TechAnthropic publishes AI-led R&D metrics showing Claude involvement grew to 26% by August 2026
Anthropic shared transparency metrics showing AI involvement in its own development accelerated—Claude led ~26% of AI R&D work as of August (up from ~1% in March), with ~30,000 agents actively researching and over a billion agent decisions reviewed monthly.
TLDR
The data provides rare public metrics on how close leading labs are to AI systems accelerating their own improvement and recursive self-improvement (RSI). This sparks discussion on RSI timelines, oversight effectiveness, and whether progress justifies or contradicts slowdown proposals. Positions Anthropic as more transparent on capabilities and risks.