METR audit: Anthropic may have crossed its own AI acceleration threshold in July
The Decoder · rss · 2026-11-06
METR spent 11 weeks auditing Anthropic's internal research data and concluded model development pace has roughly doubled, 80% of the speedup driven by AI-assisted research — possibly crossing Anthropic's own warned-about acceleration threshold as early as July. Key evidence was redacted, forcing the public to lean heavily on METR's judgment, with skeptics inside the company.
More from Companies & People
- Asked in a 2021 VC interview to pick one company, he chose OpenAI at $10B — curious_vii · 2026-11-06
- NUS professor says Singapore's 30% nationality quota hurts PhD recruiting — xuanalogue · 2026-11-06
- NUS professor shares workarounds for Singapore's PhD nationality quota — xuanalogue · 2026-11-06
- AI's Biggest Cultural Risk: Turning Employees Into the Enemy — DavidLinthicum · 2026-11-06
- TypeSafe spent 2 years in stealth on Jev; open ecosystem copied it in under a week — helloiamleonie · 2026-11-06
- You can't copy Jev: new architecture plus RLCD training method — helloiamleonie · 2026-11-06