All of this strikes me as sufficient evidence to conclude that there is a good chance (P ~ 60%?) that he is actually a sociopath, in which case learning to like him is exactly the wrong thing to do.
If this person is not “actually a sociopath”, would learning to like him be the right thing to do?
Yes, if he’s not actually a sociopath, it’s probably worth learning to like him.
But the odds of him being a sociopath are high enough that the expected utility doesn’t point that way at all. The disutility of being exploited by a sociopath is far worse than the opportunity cost of not liking this one person.
It sounds like the reason you’d want to not like him if he’s a sociopath is that then he’d probably exploit you—but don’t you already know that he’ll exploit you anyway?
If this person is not “actually a sociopath”, would learning to like him be the right thing to do?
Yes, if he’s not actually a sociopath, it’s probably worth learning to like him.
But the odds of him being a sociopath are high enough that the expected utility doesn’t point that way at all. The disutility of being exploited by a sociopath is far worse than the opportunity cost of not liking this one person.
It sounds like the reason you’d want to not like him if he’s a sociopath is that then he’d probably exploit you—but don’t you already know that he’ll exploit you anyway?