The Cellar - View Single Post

Undertoad · 06-30-2009, 10:38 AM

Well OK, you're closer to convincing me.

The two holdups will be pattern recognition and sunlight. Both are really amazingly difficult. Right now, it can't locate the tip of your index finger of your right hand, unless you have a pen cap on it.

What we'll have next is the ability to ask questions via speech recognition. That system allows you to leave the phone in your pocket (as does the TED prototype, which costs $350 but requires $350 worth of phone). You press a button on your earpiece and ask it: What is my checking account balance? It states the answer through your earpiece.

How do you ask that question of this system? Hold up your empty wallet and draw a question mark?

Here's the prototype that you can use right now. It can only answer certain kinds of questions, and it can only answer them via screen, but you know they'll be branching that out, just as soon as they've used this system to hone their speech recognition with a broad spectrum of users.

Oh and also Compare this to our recent habit of pulling little boxes out of our pocket and punching little buttons on them? We're past buttons now, we're touching the screens of high-quality motion-sensing displays...

06-30-2009, 10:38 AM	#19
Undertoad Radical Centrist Join Date: Jan 2001 Location: Cottage of Prussia Posts: 31,423	Well OK, you're closer to convincing me. The two holdups will be pattern recognition and sunlight. Both are really amazingly difficult. Right now, it can't locate the tip of your index finger of your right hand, unless you have a pen cap on it. What we'll have next is the ability to ask questions via speech recognition. That system allows you to leave the phone in your pocket (as does the TED prototype, which costs $350 but requires $350 worth of phone). You press a button on your earpiece and ask it: What is my checking account balance? It states the answer through your earpiece. How do you ask that question of this system? Hold up your empty wallet and draw a question mark? Here's the prototype that you can use right now. It can only answer certain kinds of questions, and it can only answer them via screen, but you know they'll be branching that out, just as soon as they've used this system to hone their speech recognition with a broad spectrum of users. Oh and also Compare this to our recent habit of pulling little boxes out of our pocket and punching little buttons on them? We're past buttons now, we're touching the screens of high-quality motion-sensing displays...