I'll say it straight: You really do not know what this spec is talking about, that is why you are indeed misunderstanding this. And that's the positive interpretation. By now your points amount to nonsense as they support a unsupportable position. Not one of those points has any relevance to this discussion:
- This would not change
- WebSub has nothing to do with mobile devices
- This is not up for you to decide, and the server powering online feed reader or platforms like facebook do have a lot of data sources
- That's not a point? For the type of RSS readers you are thinking about that would not change.
- A CDN has nothing to do with this.
- No. It is a decentralized/federated protocol, no middle man can do that.
Your bonus point is invalid: The 3rd party is already there (pubsubhubbub is used in production), this is not primarily about this type of traffic reduction you are thinking about, but yes, the ressources needed to do without websub what can be done with websub on a big scale are enormous.
Again, this is about enabling push architectures for server to server communication and a scalable way to achieve real time notifications. The tiny feed reader application on your smartphone is not directly related to any of this. It would not use this protocol, it can't use this protocol, and it would not stop working because of this protocol.
Edit: I made this sound a bit nicer than initially. It really feels like you want to misunderstand this spec, this annoys me a bit.