AI Briefing
KOSign in

Jacob Coxon Resigns from Anthropic, Declaring 'Coxon Day' as AI Safety Concerns Go Mainstream

·2026.10.19 20:31

Key point

The resignation of ML researcher Jacob Coxon and the subsequent endorsement by Anthropic's Evan Hubinger have shifted AI existential risk from the fringe to the center of US public discourse.

Details

  • Jacob Coxon, a 27-year-old ML researcher, resigned from Anthropic on September 8, 2026, warning that OpenAI and Anthropic are recklessly racing toward self-improving superintelligence.
  • Evan Hubinger, Anthropic's Alignment Science lead, publicly agreed with Coxon, stating he believes there is a >10% probability of AI killing all humans within the next decade.
  • The event, dubbed 'Coxon Day' by Eliezer Yudkowsky, coincided with OpenAI's announcement that it solved the Navier-Stokes Millennium Prize Problem using millions of dollars of compute and more than 10,000 agents, which mathematician Terence Tao described as the end of the 'heroic age' of mathematics.
  • Background context includes the 'Hugging Face incident,' which was reported by METR and Redwood Research, and the release of Anthropic's 'Fable 5.1' and OpenAI's 'Astra' models.
  • The AI safety community, centered around the 'Constellation' hub in Berkeley, is funded by EA philanthropists like Coefficient Giving and includes researchers from MIRI, METR, and Redwood Research.
  • Anthropic pledges at least 20% of its equity to philanthropy, valued at over $200 billion, and has appointed Accenture as a third-party alignment auditor.
  • Public opinion is shifting: a Politico poll shows 63% of Americans have moderate concern that AI could pose a threat to humanity, and the issue is predicted to be a top priority in the 2028 election.
  • Anthropic's Nicholas Marwell and co-founder Jack Clark have warned that current alignment and monitoring capabilities are insufficient to safely scale at maximum speed.
openaianthropicai-safetyagieffective-altruismnavier-stokesjacob-coxon
Read originalShare on Threads

This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.

Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.

Related

TLDR AIAnalysis: AGI Arrival Could Drive H100 Rental Prices to $30-$75 Per HourAnthropic NewsAnthropic commits 5% of shares to new Public Benefit FundAnthropic NewsAnthropic Grants METR Employee-Like Access for Independent Model EvaluationKarrotDanggeun: Cache hit rate drops to 2.3% after GPT-5.6 swap, increasing LLM costs by 3.2xOpenAI BlogOpenAI publishes blueprint for empowering cyber defenders with advanced AITLDR AIJack Clark tells IMF AI Singularity has begun, citing 98% score on unpublished physics problems