Nobody Was Fired

Every layoff I have ever run had a date. There was a meeting, a list, a script, a room. People walked in as one thing, an analyst, a manager, a designer, and walked out having to become something else, and however badly it went, everybody in the building knew what had happened. There was a loss, and it had a name everybody understood: a layoff.

I want to talk about a different kind of loss, the kind with no date and no room, because the most careful piece of evidence I have seen this year describes it and almost nobody noticed.

In August, Stanford's Digital Economy Lab published an update to its Canaries in the Coal Mine research, using ADP payroll data through June 2026. The headline finding is reassuring: no widespread, economy-wide job displacement associated with AI. If you stop reading there, the doomsday people were wrong again, and the people who said it would all be fine can point at the chart.

Keep reading. Among workers aged 22 to 25 in the occupations most exposed to AI, employment now sits about 19% below where it would be if it had kept pace with their peers in jobs AI touches less. A year earlier that gap was 15%. It is widening. Stanford is precise about where the decline sits: in occupations that rely on codified knowledge, the formal, documented kind you learn from education, textbooks and written procedures. In plain words, knowledge work. The office. And here is the sentence that stopped me: the adjustment operates primarily through reduced hiring of young workers rather than increased separations.

Nobody was fired. The door just stopped opening.

Sit with that for a moment, because I have been on the other side of that door and I know exactly how it happens. Nobody in a leadership meeting says "let us stop hiring twenty-four-year-olds." What happens is quieter than that. A role opens, someone asks whether we really need it, someone else says the team is already using AI for that work, and the requisition sits for a quarter. Then another quarter. Then it is quietly closed because nobody remembers why it was opened. There is no list, no script, no room. There is just a job that would have existed and now does not, and a person somewhere who would have got it and never knew it was theirs.

I have signed off on that more times than I have signed off on a layoff. It never felt like a decision about a person, because there was no person yet. That is exactly why it is so easy to do.

The thing about a layoff is that it produces a story. You had something, it was taken, you are allowed to grieve it and the people around you understand. There is a vocabulary. Redundancy, severance, restructuring, the whole grim dictionary. It is ugly but it is shared, and a shared vocabulary is what lets a person say what happened to them and be understood.

The twenty-four-year-old who did not get hired has none of that. Nothing was taken from her, because she never had it. She sent applications into a silence, she was told the market was tough, she was told to reskill, she was told to be resilient. Nobody told her that the door she was standing in front of had been quietly bricked up from the inside by people like me, for reasons that had nothing to do with her and everything to do with what we believed AI could do.

That is the loss without a name. And I think it is harder to carry than being let go, because you cannot mourn what you cannot point at.

Here is where it connects to something bigger than one age group. Everything being written about AI and identity, including by me, is about people losing what they spent years building. The forty-five-year-old analyst who was the analyst, and now is not. That is real and it is the story I usually tell. But the Stanford data points at a second story underneath it, a bigger one.

Identity at work gets built, slowly, through the early years, through the first job where you were bad at something and got better, the first manager who trusted you with something that mattered, the first time somebody in a meeting turned to you because you were the one who knew. That is how "I work in finance" becomes "I am an analyst" becomes something you would defend if someone tried to take it. It takes years, and it takes a door that opened when you were twenty-three.

If the door does not open, the identity does not get built. Not eroded, not repriced, never built in the first place. And the Stanford paper is careful to say where this concentrates: in the occupations where AI automates the work rather than complements it. Which means the jobs that are disappearing for the young are precisely the entry-level, learn-by-doing jobs where the building used to happen.

Look, the people making these decisions are not villains. I am one of them. When AI can do the junior work, the arithmetic on the junior hire changes, and the arithmetic wins, in my experience, every single time it is allowed to be the only thing in the room. The point is that we have found a way to take something from a whole generation without any of us ever having to look one of them in the eye. There is no room, no script, no date. It is the cleanest restructuring ever done, and that should worry us more than the messy ones.

I keep coming back to the phrase. Reduced hiring rather than increased separations. Two dry words for the difference between a loss you can name and one you cannot.

What do you say to someone whose loss has no name? I honestly do not know yet. But I know that "reskill" is not it, because you cannot reskill your way through a door that is not there.