Also, applications tend to use RAM when it's available. Think of all the cache, typically that's why back is snappy (it doesn't re-render the page from scratch). Maybe it would make you feel good to see a lot of unused RAM, but since it's there, it's better to use it than not.
Hackernews could take 100's of megabytes to load (like pretty much all of the other similar websites do). RAM, page file or otherwise. But it doesn't. It could be a complex application. But it isn't. And it shouldn't be.
Also, I don't think having "too much RAM" is the issue on my 16 GB machine. I routinely have to just shut down the browser entirely to be able to properly do CAD work. Party like it's 1995.
Even if these websites are "legit applications" for good reasons, putting the processing burden on the client is still a choice. There are efficient ways to leave most of the processing to the server. Example: https://www.phoenixframework.org/
Disclaimer: correct me if I'm wrong. I'm just an EE guy with shallow limited knowledge on web stuff.
Open the cpu performance tab of task manager and set a page to reload every few seconds.
I don’t know why that happens or how to make it stop.
I wonder if it has to do with the scheduler? Desktop linux users sometimes install patchsets to their kernels like zen[1] or ck[2], which effectively enable soft-realtime in order to improve desktop responsiveness and latency, but reduce net throughput. iOS also generally feels more responsive than android, given comparable hardware, for similar reasons: more optimal scheduling algorithm.