TechAnthropic Reports Claude Leading 26% of Internal R&D with 30,000 AI Agents
Anthropic disclosed Claude models now lead ~26% of internal AI R&D (up from <1%), with 30,000+ AI agents running internally and >90% of measured R&D involving AI collaboration under human oversight.
TLDR
This signals AI accelerating its own development at a frontier lab, raising both excitement about progress and safety/alignment questions regarding oversight sufficiency. It represents a concrete benchmark of "AI building AI" in production and fuels ongoing debates about whether current safeguards are adequate as systems gain autonomy in critical research functions.
Combined views
2.3K
4 Sources, first seen 20h ago
Anthropic Reports Claude Leading 26% of Internal R&D with 30,000 AI Agents
Anthropic disclosed Claude models now lead ~26% of internal AI R&D (up from <1%), with 30,000+ AI agents running internally and >90% of measured R&D involving AI collaboration under human oversight.