Thanks for the ideas - it’s a tough problem, and one that’s near and dear. A few comments:

• If you want to get a handle on any information source, I think there are only three things you can do: Get fewer, Get faster, and Get control. For RSS: Subscribe only to the ones that matter (i.e., what Koch calls the vital few, not the trivial many) and cull mercilessly, get very efficient at processing them, and put the source in its place (e.g., don’t check it all the time). More about this from the email perspective in Got the email blues?
• To get fast, use a system. Mine is a GTD-like one - more at Afraid to click? How to efficiently process your RSS feeds. Here’s a benchmark: my average seems to be about 10-20 seconds/post - pretty good, I thought… Still a lot of work, though.
• Mark Hurst’s Bit Literacy has a nice chapter on going on a media diet. His model: Create a media portfolio with two main components: Lineup and Tryouts. Lineup: Those that’ve earned their place as your most valuable sources. (The three types: Stars, Scans, and Targets.) Candidate sources get into the lineup by going through a tryout phase. (Guidelines: Be discerning, be intentional, and be biased toward rejecting.)

The big problem (and this means big opportunity) is for the tool makers (Bloglines, Reader, etc.) to support quantitative measures of value, i.e., to help answer definitively the question “Which sources are most worth reading?” It’s really the only question that matters when there’s so much to read. There are two parts: Trusted recommendations (finding pre-filtered candidates - those worth trying out) and concrete evaluation (just mentioned). The former is a research problem, but collaborative filtering should provide a good start now. (If Gmail can use collaborative SPAM filtering, why can’t…)

The latter should be possible now - the tool would track percentage of a feed’s posts I “investigate,” down to one or two levels (expand body, click URL, etc.) As envisioned, I think this does go beyond Reader’s Trends feature. More at Information provenance - the missing link between attention, RSS feeds, and value-based filtering.

Just my 2c (OK, maybe 3c :-)