The first is learnability. A big problem with gestures is that there is no clear affordance as to what kinds of gestures you can do, or any clear feedback. For feedback, one could couple Soli's input with a visual display, but at that point, it's not clear if there is a big advantage over a touchscreen, unless the display is really small.
The second is what's known as the Midas touch problem. How can the system differentiate if you are intentionally gesturing as input vs incidentally gesturing? The example I used was the new Mercedes cars that have gesture recognition. While I was doing a test drive, the salesperson started waving his hands as part of his normal speech, and that accidentally raised the volume. Odds are very high Soli will have the same problem. One possibility is to activate Soli via a button, but that would defeat a lot of the purpose of gestures. Another is to use speech to activate, which might work out. Yet another possibility is that you have to do a special gesture "hotword", sort of like how Alexa is activated by saying it's name.
At any rate, these problems are not insurmountable, but it definitely adds to the learning curve, reliability, and overall utility of these gesture based interfaces.