I don't think voice recognition technology will be really viable for decades. While it can be useful for those with limited or no use of their upper extremities, it's just too complicated. You have to train the software because of accents, unique pronunciations, etc. What happens when two or more people are talking, how does the device distinguish the commands it should obey? If you train it to only respond to a particular voice, only one person could use it. They've been working on this for decades and there are really no significant mass appeal products out there.