it will be a long long time before this tech can actually be used in any applications. taking 8 seconds to map a single command isn't going to fly with the public.
this kind of reminds me of early speach reconition programs that required you to speak each word in order for the software to reconise it. it has been about 10 years now and the algorithums are sophisticated enough to not need that which makes it finally useful. this tech is going to have to take in years worth of input before they can really apply it to any user and get expected results.








