If I lose internet connection I no longer have access to new links or past sites i've visited.
The sites I've visited had the HTML rendered, and that can be saved for later use.
Browsers should offer re-visiting one of the 10 last cached versions of visited sites instead of offering nothing.
New content should be fetched periodically to have something to read on times of network issues.
Having nothing to read can be boring.
So many sites rely on backend related functionality such as XHRs, session management and auth etc to render the front end. Sure I can cache and re-render the frontend from some historical snapshot but the same javascript that came down in the original load will execute again - with uncertain and probably not great results.
Providing functionality like that seems like a good idea on its face but implementing it would be yet another set of test cases and problems for site/webapp developers and nobody wants that in 2024 when internet connections are ubiquitous in Western, developed nations. In the case of non-developed nations, the money to implement such a thing just isn't a viable spend.
Doable? Sure. Worth the expense relative to other clear and present customer facing problems? I couldn't justify it.
JS serviceworkers can do a great job of intelligent cacheing and I have stuff of my own I can use on the tube (no internet). I don't know why they are not used more. Sadly they only intercept own site.
I guess the reason is that Internet works so well nowadays.
A long time ago I was running a proxy at home behind a dialup (or was it ISDN?) that would cache the contents of websites and offer the old contents if it was unable to connect to refresh its copy. It worked ok, for the time, before the proliferation of SSL. I also modified it to first give the cached copy while updating the page in the background, which sometimes lead to breakage, but things weren't rock solid otherwise at the time ;-).
I guess something like that would be nice even today, but it would probably need to be built as an extension into a browser to most conveniently handle the SSL issue. I think I would use it, to archive all the pages I've visited and then easily visit older versions of them, in addition to having a way to search them.