Casey Muratori
I don't follow this explanation. Could someone at one of these labs complete the thought publicly, so I can understand exactly what is being proposed? What difference would it make if your company "gets there first"? What specifically are you going to instruct your first-mover human-extinction-capable ASI to do that will prevent competitors from also getting to that same point shortly afterward, and therefore still having the human-extinction-capable ASI you claim you must prevent them from getting? In other words, if the argument for Anthropic building ASI is that it's too dangerous to allow OpenAI to build ASI, what specific steps would Anthropic take using their ASI to prevent OpenAI from achieving ASI sometime thereafter?
Jacob Coxon
A common response is “if they truly believe this, why are they still building it?” At OpenAI, many have not deeply internalized the civilizational stakes. At Anthropic, the stakes are well-understood, but they are locked in a race to get there first - they believe no one else will act responsibly, so they must do it themselves, despite the risk.