If Anthropic has really developed an LLM that can suss out security weaknesses better than any other AI, the US government would be foolish to continue shunning them. Oh, right, it is foolish.