Over 4 years back, Walmart ended its movie suggestion product, when its algorithm suggested "Planet of the Apes" to folks who bought "Martin Luther King" DVD.
Now, Netflix is reporting that some of its users have rated over 50,000 movies. Even if a person, watched a movie per day, it would take 136 years to watch 50,000 movies.
Also, digg is facing problems sustaining due to "power diggers".
All this indicate, that relevance streams generated by machines, would also follow, the Forer effect.
Suggestions, would appear to be right on, at the very beginning and with time, it would become totally useless.
Connected Consumption with friends do not scale up, as we experience with Facebook. Pool of people, is such a bother to maintain, due to the low signal-to-noise ratio.
Expert curators are too narrow in their domains. The solution ought to be a mashup of all these : Algorithm, Curator Pool, Social Recommendation, Expert Recommendation with the right proportion.
