Elon Musk makes a pretty simple point about AI safety: AI companies shouldn’t be the only ones deciding if their own models are safe. Let rival labs test them before launch. They have the expertise to find problems and the incentive to call them out. And if AI is becoming a