Been working on a proof of concept that has involved procedurally generating Models in RealityKit for VisionOS.
It's been a lot of learning, and I can't help but feel we need one more framework on top of RealityKit to make some of this stuff simpler.