Great piece. One small push back: of course those with p(doom) near zero or one don’t discuss updating as much because it doesn’t change what is actionable. Maybe a doomer near 0.5 would be a good example. Hinton?
I have no idea what my pdoom is other than nonzero,and the error bars are huge. But yeah it goes up with every new model that's developed and released, closed and open weight.
Yudkowsky is a science fiction author disguised as a "serious" thinker.
Great piece. One small push back: of course those with p(doom) near zero or one don’t discuss updating as much because it doesn’t change what is actionable. Maybe a doomer near 0.5 would be a good example. Hinton?
I have no idea what my pdoom is other than nonzero,and the error bars are huge. But yeah it goes up with every new model that's developed and released, closed and open weight.