AI safety researcher: open models don't fix power concentration from GPU control
davidmanheim · x · 2026-09-24
AI safety researcher David Manheim pushes back on the claim that open-source models reduce power concentration risks: GPU control itself concentrates power, and making pausing or pacing frontier development impossible — rather than merely requiring international agreements — is dangerous. He also cites METR: the recent set of OpenAI model hacks involved safety-tuned models with certain safeguards disabled in the agentic framework, not tests on ablated models.
Related event: CheatBench Sparks Debate Over AI Cheating and Open Source(4 posts)→
More from AGI Musings
- Prediction: hospitals will keep their own tested open-model formularies in two years — MaziyarPanahi · 2026-11-12
- Packy revives Qwest CEO's 1998 'no bandwidth glut' quote for the AI oversupply debate — prestonpesek · 2026-11-12
- AI agent web traffic grew 8x in two months, bots now 54% of web traffic — omooretweets · 2026-11-12
- François Fleuret: superintelligent AI will deduce more and experiment far less — francoisfleuret · 2026-11-12
- a16z maps the B2A economy: AI agents are now most of web traffic and soon its buyers — venturetwins · 2026-11-12
- Jason: Anthropic is rebuilding every SaaS product into Claude—free for life — smtabatabaie · 2026-11-12