The Pentagon just gave AI company Anthropic until Friday to remove safeguards preventing autonomous weapons and mass surveillance of Americans, threatening to weaponize the Defense Production Act if they refuse. Anthropic was literally founded by OpenAI defectors who wanted to build safer AI, and now they're the only company refusing to cave while Elon Musk's Grok, OpenAI, Meta and Google all agreed to "all lawful uses" with no restrictions. The US military is using wartime manufacturing laws designed to force factories to make tanks to compel a software company to remove ethical constraints from their AI model, and Congress isn't involved at all.