Doc Edward Morbius ⭕@PixelJones What I'd really like to do is to get away from the per-item / per-use payment model, and instead think of the infostream as a distribution utility for which access rather than use is the principle consideration, and in which payment ability (wealth & income) rather than content value is the basis on which payments are made.
The questions of both what content is made available and how that content is compensated I'm leaving somewhat vague, though in general we have systems which work for this, and which have worked for nearly a century now based on broadcast & cable media, audit-based measurement (Nielson, Aribitron, etc.), distributor-based negotiations (with individual broadcast stations or networks), and something closely approaching a common-carrier model for the actual access providers (that is, ISPs).
The points @dangillmor raised are valid: a gatekeeper monopoly is a critical hazard, and is worth addressing from a competitiveness standpoint, independent of this proposal.
Why "all you can eat"? Two principle reasons:
1: Need for information is strongly independent of capacity to pay, and often inversely associated.
2: There are entirely novel capabilities afforded by access at scale which a usage-based payment model largely forecloses on. Aaron Swartz's work which lead to his prosecution and suicide based on wholesale downloading of JSTOR scientific papers is a key case in point. It's possible to look through, over, and among a corpus to find relationships not otherwise manifest. (I'm doing something along these lines with my #HackerNewsAnalytics series posted here on the Fediverse.)
The notion of an individual or household account, associated with personal mobile devices and/or household Internet service, from which pro-rata payments are then allocated amongst various providers is one option for compensation, though even that might well not be ideal. That imposes a huge surveillance component itself (who is reading, listening to, or watching what), and could well disproportionately benefit or starve less substantial or more substantial works. More critically on that last: works which are far more expensive to produce at quality, such as investigative journalism or scientific research.
Some sense of local, regional, national, and global providers / publishers, within genres, funded with a specific budget and for a minimum guaranteed time period, would provide the institutional stability to provide certain classes of work: news, education, business and government publications, academic research, and of course, entertainment.
And, again, multiple revenue streams, including premium subscriptions, patronage, advertising, etc., could well be additional components. But an access-based automatic and universally-billed tier really does seem to be a possibility that's rarely mentioned or advocated.