Basically, in order to demonstrate anything useful, this thing needs to be hooked up to a dynamometer and have a load applied. A laboratory dynamometer (and surely MIT has one of these available) can apply variable load and measure the torque and power output. At the moment it doesn't seem to have any external load applied, it is just a free wheeling device with self coupling via the various magnets etc. In free wheeling mode there may well be ways to remove hysteresis and braking effects by the coupling. But this is irrelevent in the context of all the speculation on this device as a "perpetual motion machine" or even a more efficient motor.

Until the performance under load is verifed, any claims as to enhanced efficiency are about as meaningless as the claims in spam email about *enhancement*.