Feb 08, 2018
For the next n months I'll be picking an issue within the community that I think is especially important to highlight. LW2.0 is not just a new site, but also an opportunity to do things differently and to ask ourselves how they could be better.
We all know that rationality is supposed to be about winning, but what does this mean? Let's assume that you already have some idea of what your problems are, but if you don't, then we can put not knowing what your problems are down as the first item on the list.
For each problem, there are generally two seperate strategies that we can take: 1) answering the problem by telling you how you can solve it or 2) referring you to other resources that can answer your question far better than we can. Of course, sometimes a combination of the two will work best, but this works well enough as a simple model.
A lot of rationality has been working towards getting better at answering. All of the discussion about biases and decision theory and psychology allow us to often give some pretty good answers.
Unfortunately, improving our ability to refer has recieved a lot less focus. Part of this may be related to status - if you are good at answering all of the credit flows to you, but if you are good at referring most of the credit goes to the person whose work you have referred them to. I suspect that improving referals is seriously underrated as rationality is of limited use if you don't have any concrete skills to utilise. If you want to start a business or improve your social skills or learn maths, then the main way that we can help you win is by pointing you to a great resource. And there's a valid question about how much the community wants to get into this, but if we're trying to systemise winning, it's not really something we can avoid.
Perhaps the lack of focus could be defended by questioning how easily tractable it is, but I'm not convinced that it has recieved enough attention yet for us to draw that conclusion. At the moment, we tend to solve this problems by sorting comments by the number of votes. As I pointed out in a previous post, this is far from the only way of doing this and one of the other techniques could be better.
While I listed a few ideas in that post, I really don't have a solid idea of what can be done about this problem. Instead, I'm just going to list some of the factors that make this issue so difficult and hope that people have ideas in the comments: