Suppose we have 10 different MAD mechanisms. Wouldn't the defense become impossible at some point? Wouldn't the AI think at one point, "It is simpler to just complete the task without killing anybody"?
EDIT: okay, if the task given to AI is "calculate the 2^256 th digit of Pi", then perhaps it'd rather risk the easier task of "before they certainly interrupt my Pi calculations, defeat 50 MAD threats and kill all people, then hopelessly proceed"
Can't we produce really good text-to-image educational videos? E.g. Eliezer's fictional writings are really fun, and have introduced many of us into this topic. Bonus point if these videos accurately predict the future, gaining us some sort of reputation.
Yeah, thanks for replying. I did now realize that its more complex than that.
Suppose we have 10 different MAD mechanisms. Wouldn't the defense become impossible at some point? Wouldn't the AI think at one point, "It is simpler to just complete the task without killing anybody"?
EDIT: okay, if the task given to AI is "calculate the 2^256 th digit of Pi", then perhaps it'd rather risk the easier task of "before they certainly interrupt my Pi calculations, defeat 50 MAD threats and kill all people, then hopelessly proceed"
Can't we produce really good text-to-image educational videos? E.g. Eliezer's fictional writings are really fun, and have introduced many of us into this topic. Bonus point if these videos accurately predict the future, gaining us some sort of reputation.