#AI safety

Anthropic Is Gating the Cyber-Capable Model It Disclosed Last Week

Anthropic launched Project Glasswing, restricting access to Mythos Preview. The "thousands of zero-days" claim deserves careful reading.

Nadia Rao