The problem, to me, looks like getting 90% of the way there is possible. However, getting from 90% to 99.9% (to get rid of that "wonkyness"), seems to be way way way harder. Magic Leap has made little to no progress in working out the kinks in their product despite raising literally billions of dollars.
The 90% statement is both true and is one of the things that really sucks people in. There are any number of technologies that kinda/sorta work as a PoC. There's a gaping chasm between that PoC and a viable working product but a lot of people look at the PoC and can imagine what the working product looks like--and think to themselves that it's only about some refinement.
I wrote a camera/IMU tracking software (VIO) before. I can confirm this statement. Get it working is difficult and requires getting all the algos right. Getting it work reliably is a lot of rigorous testing, collecting datasets, evaluating slightest changes, tuning numerical stability, experimenting with different techniques which might work better in low light and optimizing for the hardware (VIO is intensive but the best VIO is useless if there is no computation power left for the app itself).
Neither of this is required for a slick demo in well lit room and perhaps on better HW.
