Fourteen
Nobody hired a thousand people because the work needed a thousand people. Headcount was a measurement of friction. That is the thing collapsing.
The layoff argument is about tasks. Which tasks a model can do. That is the wrong unit.
The unit is the company.
Nobody ever hired a thousand people because the work required a thousand people.
They hired because coordination was expensive.
Every layer, every handoff, every status meeting, every document written so somebody else could act.
All of it is the cost of moving a decision from one person to another.
A company exists wherever doing the work inside is cheaper than buying it outside.
The edge of a company is a price. It was never a headcount.
So the number of people was not a measurement of the work.
It was a measurement of the friction around the work.
That is the thing collapsing. Not labor. Coordination.
The work does not disappear. The distance between the pieces of it does.
Which is why the honest number is not twenty percent fewer.
Take a drop shipping company at a thousand people.
Catalog, merchandising, ad operations, tier one support, returns, supplier email, forecasting.
Every one of those is a queue between two people who never meet.
Rebuild it as agents doing the function and a human owning it.
Fourteen because that is roughly how many decisions still need a person on them.
I do not think that is obviously good and I do not think it is fast.
Most companies will do it badly first, and the badly will last years.
But notice where the difficulty actually sits.
The engineering is the easy half and it is close to finished.
The hard half is who decides. Who carries the accountability.
What the remaining people are actually for. What you owe the ones who leave.
It is economics. It is sociology. It is a question about people.
Which is the part almost nobody is working on, and the part that decides how this goes.
WE WERE NEVER PAYING FOR THE WORK. WE WERE PAYING FOR THE DISTANCE.