Radio guy here. You are correct. I just finished reading their paper, and they describe a technique that creates a spatial null by changing the orbital angular momentum. In other words, they push the intensity away from the line-of-sight (LOS) ray out slightly (in space about the ray) so that two antennas can receive the different signals by spatial diversity. This requires A) precise tuning and setup B) a decent distance away from the transmitter source since you need the wavefront to spread out wide enough for antenna location and C) a re-calibration/tuning anytime the setup or possibly multipath changes.

This technique should work well for static point-point comms (not with the claimed infinite bandwidth though since you run into a physical problem of area and precise location of the receiving array as well as near-field antenna effects affecting receiver patterns) but in its current form could not be implemented in a mobile device. Hell, a bunch of the time in WiFi or cell reception, multipath is your best friend and can be the only way you receive a signal and this work does not seem to address this issue.

Still though, this is an interesting idea and should not be quickly discounted, although the article gives it much more hype than it merits imho.