At some point I started to split them into technical ideas on the one side and business ideas on the other side. That helps to deal with them in different ways.
The technical ideas are usually stuff that would be helpful to have and that solve particular small problems, but on closer inspection they just turn out to be things you can just as well do with existing tools (usually like excel or other tools). Or they are really just ideas for some kind of library or service that could be a small OSS project. Most of those ideas are fun to play with, but the fun would go away if I looked at them in an economic way.
The business ideas are the ones where I would immediately think about competition, costs, etc. They are usually more fun to research and reason about. And for most things that you could think about that don't exist yet, the reason for that nonexistence is often much more enlightening and interesting than the original idea.