“An effective international agreement on AI safety between the US and China will require hardware verification technology, as mutual trust and goodwill are insufficient.”