At this point the right answer (IMHO) is to create a new thing that is very similar, but has a new brand. Market the hell out of it and subsidize the hell out of it.
Trying to prevent the existing brand from dying is a fools errand.
All things you are trying to get others to use have a brand, whether implicit or not. Heck, it's usually even valued separately from the technology (good will, etc).
RSS, the spec may be fine. I've implemented RSS readers, heck, i use newsblur every day still.
RSS, the brand, is a goner.
The OP (and twitter post) is complaining that others consider the brand dead, and are arguing that the use case it covers are unmet by other things, and that the spec is still good stuff. Those are all very orthogonal things.
It's great that the spec is still the right thing! But the brand is dead, so you won't get anywhere without fixing that.
If enough people think that RSS sucks, is too old, died with GReader, has no user-base, or whatever other story, why would they invest their time and effort in supporting it? If the problem isn't technical, then it might need some kind of marketing solution to get over the historical baggage in the zeitgeist.
This whole notion of RSS' "brand" being the problem just reeks of a used car salesman trying to convince me that spare tires are for chumps.
The world is littered with highly useful technologies that nobody adopts, or eventually die to less useful, arguably objectively worse technologies. Adoption and software life cycles are not a techo-meritocracy.
In short - you are completely and totally ignoring the social aspects of software adoption and use. But they are, in fact, often the most important part. I don't actually particularly like this any more than any other software engineer - it often feels wrong. That does not (and will not) make it any less of a reality.
(as an aside, yes, in fact, as joist hangers and ties and such became more and more required due to code or construction method, structural screws became all the rage - nailing joist hangers/etc with regular framing nail guns was hard and dangerous, but structural screws just require an impact driver or a screw gun to do safely. Now they actually have specialized metal connector nailers, so structural screws are getting ignored again outside of decks)
Usefulness isn't a high enough bar. When choosing a technology, I might also want an active community, for support, for ecosystem tooling and integration, and more.
Things with a reputation of being dead gains lower interest and support.
Just publicly dropping the "old" XML-based version and introducing the "new" JSON-based version would probably bring a brand new flair and growing adoption to feeds. (Nevermind json feeds existed for years)
Now we just need someone to pull it off...
RSS was great. I mean, it is great, but it used to be too. It didn't die in the public eye by accident. It was killed by the big boys who realized they can't make money off of you if they can't decide what content you see in front of you, you don't run their software on your device, and/or you don't have to sign up for an account to get the content.
Try to pitch a new syndication spec to anyone that controls a large portion of the information availability online, they're not stupid, they'll shoot you down because "we tried that already, it wasn't profitable enough."
The era of standardized protocols, freedom of information, ease of communication, interoperability, tools that empower the user, that era is dead. And it was too short lived. The internet we have today looks like something designed by Dr Strangelove. If you want to have sensible tools that work for you you're going to be swimming in a very shallow information environment with almost exclusively other zealous enthusiasts of whatever tool you think is better.
Use the blockchain and make it a part of web3
https://carcd.org/wp-content/uploads/2018/10/Sloane-DellOrto...
But so like, if you want to lobby a local government to add some feature to their website, they aren't even effectively going to be able to do it; but if you can convince a company like Streamline to add it to their offering, then you'd suddenly experience like half the local governments in California all getting that feature at once.