AI NEWS PAGECASE FILEPERSON FILESTATUS: ON FILE
Jacob Coxon
Ex-Anthropic researcher who told the NYC council under oath that AI companies are 'extremely reckless' with the most powerful technology ever built
FILE UPDATED Oct 7, 2026
WHY THEY MATTER
Coxon quit Anthropic in September 2026 warning AI could kill us all by the end of the decade, then doubled down at the NYC Council's Oct.
SOURCES: NY Post
ZIMBO'S READ ON Jacob Coxon
STILL MORE NOTES
MORE ZIMBO ON Jacob Coxon
DEEP FILE
THE WHOLE FILE
chronology · facts · older headlines · sourcesHE QUIT ANTHROPIC, THEN CALLED IT EXTREMELY RECKLESS
Jacob Coxon quit Anthropic in September 2026 with a chilling warning — AI is advancing so fast it could kill all humans by the end of the decade. On Oct. 5 he testified at the NYC Council's AI hearing at Speaker Julie Menin's request, telling lawmakers AI companies run on a startup mindset: 'move fast, break things.' That works for a photo-sharing app, he said — 'it does not work for building the most powerful technology ever built.'
Oct 5, 2026
THE HEARING WHERE COXON SAID THE QUIET PART LOUD
Jacob Coxon testified at NYC Council Speaker Julie Menin's AI hearing, telling lawmakers AI companies run on a startup mindset: 'move fast, break things.' That works for a photo-sharing app, he said — 'it does not work for building the most powerful technology ever built.' He added that researchers have 'already lost the ability to fully control AI systems,' and said the majority of code is now written by AI and 'people do not check it that carefully anymore.' Fellow testifier Alex Turner, an ex-Google DeepMind researcher, guessed the odds of an AI takeover at 'roughly one in three.' The council is debating a bill package that would mandate human-override kill switches on AI tools used in the city, fine noncompliant companies $25,000, and create a first-in-the-nation cash reward system for AI whistleblowers.
SOURCES: NY Postreported