This kind of post symbolizes a lot of what seems wrong to me about LessWrong. Women are attracted to men who they enjoy spending time with? Fashion matters to a lot of women? Women prefer confident men? It amazes me that many extremely intelligent people are unable to make predictions that could be made by the average truck driver. It indicates, I think, that what is lacking in those people is not analytical intelligence. Because of this, I'm deeply sceptical as to what extent applying rationality techniques such as those taught on LessWrong to social interactions will really improve peoples results.

I'm a very analytical thinker; I excel at math, physics and related subjects. At the same time, I have quite poor social skills. I'd love it if I could read some social psychology books and improve my social and romantic outcomes, but I'm unconvinced, both by this post and the community as a whole. In particular, I think that there is what I'll term the 'rationalists fallacy' -- the hidden assumption, in much of thought by rationalists, that other people and the world in general are supposed to behave rationally. They're not, and by and large they don't. So, to make sense of this, rationalists read evolutionary and social psychology books. Now, armed with these explanations for the seemingly irrational behaviour of the humans around them, they're able to finally understand the cognitive biases and preferences of their fellow human beings, and are then able to successfully interact with them.

This behaviour reminds me of my behaviour in mathematics. When I have not seen a particular result proved and do not intuitively see why it must be true, I'm uncomfortable with using the result. Later, after I've seen it proved, I become comfortable using the result, even if I later forget the proof, and have not gained any insight by reading the proof.

I think this behaviour may be adaptive in mathematics - I've read more proofs than I otherwise would have done - but I think this behaviour is almost certainly maladaptive in social interaction. I don't need to understand why doing things weird will make me unpopular (it signals that I do not subscribe to the group norms), I just need to know that they do, and so only do them in private. I don't need to understand why women tend to be more attracted to confident men (confidence was a strong signal of fitness in the EEA), and men more attracted to pretty women (body symmetry indicates that the person is healthy, and does not have any significant parasitical infection); I just need to know that this is the case, and try and be more confident or more pretty according to gender. Whilst understanding that these preferences are primarily non-concious and so do not necessarily reflect peoples concious preferences is useful, not even that needs evolutionary psychology. One can use the result without proof.

I don't think I'd object to this behaviour so much if I could see real insights that people are gaining from learning more about evolutionary and social psychology -- but I don't. A lot of evolutionary psychology seems to be dangerously close to a just-so story, and there are a large number of conflicting evolutionary psychological explanations for many common human behaviours. If learning something doesn't help me predict anything I don't already know, just provide a (possibly false) explanation for a behaviour I'm already familiar with, why should I learn it?

Having said all this, I'm optimistic about what rationalists can do socially. There are few areas where systematic study as to what works and what doesn't doesn't help; it does. But, I think reading social psychology books should be a very small part of what rationalists do. A very large amount of the social behaviour we engage in is non-concious; being around people enough that you become comfortable with them may be one of the best things one can do to improve ones social skills. Furthermore, do not overestimate what one is able to achieve using deductive logic. Even as a mathematician and as a programmer, most of what I do is intuitive and involves pattern matching. One cannot learn to become a good programmer or a mathematician without writing lots of code, or working on lots of problems; likewise, one cannot become effective socially without meeting people.

tl;dr I see rationalists often engaging in behaviour adaptive in many fields (e.g. math, science), but maladaptive in other fields (e.g. socializing, romance.) I think less emphasis should be placed on attempting to find a deeper meaning behind peoples behaviour, and more on trying to find ways to benefit from peoples existing behaviour. Practice makes perfect in all fields, but especially social.