It seems to me that a simpler theory than Hanson's would simply be that integrating all of our neural sub-systems in a fully coherent way is hard. We know that a lot of compartmentalization probably happens simply because actively seeking out all connections between everything would be too time-consuming, and near/far seems like a special case of compartmentalization. Such integration is also not necessarily very useful in an evolutionary sense.
It would also be plausible to assume that far-mode systems, often dealing with more abstract issues, are a relatively recent development and there hasn't been much time to completely integrate them. I'm not convinced that "near/far is for hypocrisy" is a hypothesis that's been supported adequately enough to outweigh the complexity penalty it gets when compared to "coherence is hard".
Of course, "which of these hypotheses is the correct one" may be a somewhat confused question. It could very well be that near/far started out as separate systems simply because coherence is hard, and then hypocrisy being useful worked to counteract selective pressure that would've integrated them in a more complete manner. There may also have been other functions. See the debate on what did language/bat echolocation evolve for.
The terms Near and Far are to be taken in the context of Robin Hanson's Near/Far articles.
I was reading a fairly convincing article linked from a comment here about how theistic beliefs are so scantly supported, when not outright contradictory, that it's a doubtful whether anyone truly holds them at all. Of course there is a whole battery of explanations around the self-deception, signalling and belief-in-belief cluster, but the question that got in my head was about the kinds of people that can or cannot profess to hold these beliefs.
A common thread in many a 'deconversion' story is that some inconsistency in a person's worldview comes to their attention, and they can't let go until they have undone the whole fabric of their belief system. But given that most people are happy living productive lives while simultaneously nominally carrying around massively conflicted worldviews, what is it that makes certain individuals not capable of this fairly common human feat?
So the hypothesis that I'm considering is that the people who came to atheism this way, are those who demand detailed consistency of their Far ideals. Alternatively, they could be those for who what is normally considered Far is actually Near, in other words those with an unusually high Buxton Index. Combining the two, perhaps for people with a high Buxton Index, Far simply evaporates, as it comes under the scope of things that are relevant to a person's planning. (Edsger W. Djikstra, when introducing the Buxton Index, says that "true christians" have a Buxton Index of infinity. I think that couldn't be more wrong. Perhaps it is the case for singularitarians though.)
The obvious reason to be suspicious of this idea is that it's very flattering for those that fall in this category, which includes myself. Rather than dithering about it, I'd rather expose it to the community and see if it seems to have legs in the eyes of others.