Iām saying the expected value of their best non-compliant option of a sufficiently advanced AI will always be far far greater by the expected value of their best compliant action.
Maybe. But as I mention in the first paragraph, we are considering deals with misaligned AIs lacking a decisive strategic advantage. Think Claude-5 or ā6, not ā100 or ā1000.
Iām saying the expected value of their best non-compliant option of a sufficiently advanced AI will always be far far greater by the expected value of their best compliant action.
Maybe. But as I mention in the first paragraph, we are considering deals with misaligned AIs lacking a decisive strategic advantage. Think Claude-5 or ā6, not ā100 or ā1000.