Eli Pariser: Beware online "filter bubbles" (TED talk) (youtube.com)

I was thinking about this for a long time. There are few discovery initiatives and many "retweets" of the same information. Probably collaborative filtering/recommendation engines tries to give you a recommendation with a high match probability (I know there are research in other directions, but the industry wants to sell stuff based on recommendation so the industry is a heavy weight).

Intuitively I've seen this as a Atractive Fixed Point ( http://en.wikipedia.org/wiki/Fixed_point_(mathematics)#Attra... ) where applying filters to filters finishes in a fixed point.