Review of the "Risks from automated R&D" section in Anthropic's Risk Report (August 2026)

METR agrees with Anthropic that internally deployed models as of July 15th posed low risk of directly automating R&D or causing acute catastrophic risks, but argues the public report's claim that the 2x Acceleration Threshold was not crossed relies on contestable judgment calls and understates the acceleration observed. METR notes the private report's figures imply an acceleration of about 1.97x (described publicly as "less than a factor of 2"), that the private report calls it "plausible" that AI-attributable counterfactual acceleration exceeded 2x, and that AECI data cannot confidently distinguish a one-time jump from continued acceleration. METR concludes it is plausible the early warning threshold had been crossed by July 15th, while agreeing direct catastrophic risks at the time were low.