Nice piece Tony!
Neil Trevett

Hi Neil (and great post Tony!)

I was just having a similar conversation today with some industry leaders in this space re web bindings into the Computer Vision ARxxx stack. I’d love to port (my new cv startup) algorithms to be webXr accessible. They convinced me that it’s not the way to go (including tracking or vio into webXr) as it is so dependent on hardware spec that it wouldn’t scale, and the extent of native webassembly or similar code needed would hollow out much of the benefit of the web abstractions. There is definitely some great value in figuring out the interface/apis between cv sdks and app development environment, but it’s not an all or nothing decision. Eg vio & meshing could be done with a native sdk but some object recognition or segmentation/scene understanding might work higher up the stack. Would love to work more on this with anyone interested …