Chinese leadership might not be convinced by AI extinction risks. However, they might agree to a slowdown deal if they realize the US compute advantage is insurmountable in the short term. The deal offers them a chance to catch up on algorithms via transparency, avoiding a default loss.

The proliferation of powerful open-weight models from Chinese entities is not just a commercial move. It's a calculated geopolitical strategy to commoditize the AI model layer. By reducing the technological gap and preventing US companies from establishing an unassailable lead, China aims to dilute America's economic dominance in a field potentially worth trillions.

To prevent a reckless race, a proposed solution is a U.S.-China treaty to govern the resources needed for frontier AI. This would involve tracking and monitoring advanced AI chips in data centers and imposing a verifiable cap on the computational power used for any single training run.

A purely cooperative approach to AI arms control with China is unlikely to work due to their inherent skepticism. A more effective realpolitik strategy may be for the U.S. to advance its AI capabilities so far and fast that China feels compelled to negotiate out of self-interest to avoid being hopelessly behind.

The proposal is interpreted by some as a sophisticated geopolitical maneuver akin to Nixon-era nuclear détente. It's seen as a strategic offer to China: the US will pace recursive AI development from a position of strength if China brings its open-weight models under tighter control.

Global consensus on AI safety is unlikely soon. The most practical approach is a bilateral agreement between the US and China, the two dominant players. This can begin with informal 'track two' conversations between their respective AI labs before escalating to a formal government pact.

A pragmatic starting point for U.S.-China AI cooperation is to agree on verifiable red lines for proliferating dangerous dual-use capabilities, such as advanced cyberattack tools. This addresses a mutual security interest and builds the institutional trust and processes needed for more ambitious agreements on superintelligence.

Paradoxically, progress in AI alignment has made a global slowdown agreement impossible. Each side now trusts its own 'aligned' AI more than it trusts the other nation to uphold a deal. The perceived risk of being surpassed by a defector now outweighs the perceived risk of one's own AI going rogue, making a competitive race the rational choice.

A common argument against an AI development pause is that China wouldn't agree. However, since China is currently behind in the AI race, a pause is strategically beneficial for them as it stops the leader from extending their lead. This counterintuitive point suggests a pause is more geopolitically feasible than often assumed.