Earlier this month, the New York Times ran a profile of Medvi, a telehealth startup selling GLP-1 weight-loss drugs. Two employees. 401 million in first-year sales. On track for 1.8 billion in 2026 revenue (PYMNTS, April 2026). A dozen AI tools running everything from marketing to customer service. Sam Altman had predicted exactly this kind of company two years earlier: “In my little group chat with my tech CEO friends there’s this betting pool for the first year that there is a one-person billion-dollar company. Which would have been unimaginable without AI and now will happen” (Fortune, 4 February 2024).
The future had arrived. Then it started falling apart.
Forrester published a breakdown within days. The AI-generated marketing included deepfake before-and-after photos. Fabricated doctor profiles. The FDA issued a warning letter citing misbranding (Forrester, 6 April 2026). A class action lawsuit followed (Moneywise, 7 April 2026). The poster child for the AI-enabled one-person company had perfect tools and nothing underneath them.
This is not a story about one bad company. It is a story about a popular idea that has a serious flaw at its centre.
The idea everyone is chasing
By 2026, the phrase “one-person army” had become unavoidable. Andrej Karpathy coined “vibe coding” in February 2025 to describe building software by talking to AI rather than writing code (X, 2 February 2025). Pieter Levels built multiple million-dollar products as a solo operator. Y Combinator partners have publicly advised founders to resist hiring and see how far AI can carry them. The message from Silicon Valley is consistent: AI tools are so powerful that one person, armed correctly, can replace an entire team.
There is truth in this. I have experienced it myself. Work that once required a team of analysts, researchers, or writers can now be produced by one person with the right tools and the right process. The force-multiplier effect is real, and anyone who dismisses it is not paying attention.
But the version of this story that dominates the conversation has a blind spot the size of a building. It assumes the tools are the thing that matters. Get the right stack, learn the right prompts, automate the right workflows, and you become an army.
That assumption is wrong. And it was answered decades before AI existed.
Tarasov wrote about this first
In the early 1990s, a management philosopher named Vladimir Tarasov published a book called Technology of Life: A Book for Heroes. Tarasov had founded the first management school in the Soviet Union in Tallinn, Estonia, in 1984, and the book became a reference text for a generation of Russian and Eastern European entrepreneurs.
Chapter three is titled “The One-Person Army.”
Tarasov’s concept has nothing to do with tools. It draws on Sun Tzu and the classical Chinese strategic tradition, and it applies the principles of warfare not to armies but to individuals managing their own lives and conflicts. One person can function with the effectiveness of an entire army, but only if they have internalised certain qualities.
Start with what Tarasov calls “solid and empty.” Every situation, every relationship, every piece of information is either solid (real, verified, reliable) or empty (assumed, untested, illusory). “A combination of empty and solid always gives empty.” The one-person army never builds on assumptions. They verify. They test. They get close enough to see the situation for what it actually is.
Then there is self-governance. The one-person army does not wait for direction. They are their own general, their own intelligence service, their own logistics chain. They can generate strategy for any situation rather than follow someone else’s template.
And underneath both sits what Tarasov calls horizontal career: the accumulation of genuine capability across multiple domains, as opposed to the vertical climb up a single hierarchy. Not a specialist who automates everything around their specialism, but someone who can operate across domains because they have actually learned them.
Read that list again in light of the Medvi story. The AI tools worked. The marketing scaled. The revenue grew. But nothing underneath was solid. The before-and-after photos were empty. So were the doctor profiles and the regulatory compliance. The company was a machine built on air.
Where the two ideas meet, and where they split
Ask Silicon Valley what makes a one-person army and the answer is tools. Give anyone enough AI capability and they become powerful. Ask Tarasov and the answer is the opposite. The person becomes an army through discipline, skill, and the relentless habit of distinguishing what is real from what is not.
These are not the same idea. And the difference matters practically, not just philosophically.
I have spent fifteen years building operational capability across multiple domains: transformation, P&L management, commercial strategy, team leadership, process design. That breadth came from necessity, not curiosity. Complex businesses do not break in one dimension at a time. The person who can see across operations, finance, people, and delivery simultaneously can solve problems that a specialist cannot, because the specialist’s peripheral vision stops at the edge of their function.
When I began integrating AI tools into that foundation, AI did not make me capable. It removed the bottleneck between what I could see and how fast I could act on it. Research that took days takes hours. Analysis that required a team can be done by one person. But the judgement about what to research, what to analyse, and what the findings actually mean has not changed. That still comes from the years of doing the work.
This is the part the popular narrative misses. AI is a spectacular amplifier. But amplifiers do not care what signal they are amplifying. Give an experienced operator AI tools and you get faster, sharper, more responsive advisory. Give someone with no foundation the same tools and you get Medvi: impressive at scale, hollow at inspection.
What this means if you run a business
None of this is comfortable for anyone hoping AI will be a shortcut.
If you are building your own capability as a leader, AI will not replace the need to learn how your operation actually works. It will not replace the pattern recognition that comes from years of seeing where things break, or the judgement to know which problem matters and which is noise. But it will let you deploy that knowledge faster and across a wider front than was previously possible for one person.
If you are evaluating people, whether hiring, contracting, or choosing an advisor, the question is no longer “what tools do they use?” Everyone has access to the same tools. The question is what they built before the tools arrived. The foundation determines whether the amplification produces something solid or something empty.
Tarasov wrote that the ideal manager eventually becomes “inactive.” Not passive, but so thoroughly prepared, so precisely positioned, that visible action becomes unnecessary. The one-person army at full maturity does not need to act constantly because every action is precisely placed. AI makes that kind of economy possible for the first time at scale, but only for the person who has done the preparatory work that Tarasov described.
Capability is what you build. Accountability is what you carry. AI can amplify the first, but it cannot perform the second.
The one-person army was never about the weapons. It was about the soldier.
References
- PYMNTS. “The One-Person Billion-Dollar Company Is Here.” 3 April 2026. View Article
- Fortune. “Sam Altman wants AI to create a one-person unicorn with a billion-dollar valuation.” 4 February 2024. View Article
- Forrester. “Beware The Magical Two-Person, $1 Billion AI-Driven Startup.” 6 April 2026. View Article
- Moneywise. “Allegations against $1.8 billion Medvi piling up.” 7 April 2026. View Article
- Karpathy, A. Post on X (formerly Twitter). 2 February 2025. View Article
- Tarasov, V.K. Technology of Life: A Book for Heroes. Translated by Andrey Lempitskiy. FriesenPress, 2019. ISBN: 978-1-5255-4659-4.
