Irving questions why Anthropic doesn't pause RL training alongside OpenAI

geoffreyirving · x · 2026-08-31

Geoffrey Irving challenges Anthropic in a reply: if the company is genuinely concerned about catastrophic AI risks and race dynamics, why not extend an olive branch and pause RL (reinforcement learning) training alongside OpenAI? This touches on the tension between coordination and racing in AI safety.

Original post →

More from Safety

Safety channel →