Anthropic cofounder responds to METR report as critic warns against recursive self-improvement next year
austinc3301 · x · 2026-11-06
Anthropic cofounder Sam McCandlish published a response to METR's independent evaluation of the company's model capability growth, calling capability velocity one of the most important AI safety questions and previewing more safety disclosures in the coming months.
Reposter Max Nadeau argues Anthropic's redactions drew unfair criticism since the public lacks context to judge the tradeoff — but more importantly, the METR report suggests Anthropic could let its AI recursively self-improve past human level next year, which he calls unconscionable and a decision too consequential to leave to one company.
More from AGI Musings
- Hedge drift in AI progress debates: METR's caveats vanish as AECI scores hit 1.9x — sebkrier · 2026-11-06
- O'Reilly: Software's Core Artifact Is Shifting From Code to Prompts — luisdans · 2026-11-06
- RSI isn't about acceleration—it's removing humans as the bottleneck — menhguin · 2026-11-06
- Stop Measuring AI in Gigawatts and GPUs, Architect David Linthicum Argues — DavidLinthicum · 2026-11-06
- Forecast model: 2x progress rate in Dec 2026 implies ~4.5x capability uplift — arthurcolle · 2026-11-06
- Balaji taunts Anthropic and OpenAI: you fear-mongered, now find out — beffjezos · 2026-11-06