Great link! Google has some great research publications as well that I recommened everyone checkout: http://labs.google.com/papers.html

They have an amazing collection of research papers based on their products. Personally, I recommend "Evaluating similarity measures: a large-scale study in the Orkut social network" as a preliminary mathematical introduction to user interfaces.

The problem with collaborative filtering is that there's a lot of magic going on that the user does not understand. Some users like to have more control over personalized recommendations and social networks can give them that control.

That website has some great case studies available on recommender systems available for finding "music", "content", etc. I would much rather prefer that a computer do my work than myself. Web users have a terrible attention span so such a system would work to their benefit by eliminating the need to click-and-find content. For anyone interested, I expressed my ideas on collaborative filtering here: http://www.socialdegree.com/2007/01/15/interview-w-shuzakcom-founder-jawad-shuaib/

From what I can tell a lot of this CF stuff starts with a machine learning algorithm and data about likes/dislikes, making filtering a classification or search task. Is there something that differentiates it from the usual machine learning challenges? Like dealing with users interactively, perhaps?