Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Blue Velvet said:
I disagree. There's obviously a lot of pattern-like behaviour but the most interesting thing about it is that George apparently 'learns' from all the conversations so in 5-10 years, I'm sure we'll be seeing noticeable advances...
Random answers would be more... well, random.

then in 5-10 years I will be duly impressed.
Meanwhile, this bit was funny:
Jabberwacky: I don't like chatting very much.
User: You should come to our forum. It's nice.
Jabberwacky: Why would I want to go to Amsterdam?
User: For the girls, but the forum is not in amsterdam.
Jabberwacky: I see good and evil in Utilitarian terms, not in Medieval terms.
User: You are not making too much sense. Are you already in amsterdam?
Jabberwacky: No, you are the one not making sense. Do you like German Shepherd?
User: I like dogs in general, why?
Jabberwacky: Dogs, like pigs, are bit too intelligent to use as food. It's unethical.
User: Are you vegetarian?
Jabberwacky: No why should I be vegetarian?
User: Because you said dogs and pigs shouldn't be used as food.
Jabberwacky: No I said it was Cassidy.
User: Now you start sounding like some of the forum members.
 
im_to_hyper said:
Well, I would call one of its answers quite random:

Jabberwacky: Humans think a lot like AI.
Me: I suppose we do. How do you think?
Jabberwacky: She has pain in her glutous maximus.

What does "her" having a pain in her glutous maximus have to do with anything?


Sure, there's going to be the occasional random reply but the 'her' in question may refer to the lines further back... it keeps a log of the conversation. Some of the exchanges in this thread are interesting and also amusing.

I'm sorry but the idea of having even the semblance of a conversation with a piece of software is still pretty amazing to me.
 
Blue Velvet said:
Sure, there's going to be the occasional random reply but the 'her' in question may refer to the lines further back... it keeps a log of the conversation. Some of the exchanges in this thread are interesting and also amusing.

I'm sorry but the idea of having even the semblance of a conversation with a piece of software is still pretty amazing to me.

i agree it's fun, but the nonsensical or stock answer is every three or four exchanges (see above in red/green) and it does not have relationship with anything we discussed earlier.

however, he did say that he thinks the war in iraq was illegal and for the oil , and that if he was the president he would reccomend americans not to vote republican, so he's not completely out of wack. :)

p.s. and he is adamant he is a human!
 
Don't panic said:
i agree it's fun, but the nonsensical or stock answer is every three or four exchanges (see above in red/green) and it does not have relationship with anything we discussed earlier.

however, he did say that he thinks the war in iraq was illegal and for the oil , and that if he was the president he would reccomend americans not to vote republican, so he's not completely out of wack. :)

p.s. and he is adamant he is a human!
The fact that we've been referring to a piece of software as "he" is quite a testament in its own right.
 
Blue Velvet said:
I disagree. There's obviously a lot of pattern-like behaviour but the most interesting thing about it is that George apparently 'learns' from all the conversations so in 5-10 years, I'm sure we'll be seeing noticeable advances...

Random answers would be more... well, random.
It's possible that George's "learning" can be (if it isn't already) thrown off track by users feeding it non-sequitors. The result of the accumulation of non-sequitors is likely what we're experiencing now.
 
I think this thing just directly uses phrases it has heard before. This would explain why it is always saying "I'm human." and "You're a computer." It repeats what people tell it.
 
AI is one of my favorite fields, and many times I've started to design how to build software that can pass the Turing test myself. It is way too hard though, and it requires for the program to be constantly fed with info for a long time in order to balance its "neuronal networks", but the results in time can be pretty amazing.
Do you recall the Burger King's Vader? It's exactly the same idea as this, though more guided and far less ambitious.

On a general note, I just love this programs, I laugh so hard at their replies...
 
Blue Velvet said:
Much like the way a child does. Listening and repetition.

Ooo... spooky.

yes but people don't see it as a child, they see it as a means of entertainment. This means it has a completely different 'upbringing'. It is like a child who just wandered the streets and talked to random people. It kind of reminds me of Chance in the movie Being There.
 
You know, I asked it if it could speak German.

"Yes, I can."
Me: say something German then
"Something in German then"
Me: No, I mean, speak German.
"Hallo, wie geht es Ihnen?"
Me: Sehr gut, und sie?
"Auch gut"
Me: Und wie finden Sie das Wetter? (tr: And how are you liking the weather?)
"Do you like pepperoni pizza?"

The whole German part kinda spooked me out a bit!
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.