“Sebastian Malaby believes that AI lab leaders do not trust each other, necessitating a government enforcer to ensure collaboration on safety.”