The link and the URL are what made hypertext hyper. They turned words in a document into portals to other documents.

But that web isn’t the one we’re dealing with anymore. We’ve moved from documents to applications. Like tabindex, I’m sure browsers or frameworks will gradually find ways to make any element behave link-like in every way. Which isn’t really a bad thing, it’s just development of the core idea of the web.

The average modern SPA pulls in far more unnecessary JS than the average plain HTML site pulls in via redundant tag structure following the initial page load. Moreover, server-rendered HTML renders far more quickly because there tends to be fewer total requests and there’s no hiccup parsing the JS when it lands, much less the time it takes to execute it.

And don't forget that gzip works wonders, removing most of the network overhead of transmitting redundant tags.

Browsers are perfectly capable of prefetching pages on their own.

It is the H in HTML and HTTP!