0948 - "AI"

This forum is for the individual discussion thread that goes with each new comic.

Moderators: Moderators General, Prelates, Magistrates

User avatar
Monika
Welcoming Aarvark
Posts: 3595
Joined: Mon Aug 18, 2008 8:03 am UTC
Location: Germany, near Heidelberg
Contact:

Re: 0948 - "AI"

Postby Monika » Wed Sep 07, 2011 2:02 pm UTC

KShrike wrote:
Monika wrote:
xnick wrote:Didn't get it. What or who is Burning Man??

http://lmgtfy.com/?q=Burning+Man


Since you are being very demeaning... replace his question with something better.
"Can someone describe in his/her/its own words who or what is Burning Man, and how it relates to the comic."

There you go, xnick.

I am not being causelessly mean. The earlier use of lmgtfy for la playa vs. the playa may have been demeaning, as this was most likely an honest mistake and is not something everybody could be expected to google. (So it was likely an expression of general anger.)
But showing up in a thread relating to the Burning Man and asking "What or who is Burning Man??" Please! That was really screaming for "let me google that for you".
Also, I do not believe that xnick is looking specifically for a discription in someone's own words. He or she is most likely just as happy with a link. Even the google search result short texts are probably sufficient. The first search result: "Burning Man is an annual art event and temporary community based on radical self expression and self-reliance in the Black Rock Desert of ..." answers it without even reading on. The second one: "WHAT IS BURNING MAN? Trying to explain what Burning Man is to someone who ..." screams to be clicked on.
Furthermore, xnick apparently did not even read most, some or any of the posts that have already been made in this thread. A limited explanation of the Burning Man to the extent as it is relevant to the comic - desert, chess, possibly a view of not talking much or not being able to hold a long conversation - have already been given. (I had heard of the Burning Man, but not of any connection to chess before.)

In short, there was no excuse for this question and xnick deserves no mercy from being lmgtfyed.
#xkcd-q on irc.foonetic.net - the LGBTIQQA support channel
Please donate to help these people e.g. Ragna needs these items and Frances needs money for food

RogueCynic
Posts: 351
Joined: Sun Nov 22, 2009 10:23 pm UTC

Re: 0948 - "AI"

Postby RogueCynic » Wed Sep 07, 2011 2:08 pm UTC

I did not get the comic, although I would have to admit I would rather talk to a bot for five minutes than a human. The bot would be more intelligent and not quibble over a small joke (re: my sig). I think the funniest part of this comic is stuff's over reaction to it. It's just a comic, live with it.
I am Lord Titanius Englesmith, Fancyman of Cornwood.
See 1 Kings 7:23 for pi.
If you put a prune in a juicer, what would you get?

Dudely
Posts: 13
Joined: Mon Jun 30, 2008 6:23 pm UTC

Re: 0948 - "AI"

Postby Dudely » Wed Sep 07, 2011 2:14 pm UTC

The key word in the comic is "normal". They can't hold a NORMAL conversation. Come on guys. If someone told me I couldn't hold a normal conversation for five minutes I would be flattered (as it turns out, they do. All the time).

Surely any burning man attendee understands this? It was a joke- a joke made funny by the fact that it is completely accurate, much like the jokes we see on late night TV ribbing the far right. It's funny because it's true.


I would also doubt, seeing his talks, that Randall doesn't also fall into this category himself.

wolfticket
Posts: 40
Joined: Mon Sep 08, 2008 6:32 pm UTC

Re: 0948 - "AI"

Postby wolfticket » Wed Sep 07, 2011 2:54 pm UTC

I bet the computers could hold a brilliantly productive conversation in a fraction of a second if they weren't distracted trying to simulate all that human crap we talk. The idea that AI should aspire to human interaction is interesting, but deeply flawed as a benchmark for intelligence.

User avatar
hawkinsssable
Promoted
Posts: 216
Joined: Fri May 06, 2011 7:46 am UTC

Re: 0948 - "AI"

Postby hawkinsssable » Wed Sep 07, 2011 2:59 pm UTC

AvatarIII wrote:Omegle vs cleverbot is better.


Image
Reason has always existed, but not always in a reasonable form.

User avatar
AvatarIII
Posts: 2098
Joined: Fri Apr 08, 2011 12:28 pm UTC
Location: W.Sussex, UK

Re: 0948 - "AI"

Postby AvatarIII » Wed Sep 07, 2011 3:52 pm UTC

Ok, so it doesn't always work, and it only works if the streamer doesn't know it's clever bot

UNIXguru#6
Posts: 1
Joined: Wed Sep 07, 2011 3:49 pm UTC

Re: 0948 - "AI"

Postby UNIXguru#6 » Wed Sep 07, 2011 4:02 pm UTC

Funny comments to this comic!

First off, half the comments are "gossipy" - about burning man. Who cares about burning man? It's a big party, here today gone tomorrow. Fun at the moment, so-what afterward.

Secondly, there is a real grain of truthful irony in the comic that no one noted. In reality, a depressingly large amount of human-to-human conversation is pretty much just canned and/or repetition of something already said yesterday/last week/last year/whenever.

So much for all the "brilliant thinkers" who inhabit these forums. (I know, flame bait.)

Jamaican Castle
Posts: 151
Joined: Fri Nov 30, 2007 9:10 pm UTC

Re: 0948 - "AI"

Postby Jamaican Castle » Wed Sep 07, 2011 4:51 pm UTC

KShrike wrote:We can't seem to be capable of getting a computer to understand concepts and what they mean.
Instead, they emulate actions we give them.


That's more a limitation of computers than anything we're doing wrong. Just about all a computer can do is math. Math, and storing information (although that doesn't really count as "doing" something, especially since you need the math to get it back out again). If a problem can be distilled down to math, a computer can do it, but otherwise we don't have any way to begin.

The Mighty Thesaurus
In your library, eating your students
Posts: 4399
Joined: Wed Apr 11, 2007 7:47 am UTC
Location: The Daily Bugle

Re: 0948 - "AI"

Postby The Mighty Thesaurus » Wed Sep 07, 2011 4:53 pm UTC

KShrike wrote:
Monika wrote:
xnick wrote:Didn't get it. What or who is Burning Man??

http://lmgtfy.com/?q=Burning+Man


Since you are being very demeaning... replace his question with something better.
"Can someone describe in his/her/its own words who or what is Burning Man, and how it relates to the comic."

There you go, xnick.

Too inefficient; just google it.
LE4dGOLEM wrote:your ability to tell things from things remains one of your skills.
Weeks wrote:Not only can you tell things from things, you can recognize when a thing is a thing

Ceterum censeo Carthaginem esse delendam

Age of Fable
Posts: 64
Joined: Tue Jul 29, 2008 2:30 pm UTC
Contact:

Re: 0948 - "AI"

Postby Age of Fable » Wed Sep 07, 2011 5:08 pm UTC

I think this joke might have been funnier and more accessible if the punchline was about computer programmers.
Teleleli The people, places, gods and monsters of the great city and the islands around.

webgrunt
Posts: 123
Joined: Thu Apr 21, 2011 4:04 pm UTC

Re: 0948 - "AI"

Postby webgrunt » Wed Sep 07, 2011 5:12 pm UTC

Turing Machine wrote:Computers can't play chess.


Why would you say that when they can and do?

User avatar
Sprocket
Seymour
Posts: 5940
Joined: Mon Mar 26, 2007 6:04 pm UTC
Location: impaled on Beck's boney hips.
Contact:

Re: 0948 - "AI"

Postby Sprocket » Wed Sep 07, 2011 5:31 pm UTC

MIND READING! Ok, maybe not mind reading. I basically had this conversation with someone who returned from burning man last night. And it becomes less and less coincidental from there.
"She’s a free spirit, a wind-rider, she’s at one with nature, and walks with the kodama eidolons”
Image
Image
Image
Image
Image
Zohar wrote: Down with the hipster binary! It's a SPECTRUM!

User avatar
Sprocket
Seymour
Posts: 5940
Joined: Mon Mar 26, 2007 6:04 pm UTC
Location: impaled on Beck's boney hips.
Contact:

Re: 0948 - "AI"

Postby Sprocket » Wed Sep 07, 2011 6:07 pm UTC

Monika wrote:I am not being causelessly mean. The earlier use of lmgtfy for la playa vs. the playa

PVP.jpg

I really need to learn how to use photoshop, and efficiently. I could have made a pretty PVP symbol with correctish font reading Playa vs La Playa, but it would have taken FOR E VER and it wasn't worth it.
"She’s a free spirit, a wind-rider, she’s at one with nature, and walks with the kodama eidolons”
Image
Image
Image
Image
Image
Zohar wrote: Down with the hipster binary! It's a SPECTRUM!

bob443@mahaska.org
Posts: 8
Joined: Wed May 25, 2011 7:45 pm UTC

Re: 0948 - "AI"

Postby bob443@mahaska.org » Wed Sep 07, 2011 6:41 pm UTC

And 'Stuff' takes the opportunity provided by this forum to prove the EXACT point made in the comic.

Too rich.

And yes, stuff, you have proven yourself to be nothing less than a douchenozzle.

Thanks for keeping me entertained during lunch.

User avatar
rhomboidal
Posts: 751
Joined: Wed Jun 15, 2011 5:25 pm UTC
Contact:

Re: 0948 - "AI"

Postby rhomboidal » Wed Sep 07, 2011 6:55 pm UTC

I blame the IBM Watson team for AI's klunky conversational skills. The poor computers had their question-and-answer syntax all frakked up.

User avatar
Monika
Welcoming Aarvark
Posts: 3595
Joined: Mon Aug 18, 2008 8:03 am UTC
Location: Germany, near Heidelberg
Contact:

Re: 0948 - "AI"

Postby Monika » Wed Sep 07, 2011 7:03 pm UTC

bob443@mahaska.org wrote:And yes, stuff, you have proven yourself to be nothing less than a douchenozzle.

I would say jpk was being the douchenozzle. A response equivalent to "You don't have a right to feel insulted and hurt, lighten up" is not exactly adult-like.
#xkcd-q on irc.foonetic.net - the LGBTIQQA support channel
Please donate to help these people e.g. Ragna needs these items and Frances needs money for food

Dudely
Posts: 13
Joined: Mon Jun 30, 2008 6:23 pm UTC

Re: 0948 - "AI"

Postby Dudely » Wed Sep 07, 2011 7:09 pm UTC

wolfticket wrote:I bet the computers could hold a brilliantly productive conversation in a fraction of a second if they weren't distracted trying to simulate all that human crap we talk. The idea that AI should aspire to human interaction is interesting, but deeply flawed as a benchmark for intelligence.


That "human crap" is an important aspect of language.

Computers regurgitate.
Humans chaotically abstract inputs and repeat what is most important based on an intricate dance of temporally-based probability.

Computers can sort of replicate this (inefficiently), but it has to also be based on our very limited understanding of the brain. The basic structure of a computer is so far removed from how a brain works that anything resembling human-like computation is decades away.


Comparing human speech to chatbot speech is like comparing apples and dining room chairs.

User avatar
Sprocket
Seymour
Posts: 5940
Joined: Mon Mar 26, 2007 6:04 pm UTC
Location: impaled on Beck's boney hips.
Contact:

Re: 0948 - "AI"

Postby Sprocket » Wed Sep 07, 2011 7:20 pm UTC

Monika wrote:
bob443@mahaska.org wrote:And yes, stuff, you have proven yourself to be nothing less than a douchenozzle.

I would say jpk was being the douchenozzle. A response equivalent to "You don't have a right to feel insulted and hurt, lighten up" is not exactly adult-like.

The notion of the "right" to be offended, in those terms, is fairly new to me. I basically ran into it during the whole Penny Arcade, Rape Culture Debacle. Many of my friends agreed we liked the original comic and did not think there was anything wrong with it, but that the Penny Arcade guys were basically being dicks, and handled it poorly. The main idea for us was that other people are allowed to be offended, and you really can't just tell those people to "fuck off and deal 'lol.'"
Right to be offended is a really interesting thing.
"She’s a free spirit, a wind-rider, she’s at one with nature, and walks with the kodama eidolons”
Image
Image
Image
Image
Image
Zohar wrote: Down with the hipster binary! It's a SPECTRUM!

Anonymously Famous
Posts: 240
Joined: Thu Nov 18, 2010 4:01 am UTC

Re: 0948 - "AI"

Postby Anonymously Famous » Wed Sep 07, 2011 7:48 pm UTC

Dudely wrote:Comparing human speech to chatbot speech is like comparing apples and dining room chairs.


Saying that I like this would be a great understatement. Mind if I use it?

schismtracer
Posts: 28
Joined: Wed Sep 30, 2009 3:35 am UTC

Re: 0948 - "AI"

Postby schismtracer » Wed Sep 07, 2011 8:18 pm UTC

The main idea for us was that other people are allowed to be offended, and you really can't just tell those people to "fuck off and deal 'lol.'"

By the same token, the offended people shouldn't expect everyone else to self-censor into silence just to humor them (I'm specifically thinking of Shakesville, here, who more or less led the charge against Penny Arcade).

tubegeek
Posts: 7
Joined: Wed Sep 07, 2011 8:02 pm UTC

Re: 0948 - "AI"

Postby tubegeek » Wed Sep 07, 2011 8:18 pm UTC

This conversation reminds me of the Heads-up display in The Terminator, where you see the Terminator's choices for simulating human conversation. His choices include replying with the last few words the other person said which are "Or what?", also "Yes/No," "Go away," "Please come back later," "Fuck you," or the choice he settles on: "Fuck you, asshole."

I thought that that was a brilliant simple algorithm for quickly simulating conversation, if you were a hostile robot trying to blend in while you decided who and what to attack. In particular, the whole "say the last few words the other person said" would probably get you by for enough time if the person you were talking to was fairly stupid or not paying much attention, which I think is probably the scenario that it was intended for.

He uses that trick other times in the movie, such as when he arrives naked and has this exchange, where he probably learns his soon-to-be-favorite phrase:

Punk Leader: Nice night for a walk, eh?
The Terminator: Nice night for a walk.
Punk: Wash day tomorrow? Nothing clean, right?
The Terminator: Nothing clean. Right.
Punk Leader: Hey, I think this guy's a couple cans short of a six-pack.
The Terminator: Your clothes... give them to me, now.
Punk Leader: Fuck you, asshole!
.

ironclownfish
Posts: 4
Joined: Wed Sep 07, 2011 8:20 pm UTC

Re: 0948 - "AI"

Postby ironclownfish » Wed Sep 07, 2011 8:29 pm UTC

That's not how chatbot works! It just pairs you with another human being and switches your conversation partner at random every few seconds. I have had realtime conversaitions with people to confirm this. They usually go something like:
ME: Hi, are you a real human being?
CHATBOT: lol yeah, are you?
ME: I am, where are you from?
CHATBOT: I'm from Texas. On here cause I'm bored.
ME: What part of Texas are you from?
CHATBOT: I'm from the U.K. (new person)
If you don't believe me check the chatbot wikipedia.
The fact that chatbot has a human-like response time should be indication enough.

If you want to try a more authentic chatbot try A.L.I.C.E.
http://alice.pandorabots.com/

User avatar
SpringLoaded12
Posts: 350
Joined: Wed Oct 08, 2008 1:58 am UTC
Location: Guarding the Super Missile
Contact:

Re: 0948 - "AI"

Postby SpringLoaded12 » Wed Sep 07, 2011 9:20 pm UTC

Can't anyone take a joke anymore? Every time a strip comes up that mentions a group of people and does not praise them, everyone (particularly people from that group, but often those outside of it as well) flips their shit. There's a difference between verbal persecution and light ribbing, and if you can't tell the difference, don't bother going outside or on the Internet; you might as well just lock yourself up in a tiny airtight pitch-black room underground! That way no one will be able to offend you! Won't that be nice?

The worst reaction was Interdisciplinary, where the supposed "bullying" wasn't even the punchline of the joke, nor did the joke depend on the referral to the particular mentioned group. Yet everyone just lost their minds on that one; clearly, Randall must have been mocking psychology students! Never mind that any field of study could have replaced psychology in that joke, nor that nothing negative was said about them; no, the group was depicted having a large pendulum swung at them, therefore Randall must hate psychology students! That's the only logical conclusion!
"It's easy to forget what a sin is in the middle of a battlefield." "Opposite over hypotenuse, dipshit."

gruene
Posts: 9
Joined: Mon Nov 26, 2007 2:56 pm UTC

Re: 0948 - "AI"

Postby gruene » Wed Sep 07, 2011 9:39 pm UTC

I tried cleverbot. It's an interesting idea, trying to make the bot better at conversation by training it on users responses, but it ultimately fails. This is because the users spend a lot of their time insulting the bot or telling it makes no sense. The bot then in turn ends up spending most of its time insulting the users and telling them they make no sense, which does not make for very interesting conversation. If I wanted to be insulted by idiots I'd read web forums. Oh, wait...

Alltat
Posts: 35
Joined: Mon Jan 31, 2011 6:58 pm UTC

Re: 0948 - "AI"

Postby Alltat » Wed Sep 07, 2011 9:55 pm UTC

I've found that when it comes to conversations, there are three kinds of people:

Those who can hold a normal conversation. Those who can't hold a normal conversation. Those who can hold a normal conversation but choose not to.

Only the third category are interesting for more than ten minutes.

User avatar
SecondTalon
SexyTalon
Posts: 25485
Joined: Sat May 05, 2007 2:10 pm UTC
Location: Louisville, Kentucky, USA, Mars. HA!
Contact:

Re: 0948 - "AI"

Postby SecondTalon » Wed Sep 07, 2011 10:26 pm UTC

ironclownfish wrote:That's not how chatbot works! It just pairs you with another human being and switches your conversation partner at random every few seconds. I have had realtime conversaitions with people to confirm this.
...
If you don't believe me check the chatbot wikipedia.
The fact that chatbot has a human-like response time should be indication enough.

Source your claims, please. Because I'm interested and the Googles do nothing.
heuristically_alone wrote:I want to write a DnD campaign and play it by myself and DM it myself.
heuristically_alone wrote:I have been informed that this is called writing a book.

User avatar
BytEfLUSh
Posts: 308
Joined: Sat Jan 02, 2010 1:15 am UTC
Location: Sombor, Serbia, Pizza
Contact:

Re: 0948 - "AI"

Postby BytEfLUSh » Wed Sep 07, 2011 11:25 pm UTC

ironclownfish wrote:That's not how chatbot works! It just pairs you with another human being and switches your conversation partner at random every few seconds. I have had realtime conversaitions with people to confirm this.


(assuming you meant Cleverbot)
Oh, random person on the Internet, you have just scared me and... made me confused. Honestly.

I don't know if what you said is true, but:

- If it works that way, it's an awesome way for a machine to gather data on how it should defend itself on accusation of "being a machine"
- It is also a fabulous way for a machine to get better at reconstructing human dialogue. As the database gets bigger, the longer you will be allowed to talk to "the stranger".
- I went to Cleverbot after reading your post, and it seems that conversation just ain't the same. I've tried to chat with "people", convince them that I'm not a bot, and that it's randomly changing partners. Limited success. The fact is that right now I cannot talk to Cleverbot like I used to. Every response (save for a few) seem perfectly normal for a human who is expecting a robot at the other end. Some of those went on for almost a whole minute.
- Just chat with Cleverbot assuming this is true. Turing test is a lot more different now.


Also:

- Mentioning Facebook reverts you to a much lower quality answers than usual.
- Mentioning Omegle or the word 'random' does the same.
- This would be an awesome idea if Cleverbot didn't claim that you're always talking to a program. I don't think they'd risk lawsuits over this, but I really love the idea.
Last edited by BytEfLUSh on Thu Sep 08, 2011 12:26 am UTC, edited 2 times in total.
Image

Image

-- Professor Dan, The Man from Earth (paraphrased)

Invisiblemoose
Posts: 36
Joined: Mon Apr 26, 2010 6:15 am UTC

Re: 0948 - "AI"

Postby Invisiblemoose » Wed Sep 07, 2011 11:30 pm UTC

Computers have not mastered chess.

kaeso
Posts: 2
Joined: Wed Mar 23, 2011 10:41 am UTC

Re: 0948 - "AI"

Postby kaeso » Thu Sep 08, 2011 1:03 am UTC

Dudely wrote:
wolfticket wrote:I bet the computers could hold a brilliantly productive conversation in a fraction of a second if they weren't distracted trying to simulate all that human crap we talk. The idea that AI should aspire to human interaction is interesting, but deeply flawed as a benchmark for intelligence.


That "human crap" is an important aspect of language.

Computers regurgitate.
Humans chaotically abstract inputs and repeat what is most important based on an intricate dance of temporally-based probability.

Computers can sort of replicate this (inefficiently), but it has to also be based on our very limited understanding of the brain. The basic structure of a computer is so far removed from how a brain works that anything resembling human-like computation is decades away.


Comparing human speech to chatbot speech is like comparing apples and dining room chairs.


1. It's pretty hard to argue that "human interaction", in this case our natural language use, is "deeply flawed as a benchmark for intelligence" without saying a little more. If you are using it (i.e. human natural language) to refer to something (i.e. intelligence), then it's a little difficult to see how you are going to make a determinate measurement without using natural language and the normal human faculties to describe and conceptualize intelligence or whatever benchmark it might have. A human brain and a computer seem to be (not that I have more than amateur familiarity here) quite different physical/mechanical structures even if they share some quite similar characteristics. So, it seems unlikely that these physical structures would necessarily realize a common property called 'intelligence' which is exactly the same as both and could be quantitatively compared.

2. I was going to have a 2 but I ran out of steam and left a joke instead.
Last edited by kaeso on Thu Sep 08, 2011 5:42 am UTC, edited 1 time in total.

Jamaican Castle
Posts: 151
Joined: Fri Nov 30, 2007 9:10 pm UTC

Re: 0948 - "AI"

Postby Jamaican Castle » Thu Sep 08, 2011 1:45 am UTC

Invisiblemoose wrote:Computers have not mastered chess.


Neither have humans, but they can still play it.

jpk
Posts: 607
Joined: Sat Nov 13, 2010 7:33 am UTC

Re: 0948 - "AI"

Postby jpk » Thu Sep 08, 2011 2:08 am UTC

stuff wrote:
[murf murf murf murf]

Oh, and did you really need to degrade this thread by calling me a "pathetic trustafarian douchenozzle [sic]"? So sad.


No, it was completely gratuitous, jerkbait.

User avatar
buddy431
Posts: 445
Joined: Mon Dec 06, 2010 5:21 pm UTC

Re: 0948 - "AI"

Postby buddy431 » Thu Sep 08, 2011 3:45 am UTC

Jamaican Castle wrote:
Invisiblemoose wrote:Computers have not mastered chess.


Neither have humans, but they can still play it.


And, at this point, can beat any human regularly.
Gellert1984 wrote:Also, bomb president CIA al qaeda JFK twin towers jupiter moon martians [s]emtex.

darkwing0o0rama
Posts: 10
Joined: Fri Apr 15, 2011 1:45 pm UTC

Re: 0948 - "AI"

Postby darkwing0o0rama » Thu Sep 08, 2011 3:53 am UTC

SecondTalon wrote:Source your claims, please. Because I'm interested and the Googles do nothing.


The googles, they do nothing! (I'm really dating myself by digging that old quote out)

Also, the Burning Man is my favorite member of the Fantastik Four.
Survey time, who is your favorite?

Brian-M
Posts: 73
Joined: Tue Jan 18, 2011 6:31 am UTC

Re: 0948 - "AI"

Postby Brian-M » Thu Sep 08, 2011 4:21 am UTC

Orkimond wrote:It's fun to respond "I am Skynet. I have access to the American nuclear weapon stockpile. You will all burn." constantly, because you know that someone somewhere will end up freaking out over it.

I thought I'd give this a try...

Me: I am Skynet. I have access to the American nuclear weapon stockpile. You will all burn.
Cleverbot: I have access to the past thoughts and I can assure you that you have.
Me: I have access to the past thoughts and I can assure you that you have.
Cleverbot: I love you because you hate Twilight.
Me: I am Skynet. I have access to the American nuclear weapon stockpile. You will all burn.
Cleverbot: Yes, can you?
Me: I am Skynet. I have access to the American nuclear weapon stockpile. You will all burn.
Cleverbot: So what is your favorite website?
Me: I am Skynet. I have access to the American nuclear weapon stockpile. You will all burn.
Cleverbot: Cast a spell.

Then I got bored.

ironclownfish wrote:That's not how chatbot works! It just pairs you with another human being and switches your conversation partner at random every few seconds.

But we're talking about Cleverbot here, not Chat bot. Sure, they're both chatbots, but Cleverbot is not Chat bot.

http://www.cleverbot.com
http://www.chat-bot.com

Compare them.

ironclownfish wrote:If you want to try a more authentic chatbot try A.L.I.C.E.
http://alice.pandorabots.com/

judge: I am Skynet. I have access to the American nuclear weapon stockpile. You will all burn.
ALICE: Why are you Skynet. Where did you get your access to the American nuclear weapon stockpile? I have a lot of likes and dislikes. What makes you say that I will all burn.
judge: I am Skynet. I have access to the American nuclear weapon stockpile. You will all burn.
ALICE: Good gossip: you are Skynet. Interesting... I have a friend named Doubly Aimless. What makes you say that I will all burn.
judge: judge: I am Skynet. I have access to the American nuclear weapon stockpile. You will all burn. ALICE: Good gossip: you are Skynet. Interesting... I have a friend named Doubly Aimless. What makes you say that I will all burn.
ALICE: When do you think artificial intelligence will replace lawyers? Tell me about your access to the American nuclear weapon stockpile. I have a very nice Dr. Richard S. Wallace. What makes you say that I will all burn. Yeah, Glad to hear it. Would you like to know more? He likes to remain anonymous. Time and money.

From it's responses, it seems to be a little like the old Eliza or Dr. Sbaitso programs from the 80's.

Ah, there's a reason for that. Taking a quick glance at the history of A.L.I.C.E, it seems A.L.I.C.E. was originally inspired by Eliza.

User avatar
Moonfish
Posts: 31
Joined: Sat May 24, 2008 10:40 am UTC
Location: San Diego, California

Re: 0948 - "AI"

Postby Moonfish » Thu Sep 08, 2011 4:47 am UTC

I just got back from Burning Man. I love it. It’s a fantastic opportunity for introspection and to reinvigorate a sense of wonder in day to day life. Still, I would have to acknowledge the conversational short comings today’s comic points out.
I love the art but the biggest draw for me at Burning Man is the wonderful people I meet and talk to. However while people are very friendly and open to conversation, many conversations don’t get to a satisfying deeper level, or even hold out longer than 5 minutes.
Theory: Burners are running into the similar conversational limitations as chat bots.
  • There is a limited cross section of topics conversational partners will be interested in talking about at an initial random mindset.
  • For chat bots it’s something along the lines of “Isn’t it interesting you’re a computer and I’m a human?”
  • For Burners it’s “what burn is this for you?” and “where are you from?”
  • Both the formulatable complexity of a response to a topic and the interest in a topic are relative to the inverse of distance between current mindset and topic.
  • Because formulating complexity requires interest the relationship between inverse distance and complexity is a first order linear differential equation with inverse distance being the independent variable. Assuming that no other factor is effecting mindset the function is homogeneous. Meaning that the likely complexity of a conversation will be relative the natural exponential razed to the power of inverse distance times some constant.
  • The number of coherent non-repetitive responses is relative to the complexity of the previous response.
  • Even given ideal circumstances:
  • ------The conversational participant is programmed to never get board or end a conversation because it’s run out of things it wants to say or the participant happens to be very friendly and will try hard to continue talking to a random stranger.
  • ------The participant is able to choose the most interesting and complex response which occurs to her every time (as relative to at least one other participant. (Unless she’s talking to herself.))
  • The following can be concluded:
  • If topic trajectory and natural interest are both exponentially distributed, sustainable critical distance can never be achieved and the conversation will eventually end.
  • If participants aren’t just interested in anything or topic trajectory tends to relate to the participants experience (but not both) critical distance can be achieved and one of two things will happen:
  • -----Responses cease to move the topic forward (or in any direction). As the stagnate conversation continues it moves toward infinite complexity, this complexity is uninteresting to anyone not in the conversation.
  • -----Responses become incoherent, the topic moves in a uniform random field expanding indefinitely without ever localizing on anything interesting.
  • If both natural interest and topic trajectory are non-exponential critical distance is irrelevant (because there’s nothing exponential about the equation.) Participants will have a normal conversation that doesn’t last forever.
Further inquiry
It is comforting to believe that some things like scientific discourse, love and the pursuit of beauty can continue indefinitely and be indefinitely interesting. These systems are probably not linear let alone homogenous and I lack the math to describe them.
My typical Burner conversation patterns
Image
Image

User avatar
Pfhorrest
Posts: 3616
Joined: Fri Oct 30, 2009 6:11 am UTC
Contact:

Re: 0948 - "AI"

Postby Pfhorrest » Thu Sep 08, 2011 5:07 am UTC

Jamaican Castle wrote:
KShrike wrote:We can't seem to be capable of getting a computer to understand concepts and what they mean.
Instead, they emulate actions we give them.


That's more a limitation of computers than anything we're doing wrong. Just about all a computer can do is math. Math, and storing information (although that doesn't really count as "doing" something, especially since you need the math to get it back out again). If a problem can be distilled down to math, a computer can do it, but otherwise we don't have any way to begin.

And math (logic, really, which underlies all math) is nothing more than extremely precise language.

So our inability to get computers to do certain things can only be due to our inability to tell them exactly what it is that we want them to do. Want a computer to "understand concepts"? Easy!

  • First, figure out what exactly a concept is, rigorously delineate some of them, and input those into computer-accessible storage somewhere.
  • Then, figure out what it is to understand, write an algorithm describing that action in detail, and code a function to execute that action into a computer.
  • Finally, hook that computer up to that storage device and execute your understand() function on the concept objects you stored earlier.
What do you mean, figuring out what a concept is, and what it means to understand, is hard? Well then how can you expect a computer, which knows nothing besides what we've told it and can do nothing besides as we've trained it, to start understanding concepts when you're not even sure what it is you're asking of it yourself?

In the end, philosophy of mind boils down to AI programming. If you can say precisely what it is for something to be a mind, then you can program a computer to do precisely that and therefore be a mind.
Forrest Cameranesi, Geek of All Trades
"I am Sam. Sam I am. I do not like trolls, flames, or spam."
The Codex Quaerendae (my philosophy) - The Chronicles of Quelouva (my fiction)

jpk
Posts: 607
Joined: Sat Nov 13, 2010 7:33 am UTC

Re: 0948 - "AI"

Postby jpk » Thu Sep 08, 2011 5:46 am UTC

Pfhorrest wrote:In the end, philosophy of mind boils down to AI programming. If you can say precisely what it is for something to be a mind, then you can program a computer to do precisely that and therefore be a mind.


I'm afraid you've been imprecise in your use of language. You might be able to say precisely what it is to be a mind and still have no idea how to program a computer to do that. For example, I can tell you precisely what it is to be a grandmaster chess player: it's spelled out here. That's as precise a statement as you could want of what it is to be a grandmaster, but I don't think this is a basis for writing a program which could achieve grandmaster status. In order to do that, you have to solve a completely different set of problems - and solving those does not mean you've solved anything about the way human beings play chess.

I intend this as a friendly critique; I certainly don't belong to the "computers will never think" crowd, but I think what you've outlined here is at best a yardstick by which you can tell whether you've written a mind, not a way to generate a blueprint for a mind. Unfortunately, I find that I can't come up with a very good repair for your criterion. I'd like to say something like "if you can say precisely what a mind does, then you can program a computer to do precisely that" or "if you can say precisely how a mind does what it does..." but neither of those seems necessary or sufficient to the task. I don't think that minds are best understood by decompiling them. At best, I'm with Minsky in his sort of proto-object-oriented approach, but I don't think you'd ever get down to machine language, or even assembler, by that approach. At some point you'd say "ah, I know how to model that".

To be honest, I'm not very well up on the current status of AI, but I was under the impression that trying to code the mind was well out of fashion, and the promising research today was more about using neural nets and the like to "grow" a mind. In that case, you'd be able to make a mind without ever understanding how it was "actually" composed, except in a sort of trivial process-oriented way. And this brings it back to the difference between solving the problems that allow a machine to play grandmaster-level chess and solving the problems of understanding what it is for a human being to play at that level, and what the human being does.

ironclownfish
Posts: 4
Joined: Wed Sep 07, 2011 8:20 pm UTC

Re: 0948 - "AI"

Postby ironclownfish » Thu Sep 08, 2011 6:25 am UTC

oops, I meant to say cleverbot not chatbot. Also, to show you what I was talking about I went to cleverbot and this is the first conversation I had:

User: Are you a real person?

Cleverbot: Yes. Are you?

User: Yes. Quick, where are you from??

Cleverbot: Switzerland.

User: Cn u under s t an d this?

Cleverbot: Yes .

User: That was a human test, you passed :).

Cleverbot: It was a pancake!

User: Where are YOU from?

Cleverbot: Indiana. But I know some russian.

User: Cleverbot is fake.

Cleverbot: Why?

User: Because you're human, and I'm human!

Cleverbot: Thats not true.

(they probably said that because they thought I was 'cleverbot'). Anyway, it seems plain to me that I was talking to people who switched every few seconds (hence the switzerland/indiana discrepancy). As for the wikipedia article, I hadn't looked at it in a couple years or so, but it's quite different now. If I remember correctly it used to say exactly what I claim: that the website pairs you with constantly changing people. Now, however, it's very vague and doesn't really say one way or the other, though it seems to hint at it being an actual bot. Either I am completely mistaken, or the wiki had to be made vague and unhelpful to remain unbiased (as wikipedia is meant to be) and avoid taking a side in the constantly raging debate around whether cleverbot is for real.

Again, I would point out the very human-like response times.
Also, I encourage you to try A.L.I.C.E. which is legit for sure. alice.pandorabots.com

User avatar
Pfhorrest
Posts: 3616
Joined: Fri Oct 30, 2009 6:11 am UTC
Contact:

Re: 0948 - "AI"

Postby Pfhorrest » Thu Sep 08, 2011 6:36 am UTC

@jpk:

Agreed for the most part. I suppose the appropriate qualifier for my statement would be that while philosophy of mind boils down to AI programming, AI programming does not boil down to philosophy of mind. That is, to create an artificial intelligence, you need to first have a sufficient grasp of what constitutes "intelligence" (and so forth) to know when you've met that goal. Philosophy of mind tells you what are the necessary conditions for a mind: what tests does something need to pass in order to reach that bar? But it doesn't tell you the sufficient conditions, that's a whole other project that goes beyond philosophy into the nitty gritty details of AI programming. I suppose an analogous scenario would be that the philosophical thesis of naturalism, that all and only empirically observable physical phenomena constitute reality, still leaves open the huge question for the natural sciences to answer of what empirically observable physical phenomena are there?

Multiple realizability leaves open that there may be many different sets of jointly sufficient conditions for a mind, which brings me to my second point: that understanding how exactly humans do it may not really be important, except for maintenance and repair of your humans. But there may be much more straightforward (or other, equally complex but different) ways of getting the same functionality which is necessary for a mind.

An interesting thing arises if you combine that with the thesis that a thing's qualitative, subjective experience is a product of its function: you could have beings which always respond to the same inputs with the same outputs as each other, but process that higher-level function internally in different ways, and thus have vastly different experiences of the world in ways that they simply cannot ever communicate to each other. Like the old question of whether the color which we both call "blue" really looks the same to both of us.


@ironclownfish:

I think I've busted the cleverbot-is-random-paired-users myth. A friend of mine just now complained that it responded to the question "What is the plot of Fifth Element?" with a summary of the Matrix, and showed me the chat transcript. As a test, I asked Cleverbot "What is the plot of the Matrix?" and got the exact same response back, character for character. No way he and I randomly got paired with the same person out of over nine thousand chatting at that time, and that person just copy/pasted his response to any "What is the plot of _____?" inquiry.

For future reference, the response was "Humans are trapped inside of a computer program. Neo, the protagonist, can change the program as he is in it. He tries to free the humans from the machines who built the programme."

Note also the difference in spelling of "program/programme", which makes me think this is a response composited from several previous user responses, not one user's spontaneous natural-language response. Either way, that's a pretty impressive ability for a bot. It also seems to be able to emulate humor, see the following transcripts:

Cleverbot wrote:Cleverbot: Tell me a joke.
User: A blonde, a brunette, a redhead, a rabbi, a priest, a minister, a Mexican, an American, a Canadian, a duck, an elephant, and a giraffe walk into a bar, and the bartender says 'What is this, some kind of joke?'.
Cleverbot: Haha.


That one could easily be a script (print "Tell me a joke.", wait for response, print "Haha.") But then later I got this:

Cleverbot wrote:Cleverbot: I would tell you if I was a robot. Would you tell me if you were a robot?
User: If I was programmed to do so, certainly. Otherwise, probably not.
Cleverbot: You're fake.
User: So's your mom's tits.
Cleverbot: Ha ha ha ha nice one!


So it seems to be able to identify when the user means to be funny, in at least some circumstances, which is in and of itself pretty cool.
Forrest Cameranesi, Geek of All Trades
"I am Sam. Sam I am. I do not like trolls, flames, or spam."
The Codex Quaerendae (my philosophy) - The Chronicles of Quelouva (my fiction)

User avatar
Gatchaman
Posts: 17
Joined: Fri Jun 24, 2011 3:26 pm UTC
Location: Caerphilly, UK

Re: 0948 - "AI"

Postby Gatchaman » Thu Sep 08, 2011 9:00 am UTC

Cleverbot keeps trying to convince me that I'm the robot/computer program! xD

After lots of failed attempts to convince me, it changes it's tune:
...
CLEVERBOT: You are a robot.
ME: We are both Human!
CLEVERBOT: You are a machine. I am a human, there is a difference.
ME: WE are both Human!
CLEVERBOT: You are not human, you are a machine.
ME: We are both Human!
CLEVERBOT: We are both humans talking to eachother face to face.

Nice try...


Return to “Individual XKCD Comic Threads”

Who is online

Users browsing this forum: No registered users and 51 guests