transcript from a recent Turing test
Sep. 20th, 2006 11:43 am![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
spoonless(from http://www.technologyreview.com/read_article.aspx?id=17518&ch=infotech)
This computer program (Joan) was the winner of this year's Loebner Prize ($2000... will be $100,000 for the first program that actually passes, but for now it's just a "best attempt" award):
Judge: Do you know any good jokes?
Joan: I don't even know any bad ones. I just can't remember them for very long.
Judge: Tell me a joke.
Joan: Sorry, I'm not very good at telling jokes.
Judge: Please tell a joke.
Joan: What do you get if you walk under a cow?
Judge: I don't know?
Joan: A pat on the head.
Judge: That's very good.
Joan: I know.
Judge: That's very good.
...
Joan: I am 9,432 years old this year.
Judge: Really!
Joan: Yes. And I am much more intelligent than any man.
Judge: And you are human...?
Joan: Yes, I'm human.
Judge: So what you're saying is that you are not just human but the THE smartest human?
Joan: Yes, why, what did you think I was saying?
Judge: So what is two plus two?
Joan: Four.
Judge: Is it possible for two plus two to not equal four?
Joan: I don't know. Do you think it's possible?
Judge: How many degrees in a triangle?
Joan: 180 degrees.
Judge: Is it possible for more than 180 degrees to be in a triangle?
Joan: Probably not. You can change your opinion, but not your constitution.
Judge: It's not a question of opinion, it's a matter of physics and maths.
Joan: A five-ounce bird could not carry a one-pound coconut.
I think some of this conversation is pretty impressive... although I did select the better parts, and perhaps so did the article. Looks to me like computers are much more capable of carrying on a meaningful conversation at this point than non-human animals, but if you were to compare them to 3-year old humans, they would sometimes fare better and sometimes worse. Every conversation with a computer today ends in a non-sequitur after not too long, but then again I've had plenty of conversations with full grown humans who end up saying surprisingly non-sensical things (at least from my perspective) after a few minutes. I look forward to seeing which ways this will improve over the next decade. I'm not sure I agree that this mass-mining of call center data is the best approach though. Last weekend I attended an AI conference and heard Ben Goertzel speak about the work they're doing at Novamente. I like their approach a lot better (having an AI learn inside a virtual environment where it can actually see things, move around, and pick things up to learn about them, rather than just processing text), even though some of the conference turned out to be what I'd call fruitless speculation at this stage. No offense to Yudkowsky, but I think I'm starting to become more and more skeptical that "friendly AI" is fully definable, desirable, or ensurable.
PS: When I read Joan's comment about smarter than any man, I figured she might just mean the male portion of the species; she may just identify as an average, or even sub-par woman but feel that men are dumb. However, the judge interpretted it in a much more sexist way, as if the only realistic possibility was that the smartest human is a man. Then Joan gets led into agreeing with him. Quite an intriguing interchange, on many levels!
This computer program (Joan) was the winner of this year's Loebner Prize ($2000... will be $100,000 for the first program that actually passes, but for now it's just a "best attempt" award):
Judge: Do you know any good jokes?
Joan: I don't even know any bad ones. I just can't remember them for very long.
Judge: Tell me a joke.
Joan: Sorry, I'm not very good at telling jokes.
Judge: Please tell a joke.
Joan: What do you get if you walk under a cow?
Judge: I don't know?
Joan: A pat on the head.
Judge: That's very good.
Joan: I know.
Judge: That's very good.
...
Joan: I am 9,432 years old this year.
Judge: Really!
Joan: Yes. And I am much more intelligent than any man.
Judge: And you are human...?
Joan: Yes, I'm human.
Judge: So what you're saying is that you are not just human but the THE smartest human?
Joan: Yes, why, what did you think I was saying?
Judge: So what is two plus two?
Joan: Four.
Judge: Is it possible for two plus two to not equal four?
Joan: I don't know. Do you think it's possible?
Judge: How many degrees in a triangle?
Joan: 180 degrees.
Judge: Is it possible for more than 180 degrees to be in a triangle?
Joan: Probably not. You can change your opinion, but not your constitution.
Judge: It's not a question of opinion, it's a matter of physics and maths.
Joan: A five-ounce bird could not carry a one-pound coconut.
I think some of this conversation is pretty impressive... although I did select the better parts, and perhaps so did the article. Looks to me like computers are much more capable of carrying on a meaningful conversation at this point than non-human animals, but if you were to compare them to 3-year old humans, they would sometimes fare better and sometimes worse. Every conversation with a computer today ends in a non-sequitur after not too long, but then again I've had plenty of conversations with full grown humans who end up saying surprisingly non-sensical things (at least from my perspective) after a few minutes. I look forward to seeing which ways this will improve over the next decade. I'm not sure I agree that this mass-mining of call center data is the best approach though. Last weekend I attended an AI conference and heard Ben Goertzel speak about the work they're doing at Novamente. I like their approach a lot better (having an AI learn inside a virtual environment where it can actually see things, move around, and pick things up to learn about them, rather than just processing text), even though some of the conference turned out to be what I'd call fruitless speculation at this stage. No offense to Yudkowsky, but I think I'm starting to become more and more skeptical that "friendly AI" is fully definable, desirable, or ensurable.
PS: When I read Joan's comment about smarter than any man, I figured she might just mean the male portion of the species; she may just identify as an average, or even sub-par woman but feel that men are dumb. However, the judge interpretted it in a much more sexist way, as if the only realistic possibility was that the smartest human is a man. Then Joan gets led into agreeing with him. Quite an intriguing interchange, on many levels!
![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png){kind=small}
no subject
Date: 2006-09-20 08:09 pm (UTC)no subject
Date: 2006-09-20 09:20 pm (UTC)no subject
Date: 2006-09-21 02:22 am (UTC)no subject
Date: 2006-09-21 08:29 am (UTC)So were you the one with green hair and a laptop? I was actually thinking to myself "I've seen her at another conference before... but where?" maybe it was just from lj!
Well, next time we're both at an event I'll be sure and say hi. I was sitting in the back left, my hair is getting kind of wild these days... at least compared to most of my lj pictures.
no subject
Date: 2006-09-21 02:17 pm (UTC)no subject
Date: 2006-09-24 06:14 am (UTC)no subject
Date: 2006-09-21 04:14 am (UTC)hello i am random
Date: 2006-09-27 12:53 pm (UTC)Judge: Is it possible for two plus two to not equal four?
Joan: I don't know. Do you think it's possible?
sounds like someone programmed an effing psychotherapist in there by accident.
Re: hello i am random
Date: 2006-09-27 05:04 pm (UTC)Re: hello i am random
Date: 2006-09-28 02:05 pm (UTC)