Speaking of RSS readers, I’ve just found (through Scripting News) that there is an Olympics website on twitter (it sucks too). Dave made a River for some Olympics related newsfeeds that show the power of using RSS the right way. If you’re a sports fan take a look at it (http://olympics.newsriver.org/).
The only problem (for companies) of these aggregations, is that many times the mashed up versions are better than those produced in house. Aggregation gives you better and faster information (the lemma shit in shit out still applies). Many companies have tried to give you an easy way to make your own rivers (yahoo pipes comes to mind), but few services lived to the promise.
Many users converted themselves to using a RSS reader but that isn’t the same thing as automated content curation. Others found curator services like scoop.it or pinterest useful, but both aren’t automated. It seams that there’s no middle term (yet) between these two extremes. Something that is automated as in RSS fetching and has the selectivity of manually curated content. Obviously this is a challenge for machine learning researchers. Maybe RSS feeds can be used as input corpora and then some kind of algorithm will be able to do the curator part to produce results for rivers that make it interesting.
Unfortunately I don’t think that we’ve seen many services that are good enough for widespread adoption. This means that we end up relying on personal hacked versions that never get published beyond our Git repositories.