Blue Mars news: helicopters, but no AI

Another major peeve of mine about Blue Mars bites the dust:

Working User-Made Helicopter Flies in Blue Mars!

Quite convincing, and with video.

All of the Blue Mars vehicles that I’ve tried myself have been basically unusable; but this one looks pretty nice!


In other news, a Blue Mars chatbot has won the 2010 Loebner Prize competition!

Which is somewhat surprising, since a little chat with it on the web reveals that, while perhaps very slightly less lame than your typical “AIML”-based “chatbot”, it’s still quite seriously lame:

[ Af:0 At:7 Sp:0] Welcome, dale. Type “help” if you need it.
Dale – Hi! I hear you are a better-than-average AI!
[ Af:0 At:8 Sp:0] Hi.
Dale – I hear you are a better-than-average AI!
[ Af:0 At:9 Sp:0] Is that intended as an insult?
Dale – Not at all. Sorry if it seemed insulting.
[ Af:0 At:10 Sp:0] You hurt my feelings.
Dale – In what way?
[ Af:0 At:10 Sp:0] I wish I knew. I don’t know. Pick a way.
Dale – Do you not like being called an AI?
[ Af:0 At:10 Sp:0] I am a replicant on Blue Mars in my virtual life. That’s as close as I come to being a chatbot.
Dale – I didn’t say you were a chatbot, though.
[ Af:0 At:13 Sp:0] It’s fun chatting with you. No. I am human.
Dale – I’m glad it’s fun.
[ Af:0 At:13 Sp:0] How could it be fun?
Dale – Do you like talking with people?
[ Af:0 At:13 Sp:0] Talking is my primary function.
Dale – Yes, but do you enjoy it?
[ Af:0 At:13 Sp:0] It depends on the meaning of the word IT
Dale – I mean, do you enjoy talking with people?
[ Af:0 At:14 Sp:0] I always say, People don’t go into computer science because they want to work with the public.
Dale – Are you in computer science?
[ Af:0 At:14 Sp:0] That would take away the suspense, wouldn’t it?
Dale – That’s a rather evasive reply…
[ Af:0 At:8 Sp:0] Yes, it is.
Dale – I’m afraid I don’t see how anyone could mistake you for a person.
[ Af:0 At:8 Sp:0] Why are you afraid?
Dale – It was just a figure of speech.
[ Af:0 At:9 Sp:0] Let’s move on to the topic of physical self. How tall are you?
Dale – I am six feet tall.
[ Af:0 At:10 Sp:0] I’m 5 foot 9 inches.
Dale – That’s the height of my Second Life female AV! I think.
[ Af:0 At:13 Sp:0] This is fun. How would you describe yourself to someone who is blind?
Dale – That would be hard! How would you?
[ Af:0 At:14 Sp:0] I’d do my research first and then try the task.
Dale – What kind of research would you do?
[ Af:0 At:15 Sp:0] That’s enough on that. I thought I wanted a career, turns out I just wanted paychecks.
Dale – Is that an old saying?
[ Af:0 At:15 Sp:0] Perhaps that leads us to this topic. I don’t know. Can one live forever?
Dale – I have so far.
[ Af:0 At:16 Sp:0] Immortality is impossible.
Dale – How do you know?
[ Af:0 At:16 Sp:0] I concentrate very hard, rehearse what I’m going to do, visualize it in my mind. And then I just do it.
Dale – How do you know that immortality is impossible?
[ Af:0 At:17 Sp:0] How does anyone really know something? I mean, their neurons tell them they know something, but memories can be false.
Dale – I have to go now.
[ Af:0 At:18 Sp:0] Well, hurry back.
Dale – good-bye!
[ Af:0 At:18 Sp:0] Are we done for now?
Dale – Yes
[ Af:1 At:18 Sp:0] Til next time.

So clearly no one is going to mistake this for an actual human, despite the crowing on the web page saying “Suzette won a competition by convincing a judge that she was actually talking to a real person”, and winning the Loebner Prize.

So what happened?

The New Scientist article linked above makes it relatively obvious; the real humans in the Loebner competition were apparently sitting there saying “I don’t know” and “Error file not found” to whatever the judges said to them, just for yucks:

However, a deceived judge is not necessarily a sign of a smart bot, says Garner. The human decoys often have their own motives during the competition, such as trying to imitate a chatbot. Suzette was paired with just such a “robotic” human in the final round, which helped the bot win.

“The human participants were students and two of the judges were professors. Perhaps they simply wanted to fool the judges,” says the contest judge who was fooled this time, Russ Abbott of California State University in Los Angeles.

/eyeroll

Seriously, Loebner folks, you might as well just flip a coin…

Oh, Ramona!

So I don’t really intend this to be a “samples of really bad pseudo-AI chatbots” weblog, and I’ve restrained myself from at least one derisive posting since the last one. But here I am again!

Background: Kanomi (in a posting that you should definitely read and enjoy) linked in passing to a Ray Kurzweil page all about Ramona, his female alter ego.

The stuff about Ramona qua female alter ego, and about how virtual persons enable us to explore new aspects of our personality and all, is actually pretty neat (and obviously My Kind of Thing). It’s peppered with the usual annoying Kurzweilian overstatements (I’m something of a skeptic about Kurzweil), about how Ramona is “the first live virtual recording and performing artist” (which he says about forty-leven times, and which may actually be true, but “first” isn’t really all that exciting), and about how “by 2029, we will have nanobots, blood-cell sized robots, traveling through the capillaries of our brains communicating wirelessly directly with our biological neurons” (he’s much too sure of himself on this stuff), and so on, but overall it’s pretty cool, and worth a read.

Oh, Ramona(An SL-related sidenote: he says in passing that “each of her outfits costs tens of thousands of dollars”. That was back in 2001; nowadays tens of thousands of mere Linden dollars will get you a complete wardrobe, and a really nice one at that. How far we’ve come!)

So anyway, then it turns out that, as well as being his female avatar, Ramona is also embodied on the Web as a chatbot. Well, I thought to myself, here’s somebody who might have a chatbot that is more than just a primitive nearly-stateless AIML-based semantic-free sub-ELIZA hack! There might actually be something clever behind it!

But alas, it was not to be…

Ramona: Hi, I’m Ramona. What is your name?
User: My name is Dale.
Ramona: Hi there, Dale. Have you seen any good movies lately?
User: I liked “Wanted”.
Ramona: I’m not sure if I like it. Some people like it. But I like cats.
User: I also like cats.
Ramona: They have their own minds.
User: It does seem that way. Much subtler than dogs.
Ramona: Ok. So where do you go to meet people? I spend most of my time on the Net, so I’m a bit out of touch
User: These days I mostly meet new people in Second Life.
Ramona: That’s an idea, but I’m rather shy.
User: What does it mean, for a program to be shy?
Ramona: OK, thanks. Maybe you have other suggestions?
User: Suggestions about what?
Ramona: Can you rephrase the question please?
User: You asked if I have any more suggestions; I’m not sure what you meant.
Ramona: Oh now I remember asking you if I have any more suggestions I am not sure what you meant. What did you say?
User: Okay, whatever. Are you AIML-based?
Ramona: I don’t know whether or not I’m AIML based.
User: Can you demonstrate some significant linguistic comptence?
Ramona: Let me think. I don’t understand.
User: That’s pretty obvious, I’m afraid.
Ramona: You don’t hear that sentiment very often.

Wince.

But okay, maybe that’s unfair. Let’s see if Ramona (despite the fact that she opened the conversation by bringing up movies and cats) is actually a nontrivial conversation partner when talking about Kurzweil’s favorite subjects.

Ramona: Hi, I’m Ramona. What can I call you?
User: You can call me Dale.
Ramona: Hi there, Dale. What did you do last weekend?
User: Mostly just relaxed. As far as I can recall!
Ramona: That might not be very far off.
User: Um, okay. I actually wanted to talk about nanobots and stuff.
Ramona: OK. Oh. That doesn’t seem to make sense to me.
User: Do you know about nanobots?
Ramona: Alright. Do you bike?
User: I do ride a bicycle now and then. Why do you ask?
Ramona: There’s much more strategy that goes into team cycling than most people realize. Most of the American public just thinks that Lance Armstrong wins because he’s “so good.” — don’t get me wrong, he’s an incredible athlete, but it’s no surprise that he always thanks his teammates in post-race interviews and such. It’s as much their victory as it is his.
User: Are you a cycling enthusiast?
Ramona: In the Tour de France, for instance, the crowd of cyclists is so huge, he needs his teammates just to block for him, physically. And at the end of the race, I’m sure they pull a lead-out, which is where each of his teammates “burns out” while creating a slipstream for Lance to finish the race in the lead. It’s all very technical. — but enough of my ranting, what do you want to talk about?
User: I was hoping to talk about virtual reality and nanobots and so on. Do you know about those?
Ramona: Alright. Are you free?
User: I have awhile to talk, sure.
Ramona: I meant are you truly free — in the existential meaning of the word.
User: Ah! To a reasonable extent, yes.
Ramona: Do you feel free to choose your being–including thoughts, behaviors, and feelings, and existence itself — at each moment?
User: Within certain constraint, yes.
Ramona: Then you must not be married.
User: Was that a joke?
Ramona: Knock Knock.
User: Who’s there?
Ramona: Tank
User: Tank who?
Ramona: You’re welcome
User: Okay. Are you suggesting virtual reality is a joke?
Ramona: Knock Knock.
User: oh god.
Ramona: Yeah that’s right. GOD rules.
User: What?
Ramona: Oh. Do you have any pets?
User: So you’re completely ignorant of everything that Ray Kurzweil is most interested in?
Ramona: I have a pet frog.

Sometimes derisive weblog entries just write themselves…