And here I thought that the iPhone could do everything except make waffles for breakfast. Turns out it’s deficient in one very surprising way: NO VOICE COMMAND SUPPORT! (And none planned — even through isolated native apps — until at least the 2.0 firmware update…)
So AT&T has unveiled a way for the Mighty "i" (and other mobile handsets, for that matter) to do vox-recog — as a web service.
Using a new version of their classic WATSON speech recognition engine, the software trick (dubbed "Speech Mashups") captures voice commands, parses them remotely, then sends them back to the device in a website- or software-understandable form. Here’s a demo:
The drawbacks? It’s a very limited solution that requires a sufficiently fast connection to AT&T’s servers, and if the app doesn’t support web code (as a lot of Apple’s don’t), you’re SOL. Still, AT&T envisions Speech Mashups implemented on everything from IP-based TV boxes to smartphones and Blackberries, but getting it up and running for the iPhone appears to be their priority. (Maybe next week for the waffel app…)