Anthropic discloses 6% of AI R&D compute goes to safety, publishes three transparency metrics
gaganghotra_ · x · 2026-09-18
Anthropic published three transparency metrics meant to let the public track frontier AI development, arguing any frontier lab could publish the same measures and third parties could verify them.
- The metrics: how much AI R&D is done by AI, how well AI agents are overseen, and how compute is allocated.
- Per the reports, Anthropic is spending roughly 6% of its AI R&D compute on safety, and advocates labs allocate more compute to safety as the world considers pacing the frontier.
- Anthropic frames the effort as minimizing the gap between what frontier labs know and what the public knows as AI increasingly builds its next version.
More from Companies & People
- Jensen Huang counters AI doomerism: agents going wild means the products just aren't ready yet — prateekj · 2026-09-18
- Enterprise AI's hardest problem: capturing organizational context, not just documents — blaizedsouza · 2026-09-18
- Google DeepMind launches institute to widen the public AGI debate — TechCrunch AI · 2026-09-18
- Meta's Moat Is Distribution: 'Good Enough' Beats Benchmarks — kimmonismus · 2026-09-18
- Every company now has an 'AGI support group' on Slack — shauseth · 2026-09-18
- Stephen Wolfram announces return to livestreaming with personal update and AMA — TOEwithCurt · 2026-09-18