In my essay 'The Silent Externalization of Thinking,' I described how language models begin to take over parts of human thought, much like the printing press once externalized knowledge storage and pocket calculators externalized mental arithmetic. What was treated there as a cognitive phenomenon now has an economic counterpart: we are no longer externalizing individual thinking steps, but entire organizations.

When people can generate entire companies with a single click, the idea remains the only thing of value - as long as nobody knows about it.

— Steven Broschart

In other words: today, with the right selection of AI tools, a single person can act like a company with fifty employees. This is not science fiction, but reality. And it is the foundation of a term that is currently making its way through every business section: the One-Person Unicorn. The term sounds seductive ...

Two examples, one question

Maor Shlomo, a 31-year-old developer from Tel Aviv, started a side project in February 2025: a platform called Base44 that allows people without programming knowledge to build small applications through text input – what is now known under the term Vibe-Coding. Three weeks after launch, Base44 reached a million dollars in annualized revenue. After six months, the platform had several hundred thousand users, including companies like eToro and Similarweb. In June 2025, the publicly listed Israeli company Wix bought Base44 for 80 million dollars in cash, plus earnouts tied to revenue targets through 2029 that, according to Wix's quarterly reports, are likely to grow to around 90 million additional dollars. Shlomo built the whole thing alone, from code to customer service. His startup capital was around 15,000 dollars.

In February 2026, a British millennial named Jonathan Laramy launched the TikTok and YouTube channel Chloe vs. History. The titular Chloe, a young American woman who travels through time and finds herself in historical catastrophes, does not exist. She was created with AI video generators and has a consistent character reference that remains stable across dozens of clips. The channel accumulated millions of views within weeks.

Two very different stories. Both can be discussed under the label 'One-Person Unicorn.' Only one of them is one in the strict sense. But both show the same mechanism: a person appears to the outside world like an organization, or can accomplish what usually only an organization with multiple employees can accomplish. And in both cases, exactly what makes this impression possible is also its greatest weakness.

What is actually new

Sam Altman, founder of ChatGPT, told Reddit co-founder Alexis Ohanian in an interview in autumn 2024 about a betting round among tech CEOs over exactly when the first one-person billion-dollar company would emerge. Dario Amodei, CEO of Anthropic, gave the concrete answer in May 2025: 2026. With 70 to 80 percent probability. At the 'Code with Claude' conference in mid-May 2026, Amodei repeated the prediction – with the sober remark that with seven months remaining in the year, it still hadn't happened. What seemed like overblown industry rhetoric at the time can no longer be dismissed as hype without further consideration. Base44 hasn't become a billion-dollar company – but a solo founder pulled off an eight-figure exit within six months, demonstrating that the magnitude is no longer a theoretical construct.

What has shifted is the relationship between revenue and employee count. While a classic startup a few years ago generated about 150,000 dollars in revenue per employee, solo founders today are reaching magnitudes that can no longer be represented in any calculation because the denominator approaches one. Beyond spectacular individual cases, the shift is visible: Midjourney, one of the most profitable AI providers overall, generates over 200 million dollars in annual revenue with roughly 40 employees. That's about five million per person. For comparison: Meta is at roughly 1.6 million, Google at roughly 1.8 million.

The mechanics behind this are not magical. They consist of three layers:

  • AI-powered code generation, with which a single developer delivers in days what teams previously needed months for
  • AI-powered content and design production, with which one person generates marketing, visual world, and copy
  • AI agents that take over operational tasks like customer service, data analysis, or bookkeeping. A complete solopreneur stack costs between 3,000 and 12,000 dollars per year today.

Yes, this is real. And it sounds impressive too. But it is only one side of the coin.

The built-in half-life

The real paradox of the One-Person Unicorn is this: what lowers the barrier to entry for the individual also lowers it for everyone else. If someone can build a company with a few thousand dollars and a few tool subscriptions, they have a thousand imitators who can do the same.

Chloe vs. History is an instructive example here. The channel established a format (modern young woman with smartphone in historical catastrophe) that was copied within weeks. On specialized platforms like alici.ai, there is now its own page for every Chloe clip where you press a 'Remix This' button and get the image composition, character reference, and a reconstructed prompt formula – executable on an interchangeable stack of modern video models like Veo, Kling, or Seedance. Meanwhile, the same platform is now even distributing an independent product under the name 'AI History Generator' that promises users to place themselves in the format in fifteen seconds. What Laramy developed through months of iteration has been reduced to a single click. What was an original format has become a template, from the template a genre, and from the genre a third-party product in which Laramy himself is merely the eponymous pioneer.

With Base44, the mechanism looks different, but it still works; and Shlomo himself has openly named it. In an interview, he said in essence that he decided to sell to Wix because despite impressive growth, he was no longer able alone to achieve the necessary scaling. It is a remarkable justification: it was not the product that reached its limits, but the capacity of a single person. And parallel to Base44, an entire field of competing Vibe-Coding platforms grew within months, like Lovable, Bolt, Replit, Cursor, and dozens of smaller offerings. The lead is not what Shlomo built, but when he built it. And it is no accident that the exit came before that lead was exhausted.

This is a remarkable finding because it contradicts the marketing logic of many official voices that present AI as an indispensable competitive advantage. The reality, however, is that access alone to such tools cannot provide a lasting advantage because no one owns it exclusively.

Which competitive advantages actually count

If the tools are not the advantage, then it must lie elsewhere. An honest inventory yields a shorter list than the hype literature suggests.

  • Timing works, but is limited. Whoever first occupies a format or market is rewarded by the algorithm and identified with the thing in people's perception. This lasts months, sometimes years, but it does not last forever.
  • Distribution is the most underestimated advantage. Pieter Levels, one of the most famous solo operators, has built over 800,000 followers. This reach is not copyable. Not because it is technically protected, but because it took years to build. Whoever has an audience can test every new product immediately. Whoever doesn't builds in silence. AI radically lowers the cost of production, but it does not lower the cost of attention. On the contrary.
  • Style and the judgment of the masses cannot be scaled via templates. Which idea, which moment, which punchline – that is the layer in which Laramy remains superior to his imitators. This advantage is real too. Unfortunately, it is also not reproducible: it depends on exactly one person and scales precisely as far as their working hours.
  • Regulatory and contractual moats are usually overlooked, but are often decisive – and this applies precisely where the solo model is supposed to work long-term. A solo founder building a product in a regulated market (health, finance, law) needs either years or outsourcing to specialized third parties for licenses, certifications, and partner contracts. This seems unsexy, but it is the only advantage in the list that cannot be bypassed through money and speed. In unregulated markets – where most one-person unicorns are currently emerging – this protective layer is completely absent.
  • And finally, the brand – that term which is simultaneously used inflationarily and massively underestimated. Brand in this context is not a matter of logos and tonality, but of credibility: who has demonstrated over years that they know what they are talking about? Who do you trust to make the next thing work? This form of trust cannot be prompted. It is slow, expensive, and uncomfortable for anyone who sees speed as a central advantage. But it is one of the few aspects that become increasingly scarce in a world of easy replicability.

The dark side no one talks about

There is a second side of the One-Person Unicorn that either goes missing in most texts on the subject or appears merely as a footnote: the one person becomes the only instance that can catch machine failures.

Shlomo himself has spoken about this publicly. In the first weeks after Base44's launch, he woke up every two to three hours to check whether the platform was still running. At an industry event, when asked about the solo experience, he simply said: 'Yes, absolutely, I felt lonely.' That is a remarkably honest statement in a discourse otherwise characterized by vibe CEO romanticism.

A self-observation

Some time ago, I engaged intensively with vibe-coding. This resulted in a small tool that tracks changes to websites: visible content, job postings, and also comments in source code, which often reveal more about a company's development than any press release would. When searching for a suitable name, I came across a comparable project from another provider. And it was clearly recognizable as having been created with the same method.

What does 'recognizable' mean? It was the unused default structures, the unconfigured standard elements, the typical visual patterns that emerge when a language model generates a UI and no one bothers with individualization. Vibe-coding leaves a patina if you don't polish it off. The function selection also seemed reconstructible – as if one could roughly guess the conversation that led to it. It was a product that did not look like the result of understanding, but like the protocol of a conversation with the AI.

This is not about whether my own tool would be better. But about the question that arose in the comparison: which product does a user trust more? The one where the configuration decisions, the UI peculiarities, the selection of functions are clearly made by someone who understood the problem – or the one where the same decisions are clearly not made? If a product looks like a fleeting conversation with an AI, it signals exactly that: that the person behind it didn't spend long enough on the matter to make their own decisions. But trust arises precisely where such decisions become visible. Where the investment in the later user base becomes visible. From the perspective of UX and design. In other words: with the development of one's own brand.

Democratization with a price tag

How should all this be assessed? A comfortable answer would be: positively, because more people can build more. Another quick answer would be: negatively, because a wave of interchangeable solo operations is washing over us. But both answers fall short.

What is actually happening is a democratization of possibility, not of success. The barriers to building something have dropped dramatically. The barriers to remaining permanently visible and relevant with what you've built have not dropped – quite the opposite. Because whoever enters a market where everyone can do everything no longer competes with what they can do, but with what they are. Through credibility. Through personal preferences. With a history that cannot be built in four weeks.

That is the uncomfortable realization for all those who see the One-Person Unicorn as a universal business model: the very characteristics that protect a solo operator long-term are the ones that cannot be scaled and cannot be packed into tutorials. They depend on one person, on their time, on their history, on their judgment. Those who have them will go far with AI. Those who don't will attract brief attention with the same tools and then disappear – not because the tools fail, but because they work. For everyone.

The real scarcity good of the coming years is not production, but the credibility of the person behind what is being produced. That may sound romantic. But it is a sober economic view of a market whose means of production are just losing their value ...