With AI, what's more important for the U.S. — winning or safety?

Everyone is talking about the spectre of AI catastrophe. Can the U.S. and China agree on guardrails to blunt the biggest risks when their leaders meet next week?