Here’s one: orthogonality thesis. People who disagree with you are not just dumber versions of yourself, and making them smarter without addressing their hostile values just gives those values more effective people to attack you with.
Setting aside concerns about the orthogonality thesis as applied to artificial agents, do you really think it’s reasonable to bring it to bear when the agents in question are all still basically structured by the same evolutionary and cultural pressures and the intelligence in question is nowhere near the “you can achieve any goal you could possibly pursue” threshold?
I think extant neurotypes and values provide enough examples of goal-diversity that humanity itself is, to me, the most persuasive argument in favor of the orthogonality thesis. (Though, it’s negation was never persuasive to me to begin with.)
Considering how heavily loaded crime/etc are towards low IQ, I think there’s a significant gain to be made from using genetic engineering to bring up the low end.
