Correct about the current state of SLAM. However the next wave for computer vision focused tracking and mapping is heavily based on ML - though there isn't a single combined system that gives robust SLAM/PTAM level results yet.

There seem to be some nice developments into this direction recently? Something like the SfMNet (https://arxiv.org/abs/1704.07804) or the Correlation Filter Net (http://www.robots.ox.ac.uk/~luca/cfnet.html) for tracking.

With the SLAM algorithms I am familiar, they are mostly statistics and abstracted above the camera component. They are mainly concerned with where am I most likely to be given I am witnessing these previously seen landmarks and have this noise model. Moreover, the chosen statistical model has to be such that it can be updated rapidly as new landmarks and state estimates are added (which typically presents as an increase in dimension until you hit some limit you impose).

Most (all?) SLAM approaches are naive and produce point clouds. True geometry approach would result in purely 3d vector output. Kinect, Tango, Hololens all fail at recognizing flat surfaces and straight lines.