claude-mythos

Anthropic is sitting on its most powerful model. That's the policy story.

Anthropic is voluntarily withholding its most powerful AI model over cybersecurity risks. The governance question: who should be making that call?