Good to see you on this thread Brad...that video has 261 Dislikes, but i'm sure 260 of them have never tried such couplings.
The drag is only reflected upon the prime mover, if the coupling isn't tuned. It's a lot like the Primary and Secondary of a Tesla tower (and just about the only reason I can think of as to why Tesla got magnet strength named after him

).
Slightly too near = better cogging and easier starting, but a slow down of the powered rotor.
Slightly too far away = rotation coupling may drop out.
There can be a loss, something like 10RPM from 300RPM, but the distant rotor will now be spinning at 300RPMN or 600RPM, as Steve showed.
If the powered rotor is doing 300RPM and the same weight mass distant rotor is doing 600RPM, that doesn't make a lot of sense does it ! Torque would presumably have to be half of the powered rotor, minus natural losses from not being powered - which wouldn't even feature if aircore coils picked off the energy of the spinning rotors.
The Faraday Paradox proves the need to think of magnets as gears. Whichever way they link is just the same as teeth on a cog.
In the video above, the magnet didn't spin when under the other one, same as a cog won't turn if above another.
I believe the lines of force are cone shaped, giving Steve's ability to put the distant rotor on books (enjoyed that vid as i'd never tried it).
Here's my best ever run so far:
http://www.youtube.com/watch?v=D9V8BEtqRbg4 rotors, only 1 powered.
Feasibly, rotor after rotor could link from each other and have collector coils tapping the energy from each!
In such a way, the 261 Dislikes video doesn't strike the mind as impossible for what the guy claims.
Here's a fun thought - consider a large circle, with say a dozen rotors spinning.
What happens if the 12th rotor is placed to interact with the first rotor ?