Most smartphones today do real-time face tracking, object tracking and other special effects like a cakewalk. I feel like hiding under the table when my novice clients show these on their Apps and I keep struggling for hours between AE, Mocha, and PP.
IMHO Those chat filter/AR tracking gimmicks aren't good enough for final quality fx work. Also a lot of that is using information that you can't get from just a footage file (eg. depth information from multiple cameras).