I’d love more axes along which to filter my “feed” content. I think user-facing analytics is one solution for users like me who are willing to work for it: Google Reader is actually a wonderful product in that regard. You can see which feeds you keep up with, what percentage of stories in different categories you’re reading, which days you read the most, etc. Helpful for identifying gaps in what you’re seeing and remindinding yourself of other feeds you’ve neglected.

But for services or moments or whatever where you want to be kind of lazy, perhaps you can also add value by creating “seeded” feeds — think Apple’s Genius playlists instead of shuffle. The “sort by magic” feature on Google Reader is a good example, too. Some old, some new, some relevant, some surprises. Pandora does this too — it replays songs you already said you liked, but mixes in new stuff you haven’t rated, and let’s you exclude both songs and artists on a station-by-station basis.

So, in summary, maybe my answer to the question is: more real-time feeds that give users an optimal mix of things they know they like, and things they don’t know anything about yet. Music might be an interesting industry to draw on for ideas.