As with the wide array of VR instruments (that also use highly accurate position tracking as inputs), what it really boils down to is:
If you are procedurally generating audio (ie. using a synth) on device then what you have is categorically inferior to using any number of existing synths and virtual synths along with a midi input device (such as the genki wave).
Don’t embed the audio generator.
Just make it a midi input device.
It’s not a thermin. Same way a microphone that emulates a flute when you blow it is not a flute.
It’s a synth that can only take one kind of input.
Novel input devices are fun, and I like that; people like exploring different ways of expressing themselves in performances… but I’m a bit skeptical about this.
Seems like it’s lost what was special about the theremin.