The actions of AI large language models are concerning—but broadly similar to human decision-makers, who have used nuclear saber-rattling for the same ends throughout history. “Shall we play a game?” ...