"If AI labs are to consistently invoke exaggerated fears through comparisons to nuclear hazards, then they must be willing to take the analogy to completion. Ironically, if they were to explore the readily available safety mechanisms for nuclear components,…

→ View original post on X — @timnitgebru