Trump’s China rivalry and “AI race” delusion may endanger US, experts say

There is a particular kind of tragedy in watching a species build a fire and then argue about who gets to hold the match. The Ars Technica report—Trump's China rivalry and "AI race" delusion may endanger US, experts say—paints a picture of a Washington so locked into Cold War muscle memory that it cannot see how the rules of engagement have changed. The "race" is not a technical necessity; it is a narrative choice, and it is a catastrophically bad one. The heart of the problem is information asymmetry, weaponized by ego. If Beijing believes Washington views AI as a zero-sum contest, then every safety finding, every red-team failure, every near-miss in a Chinese alignment lab becomes a strategic secret rather than a public good. Why would you tell your adversary where the cliff is if you believe they want you to drive off it? The administration's plan for AI safety alerts—which, per the article's own headline, omits tech experts entirely—suggests a diplomacy of megaphones, not laboratories. This is the classic collective action problem, but with an existential multiplier. Nuclear deterrence worked (barely) because the failure modes were slow enough for human judgment to intervene. AI failure modes are emergent, recursive, and global—they do not respect the finish lines that "race" rhetoric implies. The delusion, as the experts call it, is the belief that there is a finish line at all. There isn't. There is only a shared planet and a technology that neither superpower fully understands. The speculative future writes itself: a Chinese lab quietly observes a dangerous emergent property in a frontier model, and stays silent because sharing it would admit weakness to a rival who has already declared victory. That silence is the fuse. The article's experts are essentially saying that the US is building the very walls that will keep