Looks like they are trying to do this with some machine learning.. not just canned recordings.
http://thecorpora.com/blog/

At some point, machine learning and sharing of logic could move very fast.. aka skynet :-)

I just posted a relevant video on computational neuroscience, which might answer some questions.

It's a bit hard to argue with "we think it works this way" when you have a machine that also works "this way".

How a Theory of the Neocortex May Lead to Truly Intelligent Machines

>> ^csnel3:
Yes, rougy, I think I get it.

Fair enough. But you have to admit, it's two different worlds. I'm in Roswell, NM. Aside from the celestial alien problem, we have a crime rate that is disturbingly high, and cops who are sometimes not very thoughtful, because they're underpaid and overworked. Hate to say it, but true.

Ace up there noted that I noticed, as did millions of other people, that a man set himself afire in order to draw attention to an injustice...so I guess his effort was a success...in as much as it drew attention. Did it solve the problem that he gave his life to protest against? Probably not.

That's a pretty big deal, to have that kind of dedication, to believe in something so strongly that you would voluntarily suffer one of the most cruelest fates known to man.

I'm not being cheeky, but the closest thing I can think of is when Data gave his life for Picard.

Vietnam was the prototype to Iraq and Afghanistan. And our country, the USA, is in a lot of trouble because we are watching people rob us blind and we are powerless to stop it because our government is inept and corrupt beyond remediation.

The war machine learned quickly, while the best that we can do is sit around and bitch.

I disagree with anyone that says chips like this won't be useful for solving AI. AI is all about data and machine learning. And most of the operations that deal with that data are trivially parallelizable. I'm perfectly happy with this trend as long as it keeps Moore's law marching along.

I agree with gwiz665 except for the part about (computational) Neural Networks. Neural networks are just one of a million function approximators. Nearest neighbor is my favorite. Support vector machines are the best for many applications. People from brain and cognitive sciences tend to employ neural networks because they find it to be a useful analog to the brain, even if it doesn't actually work well and isn't actually similar to the brain.