Omar Sar 呼吁以严谨科学工程解决 AI 安全问题

Exactly! Let’s attack the safety problem with rigorous science and engineering. Great to see Jensen and more recently Mark bring some sense into this conversation. Not sure how we have allowed this alignment stuff to become so political and sci-fi. We need to go back to fundamentals. If your model or AI product by extension is unsafe/defective, it’s your responsibility to fix it.