Mostly I am in agreement with you, but I would say that if I have a cell phone, it will not need to be speaker independent. It will be my phone and it can be dedicated to me. In that case, it makes more sense that the device could recognize my speak with an on board process, but then that's just ASR.
As you correctly bring back the focus to MagicTalk, a process of dialog, you see the meat of the problem is not turning voice into machine code, but making the systems perform functions in a powerful and friendly manor.
Still, I have been lead to believe there is a gray area where MagicTalk is separated from Nuance and I would be very pleased if ever I could get a good definition of what is what.
But again, I am in agreement with you that it's not a hardware issue, but one of service and strategic positioning. A good example would be RealAudio. I can name several companies that do similar things, but the lead that Real established has made them the only solution.
I certainly would like to see these reporters calling out comparisons to GM when they do new product reviews.
Still long and still fascinated by the whole thing.
Regards,
Mark |
