Artificial Intelligence · June 15, 2026 · 3 min read

The Small Team Is Not a Saving; It Is a Technology

Brooks's law was always right: every new person adds not speed but communication lines. In the age of AI leverage, headcount has stopped being a show of strength and become a cost item.

Document cards stacked layer by layer with module badges linked to them — a structure built from parts

The small team is not an austerity measure; it is the technology of a speed the crowd cannot carry — and artificial intelligence has, for the first time in history, pushed that technology’s leverage towards the unlimited.

The corporate reflex still comes from the old world: when the work grows, the team is grown; when the team grows, success is declared. How many people you manage is still a status sentence. Yet the mathematics of that sentence has been on the table for fifty years, and it is not favourable.

The Coordination Tax

In The Mythical Man-Month, written in 1975, Fred Brooks gave the formula of the tax we still pay: in a team of n people, the number of possible communication lines is n(n−1)/2. Five people means ten lines; twenty people, one hundred and ninety. Quadruple the team and you multiply, first of all, not the production capacity but the need to talk. Brooks’s famous law follows from here: adding people to a late project makes it later — because the newcomer consumes understanding before beginning to produce.

I may not see this tax on the balance sheet, but I see it in the meeting calendar. The alignment meeting, the sync meeting, the pre-meeting meeting… In a crowded organisation, people’s job increasingly becomes explaining the work to each other. Production gets squeezed into the time left between.

The Age of Leverage

Here is what changed: throughout history, the small team’s limit was the work its hands could not reach. Strategy is built by three people, but execution — production, variation, measurement, repetition — demanded a crowd. The agent age removes exactly this limit. I have written it before: the labour column can now be delegated generously. Today a team of three can reach the output of yesterday’s department of thirty — without paying the tax of one hundred and ninety communication lines.

This means the equation has quietly inverted. Once, the crowd was mandatory for scale; the crowd’s tax was the price of scale. Now the machine carries the scale, and the crowd has become a choice with a price. The organisation doing the same work with thirty people is, against the one doing it with three, not merely expensive — it is slower, blurrier and more fragile.

The crowd is no longer the condition of scale; it is a tax paid in spite of it.

The New Definition of Growth

This does not yield the conclusion “hire no one”; that is the wrong question. The right question is: what leverage does each new person bring? My hiring test has three items:

  • Will this person share the existing work, or add a new capability?
  • Does their contribution exceed the tax of the communication lines they add?
  • Will this role still require a human in a year, or does it fill today’s temporary gap?

The measure of a business that sells intelligence cannot be headcount; it must be solutions produced per person. In the chain of thought, structure, solution, the crowd is very often the dressing over a missing structure: whoever cannot build a process, adds a person.

The most interesting companies of the coming decade will surprise not with their revenue but with their revenue-per-person. Size is still impressive; but it is now sought in leverage, not in the organisation chart.