One of the downsides of so many people publishing so much of content (Blogs, Twitter, Microblogs, Flickr etc) is that keeping your head above all the noise has become an impossible task. The situation is so bad that often times the only recourse available is to shut off most of these channels. And this is a problem that will only get worse as the number of people who can produce content online is only going to keep growing.
Most of our current filtering mechanisms like aggregators and ranking systems work by applying the preferences of the many to the individual. This does allow for a dynamic data set compared to the static ones if you were to use the individual's sources as the data set being operated up on. But it also results in the dilution of context, leading to massive generalization and approximation.
The main reason why personal context is a non-starter in most cases is cost. It is simply way too expensive to index and group content on a per-user basis. One way to work around this is to offload the processing and storage to the end user, like what Mailbrowser has done. I have been an advocate of a hybrid model based on this for a while, but, unfortunately, I don't have the technical chops or the required infrastructure to build it.
This method also goes very much against the “cloud” school of thought. So you can't expect any of the existing big players to make much inroads into using this approach. But I do believe that we are reaching the limits of what a generalized context can provide. If you sit out the noise for a week, you will realize that there is little that you have missed and that the ROI on processing information churned out by existing filters is near zero.