[retsibsi]

I think this is mostly true for me, but I wonder if the best approach involves a little bit of the 'semi-random' too. The internet is an amazing way to discover books and authors, and I feel like I'm exposed to a pretty eclectic mix, but I assume my discovery process is biased in some unhelpful ways. The payoff of discovering an entirely new (to me) and excellent author can be pretty great, perhaps even more so if their work is completely different from most of the stuff I read -- so maybe a bit of unguided exploration in unfamiliar territory is worthwhile, even if it's mostly unproductive? (Either way, I bet someone has analogised this process to some well-studied class of CS problem, and drawn interesting but practically dubious conclusions!)