Eirikr70 The benefit to app users would be to have a service that "just works" instead of searching the web trying to find why they don't get their notifications in time, and finally potentially making the wrong decision in installing GSF.
"Just works" is indeed a situation that would be great.
But that situation happens because of #1 through #3, at which point #4 is a faint optional extra bonus. #4 by itself does not achieve "just works", and doesn't even make #1 easier by 5%, or #2 easier by 5%, or #3 easier by 5%.
It is true that Google has "integrated" their push code, and it is largely true that notifications "just work" on their stock OS. But the integration of the client-side libraries (#4) is not the primary cause of "just works". It's not required.
Google integrated the client-side libraries. They also spent at least tens of millions of dollars on server infrastructure (and are still spending on that), and probably a large fraction of a million dollars on publicizing their solution and signing up app authors. If they had spent that money and time but not integrated the client-side libraries, the difference would be imperceptible to users.
But you don't need to take my word for it! Ask an app author, ask the UnifiedPush authors, ask around to see if anybody plans to volunteer an open-access ntfy server for a billion people to use: ask them what percentage easier the problem would be if one Android variant (or even all Android variants) integrated the UnifiedPush client-side libraries.