Anthropic Called Its AI a Weapon for Months, So the Government Seized It Three Days After Launch

Anthropic spent months telling everyone Mythos was so dangerous it could crack every operating system on earth, locked it behind a vetted-access program, and built its whole IPO-bound brand on being the safety grownup in the room. Then the Commerce Department took the marketing literally and used national-security export controls to order both Fable 5 and Mythos 5 shut off for every customer worldwide, three days after Fable went public, over a jailbreak Anthropic insists is narrow and present in rival models too. Sam Altman, whose own company faces no such order, had already called the routine “fear-based marketing,” the digital equivalent of selling a bomb shelter for the bomb you keep bragging about. Brag that your product is a munition long enough and a government will eventually agree and confiscate it. TechCrunch on Anthropic’s safety warnings backfiring