The Berkeley team proposed aerodynamic suspension. There's an analysis with ANSYS which indicates that the design can be tweaked so that at no point does the air go supersonic.[1] But nobody seems to have published a stability analysis for the vehicle in the tube.
The "flying height" in the original proposal was to be about 1mm, so this thing has to be really well behaved aerodynamically. And the tube has to be really smooth. (Expansion joints, emergency escape hatches, pressure doors, and switches would be problems.) The whole point of the Hyperloop is that it's cheap because the tube is simple and dumb.
When you look at the MIT design [2] it's really a maglev monorail that operates in an evacuated tube. (Or outside of one, as they've demonstrated by sending down a rail). It doesn't have any of the fan propulsion or air cushion lift or tight tolerances from the original Hyperloop proposal.
[1] http://www.ansys-blog.com/20130925hyperloop/ [2] https://www.youtube.com/watch?v=yiyPQU48Ef0
Anything good you've read that explains why they've abandoned that route?