This has been a real issue on my side for several months. Perhaps much longer than that but I only started complaining about it in a recent time. YouTube, for some reason, becomes really hungry in term of memory usage when I watch videos on Firefox. This is the worst my machine can ever reach:
Yes. That amount of memory. Even games like Half-Life 2 with a shitload of mods running on Linux wouldn't even ask for that much. Though usally it would use about 2GB of memory. Yet still, I was like, "dang it that bloody hell, is it trying to compete with how Yandere Simulator was badly optimised???"
I've heard and tested myself, YouTube doesn't seem to behave like that on Chrome (or Chromium-based browsers). Perhaps it's only trying it's best to simulate Cyberpunk 2077 v1.0 on platforms Google don't bother giving a shit into making them good.
But... sometimes Google isn't the actor of everything behind. Firefox is said to also have memory leaking problems. There are a lot of users across forums complaining about it (see this Reddit thread).
Talking the said Reddit thread, I found a comment that explained something:
Don’t ask me for a source, but I believe a factor could be that YouTube kills your system when uBlock is detected. It can’t show the ads, but it can make YouTube use like all your motherfucking RAM.
Prestigious_Cherry_4 (Comment on Sat, 2 Nov 2024)
I myself also did a test by disabling uBlock and see. It actually did reduce a bit, however I am enabling it back in and browse for a while, it's keeping around 350MB. Perhaps there is actually an issue, but in another aspect to say.
After that 4GB surge, I had to temporarily migrate to a Chrome-based browser (BrowserOS) in order to watch YouTube (I actually felt safer there with an average RAM usage of 300MB). Now I just went back to Firefox and it appears that it just went back to its "ordinary" state.
yes, now with two of them.
So probably YouTube will not hoard my poor laptop's resource for now. Still, I will have to constantly monitor its memory use for a while when possible to see that what is really causing it to consume way too many RAM.