> Maybe it's a problem with empathy? Of "putting yourself in someone else's shoes?" Maybe it's not even possible for extremely intelligent, relatively rational

> Maybe it's a problem with empathy? Of "putting yourself in someone else's shoes?" Maybe it's not even possible for extremely intelligent, relatively rational people, to adopt the mindset of people who (by choice, culture, or means well outside their control) live more emotional, instinctual lives. How do people predict other people? The natural way is to imagine yourself being in their situation. For example, if you know that if you hit your head you will be in pain, it is easy to predict that when another person hits their head, they will be in pain. If you know that you would be scared of something, it is easy to predict that other people would be scared of the same thing. The problem with this natural heuristic is that it only works when the other people are sufficiently similar to you in the relevant aspect. When they differ, the strategy stops working. For example, being hugged makes one person feel happy, but another person hates it; this makes it difficult for them to predict each other's reactions. I believe this is what causes big problems to many highly intelligent people. They naturally (when they do not think explicitly about what they are doing) assume that other people are just as highly intelligent. They assume other people would see the "obvious" connections, and would avoid doing "obviously" stupid things; when in fact the "obvious" thing may actually require IQ 150 to notice. So they constantly predict people wrong. Of course, there is also another way of predicting people, by looking from outside. Instead of imagining yourself in their situation, you just use your knowledge about how people in given situation behave. It may be your experience, or prejudice, or something you read in a book; the important thing is that you are NOT imagining yourself, but rather trying to construct the answer in a way that feels impersonal. This other method has a few disadvantages. First, it takes more attention and time, so you may forget to do it, or you may do it too slowly. Second, because it is more complicated, you will probably learn it at later age (so it is specifically difficult to use for highly intelligent kids). Third, there is a cultural taboo against using it; it will be called manipulative. (There are some good reasons for this taboo in general, but sometimes it just hurts some innocent people.) Fourth, to use this method, you need to have the data. This was one of the most important lessons in my life that took me too long to learn. Mostly because the society was always explicitly and implicitly telling me to do the opposite: to use the natural strategy, and avoid explicit modeling. This may work for an average person, for whom the natural strategy mostly provides good results. In my case, when I treated people the way I wanted to be treated, it mostly pissed them off. (For example, if I am wrong about something, I prefer when other people tell me about it. Coincidentally, telling other people that they are wrong, especially when they actually are wrong, is a very effective way to make enemies.) In contrast, using even very simplistic and very cynical models provided much better results. (Tell people what they want to hear, which mostly means that they are the best and perfect, and they will classify you as a wise and good person.) I feel bad about using this method, but it works, and it makes the other people happy. There are also a few people with whom the natural method works for me; but they are very rare.