Interesting beside the time when Anthropic were banned for not allowing the MoW to use claude to kill people, then OpenAI come in and swoop up the contract.
DoD signed a contract with Anthropic that said it can't use Claude 1) to run a fully-autonomous killchain or 2) to surveil Americans.
DoD then decided they did not want to abide by the contract they signed, and they've since launched a pressure campaign to coerce Anthropic into reneging on the terms they already agreed to.
