I've had zero "system breaks", neither has anyone who used a distro supplied driver package. And if you installed from the official SH file, well, you did it at your own risk and you had to understand what you were doing. In short, you're lying.
> kernel revisions breaking NVidia drivers, leaving you with a non-working desktop.
Again, didn't affect those who used a distro supplied package. Those who used the installer patched the kernel, which I've done countless times but over the past four years I've done it zero times. In short you're lying again.
> How about all the times lightdm/sddm/gdm would make a tiny tweak, and Nvidia users would have to downgrade in Arch or Ubuntu or Gentoo because the driver's house of cards completely collapsed?
Never ever happened to me or anyone who I know. Show me actual fixed bug reports where NVIDIA is blamed for breaking the DM or bust. I think you're now lying again. Xorg support in NVIDIA drivers have been stellar aside from older GPUs and older drivers not supporting newer Xorg releases but then HW companies do not owe you infinite support. People who expect their more than 15 year old GPUs to work with modern OS'es and support all the features are lunatics.
> How about all the times that XF86.conf or xorg.conf would get completely corrupted upon a system freeze, and you'd end up having to reconfigure X in order to use your login manager again?
What?? Did you come up with this BS in your dream? I've been using NVIDIA since late 90s. Never ever in my entire life I've got the Xorg config file lost. In fact, and here's you lying again (!) Xorg reads this file and closes it right away. It doesn't keep it open. Under no conditions this file could have been lost now or ever. And Xorg has used auto configuration for almost a decade now if I'm not mistaken, so there's no xorg.conf file to speak of. I have a small file in /etc/X11/xorg.conf.d to enable CoolBits in the NVIDIA driver, that's it.
> How about every time XF86 would refuse to even recognize your card occasionally for absolutely no reason?
You are making stuff up. Again, there will be no bug reports, citations, anything.
> If you didn't experience any of the above, at least once every 6 months, you either never upgraded your distro, or you weren't using Nvidia cards in linux 20 years ago.
I've been using NVIDIA since Riva TNT2, thank you very much. And the list is a lot longer than you could imagine including MX440, FX 5600 (it was horrible), GeForce 6100 (integrated GPU), 7600 GT, 8800 GT, GTX 660, GTX 1060 and now GTX 1660 Ti.
I don't remember your nickname on nvnews.net (do you even know about this website? nah, I don't think so) - NVIDIA's Linux support forums in the late 90s/early 00s which tells to me everything.
Almost everything you wrote (sans occasional kernel incompatibilities which did _not_ affected normal people) is just pure made up crap.
I can show you nvnews.net and forums.developer.nvidia.com threads with actual issues but you've listed none.
I'm the person who filed this https://github.com/NVIDIA/open-gpu-kernel-modules/issues/511 and it affects countless people, mainly Turing users.
I know how none other than kernel hackers _intentionally_ broke NVIDIA drivers recently by breaking EFI console, a malicious patch which was later reversed because it affected more than just NVIDIA drivers.
And I also frequent kernel bugzilla and AMD/Intel bug trackers. They have far more serious bug reports.
AMD drivers for certain GPUs have been broken for a large number of people in the past three "stable" releases, including 6.3.9, 10 and 11: https://bodhi.fedoraproject.org/updates/FEDORA-2023-9b7e0890...
AMD open source drivers have become semi-decent maybe in the past two years at most. Oh, except, RocM. Intel fares better but they are not strangers to deal breaker regressions.
No one is perfect, and trying to say that NVIDIA is hostile to Linux while offering a high quality driver is either malevolent or just mean but it's definitely not true at all.
I know this crap dearly, sorry.