rocketpig said:
HappySqurriel said:
ImJustBayuum said:
5. Voice input
We're loath to call this voice recognition, because it's bigger than that: natural language systems such as Apple's Siri are closer to virtual assistants or intelligent software agents than traditional computer voice recognition, not least because you're taking advantage of enormously powerful servers rather than the processing power of your device.
|
Voice input is still not good enough for the average person (people tend to find it impersonal and error prone), and expect people to continue to favour conventional systems for the next several years
|
Have you used Siri yet? It's a HUGE step forward. And due to its server-side nature, it's only going to get smarter over time. Being able to say "wake me up at five in the morning" and have it set an alarm immediately (no further input needed) or say "I want pizza" and have it list you a half-dozen places to call 2 seconds later or any of the other crazy shit I've seen it do is a giant leap forward for voice tech. And it just launched. Where it will be in 12 months is anyone's guess.
Apple moved the bar forward (again). It's enough to make me strongly consider going back to iPhone unless Google steps it up before my contract expires.
|
The technology behind Siri is nothing new and we have been using similar voice recognition software on automated DA services at work for several years ... As long as you keep it to a limited number of well defined requests the results are decent but, even with very good results, automated voice recognition systems rank at or near the bottom of people's prefered ways to deal with anything.
I have no doubts that Siri is great for a limited number of use-cases, but I also have no doubts that Siri is (mostly) just smoke and mirrors hiding the limitations of the system.
