TechAnthropic releases transparency metrics showing Claude leads 26% of internal R&D, up from <1% earlier in 2026
Anthropic published metrics: Claude leads 26% of R&D, >90% collaboration, ~30k agents with real-time/post-hoc monitoring blocking ~1 in 47k actions, ~6% of R&D compute allocated to safety. No fully autonomous (AL5) work measured.
TLDR
Provides concrete data on AI speeding its own development. Part of broader transparency and pacing push advocated by safety researchers. Contrasts with concurrent hack/autonomy stories, showing rapid capability growth alongside oversight controls. Relevant to recursive self-improvement and supervision debates.
