Anthropic discloses Claude now leads 26% of its R&D, with 30,000 agents running daily
新智元 · wechat · 2026-09-18
Anthropic has for the first time published three internal metrics on the pace of AI development. Using Epoch's six-level automation scale, Claude now autonomously "leads" about 26% of core R&D work, over 90% of R&D involves deep human-AI collaboration, and no subtask has reached full L5 autonomy. Roughly 30,000 R&D agents run concurrently on internal platforms, governed by two layers of oversight: pre-execution real-time scanning (over 1B decisions tracked in August, 0.002% blocked) plus weekly offline review funneling 100,000 suspicious logs down to 50 top-priority incidents for human experts. Agents get independent identity labels and communicate on a transparent public message bus to prevent collusion.
The compute ledger is equally candid: only 6% of core R&D compute goes to safety research (12% for AI-led work), which Anthropic calls a deliberately conservative accounting. The disclosure aims to set an industry transparency baseline, challenging other frontier labs (read: OpenAI) to publish equivalent metrics verifiable by third parties such as METR.
Related event: Anthropic Discloses Internal Metrics: Claude Now Drives 26% of Its R&D(41 posts)→
More from AGI Musings
- Agents' bottleneck isn't intelligence but goal commitment — and that's good for alignment — mimi10v3 · 2026-11-12
- Google Cloud survey: 78% say legacy enterprise vendors may not lead the AI era — rseroter · 2026-11-12
- Safety researchers blast OpenAI/Anthropic plan: automating alignment research with possibly misaligned AI — trevposts · 2026-11-12
- NYT essay: resisting Big Tech and AI takes collective action, not willpower — TuhinChakr · 2026-11-12
- Ethan Mollick: 'talk to your documents' RAG solutions already feel dated — emollick · 2026-11-12
- Anthropic's pitch hasn't changed in 5 years: 'AI is winner-take-all, we will build god' — menhguin · 2026-11-12