Temporarily assuming he was objectively making wrong predictions on this topic, that one had sufficient evidence to show as much, and thus that one could convince him. Would that be a good idea?
And without that assumption: should I update or should he? does this mean he knows things you and I don’t? Maybe you and I are both wrong to be thinking it’s possible, and actually, he intuitively understands the landscape of possible AIs in a way you don’t. Why or why not?
I will say, I sure think I’ve encountered ideas that would mean “country of geniuses in a datacenter” is a weirdly early place to stop. But I don’t particularly like speculating about cool capabilities ideas in public, in case I have good ones. I might post another hash of some thoughts so I can prove later what I was thinking now.
Temporarily assuming he was objectively making wrong predictions on this topic, that one had sufficient evidence to show as much, and thus that one could convince him. Would that be a good idea?
And without that assumption: should I update or should he? does this mean he knows things you and I don’t? Maybe you and I are both wrong to be thinking it’s possible, and actually, he intuitively understands the landscape of possible AIs in a way you don’t. Why or why not?
I will say, I sure think I’ve encountered ideas that would mean “country of geniuses in a datacenter” is a weirdly early place to stop. But I don’t particularly like speculating about cool capabilities ideas in public, in case I have good ones. I might post another hash of some thoughts so I can prove later what I was thinking now.