I probably overindexed on “we should enhance intelligence to solve alignment problem”. In general, yes, improving towards von Neumann level gets you many benefits, just not benefits that I consider enough for problems I care about.
why not select intelligence outliers from the population and give them high personality traits
I think it’s also intuition that you can imagine coherent notion of “yourself but smarter”, even if being smarter changes your personality, but “yourself but with different personality traits” is much tricker and at some point you run into “what is ‘personality trait modification’ vs ‘killing person and replacing it with another’”.
I probably overindexed on “we should enhance intelligence to solve alignment problem”. In general, yes, improving towards von Neumann level gets you many benefits, just not benefits that I consider enough for problems I care about.
I think it’s also intuition that you can imagine coherent notion of “yourself but smarter”, even if being smarter changes your personality, but “yourself but with different personality traits” is much tricker and at some point you run into “what is ‘personality trait modification’ vs ‘killing person and replacing it with another’”.