This whole thread is full of people wildly speculating about applications which, without amazing platform level support are each going to be a small miracle in state of the art computer vision applications.
Can they offer full 3d image detection / segmentation to give you a full spatial model of the user's environment? If not, a lot of people are going to be sorely disappointed by the amount of computer vision expertise required to get even basic apps they are envisioning to work. On the other hand if Apple does offer this type of platform support, it will be quite revolutionary.
Meta has offered a range of spatial computing APIs for a while but they have been slow to take off. The Quest 3 should really push this along though. It will be fascinating to watch the two platforms compete on this front.