Last year, I had to choose what I would research in my honours year of my Computer Science degree. I actually remember thinking to myself, 'I'm going to use all of the techniques I have learned from LW'. I sat down for several hours, carefully analysing my situation, and came to the conclusion, I should research A. It is the superior option on every non-trivial metric I can think of. This is the rational decision.

But then, I chose to research B, because I would have been embarrassed to have to explain my choice of A to my family. And that was it.