If a publisher had insanely high quality articles on everything, behind a paywall, I'd subscribe to $15 a month pretty quickly.
Every hour saved would pay for itself many times over.
The problem is having a wide enough array of content that such a site becomes the single go to. O'Reilly had that catalog, but I think the tech world had gotten much more complicated since then.
But I’m unsure how impactful revenue wise for an author can make or how sustainable is it long term.
Most of the books I see on Unlimited are either older books (there are some good gems), books that are already freely found on the internet or books that are scrappily put together.
On the one hand, there are no 10 hour songs with some great part spread out randomly throughout that listeners have to find (obvious, and I think such a book wouldn't be successful either).
On the other hand there are tons of generated crap on Spotify, than sometimes sneaks into your playlists (even things contained text-to-speech advertisement).
I think all such platforms would need some kind of spam filter functionality and a good rating / statistics system so that quality content could be surfaced.
So maybe the economics would disincentivize authorship.
The "like spotify" was more in reference to pay once, get everything imaginable.
This also explains the difference between publishers and distributors, and the dissimilar risks they accept.
I'd suggest something more like Netflix as a model; they are primarily a distributor and only very rarely a publisher; their library has a selection of standard classics, recently(ish) popular content for which the original publisher has more incentive to license than sell direct, plus rotating items of interest to keep the catalog feeling fresh for ongoing customers.
(note that even so-called "Netflix Original" content is mostly produced by other companies).
It's been thought of, of course. The search history of "Netflix for e-books" is in practice a mixture of junkyards like Scribd and "what-if" questions on Quora. There's likely a causality dilemma with achieving the catalog scale necessary for success, mediated by the conservative strategies and revenue fears of existing publishers, all of whose investors/owners are entrenched incumbents that see their businesses as perpetuities (steady, low-margin cash printers) rather than growth ventures.
Basically, they won't do it, until they see it can be done.