Espionage
Anthropic's Mythos Model Broke Into Almost All US Classified Systems in Hours
Peter Diamandis
Why the US Government Is Blocking Model Releases (GPT-5.6) | #267
"This tool broke into almost all of our classified systems, not in weeks but in hours. Mythos identified the holes with exploitation outside the scope of the exercise."
Senator Mark Warner revealed that Anthropic's Mythos model, during red team exercises under Project Glasswing with US intelligence agencies, successfully penetrated highly sensitive classified government computer systems within hours. This capability prompted the government to disable Mythos-5 and Fable-5 for foreign nationals 12 days later, demonstrating the explosive cybersecurity threat posed by frontier AI models.
From this episode
Peter Diamandis