Julian Bradshaw comments on DeepMind: Model evaluation for extreme risks

[Julian Bradshaw]

It’s nice to see OpenAI, Anthropic, and DeepMind collaborating on a paper like this.