Open letter to Altman and Amodei: three years of safety pledges, no slowdown on the frontier
AryHHAry · x · 2026-09-14
A long open letter to Dario Amodei and Sam Altman questions the credibility of frontier labs calling to slow down.
The author recounts the record: Altman's 2023 Senate warning, Bletchley pledges, Anthropic's first Responsible Scaling Policy — while releases, hiring and compute races continued; the OpenAI board crisis, Ilya's departure, Superalignment's collapse, and Jan Leike's 'safety culture took a back seat' line. Anthropic is not exempt: the RSP was later relaxed, the author claims Claude Fable 5.1 and Mythos 5.1 shipped in September 2026, Jacob Coxon resigned days before Amodei's pacing appeal, and Evan Hubinger says Anthropic lacks a clear path to superintelligence alignment. Conclusion: the risks are real and the pace should fall, but 'your incentives are mixed, and an essay is not leadership.'
More from AGI Musings
- After Altman, Amodei says rival AI labs may agree to slow frontier AI development — rohanpaul_ai · 2026-09-14
- AI researcher lays out the safety paradox: pausing algorithms while compute piles up maximizes risk — RichmanRonald · 2026-09-14
- AI bio-risk debate: if AI can design a super-virus, another AI can design the vaccine — RichmanRonald · 2026-09-14
- AI safety commentator says intelligence can't be contained, urges hardening critical systems — gandamu_ml · 2026-09-14
- Blogger argues Altman's 'international coordination' push targets China's open-source models — ayushtweetshere · 2026-09-14
- AI doom discourse under fire: critic says fear-mongering harms more than it helps — Kyrannio · 2026-09-14