Jacob Coxon

Chercheur en pré-entraînement qui a passé trois ans chez OpenAI puis Anthropic et a démissionné publiquement le 9 septembre 2026, affirmant qu’aucune des deux entreprises n’agit de manière responsable et que les scénarios les plus agressifs pourraient échapper à tout contrôle d’ici fin 2027.

Fonctions

2023 – 2026Pretraining researcher, OpenAIThree years of pretraining research across OpenAI and Anthropic, by his own account; left OpenAI earlier in 2026.Source →
2026 – 9 sept. 2026Pretraining researcher, AnthropicResigned publicly, saying neither Anthropic nor OpenAI is acting responsibly.Source →

Au registre

9 sept. 2026 · Entretien · Fox News
This is possibly the most dangerous technology that humanity has ever created. It's basically out of science fiction. I think we have no other choice but to cooperate internationally because an arms race would be disastrous in a way that no other human activity has been in the past.
Pretraining researcher, Anthropic à l’époque
To Bret Baier on Fox's Special Report, Wednesday 9 September, the day after he posted his resignation (Tuesday evening US time; the register dates the thread in UTC); Fox prints the passage across an attribution tag — ," he continued. " — after the first sentence, and its only "superweapon" is one word inside the reporter's own sentence.
Researcher who departed AI role over fears of racing into extinction says people are 'begging' for regulation →
9 sept. 2026 · Entretien · TIME
It's not like some weird, distant, far-flung concern. It is the default trajectory in the next couple of years, unless people start taking some sort of action.
Pretraining researcher, Anthropic à l’époque
On losing control of AI, to Harry Booth of TIME the day after his resignation.
He Helped Build Powerful AI at OpenAI and Anthropic. Now He's Afraid It Could Kill Us →
9 sept. 2026 · Entretien · The Wall Street Journal
We're on track for a lot of the most aggressive of these scenarios where by the end of next year things could be out of control already.
Pretraining researcher, Anthropic à l’époque
The Journal's interview is paywalled; the sentence is reproduced identically by Futurism, Business Standard and Yahoo News.
Anthropic Researcher Quits Over ‘Out-of-Control’ AI Fears →
9 sept. 2026 · Publication · X
The people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt.
Pretraining researcher, Anthropic à l’époque
From the same thread, as quoted by NBC Los Angeles.
Anthropic researcher resigns with warning about the dangers of AI →
9 sept. 2026 · Publication · X
I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives.
Pretraining researcher, Anthropic à l’époque
The first post of his resignation thread, as quoted by NBC Los Angeles; the thread is at x.com/hilbertspaess/status/2097476196791709843.
Anthropic researcher resigns with warning about the dangers of AI →