The funniest part of the U.S.–China AI rivalry is the difference in temperament. Chinese AI companies don't bother packaging themselves as civilizational saviors, and they certainly can't be bothered to trade daily barbs with American firms. That's why the vibe of the U.S.–China AI rivalry is fire and ice. Anthropic publishes threat reports. OpenAI talks about leadership. Washington talks about containment, distillation, security and who is “ahead.” Chinese labs mostly keep their heads down and build. New models. Cheaper inference. Open weights. Domestic chips. Bigger clusters. It was precisely these unassuming Chinese STEM guys who single-handedly closed the gap on America's early-mover advantage in AI. That may be the most dangerous kind of competitor: the one too busy engineering to argue with you.