I'm a React dev and have been mentoring a junior. React's hooks, when and why they run, is very unintiuative for them. Skill issue is part of it, but React's immutable prop diffing forces the use of hooks and understanding how the framework uses them. Moving to a re-render model which always rebuilds the vdom tree allows callbacks, state to be defined outside of the render method without framework abstractions wrapping them. I'm not familiar with what Remix is doing but it looks a lot like Mithril.js, and working with Mithril is really enjoyable after working with React for as long as I have.
I like the safety aspect of QUERY. Having CDNs cache based off the semantics of the content might be a hard ask. I wonder if this might lead to a standards based query language being designed and a push for CDNs to support it. Otherwise you probably need to implement your own edge processing of the request and cache handling for any content type you care to handle.
GET is a keep things simple stupid approach to caching. The URL is the cache key plus any headers touched by the vary header. Adding the requirement to vary on the body and understand the body content semantics brings in a whole lot of complexity that GET avoids.
I find the semantic web stuff is really interesting but it seems to have more success where there are complex up-front data modeling and collaboration requirements, maybe between orgs. There is a Genome database out there for semantic web and I've heard of banks using it in complex fraud detection systems.
These elements have non-presentational definitions in the current spec. A screen reader could make use of the semantic difference between <strong> and <b>, even though desktop browsers default to presenting them the same way.