I think voice recognition will become more popular and reliable. I can also envision devices like Wacom tablets that are also touchscreen keyboards and trackpads. I also know plenty of people (most of them children) who can type at amazing speeds using touchscreens, and find keyboards clunky. It might take awhile (10-20 years), but I think the keyboard is ultimately doomed.
Search my buzz for voice recognition, quality read, though it's just one mans opinion and I don't think I'll agree in a few years time - but that is where we are at the moment. While you're at it look at the augmented reality article, it's pretty cool (but more distant-future).
http://www.google.com/profiles/piratepenguinI believe (unlike others) that voice recognition will become pretty good (and it will take a while) - I can see Google being a major innovator here (and not Apple, Apple are marketing smart, Google are far far more book smart) since a) they've entered (Google voice, automatic youtube transcription which currently
sucks) and more importantly b) they have youtube videos and a community writing subtitles for audio, and they normally do interesting things with what they've got (so the transcription which sucks, will learn!) - also I can only think it is this way that voice recognition will match, and eventually beat, human accuracy. Google may have a plan to do this soon infact, who knows what they're up to. I do know that the problem of voice recognition will require new techniques, and someone at google is probably smart enough to be onto something.
edit: there is still a lot a lot wrong with the idea of using primarily your voice for computer input - I think a lot of people would never use it in public, and it would probably be bad manners (something like that can change), on top of the reasons in the article I referred to. Possibly the main uses for it are, hands-free operation necessary in cars (where the technology is already there or thereabouts already (if you're setting up a call, not if you're writing a message)) and automatic subtitling.
I could be writing my own voice recognition app before the summers out
(it seems like this isn't a big task) At least, this is super high on the list of things I want to have a very good understanding of.