I’ve talked with them before. The idea of incorporating an Arm processor into the Flash is not new. Exposing that processor to an end-user is, however. The issue I have with what they’re doing is that you have to access that CPU via a driver. So, you’re trying to offload the main CPU by issuing a function call to an aux CPU, but you have to go through the main CPU to get to aux. Make sense? Basically, you’ve made the main CPU the arbiter for the offload, which works, but is less than ideal, since you still typically have just one main now processing requests for a whole bunch of aux processors in the media. Just my 2 cents. Would love to hear from others.
On the other hand, we could potentially combine what we’ve done with what NGD has done. That combo could be interesting.