Latest / Growth After Headcount / There's A New Chasm Before The Chasm — Yannick Dillen (Vlerick) #006

Transcript

There's A New Chasm Before The Chasm — Yannick Dillen (Vlerick) #006 · Growth After Headcount

• Haris: Yannick, thank you for joining the Growth After the Headcount podcast. Yannick: My pleasure, absolutely. Haris: I read you wrote a piece and where you actually argue that startups with no employees are not naturally cheaper ones. If you can explain that. Yannick: Absolutely, it all has to do of course with the cost of tokens that we all know with AI. And I think many budding entrepreneurs, so people that want to start a business, they don't realize in the beginning that after their first vibe-coded product that they can make in Lovable in Replit, in base base44 they don't realize that after a while, yeah, there is a cost of development, there is that token cost, and yeah, that might lead to start up stop after a couple of weeks. I've seen also because I'm a professor of entrepreneurship and I've seen some of my students quitting their startup idea after a couple of weeks because they realized like hey if we want to develop this it's gonna cost us maybe not a lot of money in in in in absolute terms 1000 euros but for a student budget of course that is yeah quite a quite a sizable amount and that left them with the decision not to continue with that startup so that's a pity Of course, because in essence the idea can be good, but the cost of development is still there, and that is something I think many AI native startups they don't realize yet that yeah, building is still something that costs money, they might have still that narrative, okay, now we can build everything. us you used to develop something with and then there was a salary cost of a developer. Now of course that salary cost has been changed to some extent of course to yeah what they call the tokenization of development and that is that is something we we frequently see. Haris: And what's your view then on the use of lower models, right? We have like but they're also like the previous model, four port six, and then we have the whole category of open source models. Yannick: I think we will see a natural shift towards those open source models because in the end, yeah, I would say especially for budding startups, young startups, yeah, cheap always wins in the beginning to get to an MVP. I'm not talking about yeah making your full-fledged product that that that big clients will use, but to get your decent working MVP, yeah, in the beginning cost is is is is a very bad thing for every startup who wants the bootstrap. And is not funded, so then we could we will go naturally to those open source models, I think. And that will be that is already a shift by the way that we see. for bigger companies, it's a different story, of course. They have those budgets to still work with the the big boys in the in the AI scene, and then yeah, they also sometimes have bills of millions of euros a month, and and that is something I also don't know how how long-lasting that can be because, of course, every budget now. it's it's all still it's 2026 but it's still new and shiny yeah what will that be in 2029 will even those big corporates say like we're still having a budget of a couple of millions a month for developing AI tools I don't know Asking the question is answering the question, I don't think so. I think also there we will see a shift and we will see a sort of a democratization in in the prices of everything. And that is yeah that is good for the users. There is maybe more bad for the yeah the developers currently or I should say the AI companies who are developing the infrastructure. Haris: yeah, I think it's a complex topic. I think as long as the return on investment calculation for a company is positive, but it's also I think a more complex calculation. The one side is can the AI first automate this function, right? Is there maybe some other factors? maybe personal touch is still important, like saying B2B sales. Maybe a person does not want to talk to an AI. So even if the technology is good, maybe there might be other person for not using an AI and rather even if there's a cost advantage it still go with the person. So I think the token cost is one of bigger pieces. Yannick: That's well I think that's correct indeed. Although we see a shift already happening with with many software companies, also smaller scale ups but the bigger ones as well. Here for example in Belgium, I live in Belgium, that they fired software engineers and now they're having vacancies on the website again for software engineers because they feel like hey we might have been too fast with with with firing them and and having that idea of okay, it's cheaper. We don't Need them anymore, but now they realize first of all it might be not cheaper, and secondly, to build again that full-fledged product that we need that is like perfect to ship. Yeah, we need those software people coming back because we can't we can't rely only on yeah on cloud code for example. So I'm you feel it already a bit there. I'm I'm I'm not I'm a believer in AI for sure, but I'm not blindly following I think all the yeah all the new stuff that is out there because I Still think that that in the end it's a bit like a pendulum, huh? You know, like a a clock is going from left to right. That pendulum was once not so long ago, a couple of months ago, was fully to that AI side, not to the software side. I feel now, and we're talking, yeah, let's say mid-26, we or even even fall 26, we feel that the the the the the pendulum is now going back to that software side, and in the end, if a clock is not Anymore, where does it end? In the middle, and I think that is exactly where we were gonna end in maybe one and a half year in the middle where we're gonna have to balance AI and software, and that is something I think we don't see enough. It's still the pendulum shift that we see. It's either software or it's either AI. No, in the end, it will be it will end in the middle. That is, by the way, something we exactly also see in in entrepreneurship versus management. Companies that that that that the big debate between do we need more entrepreneurship. Entrepreneurship in a company, or do we need more managers? Yeah, in the end, the pendulum stops and it always ends up in the middle. We need a balance between entrepreneurship and management. And the most successful companies they have that balance between entrepreneurship and management. And I believe that the most successful companies in the future, and that's not a distant future, will have that balance between being good on software, having the capabilities to develop software, and having the capabilities. capabilities to do great things with AI and that goes hand in hand of course. Haris: Mm-hmm. Yeah. I think a big piece of there is also the pace of AI development and where is it gonna go? Right. I think there's always the not sure if you read the book of Max Steckman, Life Three Po. And here's this metaphor of AI development where you can imagine like oceans and a continent and the mo mountains going up. sticking out and then like now imagine like what happened to Atlantis. At some point the water level went up and up until everything was soaked up by the water. And the water is like the AI, and then we have different peaks like generating an image, generating a video, generating code, and all of those different peaks, mountains have a different height. And the AI is rising and rising and rising And at some point, once it's over the water, it basically surpasses the human capabilities in that category, which we, for example, see in image recognition, like AI is like probably already above that mountain, they're really good at image recognition. So the question is: in what areas will the peak, the water basically reach the peak and go beyond? And I think that will also determine. Where AI can be deployed or not. Yannick: I think if we wait long enough in in in all areas that that's a singularity moment that we might we might pass sooner than later. and that is a reality we have to live in. But then again, what I was just saying, it's about having that balance between AI and and it goes hand in hand with with with software development which still which will still be needed. So I'm not a believer in AI who takes over let's say entrepreneurship because that's also yeah we we we we We spoke about it also that that that you have that phenomenon of of entrepreneurs who think that I set up an AI company and I can sleep and everything will be done for me. Yeah, I'm a huge non-believer in that type of companies. If you think that you can be such an entrepreneur, yeah, that's what I called in an opinion piece in the Finan in the Financiël Dagbot the beautiful weather entrepreneur, the mooi weer ondernemer in Dutch. Yeah, they think like we can just be on the own. Terrace sitting in the sun, and the AI will do the stuff for me. And if I come back home, yeah, I have some clients and I will have some cash on the bank because AI did that stuff. Yeah, it will not work like that. And sorry, it will also not work like that in five years from now when AI is even better. So, entrepreneurship is a hard thing to do, and that will always be the case. And even more, that's a sad story, I think. Even more in AI, in the AI era, because now it's even more difficult. we get the attention of customers to to yeah to stand out with a product that is that is more different because we will get an overload of I would say tech infused products and it's gonna be much more difficult to stand out in the crowd with a yeah with a product that has a mode and that is something I think that those beautiful weather entrepreneurs underestimate because they will have to do it themselves with their boots on the ground and that is not Much different, I think, from entrepreneurship that we have seen pre-uh AI that is still going out, talking to customers, fixing your problems yourself, being very close to the daily operations. And if you don't do that, I think you might you might run behind. And of course, AI will will will Many efficiencies that that can that it can take away. That's what every entrepreneur is doing these days. But if everyone is doing it, where is your mode? That is, of course, then again the question. So you will have to do extra things that other entrepreneurs are not doing to get to a situation where you have some uniqueness. Because if all AI does the same, there is no uniqueness anymore. That is, of course, also an open door. But I like to to yeah to open some open doors let's say Haris: Yeah, no, I hear you. And I talked to one founder, his name is Amos from Swan AI, and they're like three co founders and they scaled it to two hundred customers and multiple million in revenue, so more than one million AR per headcount in that regard. And but they were working seven days per week. Like they were really hustling as well, right? So the level of entrepreneurship, and I think he also talks about there are things I will never let AI do, and that is doing LinkedIn posts because that's his personal brand. He does not trust AI and it ever. So and their philosophy is also to amplifying the human. So if they hire somebody, how can we make that person extremely efficient? So I think there's this change. In that entrepreneurship will be extremely important, but how much you can accomplish with the right skill set using all of those tools, I think that is the mode basically that AI enables. And a counterexample, it's like the episode is gonna come actually out this first day, is Adam Robinson from RB2B. And their story is quite interesting because They set out to build an autonomous business, something that runs itself, and they developed it, grow it to six million AR, also with three co-founders, and then they did an experiment. What happens if we don't touch anything for seven days? They did it and everything went growing. And then they went like, okay, how far can we get this? And then they didn't touch. Anything almost for seven months. What happened? The company grew from 6 million to 8.4 million era. So I think it's a little bit a counter example. His case is quite unique. They made very strategic first principle design decisions to be able to allow for that. That is still needed, but you can create something that at some point with your hard work can be autonomous to some extent. Yannick: I I I sometimes compare it with that's not that long ago, four or five years, we had that hype with the drop shipping entrepreneurs. You know, everyone was like, Yeah, you need to set up your own webshop, you can do drop shipping, and basically you can also become yeah, a wealthy entrepreneur without doing a lot because you just have to set up a website, drop shipping happens and everything runs on itself. That was even in the pre-AI era. Today I think also there we have seen. Some successful people who do that for sure, but we have I think if if you compare it on on 100%, probably 3% became successful and 97% became unsuccessful because they start believing too much in that drop shipping story and then it doesn't work. I think we will see the same with AI. And and and let's say the AI runs on itself, yeah, for sure there will be unique stories where it will work, maybe the 3% and the 97% it will be. disappointments. That's that's a bit my my gut feeling, but that gut feeling is also based on what I see happening yeah in in in the startup world as as we speak today. Haris: No, no, I agree there. I think traditionally startups the majority always fails. I think I think that will always be a fact. I think even more will fail now, because if everyone can build, if building it's getting commoditized, right? Every like you mentioned in the beginning, you can start something with lovable. But how can you cross the chasm, right? The big mm yeah strategic framework there. And I think you also have a take on that. Yannick: Yeah, sure. Well crossing a chasm from Geoffrey Moore, by the way, one of my favorite books ever in entrepreneurship and marketing. It's a bit on the on the intersection between marketing and entrepreneurship. But I I I love the book, I love the concept, even if it's from the early nineties. Haris: Mm-hmm. Yannick: basically you always if you have a product, you will always find some innovators who love the product, that people that like to take risks, then you have those early adapters that also want to buy the That product because they want to take an advantage on the competition. But the chasm is that period when you have to basically move from early adopters to the early majority, and the early majority they are risk-averse. But in order to be successful, you have to cross the chasm, the gap, because of course that majority is the biggest chunk of the market. 68% of the customers are in that early and late majority, but they are typically risk-averse, they don't like to take risks, they want to see other people. Other companies using the product, and then they will start also maybe having interest in that product. Okay, Britain 1993, I think, or 1991 by Geoffrey Moore. So AI existed, of course, AI exists for a long time, but not AI as we know it. What happens, let's say, since the ChatGPT moment, and especially since I would say the vibe coding moment, which is not that long ago, since Lovable and all these things became popular. As you mentioned, yeah, building became super super easy. What also happened, we saw a bunch of new tech products coming out. So, what's now the difference? You have, I would say, an extra chasm in the beginning, pre-AI, you would always have some innovators, some clients that would say, Yes, let's try this. Now, in the AI era, that is not even a given. Having those innovators, having those thirds. clients will become difficult because you have so much choice. In German they say die qual der much Haris: No. Yannick: that you basically yeah if you are a customer a B2B customer and you have let's say an Hart tech solution yeah you might have 10 12 11 similar companies to choose from so if you are one of those 11 or 12 companies how will you stand out so I think yeah that crossing the chasm in AI you have now a pre Chasm before the innovators. So if you want to stand out, yeah, you are lucky already. I think if you can have some innovators buying your product and then the same things still exist, you still have to cross the big chasm going from early adopters to early majority. So life hasn't become easier for an entrepreneur in tech because there's so much out there. And of course, that's comes basically then my story comes back where I said, okay, the pendulum between Software and and AI, yeah, you need to find the balance there. I really believe that the best startups will be those who have yeah, good software development and good AI capabilities hand in hand. And those will be probably the ones who can cross that first chasm to get to those innovators, and they're also probably more successful to to go to the majority. I think that is a bit what I want to say there. I also said once in In an in an article, like it has never been that easy to build the wrong product. With vibe coding, you can build everything. In 10 minutes you can build a product. But you can probably, or you will probably, build the wrong product. What I mean with that is you build a product that nobody wants. In my my classes I I think I say five times every class, talk to the customers. If you don't know what to do in entrepreneurship, talk to your customers. And that is a step, yeah. We have forgotten or largely forgotten in AI. We start building, building, building. We start going to the market very fast and all of a sudden we forgot, hey, but that step where we have to talk to customers, we didn't do that. We built something we think that is valuable. But yeah, we didn't really talk to potential customers. And then you have that famous SISP, yeah, a solution in search of a problem, which is now I think many of those tech products have that problem. They have created a solution where there is not really a problem to be solved, and that again is because we we forgot the biggest I think step to take in entrepreneurship talk to customers, talk to potential customers to know do we have to build some. Yes or no, and that is again what we see now. Slowly, we see a shift that even with vibe coding, the first step is again: hey, we have to go to the market, we have to talk to potential customers, and we're gonna build what they want. But for some months, I saw it also happening again with my students, and in the startup scene, people were building a lot of things that basically were not checked with the market, were not checked with potential customers. Haris: Yeah, I think definitely agreed. And I think also what adds more complexity. I think the recipe you outlaid can work if you have an extremely phenomenal product that is cutting edge or deep deck, something that did not in that realm exist, a massive value add, like base 64, in lovable, etc. Like they created almost a category or revolutionized something. The likelihood that Anyone listening to this, or any of your students, or everyone, most people will not even get such an idea that they can even accomplish that. Which brings me to another point. I think it's not the pendulum only between AI and software. I think the bigger pendulum what's also evolving is probably the the role of marketing, distribution, and personal brand. Like what Yannick: Mm-hmm. Haris: I've seen now, many of those who have success with building an autonomous business or an AI native company, many of them are serial entrepreneurs. So they had something before, they develop for themselves a personal brand, they keep investing into their own personal brand. They usually also bring a lot of domain expertise into the product. So basically almost they productize the domain experience in a way and then keep pushing on the personal brand because they all know as well building is commoditized. How can I stick out? If I have an equal product like somebody else, what is my mode? Okay, I I have a personal brand in this space, people know me, or they know some other ones, and I'm famous for delivering. this quality results already, right? And I think the focus on marketing and personal brand, I think that's one thing that's gonna be more important, which is very difficult for very new entrepreneurs, because they don't have to. Yannick: I I tend to disagree a bit there because what I sometimes say also in my classes is the most Haris: Mm-hmm. Yannick: successful entrepreneurs or the most quiet entrepreneurs, which is not always true. I get it, that heavily depends on the sector. But I will give you an example. I'm I'm I'm involved in a startup called Marple Data, and they are building basically software for data analysis, some deep tech solutions for engineers. Who is their customer? Yeah, engineering company. But also let's say those engineers. Those engineers, and I'm now putting it very blunt, those engineers they are not on LinkedIn. Or many engineers, maybe they check their LinkedIn once a month. So the founders of Marple Data, Nero and Mathias, they are not really building that personal brand. They don't really care about that. What they do is, yeah, going to their customers because they know yeah, that's a sort of like an ICP, a persona that is not. Caring about the personal brand of the founder, they care about how good is that product that we're gonna use. So they go all in on making a top-notch product, and they don't really care about the personal brand behind Marple Data, which is then Nero and Mattia's the founders. So you have also that type of I would say successful startups. And on the other hand, I totally agree. If you are in a I would say a more closer to that seaside, closer to the consumer. side yes I guess then indeed the personal brand becomes becomes a becomes an interesting one yeah Haris: I think the tricky part is reaching your customers nowadays, because one other thing that AI accomplished is it is as easy as possible to spam people, like email automation, any type of automation, call automation, getting to that, which is part of the distribution. So even to sticking out, to figuring out, even if you build something that is amazing, to activate word of mouth. All of that is becoming more important. That's why I like the topic of distribution and brand. Except again if you have something exceptional, but the bar to create something exceptional, I think that is keep rising and rising. Yannick: Yeah. That also relates it. I'm I'm sometimes a bit a contrary and I like if you if you mention indeed that and I feel that as well in my mailbox, how many yeah automated emails you get, I don't barely read them. So what I s yeah, what I think is also interesting and I'm I'm I'm going to three or four fairs in this in this fall, just because you reconnect with people, you see what's happening there, you can have like indeed a r a face to face chat with the founders. And I think that is then the b a bit Like the strange thing happening, you see all the AI coming there, but what is again becoming more popular? Traditional fares. after COVID they had a they had they struggled. Now you see again that a lot of B2P companies they go to to Frankfurt, they go to London, they go to Las Vegas even just to be there because they feel like okay this is one of those occasions where we can really grasp potential customers and stand out of that let's say crowded field of yeah what I said the pre innovators chasm and if you just do that by sending emails it will not it will not work. Yeah, pick on. Haris: Exactly. Like that's again coming to the point, the figuring out marketing and distribution, which going to conferences it's just also another way of your distribution channel. So that is gonna be key. I think somebody also mentioned earlier the rise of analog, like automating everything, but okay, I need to be on the road. I need to meet the people, especially I think B2B because it's such a trust based business in a way. And Yannick: That's good. Fun story, yeah. You know, the company Aikido it's it's the is the fastest growing unicorn in Belgium, and they do security solutions for software companies, and they now have a yeah, they want to grow in the States in the USA. And what do they do there? Also, billboards. You can say, like, why do they do billboards? We see as well other companies do that. But if you have a foot in the door in the US, you might think, okay, let's start with the traditional. way the fully digital way of of marketing. No, they do the opposite. They go again to some extent I would say yeah with traditional marketing on the streets and that is I think the fundamental not not fundamental but that is like a strange but funny shift we see in a fully digital era. You go again to some extent to yeah to the classic old school ways of doing things and that's I think a fun way to to look at it. Haris: They probably got expired by an all other Belgian company, Udo, Odo. And I see their billboards when I travel to Thailand and Asia or Portugal, they have billboards everywhere, on the highway, like twenty meter billboard, Odo. Which is quite funny. Yannick: And then the I was thinking like, why do they have to do that in such a B2B space? But it makes sense. And they did that pre-AI, but now I think indeed, yeah, to stand out of the crowds, you have to think about yeah, guerrilla marketing, and that is not even guerrilla marketing, but guerrilla marketing will be one of those ways to yeah, again, to make your way into that, into that yeah, to cross the chasm, to cross the two chasms that we have now, the pre-innovators chasm. And then the real chasm to get to the majority. And then I think you will have to do maybe some crazy marketing stuff that is not in the textbooks. Haris: Exactly. People need to get creative. Which is that's entrepreneurship coming back to your point, right? Figuring that out. Like I think that's where the fun is as well. coming slowly to an end. Like do you maybe have few recommendations for new entrepreneurs for the time? Anything maybe recap what we talked to something new. Yannick: But Maybe something something new but you also referred to that I I like that the word complexity. I and and I one of my favorite authors is Chris Suk. And Chris Suk is an American yeah author, consultant, he was linked to Payne. And what he said frequently was complexity kills. Every layer of complexity that you can that you can basically erase from a company, that you can erase product those companies will be the winner and that's what I also again see happening now with AI. You can do so many things it's so easy again in lovable in all those vibe coding tools to add things and add things and add things which creates complexity and the Companies that will win, in my opinion, are the ones that will reduce complexity, that are easy to use, that are simple, and that are just yeah, making things easy in a non-complex way. And that I think is yeah, another trap that AI creates. You can create easily a very complex environment, but complexity kills, so yeah, I would just try to avoid as much complexity as possible. Haris: Okay, that's a good parting word. Yannick, thank you very much, and was a great session. Yannick: Thank you.