78. We need to talk about The Vibe Chasm

0:00Welcome to The Difference Engine, the show for tech founders, investors, investors and innovators.

0:11this time we're going to talk about an entirely new concept the vibe chasm

0:20so today we're going to talk about something entirely new something we're calling the vibe chasm so listen up this could be our most important episode thus far we're going to reveal the hidden issue behind the breathless valuations that's seen by some non-foundational models and why the world's going a little bit crazy with AI and what to do about it. Okay, so a little bit of background to this thinking. If you think about it, for the last couple of decades, there's been a startup mythology and that's revolved around essentially one idea really. Growth requires headcount and venture capital to pour money into firms and that would quickly convert into salaries and benefits.

1:04There'd be more engineers, more managers, more meetings, more layers, more spend. The thought was there was enough of a moat with software viewed as a high profit, winner-take-all type category winning game. And even in post-growth, once the growth finished, there was still this annuity revenue to be had. So there was plenty of time to put those engineers, managers, and meetings together and layers together. But not anymore. Yeah. I mean, essentially, it was a slight acceleration of industrial pace, which we've had since about 1760 or something like that. But now, you know, those of us who haven't been asleep have noticed that there's a new type of company emerging.

1:47And it really goes like this. It's essentially a single founder armed with AI tools and particularly Vibe Coding. They can build products, acquire customers, automate operations, and grow revenues to levels that previously required entire teams. And, you know, these companies are not theoretical. They really are not theoretical anymore. Certainly with M &A hat on, being approached daily by these companies, believing that they have created something value to be acquired. Well, there are some values. The three that come to mind are here in Europe, Base44, Choco, and Lovable. The most famous of those is clearly Lovable, the Swedish vibe coding company.

2:34Nothing to do with Vibe Chasm, just happens to be about coding. And this is started by a Swedish founder, 35, 36, I think, today. And it describes itself as an AI native coding platform. headcount's still very low and the valuation has grown very very rapidly where it's going to go from here to be decided choco is the german british founder the sweetheart of the berlin scene and that's you know rocket internet former employee and base 44 the literally solopreneur so we can really see that these things are live and out there so basically they're not theoretical anymore and you know if you look out there closely enough they're in software media e-commerce consulting education and and some niche ses categories so from our point of view the pattern is becoming increasingly clear you have one founder tiny operational overhead and support team, AI native workflows, multimillion dollar turnover in a matter of months, and extraordinary profitability because of the lack of overhead.

3:49But nothing stays the same forever. So the something that happens with these companies is that growth starts to slow, complexity starts to compound, systems break, and the founder becomes the bottleneck. Now, that is when what we call the vibe chasm starts to open. Yeah, and I think the vibe chasm is also very real. Let's think about Lovable. Lovable raised in July 2025 at 1.8 billion. And in December of that year, it raised 330 million at a 6.6 billion valuation. What could go wrong? What, oh, what? Well, if we speak to Mr. Stebbings and co, I think you'll find there's a doubt in the valuation.

4:33So the chasm has opened. it's there to be leaped and I guess behind all of this chasm is the notion of vibe coding. Those of you listening I'm sure none of you will not have heard of vibe coding but let's try and get a common definition on what it actually is. I mean what it isn't is using AI to write code right. It's much more than that. It's an operational philosophy where founders work conversationally with AI systems to create and run businesses at unprecedented speed, right? So, and, you know, that just sounds very informal. Well, it sort of is because what it's replacing is formal specifications, large engineering teams, structured project management, and traditional organizational processes, which have, you know, remained largely unchanged, certainly since the second world war now you've got these lovely lovely lovely fast-growing companies and it's about the founder and the founder operates through because of the vibe coding rapid iteration their intuition the way they structure their prompts the automation which they're being delivered you know relatively cheaply although you can burn through a lot of tokens very quickly and in all of this it's about AI assisted execution right now what is different here is the products emerge through dialogue the actual conversation with a system so the infrastructure is assembled rather than engineered in a formal process marketing we've seen layers and layers of marketers trying to academize their craft to build up models.

6:27Well, in this environment, you've got rapid iteration through automated experimentation. And even operations in this environment become prompt-driven workflows. Simple as that. The founder acts less like a traditional executive and more like an orchestrator, a conductor, of intelligent systems. So this dramatically compresses the cost of creation. So a company that, for instance, which you'd think of as a pretty normal startup, only months ago, really, maybe have 20 developers in it, a product manager, designer, support team, and, of course, the operational staff to make it all work. But that, the expression of all those people can be built and operated by one highly capable individual with the right AI stack.

7:23That's incredible, isn't it? It's pretty not to like. But then let's talk a little bit more about the rise of the one-person, million-dollar-plus company. And some of our friends in the States are talking about one-person unicorns. Well, we'll see about that. So historically, there were some solo businesses that existed. We know that. Consultants, freelancers, creators. The trouble is they rarely scared globally. Look at Johnny Ives with his new Ferrari design. He's got a whole team. Steven Spielberg, he doesn't make a movie alone. And there was a lovely picture of Taylor Swift sharing her spoils from a recent tour with their team.

7:59So the trouble with these sorts of individual contributors, if you will, is their output is constrained by human bandwidth. And even when they sold time and attention, which you could say Spielberg and Swift at the very least did, they did need to change and build out the team, build out the company, because that's what these things are. And AI, though, has changed all of that. It changed the equation because software generation, customer support, content creation, analytics, even sales outreach and operational tasks can now be done at least partially at what seems like, on the face of it, near zero marginal cost.

8:37That brings a new economic structure, right? You know, revenue scales way faster than headcount. Some profitability appears earlier. And the pain of organizational complexity is delayed. Hurrah. You know, so in theory, a founder can now launch MVPs in days or even hours. They can iterate continuously, 996 if you're on the West Coast or even 24-7 if you're European. in, automate support such as it may be at such a stage in the development, generate those marketing assets instantly, not have to work out what to do and then try and find an agency to do it for you. This is the beauty of it. You can deploy infrastructure with minimal technical depth.

9:26So this creates businesses that previously looked almost impossible through the lens of traditional startup economics and the potential for bootstrapping, which people still talk very proudly and so they should do, about building large last-generation businesses without external finance. But, you know, you've got a single person there. Just go back to it. A single person can be generating a$1 million AIR,$5 million AIR, even$10 million plus in revenue. you know and some wild podders other podcasts are available in the excited sorry united states even think that we will see a solo unicorn or a dekacorn that sounds a bit like hype to me but those dekacorns would have margins that resemble software monopolies rather than startups sounds good sounds great is it is it true no we disagree okay because there is a hidden structural limit.

10:29There is. So founder mode, which is something that our friends in the excited states say a lot, is a well-known version of the break things and carry on mentality, go fast and break things. And as we've heard from Mr. Stebbings and elsewhere, and this is where the idea of working 996, so 9am to 9pm, six days a week comes from. early stage AI native companies operate with this sort of founder mode and founder cognition and the founder remembers the custom edge cases the infrastructure quirks she had to work around the product decisions they made the growth tactics that worked what didn't work in those early days the automation chains the undocumented workflows it's all implicit knowledge and at a small scale that works surprisingly well at larger scale it's catastrophic right why that why why is it catastrophic?

11:23Well, because if you think it through, the company is not actually systematized. It's compressed into the founder's head, right? And it's the same flexibility that enabled rapid growth becomes its source of fragility. The founder becomes the architecture, the institutional memory, the QA process, the support escalation layer, and the strategic decision engine. That's Not just the brains, but the heart, the lungs, and the nervous system. So eventually what happens is that the customers will become enterprise-grade. You move very, very quickly through the early adopters here. Reliability expectations rise because you'll have people whose job it is to police your delivery.

12:13got compliance which really really really matters if you're selling into financial services or the public sector or whatever security matters that ain't going away that's only going to get worse uptime matters financial control matters hiring matters and governance matters right so the informal ai native model collides with the realities of organizational scale that it's So that collision point that forms the vibe chasm. All right. So let's define that vibe chasm because it's an important and maybe not obvious at first, but when you think about it, this is an accident waiting to happen for a lot of these very small companies.

12:58The vibe chasm is a stage where AI amplified solo execution stops being sufficient or even efficient at continued growth. It's that point where intuition has to become processes. prompts have to become systems, improvisation has to become written down as an architecture, and solo leverage must become leverage at the organizational level. Those of you who have been listening closely to this podcast would have remembered that we have a real fondness for Geoffrey Moore, who was a man that observed exactly what was going on in Silicon Valley and turned it into models that people who weren't in Silicon Valley could start to follow to accelerate the progress of their tech businesses.

13:42Now, he designed something called the chasm. And the chasm was the difficulty most companies have in getting from the innovators and early adopters of technology through to the early majority. And the point of his theory on this was that the buyers demand an entirely different type of sales, an entirely different support. And if you are going to start to sell to them, your organization has to change fundamentally. But it was this idea of a chasm that we liked. So hence, we decided that what's actually changed is that the whole industrial structure has changed with the arrival of AI. So crossing the chasm to sell to your customers matters less than you being able to cross the vibe chasm so that your company survives in order to be able to address the much bigger companies that you want to sell to.

14:43So if crossing the vibe chasm is the new challenge, which we absolutely think it is, it's going to require founders to introduce things they often intentionally avoid it. Because after all, they're founders and they're native AI founders. So they're a completely different breed, are they? Well, they are until they collide with the rest of the world. So the stuff they've avoided, and I think we all have sympathy with this, any of us who are trying to manage a company at the moment. would be the management layers, the engineering discipline, the documentation, the operational systems, the governance, the analytics, the accountability structures, and, of course, the repeatable processes.

15:28And the irony of all of this is the success of VibeCoding, a successful company, very quickly, is that it actually ends up creating the need for more traditional company building. Right. That is very ironic because they're trying to avoid something which eventually comes to them anyway. They've just had a bit of an acceleration at the start of it. And that might perhaps explain why Base44 or Solopreneur, example from before, sold out to Wix. But of course, that has a load of implications because after the acquisition, Wix let go of 20 % of its folks. and a move that was talked about a little bit as being something to do with foreign exchange bizarrely and a lot to do with AI.

16:15But the good news there is that Wix were able to build its acquisition very, very quickly. And there is a belief out there in technology land that if you grow something very, very quickly, but you have a lack of grown-up company discipline and you don't have the routes to market and so on and so on, if only somebody would buy you, then you could build the company. Further, the problem is there are thousands of companies that believe that and thousands of companies already out there trying to find a buyer. And there's going to be more and more and more of them hitting the vibe chasm. You're right.

16:55So the playbook has completely changed that nice trade sale exit. So beloved of European companies may not be as easy as it once was. And there are other implications. I know you think that many founders will try to resist crossing the vibe chasm just because the very reason they're attracted to vibe coding in the first place, because they're motivated by freedom, right? They escape the bureaucracy, the meetings, the management overhead, and all of the stuff that's associated with the admin of slow moving organizations. And of course, you know, the idea of rebuilding those structures, you know, feels like some form of regression.

17:33Some of them are going to avoid it. So again, because a lot of Vibcodes will be relatively young, this will be their first rodeo. And it won't have occurred to them that there's going to be all sorts of changes. They'll have to move forward. So if you want to stay intentionally small, that's fine. You can just cap your growth at a certain extent. You can avoid the big enterprise customers, go around with your early adopters. And you can continually replace any emerging processes with automation. And, you know, it's sort of exhausting, but, you know, people will move on and do other companies. You know, that is going to suit certain companies.

18:11It won't create the commanding heights of the economy. Many of these people are going to realize that the complexity cannot be abstracted away infinitely. And the point is, is that AI does reduce operational friction. What it doesn't do is eliminate organizational entropy. At scale, coordination itself becomes the product. So I guess we should round this section off because we will be returning to this subject very soon with some advice. I think what we're saying is the most important skill of the next decade in tech is the ability to assess where you're at as you approach the vibe chasm. and then plot from there, jumping or staying on the side that way before you jump, don't make the leap.

19:00So for years, though, this was not what was going on. For years, the defining startup skill was just build, build, build, build, build. And today, AI increasingly commoditizes all of that building. And so the differentiator may not be, instead of knowing when to transition from free improvisation to agile skills, the winners may be, not the founders who use AI the fastest, nor the companies with the fewest employees, which seems to be some badger-monop is all at the moment, but the organizations that manage crossing the vibacasm effectively, which has a few implications, right? Yeah, it sort of does.

19:36I mean, it means preserving speed, creativity, and AI leverage whilst introducing operational rigor, scalable architectures, very useful, and some sustainable management systems. You know, in other words, you know, the future may belong to companies that combine the fluidity of AI native creation with the discipline of enduring institutions. And so for a sort of a final thought on this, what we have seen is that for decades, technologies reduce the cost of distribution. So that's fine. That's what everybody wanted to do. Now AI is reducing the cost of the organization itself. Now, the consequence is profound.

20:25One person can now build what previously required tens, scores, hundreds of people. But any form of scale introduces complexity, and complexity eventually demands structure. So the defining challenge for the next generation of founders will not simply be building with AI. it'll be recognizing when they have reached the edge of solo scalability and learning how to cross the vibe chasm before growth collapses the company under its own weight hopefully that the description of the problem has been really helpful we see it very clearly you see it every day when people are trying to silly vibe coded uh companies right next time i think we should talk about how you cope with the vibe cousin

21:16Thank you for listening. If you want to learn more about category design, head to becategorical.com. If you need help designing and dominating your category, then get in touch. Contact details are in the show notes.

21:45Thank you.