The desktop client still lacks developer manpower (see how many issues aren’t even tagged let alone replied to or even fixed) and the newer releases of the desktop client have some very troublesome bugs that aren’t fixed (some people still use 2.5 or even 2.3 either because of these bugs or because they don’t know a new version exists because the updater didn’t work till 2.6). It would be better to stop creating new features and start to fix serious issues in the existing implementation. Same goes for E2EE which is still not out of Alpha quality.
Also, it would be good to refrain from implementing features in parallel that others already did (Mountain Duck/Cyberduck/Cryptomator) to save energy for really important stuff (like reliable syncing). But I think this is a bit like KDE is done if you know what I mean.
I think this would break the workflow of a lot of people that they use now for many many years already. Removing the folder based sync would be a very bad idea for all those people and a lot of sysadmins and users alike will be very unhappy with this (let alone switching all of them to the new way of syncing).
Please listen to your users, the Gitlab backlash should be a lesson to everyone about how to not do things.