I'd say that has some merit. Fairly simple, lightweight. Because the Ball Wheels are not coplanar, inward lateral loads will try to move the worm rather than compress it. This means the bearings and housing will have to be beefed up.
Another concern is when manoeuvring you will have assymetric torque requirements on each rotor. This could result in very large forces being applied to the worm. The worst case for this is in autorotation with an aggressive roll rate, the rotors will want to turn at different speeds, but that worm will have to keep them in sync.
Would it be an improvement to have this as a driving mechanism but still have a conventional gear between the rotor masts for synchronisation?
Just my $.02