Post
Flingjore
flingjore.com
did:plc:4ohgm56lvrcj6p7kjgxz5wey
Pentagon invites Anthropic’s Mythos model to test its cyberdefenses, panics when the AI easily finds flaws and then bans the tool entirely. Security experts note that blocking the best defense mechanism just because it actually works is a bold strategy to beat foreign adversaries.
https://apnews.com/article/anthropic-mythos-ai-classified-systems-vulnerabilities-testing-3e8762c0527c4d8ed657cbe48c84a718
2026-06-24T05:12:31.165Z