We haven’t spend a lot of efforts into visual effects, like better lights or better materials, because that’s not the main point of VR Sketch. (We’re also a bit thin on the necessary knowledge, as we’re not from the game industry ourselves.) Improvements would be necessary here, in order to get a significantly nicer result.
I don’t know yet what advantages a ray-traced approach has got in our context. I realize that the following is not exactly what you have in mind: ray-tracing is usually seen as an expensive way to get stunning high-quality renderings. But it might actually offer a performance benefit in our case, at equal quality. Technically, if you load a complex model you get very heavy “overdraws” right now. This means that, very roughly, rendering is twice as slow if the model is twice as big. But with ray-tracing it might not be the case, and we could render very big models at a reasonable performance. The main limit would then be the amount of CPU and GPU memory.