Interesting report about use of various forms of recommendations in NYT. The article suggests a likely division into human-edited, friend-recommended and algorithmically recommended stories that neatly captures what Terry Anderson and I have been discussing in terms of groups, networks and collectives. The transition between hierarchical group (the editor decides) to network (your friends suggest) to collective (sets are mined for crowd opinions) mirrors the traditional classroom, the network and the collective intelligence of some Web systems in online learning.