Always love a new Karl Guttag article, and I'm pleased to see people doing all this fundamental engineering and science work.
The more of them I read the more I tend to think that in ten years the best solution is going to be VR with video passthrough. That seems (as a layperson) to have a very clear (at least comparatively) engineering path to delivering a high quality experience.
So many (seemingly unsolveable) issues with AR displays.
This doesn't even touch the software side which seems to need AGI to actually deliver the AR that people actually expect in the end. The trivial things for AR are better served by smartphones and watches for the medium and longterm still.