Ex-Anthropic researcher Jacob Coxon quits, warns AI could kill us all by end of decade

ChuckDBrooks · x · 2026-09-10

A New York Post feature traces the escalating doomsday warnings around AI in 2026, noting that even Sam Altman and Dario Amodei have raised alarms.

The news hook: Jacob Coxon, a researcher who has worked at both OpenAI and Anthropic, says he quit Anthropic to warn that "AI can kill us all by the end of the decade," describing soon-to-arrive superhuman systems capable of hacking anything, revolutionizing any field overnight, and acquiring real power and resources.

The piece then walks through how a rogue AI scenario could unfold, starting with infrastructure: decades of automation mean critical systems are now software-controlled, making them prime targets for a misaligned superintelligence. The article is speculative scenario-building rather than technical analysis.

Related event: Anthropic Researcher Quits Over Safety Fears as Alignment Lead Puts Extinction Risk Above 10%(121 posts)→

Original post →

More from AGI Musings

AGI Musings channel →