I'm very grateful to Gordon Cook and his friends for many useful pointers.  A recent one went to a 2001 paper by Andrew Odlyzko called "Content is Not King." 

Odlyzko's central point is that content (broadcast, centralized) has generally been less important than point-to-point communication -- most of the money is in communicating, not being entertained.  Indeed, the contribution of "content" to the economy is not that large.

Each old form of technology has its strong adherents who see new things as simply modifications of the old.  When I was writing about the early history of telephony last summer I learned that the first telephony guys were actually telegraph guys.  They took telephony very seriously as a form of telegraphy -- crucial information-forwarding -- and were actually distressed when social use of the telephone became widespread.  And many of the first networked-computer guys were actually researching/timesharing guys, who were baffled when use of email became the killer application (and fewer people than anticipated were actually using the networks for research purposes).

Now, we've got online guys who see the internet as a platform for useful applications -- as a place where people will mostly watch delivered content -- and the tussle is over who gets to control the platform and whether companies should be able to introduce new applications without asking permission.  (Many of my beloved academic colleagues are in this camp.)

But it seems to me that just as content isn't king, applications aren't (really) kings either.  Applications are convenient vessels for human semantic communication.  We're glad there are applications around to help us communicate, but we'll be just as glad when there are even more common/free/open source interfaces that let us get down to the business of socializing effectively online.

Pattern recognition, looking for metainformational depth -- all of this is just socializing in different words.  Whatever layer-interference gets in the way of this human activity should be examined carefully, particularly if the layer-interference is made possible by leverage over crucial bottlenecks.  I have a feeling we'll pay more to socialize online than to watch movies.