I was asking myself if Ubuntu will this time upgrade GNOME stack to 3.38 inside Focal. I know it’s probably not planned at all, but I realized that Groovy, Hirsute and Debian 11 will run 3.38. Alignment of planets.
Thus the advantages of a 3.38 upgrade seem very clear. What are the downsides, apart from LTS design, that make this a bad idea?
Upgrading the whole stack is a huge undertaking - lots and lots of dependent packages. This requires a mass of manpower for development and testing.
The risk of getting it wrong is considerably higher than just patching stability & critical updates that a LTS demands.
I agree that upgrading focal to 3.38 would represent a risk and would take too much time away from future releases. We are unlikely to ever do that.
But don’t use Groovy and Hirsute as a comparison… Groovy reaches end-of-life this July, and Hirsute reaches end-of-life next January. By that time, both 21.10 and prerelease-22.04 will be ready with GNOME 41 or higher.
One point, you’re right.
I was obviously more focused on Debian 11.
3.36 is working pretty well. I believe it is now at 3.36.9 upstream, so it is continuing to receive bugfixes and minor backports.
As far as quality of life on 20.04- I personally think, if anything, backporting Desktop Icons NG would do the most good for polish and user experience.
FYI new upstream point releases (+ some more upstream cherry-picks) have been uploaded to focal, follow the relative SRU bugs to help testing them!
Updates are hitting proposed in these days, so please help testing the SRU bugs (see mutter and gnome-shell in focal).
Testing now! 
Thanks for getting them out there.
Thanks!
Seems to work, as far as I can see.
I still get this in logs:
Window manager warning: META_CURRENT_TIME used to choose focus window; focus window may not be correct.
when swiping workspace using 4 fingers.
It’s there:
https://gitlab.gnome.org/GNOME/mutter/-/issues/1066
but does not deserve attention 
.