I've done experiments in completely audio-based "VR" experiences. I think it's possible and could even be a useful form factor for people. It definitely blurs the line between VR and AR, and I default back to my assertion that the big difference between the two is who is providing the context for the experience: the developer (VR) or the user (AR).

From a technology standpoint, the Holy Grail device will be able to do both VR or AR selectably and seamlessly. The reason I think so many extant AR apps are just crappy VR apps is because the devices we have aren't able to blend VR and AR, the developers wanted (or were mandated) to use a specific system, and then shoehorned a specific design to the system they had.

> no, overlaying stuff over image of the world captured by a camera doesn't count as proper AR

News to me. If that is the case, then how is the HoloLens AR? Without utilization of the spatial surface recognition system, conceptually the HoloLens is nothing more than "overlaying stuff on an image of the world". That the world is viewed through plastic rather than routed through silicon doesn't make a difference.

Add surface capture to a VR headset with a pass-through camera (which can be done today with off-the-shelf parts, it's just not in a commercial product yet) and I see no difference.