Hiring is a bet against the future

Every time you make a hire, you are placing a bet.

Not on the person. On the future. You are betting that a year from now, and five years from now, you will still need a human being to do the thing you just hired them to do. For the entire history of business, that was the safest bet on the board. It is quietly becoming one of the worst.

Headcount was never the asset

Start with the belief nobody questions: that a bigger team is an achievement.

We have held it so long it feels like gravity. Seniority is measured in reports. Status is measured in span of control. “We grew the team from eight to forty” is a sentence said with pride and met with applause. The size of your org became a proxy for the size of you.

But step back and ask what you were ever actually buying when you hired someone. Not the person, exactly. You were buying leverage, the ability to do more than one human can do alone. And for all of history, people were the only place leverage came from. Want to do more? Hire more. There was no other door.

So we fused two things that were never the same: headcount and leverage. We treated the delivery mechanism as if it were the thing being delivered. Headcount was never the asset. It was the rent you paid to get leverage, because renting humans was the only way to get any.

A bigger team was never a sign you were winning. It was a sign leverage was expensive.

The rent is going to zero

Here is what changed, and why it is not a fad.

AI is the first technology that unbundles leverage from headcount. For the first time, you can buy the leverage without buying the person. And the thing that makes this a trend and not a moment is the direction. The price of leverage is falling, and it falls one way only. It does not get more expensive next year. It never has, with any technology, and it will not start now.

The price of leverage, over time

it never has.

This is the same shape as “software eats the world.” That line was never a claim about what software could do on a given Tuesday. It was a claim about a direction that only ever pointed one way. This is that, for human labor.

And no, this does not depend on AI staying dumb

The natural objection is that this only works while AI is worse than us. That it is good at the grunt work and bad at the judgment, so the people with taste are safe.

It is the opposite. The smarter AI gets, the faster this happens, not the slower. If AI only ever automates the boring middle, the marginal hire weakens. If AI gets genuinely good at judgment and taste too, the marginal hire becomes pure overhead almost overnight. Run every scenario you want. There is not a single one where the case for adding a person gets stronger over time. That is the tell of a real trend, and the reason betting against it is so dangerous: it does not matter how the technology evolves, because every path leads to the same place.

Which is the whole point. Hiring is a bet that leverage stays expensive. It is the one wager that gets worse every single year, in every possible future.

The thing that’s actually collapsing

Now the bigger idea, because “you’ll need fewer people” undersells it.

The org chart itself was a workaround. It existed to solve one problem: a single person can only do so much, so you bolt people together, stack managers on top to coordinate the mess, and pay an enormous coordination tax to buy scale you could not get any other way. The whole elaborate structure, the layers, the meetings, the managers of managers, was scaffolding around one hard constraint.

That constraint is dissolving. And when the constraint goes, so does the reason for most of the scaffolding.

The unit is shrinking

the company as a coordination structure

→

leverage

a few principals + leverage

value creation is collapsing from the org to the individual.

So the real bet is not that companies get leaner. It is that the unit of value creation is collapsing from the organization to the individual. The future is not smaller org charts. It is a small number of high-agency people wielding leverage that used to require thousands, plus a long tail of one-person operations that used to require a whole company. For a CEO or a fund, the implication is uncomfortable and clarifying at the same time: most of your headcount is not a cost to trim. It is an artifact of a constraint that is disappearing.

What you can’t unbundle

So if leverage gets cheap and capabilities keep falling to AI one by one, what is actually left for a person to do?

Stop answering that with a capability, because every capability eventually falls. Writing, analysis, design, coordination, and yes, probably taste and judgment in time. Bet on a capability and you are right back to the snapshot mistake.

Bet on a role instead. Someone still has to want the thing. Someone has to decide what the company is for, carry the risk when it is wrong, and answer for the outcome when it is. Agency and ownership are not skills AI competes on. They are stakes. They are the one thing you cannot hand to a model, because a model has nothing on the line. That is the last irreducible human node, and a company is going to collapse toward it: a handful of true principals, the people whose company it actually is, wrapped around near-infinite leverage.

What is ending is not “the middle.” It is management as we have always meant it, overseeing humans who do tasks. What replaces it is quieter and harder: pointing leverage at something you are personally on the hook for.

So stop building org charts

The operating takeaway is almost rude in how simple it is.

Stop designing companies as structures of people. Design them as the fewest possible high-agency humans wrapped around the most possible leverage. Treat every hire as renting a capability you will soon own for nearly nothing, and reserve real hiring for the only thing that does not get cheaper: judgment, ownership, the willingness to want something and answer for it.

I ran the U.S. launch of a company most people assumed needed an army, with a three-person growth team and a refusal to add a fourth unless the work genuinely could not be made to run without one. At the time it felt like discipline. It was really just an early look at the only direction this ever goes.

For a century, hiring was the safest bet a manager could make. Which is exactly why almost no one has noticed it became the riskiest one. The teams that win from here will not be the big ones, or even the lean ones. They will be the ones who understood, earlier than everyone else, that they were never buying people. They were renting leverage. And the rent is going to zero.