The VR controllers are not very precise.
While I don’t own any VR controller, their precision seems to be on-par with midi hardware (128 discrete values) in terms of rotation for example. Have you checked out Soundstage I linked before? It is definitely a VR-first design but rotating knobs, sliders and making connections look fine to me from the vids. Of course there’s no reason to stop there, there might very well be other ways of control in such a VR program. Also specific hardware for just this, like a rotary encoder stuck to a VR controller that uses something like the SAIL module to adjust whatever you are pointing at. This could be done with a mouse too for 2D Rack.
why not re-think it as a VR-first UI, instead of porting over real-world metaphors?
This argument works against Rack in general too I guess the appeal of the VR idea ( and Rack itself in a similar sense ) is that there are many great modules already working, so it makes sense to port instead of reinventing (reimplementing in this case) the wheel.
how do you make an exciting VR performance?
First is good music of course then visuals and any additional theatrics. For example deadmau5 conceals his face and still manages to pull off an exciting show. This isn’t a concern that’s specific to VR and has been around since people started using laptops to perform. If nothing else, just project whatever the performer sees on their screen/headset. Idk why this isn’t a thing (apart from algorave), maybe cause people don’t really care? Or cause most daw’s don’t look that interesting. The problem beneath it all is that twisting knobs and pushing buttons isn’t considered to be an act of virtuosity, the creation of electronic music is more of a cognitive effort than a physical one, and this just doesn’t translate well to stage without some bells and whistles.