The Small Team
Strategy · July 4, 2026
The question I get asked most isn't about revenue or tokens. It's headcount. Someone hears what Easy AI actually runs, tens of billions of tokens a month, on our own hardware, on open models a fraction the size of the famous ones, and they picture a building full of engineers. Then they find out how few of us there are, and the conversation pauses for a second while they re-do the math.
That pause is the whole series, and the answer isn't that we're frugal or that we skip weekends. It's that for most of software's history the way you got more done was to hire more people, and that quietly stopped being true. Output used to scale with headcount. Now it scales with leverage, and there are only two kinds worth having.
One is a person who makes everyone around them better: a leader who turns five people into the work of fifteen. That isn't a soft skill, it's the highest-return math in the company, and it's the same thing Netflix built on when it noticed that in creative work the best aren't a bit better, they're ten times better. The other is a person who commands a fleet of machines: one human directing a swarm of agents that used to be a department. A great hire and an AI fleet look like opposites. They are the same move, leverage, in two different materials.
Headcount is what you add when you've run out of multipliers.
None of this is free, and the series doesn't pretend it is. A team this small has a single point of failure with a name and a pulse, and an AI fleet has a metabolism that eats whether or not it works. I wrote The Harness about the scaffolding that makes a small, cheap model reliable in production. This is the same argument aimed at the company: the team is small because the leverage carries what headcount used to, and keeping it that way is a choice you make on purpose, eyes open to what it costs.
The Small Team
Output doesn't scale with headcount, it scales with leverage, and a leader who makes everyone around them better is the highest form of it. Turn five people into the work of fifteen before you hire a sixth.
In creative work the best aren't a bit better, they're ten times better. A few great people beat a crowd of good ones, and 'good enough' in a small team is a tax everyone else quietly pays.
The other multiplier: one person commanding a fleet of agents that used to be a department. The new skill isn't doing the work, it's directing and checking the machines that do.
The agent that replaced the hire lands on a different line of the same budget. When the AI bill rivals the salary, 'cheaper' stops being automatic and becomes a decision.
None of it is free. A team this small has a single point of failure with a pulse, and burnout, key-person risk, and groupthink are the bill. When staying small stops being a strength.
The founder still answers the support ticket. That's not a phase you outgrow, it's the one advantage a big company spends millions trying to fake.
Six essays on the leveraged team: the two multipliers, the discipline that keeps them cheap, the day small stops working, and what all of it is for. The instinct when something hurts is to hire. The case here is to reach for a multiplier first, and add a body only when you've honestly run out.
Kha PhanCo-founder & CTO, Easy AI