Since remote login is now a thing (with this fix applied), I’m wondering if it would make sense to set it up on a server, enable RDP login, then use it like a normal computer except remotely?
I actually managed to get a taste of it today, but only briefly because the client machine has a very flaky WiFi connection. 
I’m actually using my work rig mostly through RDP.
I really hope that the SELinux issue is fixed and no workaround is needed (BTW here’s the Bugzilla issue), but apart from that, things work pretty well and fast compared to xrdp that I tried to use at first.
The only issues I’ve noticed is that the user you are logged in is not exactly the same, i.e. some things don’t work for me like changing power mode from gnome, or changing the hostname via GNOME settings, but assuming you also have SSH enabled most of these issues can be solved.
Another thing I noticed is that Software Updates don’t show up for system packages and when you try to install a flatpak package from Software, it will be installed as a user flatpak or as rpm and there’s no option to install it as a system flatpak.
All in all, it’s been a pretty good experience 
At my my former work, local POSIX systems were removed (thankfully, after I had retired) and applications migrated to RHEL running on systems in the Data Centres. Remote desktop had always been slow, but colleagues found that web-based remote tools (Rstudio, Jupyter Lab, SageMath, etc.) work well.
Man, that sucks. I hope they moved on to something like Active Directory? AD sounds like a good candidate for these kinda situations.
Seems similar to how I’m running Windows now on a NAS with a webserver 
Although I don’t really need a GUI for any of that and would still opt for CLI Server whenever I get around to re-doing it. GUI stuff in my case is extra bloat for a dedicated server, but I don’t see why it wouldn’t work. Stuff would still be GUI-invisible background daemons and systemd scripts.
Editing stuff through Terminal with GUI on a server could be a benefit, but I already do exactly that when SSH’ing into the server from Fedora Workstation with GNOME Terminal
I like cockpit (Web Console for Linux servers). Gives me web browser interface with terminal and support for administrative tasks. I have Fedora Workstation systems running in 3 builidings (have to go outside in high latitude winter to reach a different system).
I like Cockpit, but don’t like the idea of it being another entry-point to my server.
OpenSSH is tried-and-true, along with having CVEs. With Cockpit plugging into the system with its own stuff, that’s its own set of potential CVEs to be concerned with (might not be able to get through ssh, but how about on less-eyes Cockpit?)
I glance at Cockpit occasionally on LAN but would likely never entertain WAN-opening a port for it