NSA spies are reportedly using Anthropic’s Mythos, despite Pentagon feud

The more significant detail here isn't that the NSA is using Mythos — it's that an intelligence agency apparently moved forward with a model the Pentagon has flagged as a supply-chain risk, suggesting that U.S. government AI governance is not a unified policy but a collection of competing agency-level decisions with no clear arbiter.

This lands in the middle of a week of rapid repositioning. TechCrunch reported on April 18 that Anthropic's relationship with the Trump administration was 'thawing,' and the Mythos Preview release on April 17 was read by The Verge as a deliberate bid to get back into government's good graces. The NSA deployment suggests that courtship was already working at the agency level even before any formal policy reconciliation. Meanwhile, the MIT Technology Review piece from April 16 on the Iran conflict raised unresolved questions about human oversight in active deployments — questions that become sharper when the deploying agency is signals intelligence rather than a battlefield command.

Watch whether the Pentagon formally revises its supply-chain risk designation for Anthropic within the next 60 days. If it does, the NSA deployment will look like deliberate pressure from within the intelligence community; if it doesn't, this fracture between agencies becomes a structural feature of how Anthropic sells into government.