Overview
- The NSA, which sources told Axios on Sunday is using Anthropic’s Mythos Preview, is proceeding even though the Department of Defense labeled the company a supply‑chain risk and sought to cut off its tools.
- Anthropic says Mythos can find long‑hidden software flaws and link them into multi‑step attack chains across major operating systems and web browsers, so the company is keeping it behind a gate for roughly 40 vetted partners under Project Glasswing.
- Most organizations with access are using Mythos to scan their own networks for exploitable weaknesses before attackers can, and the U.K.’s AI Security Institute confirmed it has access for testing.
- Anthropic CEO Dario Amodei met White House chief of staff Susie Wiles and Treasury Secretary Scott Bessent on Friday to discuss how agencies outside the Pentagon could use the model under supervision.
- The feud grew from Pentagon demands to allow use of Anthropic’s models for “all lawful purposes,” including areas the company refuses such as mass domestic surveillance and autonomous weapons, and it now sits in court after a judge granted a preliminary injunction blocking parts of the federal stop‑use effort.