There is a tendency to downvote articles and commentaries with a political subtext with a remark on how politics is the mind-killer. I completely understand that nobody wants his mind to be killed, however, I disagree on the employed methods. I don't think anybody can really afford to ignore politics. It's a fact about any group of even a handful of people. Thus instead of shunning politics I think it's better to build one's rational defenses. Understanding that politics is a problem is only the first step. If you stop there, there will always be a big part of life where you are not rational. Therefore I suggest that, as long as it doesn't get out of hands, there should always be room for political discussions if not on the main site at least in the discussion section.
Politics is difficult to treat empirically. Let me take the risk and mention a potentially sensitive political question now to illustrate the problems.
Should we stay in Afghanistan after 2015? (By we, I mean all countries that have some units there, and by stay I mean continued non-negligible military presence. The date 2015 was chosen arbitrarily - I suppose the complete withdrawal cannot realistically be done much sooner. I am aware that a non-negligible part of LessWrong audience comes from countries not engaged in the war, but I wanted to leave the question brief.)
It is a sensible question: important (it influences relations between Islamic nations and the West), potentially interesting for all (contrary to, say, gay marriage laws in California, which would be primarily interesting to Californians, while others unfamiliar with Californian politics would have difficulties to understand the details), morally relevant (number of killed people depends on the decision), concrete (in contrast to, say, "is libertarianism good?"), and so on.
Now how shall we determine the answer? Even if the question is pretty concrete, there is vast number of things we should consider. First, the answer depends on values: how much do we care about lifes of the Afghanis? How much do we want them to be free to choose their form of government, even if the most likely choice is theocracy? Do we want to preserve some deontological principles, like sovereignty of countries and prohibition of war, even if it prevents us from maximising utilities? Of course, we could tackle these questions separately at first, and after finding agreement come back to the overall discussions of Afghanistan. But I don't expect finding agreement on the particular questions. We haven't yet agreed upon the Trolley problem, where we are free to make up all convenient simplifications.
Then, we should engage in prediction. What will be the effects of the continued presence in Afghanistan? What will be the consequences of a withdrawal? Future predictions are hard, require specialised knowledge, and worse, exactly what knowledge is relevant may be disputed. I don't expect finding an agreement here either.
Would it even be better if we instead discussed "should Brutus et al. have killed Caesar?" There is hardly any mind-killing potency left in this question, but finding an agreement is almost equally unlikely.
The obvious mind killing, i.e. biases creating strong emotions which lead people to support parties rather than seek truth is only one side of the problem with politics. The other side is that the discussion is drown in the immense sea of differing intuitions, hard to classify evidence, confounding variables and lack of experimental data.
Some political questions can be possibly reduced into solvable elements, but these reduced parts would probably no more look like politics. Assembling them together back to solve the original political problem would then be a hard work which will not fit into a blog post. The lack of political posts here may be result of the LW's no-politics policy, but also it may be result of our ability to recognise that those questions are far too much complicated to be solved here. There are no posts about string theory on LW as well.