From Part 1, which spells out how much change can take place in 20 years:
Those who stay ahead of the curve prosper. "Cambridge University Press, for example, had tens of thousands of old backlist titles set up for print-on-demand long before other publishers did and they reaped a harvest of sales and profits in the past decade as a result. Last year, Simon & Schuster shifted resources from field reps to telemarketers."
This whole series is worth reading. In Part 2 he predicts: 'I’d expect that 20 years from now, the “local” hard drive will be relatively unimportant: a relatively short-term “emergency” cache for the rare moments when you aren’t easily connected to the network (the internet.) Data — all data, including everything you think you “own” — will live in “the cloud.” '
Publishers and other media will no longer be defined by format. The price of content (what writers can make) will go down, but the value of community/audience will increase. "The idea of a general book publisher will have no meaning." In comments, he adds: "the brands that can sell verifying the truth will command revenue."
In Part 3 he talks about "the process of content as bait to attract eyeballs -- providing tools, features, and databases to monetize the community" as is done today with Michael Cader's PublishersMarketplace. He uses Oxford Bibliographies online as an example of harnessing academic expertise to deliver curated and constantly updated bibliographies by subject.