If your expertise is in low-level browser performance but you lack deep knowledge of React/etc, your skills aren't of diminishing relevance, and probably won't be for a long time.
Something like Angular.js is a more apt example of a skill people still do advertise as a primary thing in their resumes these days and it most definitely plays into the diminishing relevance thing. On the other side of the spectrum, you have newbies writing down "HTML/CSS/Javascript" in their resume, which basically screams zero experience (aka "low/no relevance until they learn to speak the cool framework lingo"). There are plenty of other quote and quote "uncool" skills in actual resumes. Flux. Sagas. Ember. Etc.
So, yes, an argument can certainly be made that recruiters/interviewers do "discriminate" against various stereotypes to various degrees, and it isn't necessarily always based on ageism. A recruiter that reached out to me recently had an interesting term to describe this phenomenon: "skill commoditization", aka looking at people as snapshots in time, and evaluating some skillsets to be intrinsically worth less than others, even if they have some amount of overlap, instead of looking at people as malleable entities. It's worth thinking about how treating developers as commodities might hurt your hiring, especially if your hiring revolves heavily around evaluating for expertise in a specific framework.
To try to shoehorn this back into the compiler analogy, modern frameworks would be like the advanced, complex and opaque compiler, and the grey beard is the guy that understands how stuff like HTML streaming and font download prioritization and hidden classes affect performance, whereas the team lead is analogous to the run-of-the-mill bootcamp grad that knows how to use the framework-du-jour but is way out of their depth if venturing outside the comfort of the framework and into the depths of those advanced topics.
And I think the gov.uk use case here might actually be an outlier in the sense that it actually considers performance of an already tight codebase, whereas a lot of jQuery deprecation efforts bring with them heavier alternatives in the name of maintainability or developer productivity or hiring or whatever. To be clear, many of these concerns are quite valid, but it seems a bit disingenuous to create a false dichotomy between understanding of low level concerns vs concerns about SDLC management. They aren't mutually exclusive.
[0] https://youmightnotneedjquery.com/