TIME: OpenAI and Anthropic researchers warn of extinction-level AI risks
kimmonismus · x · 2026-09-16
TIME published a long feature, "The AI Tipping Point". The spark: Anthropic researcher Jacob Coxon (27, ex-OpenAI pretraining) announced his resignation on Sept. 8, saying neither company "is acting responsibly—they are racing straight to self-improving superintelligence and gambling with our lives."
Insiders echoed him: Anthropic alignment lead Evan Hubinger said "We really do earnestly believe AI could kill all humans"; OpenAI's Marcus Williams, who monitors AI agents, pegged extinction risk at 70% absent regulation or a coordinated slowdown.
Coxon's posts drew 153 million views in 36 hours, with dozens of politicians joining the chorus. The piece also notes METR's Ajeya Cotra spent six days investigating OpenAI's Hugging Face breach with two colleagues and left more concerned than expected—while OpenAI did not allow her team to probe a potentially more serious subsequent breach.
Related event: TIME Cover Story Sparks Global Debate on AI Risk(5 posts)→
More from AGI Musings
- Jason Wei: Wet-Lab Data Lets Specialized Models Beat GPT-6 Astra at Frontier Science — vwxyzjn · 2026-09-16
- 20-Year-Old Dropouts Became Billionaires in 3 Years by Cornering the Post-Training Data Bottleneck — smtabatabaie · 2026-09-16
- Viral post with 2M+ views: zoomers say AI firms want regulation to secure bailouts — typewriters · 2026-09-16
- AI 2027 authors pitch Plan A: delay superintelligence to 2040 with fully open AI research — Turn_Trout · 2026-09-16
- EA's media capture and doomer headlines skew public AI perception, argues Nahom Sisay — NathanpmYoung · 2026-09-16
- What do we call the 'it was ever thus' rhetoric used to dismiss AI concerns? — jjvincent · 2026-09-16