These are deaths by a 1000 papercuts, caused by a combination of kernel/kms drivers, xorg & dri. Reporting these bugs is an issue by itself, since it's really non-trivial to understand where some of these problems are and how to report them.
One example I like to give is how for the better part of 2018, xorg+kms with (if I remember correctly) broadwell was new, caused random tearing issues that required to use the legacy i915 driver. The i915 had also other issues at the same time, causing blits in xrender to copy dirty areas of memory. Different issues popped up depending if you were daring enough to switch the dri version, glamour and kernel version.
This got eventually resolved.. but by the time I switched to haswell. Things got better on the kms front, until I realized I couldn't rotate the screen anymore. When this got resolved, the driver couldn't restore the screen after dpms blanking... This one is fixed on skylake, but on haswell still isn't.
... and so on and so forth. All this was on the lenovo x1 carbon (various versions), then on the x1 yoga (1st, and I'm now on the 3rd version).
