I claim that--to the extent to which this was ever an issue--this is due to the notion of "micropayment" really being too large still: the industry has often used this term to describe transactions on the order $1-$10.
Instead, this discussion is about hosting fees for a chat app that will add up to almost nothing for light usage, not the $1 per article news companies keep wanting to charge. My company (Orchid) has been calling these fees for incremental hosting costs--which are on the order of $0.001-$0.10--"nanopayments".
That's my main gripe with all those transactions. If you need to initialise a complex utility that involves the trust of multiple parties and the need to verify identities on multiple people. That kind of thing costs macroscopic amounts of money (the going rate seems to be O(20ct)). Unless your average transaction size is much above that, you'll loose to friction.
If I could use micropayments this way and be sure that the final price is
a) somewhat predicable
b) usually not ruinous
It would work better.
Cloud Services for companies like AWS are using this mode and it's working well for them.
But I have a feeling it's not easy/legal to do away with the immediate price tag for consumers. At the very least it would require the a constant price per "thing" and a hard spending limit.
currently it says "250 picodollars / byte-month of encoded data"
https://www.daemonology.net/blog/2018-04-01-tarsnap-pricing-...
That said, I don't think even "Web Points" would work for links. Maybe something like buying "Member Status" with "Web Points" for a month on a site could work.
I think Brave has the right general idea.
Fund an account and then split that per month among the sites I visit, or according to a breakdown of my own preference.