Anthropic Safety Report Contradicted by Claude: Understaffed Safety Teams a Growing Concern
Miles_Brundage · x · 2026-08-15
After Anthropic released a safety report, its own model Claude was given additional private information to review and disagreed with the company's decision to completely redact an incident, calling it 'among the most genuinely informative.' Former OpenAI safety lead Miles Brundage commented that Anthropic employees privately agree but are too busy to share details, highlighting understaffed safety teams.
Related event: Claude Pushes Back on Anthropic's Safety Report Edits(3 posts)→
More from Companies & People
- New England Mech Interpretability workshop kicks off — benno_krojer · 2026-08-15
- Ethan Perez Joins Anthropic Alignment Team — EthanJPerez · 2026-08-15
- Apple partners with Alibaba for China AI model, creating a split architecture — AccBalanced · 2026-08-15
- Claim: Anthropic Revenue Hits 2-3x OpenAI's — JosephJacks_ · 2026-08-15
- Critics Question Anthropic's Transparency: Anti-Open Source Stance May Damage Reputation — SkyLi0n · 2026-08-15
- Weights Unknown Conference Focuses on Scaling and Agents — DynamicWebPaige · 2026-08-15