Anthropic researcher quits over AI extinction fears
British mathematician Jacob Coxon left Anthropic over the race for ever more capable models, saying the industry is taking reckless risks.
Symbolic image: a cleared-out workstation next to a running server rack in an AI lab.
British mathematician Jacob Coxon left Anthropic because he considers the race between AI labs for ever more powerful models a reckless risk that could end in human extinction.
At a glance
- Jacob Coxon, a British mathematician, has ended his work at Anthropic.
- The report gives his reason as fear that AI could bring about human extinction.
- The trigger, per the report, was talk of crunch time and endgame inside Silicon Valley labs.
- Anthropic chief Dario Amodei recently called for a pause on self-improving models.
- The resignation date, his exact role and any Anthropic comment are unverified.
Jacob Coxon, a British mathematician, has left his job at Anthropic. He says the push toward ever more capable AI systems could end in human extinction, and that the industry is taking reckless risks to get there.
An objection from inside the lab
Coxon worked at Anthropic, one of the labs that have spent recent months leapfrogging one another with new model generations. That is what makes the departure land differently from the usual warnings: the objection comes from someone who helped build the systems, not from an outside critic. The accessible report does not say which team he worked on, or when the resignation took effect.
Crunch time and endgame
The trigger, according to the report, was the vocabulary that has taken hold inside Silicon Valley AI labs. Words like crunch time and endgame frame the present as the decisive stretch of a race already under way. Coxon reads that language not as motivational shorthand but as a sign of how seriously the people building these systems take their own timelines.
The chief executive is warning too
Anthropic chief executive Dario Amodei recently called for a pause on developing self-improving models. Together that makes for an unusual picture: a company near the front of the race is also among the loudest voices about where it leads. Der Standard notes that existential warnings double as a business model, because they cast a firm as the owner of technology too powerful to ignore.
What we could not confirm
This report rests on a single accessible source, and that source is only partly readable without a subscription. The wording of Coxon's statement, the date of his resignation, his exact role at Anthropic and any company response all remain unverified. The source gives no risk figures, and we are not supplying any.
FAQ
Who is Jacob Coxon?
A British mathematician who worked at Anthropic and left over concern about the consequences of ever more capable AI. The accessible report does not specify his exact role at the company.
Why did the Anthropic researcher resign?
He regards the race between AI labs for more powerful models as reckless. According to the report, talk of crunch time and endgame inside Silicon Valley labs was what tipped him.
How has Anthropic responded?
No company statement appears in the accessible source. What is on record is that chief executive Dario Amodei recently called for a pause on developing self-improving models.
