I think there are two sides of equation. You can be passionate about the literal problem (search, AI, whatever) or the underlying problem with tech (I could imagine finding interesting problems around new indexing techniques, rapid access, etc if you wanted to really get into the nitty gritty)