CADIE - the world's first AI - announced by Google

[Hissingsid]

On April 1st Google announced CADIE Cognitive Autoheuristic Distributed-Intelligence Entity.

What do you think?

Cheers

Sid

;-)

[tedster]

Here's a link to start exploring:

[google.com...]

[LifeinAsia]

Also mentioned here [webmasterworld.com]

[edited by: LifeinAsia at 4:01 pm (utc) on April 1, 2009]

[tedster]

Looks like they're using Cadie for gmail, too - a new feature called Autopilot that writes replies for you.

The more Gmail messages Autopilot can sample, the better. With fewer than 100 messages, there may not be enough data to calibrate Autopilot effectively. You can adjust tone, typo propensity, and preferred punctuation from the Autopilot tab under Settings.

[mail.google.com...]

[tedster]

And now Matt Cutts is saying that CADIE is to blame for bad SERPs - LOL!

[mattcutts.com...]

[Lowkei]

I checked out the CADIE youtube channel and many people thought CADIE is an April Fool joke.. :D

[tedster]

That's a pretty good bet, Lowkei.

[tonyvf]

many people thought CADIE is an April Fool joke

And it's not? It sure is. Every year G has an April Fools joke, last year it was the partnership with Virgin to colonize Mars...

[kidder]

I seem to remember a Google dating one as well... Ahhh April is such a good month to stir up the webmasters..

[Hissingsid]

Hi Tedster,

Looks like Google and I April's fooled you.

Best wishes

Sid

[tedster]

Not a chance - I was just playing along, Sid. I have no expectation that AI is going to appear, in my lifetime or ever. Consciousness is more than data.

[whitenight]

I have no expectation that AI is going to appear, in my lifetime or ever. Consciousness is more than data.

<gasp>
Tedster, you've changed since i haven't been here to play bad cop. ;)

[JoeSinkwitz]

Intelligence is nothing more than the ability to learn; learning is nothing more than pattern recognition; pattern recognition is nothing more than mathematics...it will be done, just a matter of when

[thinker]

AI will always have to be controlled and programmed by people with consciousness, morals, and human standards.

If AI was to learn morals and evaluate human standards it would theoretically only be able to do so based on huge amounts of data that was produced by humans with consciousness.

Any computer can copy, paste, analyze and select patterns via programming - I doubt this could be considered intelligent, no matter how sophisticated the software.

AI will be an April fool's joke until the end of time.

[tonyvf]

AI will be an April fool's joke until the end of time.

It all depends on what we consider AI to be. What's human intelligence? Where does it come from? I believe it's a spark of God in ourselves. In that case, AI is not possible.
However, simulating human behavior is not impossible at all.
Having a simulated "personality" will not be that difficult as well, also based on average sampling of interactions (or even based on an individual).
We saw that happen with Kasparov and Deep Blue in 1997/1998. Until now Kasparov and others believe IBM cheated, that is not possible for a computer to have human "behavior" like Deep Blue had.
Anyway, it will be interesting to watch a computer asking "why am I here? What's the reason of being 'alive'? Why do I have to work if I'm going to 'die'?" and then just shutdowns itself :)