Ultimately it appears to be software with more fans than productive users.
I love the idea of arcan, I like how as a counter point to waylands "X went to hard, the display server should be a dynamically linked shared object" there is one solitary guy doing his own thing saying "X did not go hard enough we need an even more seamless, recursive solution to move our display between different devices" and doing amazing work on this problem.
But just like plan9 I have a hard time actually doing anything with it. Achingly beautiful software, but just a little to different and obscure. And I say that as an OpenBSD daily driver, where apparently I thrive on the pain that comes from using an obscure system.
But I am currently back on the plan9 kick, seeing if I can get it to stick this time. I may give arcan another try as well.
For me Plan 9 isn't about daily driving or how useful it currently is, rather, It's for exploring a new way to build software. There are lots of pieces missing but that's the fun part, you get to build them!
As for daily driving, 9front has vmx(8) so you can run virtual machines on supported Intel hardware. I know a dev who runs Linux (I think OpenBSD too) in vmx using VNC to run a browser. 9front Drawterm also has a few tricks to work in reverse where the Linux resources are exported to the Plan 9 workstation.
Edit, I should also mention that Arcan is close to how I would envision building a web OS using Inferno: don't start with a browser, start with a highly portable OS who's user space lives in a VM.