1) there is no access to video/depthmap stream like in iPhone ARKit so it's not possible to roll your down Object Detection or ML Model or even simple QRCode scanner
2) Latency for Image Tracking seems to be around 1-2fps comparing to iPhone ARKit
3) in some WWDC videos hand tracking seems to be not as precise and also having latency
4) You cannot implement something similar like Mac Expose because iOS/iPadOS/visionOS is much more limited than MacOS
5) At least on iOS/iPad depthmap from lidar is much worse than those provided from TrueDepth (and even those got significantly worse from iPhone 14+)
Overall for productivity I find some future iteration of XReal glasses more compelling and better direction:
- much cheaper
- very lightweight
- work with connected to desktop or mobile phone
- can have lower latency since connected directly with usb-c cable
- can possibly have less software limitation if connected to desktop
Ideally I would like to have some combination of google glasses / meta ray-ban / xreal so that you can wear it outdoor for simple HUD and AI voice control but can be connected to smartphone or laptop to provide AR experience similar like in xreal but just with better resolution.