’ll force, we’re going to force people a bit to come back to the basis of their, their foundations, of their intuitions, instead of just exciting them with the flame. We’re really going to show you how to make a fire and how to sustain it, because just making little flickers, everybody can do that. It’s like in a strategy game, StarCraft or whatever, you’re going to send your forces to another planet but you’ve got three tanks and two infantry. AI is going to say, “Yeah, yeah, no problem, we’ll do that, we’ll do that.” But your system is going to say, “Wait a minute, man, you’re going to get wiped out in 2 seconds.

Stop that right now.” We want, we want to keep our clients, you know, there’s, that’s it, there’s organizational maturity, once you identify it in any organizational change management, in any governance, maturity is from 1 to 5. There are tons of models that exist, but the best-known matrix, it’s the matrices, you know, maturity 1, 2, 3, 4, 5, where the company’s capacity to learn is measured, measurable and measured. And so you know, okay, are the people, is it top-down, the company’s hierarchy, is it horizontal or vertical? At Crewdle it’s horizontal, there’s no hierarchy, there’s a direction but honestly the most recent employee Vincent hired, he has the right to say what he thinks and to completely change our opinion on a subject. Uh, the maturity to transfer, I, there’s still, there’s still someone who, who takes responsibility for the decision. Yeah, Vincent, Vincent, he’s mature enough to make that decision to let it go, but everyone really has a free and active voice.

Okay. Yeah. The second level of maturity is, are we ready to make the changes once they’re named, and there isn’t a roadmap that, that blocks a revolutionary idea, there isn’t a roadmap that blocks an opinion. Third level, it’s all the departments talking to each other. So unification of data, transversal chats, a capacity to learn.

Yeah. When I say a capacity to learn, it’s to unlearn and to learn. Fourth, it’s become systemic. It touches the finances, it touches where the company’s investment goes, it touches the products we kill or don’t kill, the communicating vessels, kind of. And the fifth, it’s the market.

It means that once an industry, there can be tons of companies like the one we’re creating right now, we touch industries through our approaches. We can change a city if we, if we decide to go after the companies that touch cities. Well then, it’s systemic. The fifth level is that the company, a bit like Patagonia, transcends the organization and affects the world.

Vincent and I, we’re, I think you could consider us humanists, even though we’re very technical. We want to help society. Mm-hmm, mm, me, I’m not a capitalist. Even though we’re surrounded by capitalists and we’re funded by capitalists, we’re fundamentally here to help humanity with AI. Then again, it depends what you call capital, you see.

Capital, it can be harmony. I meant, I just don’t want to say I’m not anti-capitalist, I just want to say we’re humanist fundamentally in our foundations. We want the tool to serve the human, and we want AI to make the human more effective. The, the goal isn’t to accumulate cash forever in a way that makes no sense at all.

We hate people who just charge their clients more when in fact they should be giving them more freedom. And, uh, you know, making a web page, if we’re able to do it for 50 bucks but we give our client a superior capacity in their thinking for other things, well, that’s what we’re trying to do, it’s, we’re trying to get out of the old model too. Yeah, in any case it’s not going to hold, the old model.

So, well, yeah, very cool, I’d say, well, that was a good conversation, thanks, it did me good. Vincent and I had started, precisely, by saying, when resistance shows up, it’s normal, and we want to integrate it even into our business processes. The emotional intelligence it takes in the organization is going to be part of our systems now. It won’t just be, it’s going to be a feature of our product.

We’re going to, never, because, uh, I’m in the middle of learning the script of a masterclass, the script of a VSL, video sales letter. You have to integrate your client’s resistances into it. And my goal is to do that but in an authentic way, you see, because I’d already learned it with the 15,000-bucks guy, but he had no ethics. So I was a bit disgusted by that thing.

Except that she, the new coach, she does it ethically and with real intent, you see. She loves the people she supports, and that changes it. It’s the same tool but the approach changes completely. That’s it.

Really the final word. Really the theme, me, I think, that comes up today is precisely the importance of human interaction in this technological change. We can’t deny it. Uh, we also can’t deny people’s fears.

You know, Jonathan made a good point, but even us, internally, sometimes we talk about new products and people are afraid, you know. And, uh, I ask them, “But have you tried it?” Oh no, no, but it, it won’t be simple enough, it won’t be, no, no, just try it. Yeah. Yeah.

But so people find reasons too not to, uh, not to pay for them. Uh, but you mustn’t, you mustn’t minimize those fears. Meaning they’re, they’re real, they exist. We have to, we have to talk about them, we have to be patient.

Not everyone is going to make the change at the same speed. Uh, but in the end, human interaction, in fact, becomes more important than ever in, in one of the biggest technological shifts that we’re, that we’re going to know. That too, getting people out of their laziness, because I don’t feel like picking up a new tool, but look, take this tool, it’s going to save you from working more afterward, or actually you’re going to work better, you’re going to be more excited about your job. Now I’m going into a new rabbit hole.

We’re going to keep that for the next one, to finish at some point. You’re right, but me, that, I have a door that opened onto, there are people who’d already made the effort to declutter their lives and to get rid of precisely all those tools that were coming into their lives, and, it’s already a bit dizzying for a lot of people, computing, but AI isn’t just here to replace computing, you understand? In my opinion, we’ll talk about that philosophically next time, but you know, AI is beyond computing, you understand? It’s the electric, it’s a new, we talked at the start, at the start, it’s the new form, it’s a new electricity, it’s a new, a new engine, so you have to see it a little more metaphorically, I think, the impact of AI, because we’re going to see it in other domains, it’s going to be the autonomous agents, which we’re barely starting to skim the surface of, that are going to discover things the human hadn’t discovered.

So we’re going to be surprised by our AIs. I can’t wait to see GPT-6 or Mythos. What does it exactly mean, what they’ve sensed there in the markets? Okay, there’s something revolutionary in the AI hardware that hasn’t been touched yet, even if most people see it or sense it, we haven’t lived through it yet.

But when it gets triggered, there’ll be no going back. It’s going to be, that’s a bit what Vincent and I are trying to put in place, just the boat capable of receiving the wave that’s coming, you know. It’s a bit like Noah’s ark, without wanting to name it, but just being able to stay on the sea, just being able to move forward on that sea, that’s already going to be not bad. But I don’t know what the size of the wave is.

Yeah, that’s it. That, that’s going to be live. Well, in any case, thanks for your input, it was very, very cool. And maybe next time live, we’ll see, we’ll maybe get organized.