One founder, two letters, eleven months apart. The first said in writing that the job cuts were not about replacing anyone with AI. The second put AI at the centre of the explanation — and cut four times as many people.
The stock rose almost 18% the next day.
The useful account usually comes on the second telling.
931 people cut · 8% of the company
Reason given: strategy, performance, flattening
None of the above points are trying to hit a specific financial target, replacing folks with AI, or changing our headcount cap.
He volunteered the denial. Nobody had asked him.
~4,000 people cut · 40% of the company
Reason given: intelligence tools, at the core
We’re already seeing that the intelligence tools we’re creating and using, paired with smaller and flatter teams, are enabling a new way of working which fundamentally changes what it means to build and run a company.
Sent after the close. The next day the shares rose almost 18%.
Nothing about the software changed in eleven months.
The audience for the letter did.
If the software were really doing the work, the companies cutting hardest would be earning more than the ones barely cutting. Somebody checked.
Gartner · May 2026 · Companies above $1bn in revenue
The external benefit a patient collects for staying symptomatic. It is why some conditions are strangely hard to give up.
I saw this perhaps four hundred times. Not in markets. In my office. The tell was never that the explanation was false. It was usually mostly true. The tell was that the explanation was earning something.
In March 2025 there was nothing to collect by blaming the software, so nobody blamed it. By February 2026 the claim paid almost 18% in a day.
The press has a label for this now. A label is not a diagnosis. Everyone can name the behaviour; almost nobody has asked what it is being paid. And to be clear, what's actually on the couch here isn't a company or a founder — it's a sentence — one that now appears in hundreds of filings.
The prognosis is not here. Neither is the treatment plan — what to watch, what would prove the reading wrong, and where this pattern has ended before in the cycles you have already lived through. That is tomorrow’s session, 0630 Eastern, before the bell.
Start the SessionTwenty-eight years spent listening to people explain decisions they had already made. Markets do the same thing every morning, with more zeros. Written for readers past 45 who already know what happened and want to know who did it — before the rest of the room works it out.
Tomorrow, 0630 Eastern. Five to seven minutes.
