Anthropic cofounder: policy proposals that skip lab compute realities fall flat

_aidan_clark_ · x · 2026-09-09

Anthropic cofounder Aidan Clark says the world lacks a good mental model of how AI labs' compute and projects work, arguing 'reach-into-the-box' interventions usually fall flat. Snarkily, he adds that proposals not mentioning Anthropic and OpenAI often aren't at the right level of detail.

Related event: Ex-OpenAI researcher Aidan Clark questions whether AI is moving too fast(7 posts)→

Original post →

More from Safety

Safety channel →