TLDR Tech / InfoSec
Anthropic's Mythos Model Found Vulnerabilities in Classified US Government Systems, Official Says
A US official stated that Anthropic's Mythos model, tested with intelligence agencies under Project Glasswing, rapidly uncovered vulnerabilities in classified systems, although it did not exploit them. The Trump administration responded by imposing directives limiting access to Anthropic's latest models, including export controls and restrictions for foreign users, leading to the company taking models offline.
Read the full story