The sound sample seemed to fit the 'vibe', but lacked any discernible definition. Could it be that it's too sonically dense to easily reproduce? Perhaps this could be improved with a more tailored training set.
However, the current top post is NSFW and the preview is rendering in the timeline, that would be something to change so folks can safely browse on the job.
Headphones are a great analogy. For me, headphones are a burden to wear long-term.
And that's coming from someone who spends a good amount of time doing studio work. I much prefer the comfort of a good set of speakers as they're far less intrusive, but for some things you just have to have a pair of headphones on hand.
This is why I have doubts that the current form of VR headsets is going to be the new standard for working.
What you describe sounds an awful lot like SerenityOS, with the exception that it's not a VM, but a full-blown Operating System. It really does embody that pragmatic sense of no-fluff UI design that peaked in the 90s, but applies it to systems design as well.