I saw the old anime "Chobits" a few days ago and normally I'm not into that kind of stuff (anime isn't my thing usually) but it really got me thinking about tech. Basically an 18 year old farm kid trying to get into college goes to a prep school in Tokyo because he can't get into anywhere. Apparently everyone has a personal robot assistant when he gets there. He has no money but really wants a computer anyway. Why he doesn't just get a desktop is beyond me but regardless he finds one lying on a garbage pile. He turns her on, begins teaching her stuff, gets all emotional, blah blah blah. (yes I saw more than I should but I had nothing else to do at the time) The whole point is can a machine have real feelings and free will or is it ALL based on programs and what people tell it. There are primitive learning programs out there already.
Many people are basically "in love" with their computers to some extent. I know I spend too much time on mine. Wouldn't it be cool if a machine talked back to you and sounded like a cute 8 inch tall or human sized person? Some people would find it creepy to have a helper like this. Others would think it would be really amazing. Others even still probably have wives and don't want to think about another "woman" anyway, lol. So what do you guys think?
There is an AI being developed, a legitimate one. I mean, think of what an AI is. It is just a bunch of reactions to various events and situations. So no, a machine will never feel love.
The idea is that the machine will essentially write it's own code based on events. For example, how to stand up. A "goal" is programmed, and then it is attempted to be achieved through a series of randomized events. If one succeeds, it is stored in the database for other robots to use.
That is the idea. It is still very much science fiction.
Kinda sounds like the Borg from star treck. Well yeah no kidding a machine can never really feel but they can learn how to respond in an emotional way, from our point of view thats basically the same thing. Too bad you can't just go to best buy and pick one up now though. In the show anyway the dude is basically "I don't care if you feel or not. I sure as hell do feel, let me be happy." It was just a cool idea so I thought I'd post about it.
Technically the robot assistants WERE the computers, and the guy wanted a robot assistant not just a computer (I've watched like 8 episodes).
Anyways, I actually thought about that idea and making a desktop helper (I'm a programmer). But it would require A LOT of effort for minimal functionality. Essentially you'd be creating an AI that can not only understand sentences (or at the very least commands), but could also speak back in a human-like tone and speech. All for the purpose of either:
A - Just having a machine to converse with for the fun of it
OR
B - Doing things you could do with a few mouse clicks anyways.
That is unless you'd try and get the entity to understand complex commands, which would be even harder. But yea, interesting idea and it would be fun to play around with if it was ever made. However odds are the only chance of it being made is by a really dedicated independent team who for some reason or other, really want a desktop assistant.
As for the creepy bit, it really depends on how it's made vs who uses it. Though I doubt most people would be creeped out by it.
Hmmm lets see here.... parts needed: like 20 intel i7 990s, enough power to run 10 Vegas homes for a year, a team of really skilled programmers (the 1 person is 5 blizzard employees kind), quality speakers, sensors and motors galore. Thats pretty much all I can think of that would be needed.
edit: and enough of a hard drive to store the library of congress digitally.
Mostly I think it would be just kinda like the worlds coolest computer plus it would route phone calls, receive e-mail, and hopefully realize your upset and say nice things :)
facial recognition is already reality, a dude made a $7,000 robot for getting laid, the worlds first personal robot assistant platform has been created, o and hey go google "Honda Asimo." I wonder what would happen if you were to throw all of it into the same creation...
one thing has been bothering me though is that an emotion or a thought is really just a packet of chemicals that's read by a mind. I know this because my ELA teacher has us read "native son" and an article he gave us on the death penalty said that criminals have a chemical imbalance. I wonder if it would be possble to make a machine able to at least read the chemicals. I have slight AS which if servere (which I do not have) leaves you being really smart but having trouble reading faces or gestures and it leaves you socially awkward. Hey maybe we can make a robot that has a really really bad case of AS but at least it tries.
I was working on a language synthesizer thing for a while a couple years ago. It was fairly primitive, but basically I was breaking down English syntax mathematically. I didn't get it very far, but it could recognize a few things, mostly things like left, right, over, under, etc, and recognize some basic shapes. Like I would make an image in Photoshop with a circle above a triangle, and could type in "What's above the triangle?" and it would print out circle. Not very impressive, but it was kind of fun. I've been wanting to pick it back up again, but I've changed my mind on how I want to go about it. But I've been busy with other stuff that's more entertaining.
well it's better than nothing. It's too bad that you can't teach a machine the same way you can teach a 3 year old. Then again, I guess for most people having a real kid is time consuming enough. Man would I like to see a learning robot though.
What I want to get mine to eventually do is interpret input to work as a kind of advanced search and index algorithm. I don't much care about making it 'feel' as that would be ultimately pointless (I mean really, do you want your computer/helper robot to get in a pissy mood and be like "No, fuck you. I just ain't gonna do it!"?). The only real applicable thing I can see it being used in is security though, for indexing and searching through information, but there are already programs that do that. So I don't spend much time with it. I'm more interested in getting my game engine and getting alpha versions of a couple of projects done.
yeah well the idea has just been rolling around in my head. I'm not really interested in government application such as security or military. I'm thinking of more like a pet you can have instead of a cat if you care about your image. Searching though... that has potential. Reminds me of Ford's sync system except way better. If I made one it would be like a companion that you can trade in a tablet for. It takes notes, searches the web, tells you it's "opinion" based on the data it takes in, etc. etc. It would also kinda cure boredom at least until it no longer captivates you. It seems much more fun than just writing on a screen to me. People could also develop maturity through taking care of it too. Have it tell you it's tired and needs to be plugged in. Plus sometimes teaching someone else is the best way to learn on your own too. Imagine trying to tell a robot the meaning of a word or what an object is. It makes you really think about what it is yourself.
You CAN do that. It would likely be an extremely expensive project, but can probably be done relatively cheaply in terms of prototyping.
Basically you would need a camera and a large hard drive and extensive programming. Rudimentary things like left/right would need to be programmed in already, and then some kind of speech recognition and possibly synthesizing software (it would be easier if it was something like a dog that didn't talk. Then it just needs to make noise).
What I'm thinking of is kind of like an expanded model of this dog thing I had when I was little (it was a robot thing and it had some little games you could play. It got old really fast though, as you couldn't do much else). But anyway, you could teach it to recognize things through the camera (it would take a picture of something), so if you said, for instance, "Sit on the couch", you would need to show it the couch and repeat it a few times probably until it indexed the image with the phrase, and would probably need to walk around it so it knows what the back of the couch looks like as opposed to the front. If it walked around your house, it would take images, and eventually create something of a map and it can figure out where it is, etc. It would be a massive programming project (on scale with an operating system or so probably), as you probably know, but definitely doable with a decent amount of work.
Teaching it words would be different. Like attempting to teach it the difference between a lake and an ocean, for instance, as they are so extremely arbitrary (they're just symbols used for recognition, really; we attach far more complex meanings to them. Like you know that a book is a very broad category of things. Likewise, you know that Lord of the Rings is a book, specifically a novel, and then you can get into genres but that's irrelevant right now. You also know that something like a text book or diary does not fit into the category of novel, but it's still in the same broader category as Lord of the Rings, being a book).
I'm hoping no one said anything about terminator because that's just old. But I would like to point out Ghost in the Shell. Although it doesn't correspond to AI directly it relates to advancement of technology. If you want to waste some time watch that. If I'm ever bored I can waste hours thinking a bout ghost in the shell.
Artificial intelligence wouldn't be any more dangerous than you or I. You have a learning computer in your head, after all. There are "bad" humans and there will be "bad" AI assuming software tries to emulate the free-thinking nature of the most powerful computer in existence which makes your desktop look like an abacus; your brain.
I think the whole concept of trying to develop artificial intelligence is a little fruitless. At the end of the day, can an AI really surpass the human brain? And do we really even want that to happen?
The question is are computers more efficient then the human brain? If computers can work better than the human brain as an AI then they'll be using 100% compared to us using 10% or whatever it was. But the AI would probably fry out quickly due to that.
Humans don't actually only use 10% of there brains that is a myth. However an AI could be much more efficient because it would be programmed intelligently and optimized for performance where as humans obviously are not (shown by the fact that despite having billions of neurons most of us still cant do 67*43 in our heads). However I don't see why you would want to make an A.I emotional as I certainly would not say emotion or even free will is necessarily for intelligence. In my opinion an A.G.I (Artificial General Intelligence) could be based on pure logic and still be able to comprehend abstract subjects but as no such A.I exists that is obviously just an opinion.
Intelligence is the ability to apply knowledge to a task. My calculator has the knowledge to work out 67*43, but no intelligence to apply it, making it as dumb as a rock. Don't confuse knowledge and intelligence.
I am not trying to imply that current programs have intelligence. You are clearly cleverer than your calculator and your computer and the most powerful supercomputer on Earth. My point is simply that intelligence does not require emotion or even self motivation nor are human minds optimized for performance (If you think about it theres no physical reason we should not be able to do 67 * 43 rapidly). If you had a machine which when asked how to say design an arbitrary object could come up with a logical plan to build that build that object based on the data available to it then it would be intelligent (imo). I think the idea that emotion is some kind of special human trait that cannot be mimicked is also just human arrogance. Although I can see no reason why you would want to build a machine that could "fall in love" I also don't see any reason why it would be impossible.
That all being said I think we should not focus our efforts on mimicking humans. We already have (nearly) seven billion humans on Earth. Instead we should focus our efforts on things humans are bad at (like tracking vast numbers of people or doing complex calculations) and not try to create "Self motivated" A.I.
Sorry to budge in but i am raising a three year old (Reaper mentioned that exact age)
and it is fucking hard to teach them stuff that are not "monkey see monkey do" or other stuff that everyone would take for granted that a a machine could/would do (given the adequate prior instructions several times over)
feelings and emotions are the stuff that are really in question when talking about ai...
and i totally agree "we have a computer for a brain":
it's really the other way around .. every thing worth a damn has been reversed engineered from our bodies (or other living beings') or vegetation (the way they are, the way they work, grow, .. you name it..)
principal axiom to know for teaching kids: they speak before they know what they are saying.. (indeed before they realize that they are speaking :) )..
and they rely on emotion (theirs = completely mixed up from the start and their caretakers' emotions while using said words = completely whatever and yet overly active to the kid's future persona/character/etc )
they rely on emotion to teach them what the words mean (for the life of me, i can't understand how the asian's do it, since their "intonations" are part of the meanings of words..??? and i mean no disrespect! i'm genuinely puzzled by this) ..
the kids then go over and over (precisely like a cp would do) the words until they are linked to others enough that they get "dispassionate" with them.. and go on to others.. same for every thing really (everything accountable for, even fears or likes and dislikes)..
fascinating shite :)
as for my kid = he's awesome ;^) [insert a million smileys] and damn worth the effort
Well doing math in a matter of seconds is a bit unnecessary. Humans have the unique ability to invent ways to get around problems. Need to clear an enemy position with limited men? Invent a tank! Need to do quick research? Invent the internet! Humans don't necessarily need to know how to do everything on their own as we are an inventive species. Besides I could do math problems out myself but it takes longer than using a calculator. The sad thing is that eventually we will forget how to do things because machines are made to be able to do it. In the end we will forget how to make machines too. That's why a balance of both is needed.
As to the "if it's worth it" part, think about this for a second. It is much easier to take all people know and dump it into a machine, but the machine won't really understand what it has. A security robot has a hard time of determining the difference between friend and foe. People however, have a much easier time telling who the good guy vs the bad guy is. Also with all this technology that can do your match homework but no tech for socializing, it's an open market. Some people don't like being around others as much as some people. I for one am not a people person. I'm not a recluse but it's harder to be around people than computers and stuff. People are unpredictable, you can't always tell what someone is going to do. Current machines are boring. They are just boxes sitting on tables.
exactly. Although I may not be a parent or necessarily want to be right now (I'm 17 ewww) i still would like to be able to educate something to it's advantage. Even if that thing isn't exactly alive it sure would be cool to teach it how to do things. It's just fascinating. Another idea I have been thinking of is how it would be possible to make a synthetic brain. With enough knowledge of how humans really think we could replicate it as best we could out of non-living materials.
Emotions in many ways could be described as "thinking shortcuts". A few billion years of evolution (or whatever you happen to believe created humans) have taught us the things that we emotionally consider disgusting cause disease. Even though we don't necessarily know why they are disgusted(The germ theory of disease is only a few hundred years old people have been disgusted for thousands of years) there is still a rational reason for it. This gives all humans a baseline of common sense that we don't have to learn (Even someone who was never taught not to they would be unlikely to walk into a fire or over a cliff). However I think this would not be particularly hard to mimic with programming. You would simply have to have a learning program in the machine as well as some human built programs which would do things like make sure it does not walk off a cliff.
The second role of emotions is to provide motivation. If you consider the world in a 100% logical way there is no reason to do anything. Even living and reproducing are "pointless". Therefor our main motivator are things like happiness and avoiding things like depression. However this would be unnecessary for a robot because presumably for it to be useful humans would be giving it its goals. For example you would not want a robotic car to be able to go Sunday driving because it feels like it you want it to go where you tell it to. This removes the need for self motivation.
As for the argument that humans are innovative I don't think there is any magical secret to innovation. Innovation is imo simply considering all the information you have and trying to use it in new ways. This is demonstrated by the fact that throughout history when the information necessarily to build it became available many things were invented by multiple people working independently almost simultaneously (eg the telephone). The fact that multiple people can invent the same thing at the same time shows that it is not some magical creative spark but simply that they happened to have all the information necessarily to invent the item in question and they happened to make the put it all together. With some very sophisticated programmatic I think that a robot which was capable of understanding arbitrary concepts in generalized logical ways would be able to innovate. (obviously no such machine exists and probably wont exist for allong time).
Finally about math. Although some humans are able to do that problem in our heads it takes us at best 10 seconds (if your really really good at it). A machine could do it in much much less than a second. The point of this is not to show that machines are smarter than us (they aren't) it is simply to point out that the process by which humans do arithmetic is very sub optimal therefor there is room for improvement if you were to attempt to build a machine with human like intelligence.
Rollback Post to RevisionRollBack
To post a comment, please login or register a new account.
I saw the old anime "Chobits" a few days ago and normally I'm not into that kind of stuff (anime isn't my thing usually) but it really got me thinking about tech. Basically an 18 year old farm kid trying to get into college goes to a prep school in Tokyo because he can't get into anywhere. Apparently everyone has a personal robot assistant when he gets there. He has no money but really wants a computer anyway. Why he doesn't just get a desktop is beyond me but regardless he finds one lying on a garbage pile. He turns her on, begins teaching her stuff, gets all emotional, blah blah blah. (yes I saw more than I should but I had nothing else to do at the time) The whole point is can a machine have real feelings and free will or is it ALL based on programs and what people tell it. There are primitive learning programs out there already.
Many people are basically "in love" with their computers to some extent. I know I spend too much time on mine. Wouldn't it be cool if a machine talked back to you and sounded like a cute 8 inch tall or human sized person? Some people would find it creepy to have a helper like this. Others would think it would be really amazing. Others even still probably have wives and don't want to think about another "woman" anyway, lol. So what do you guys think?
@Reaper872: Go
There is an AI being developed, a legitimate one. I mean, think of what an AI is. It is just a bunch of reactions to various events and situations. So no, a machine will never feel love.
The idea is that the machine will essentially write it's own code based on events. For example, how to stand up. A "goal" is programmed, and then it is attempted to be achieved through a series of randomized events. If one succeeds, it is stored in the database for other robots to use.
That is the idea. It is still very much science fiction.
Great to be back and part of the community again!
Kinda sounds like the Borg from star treck. Well yeah no kidding a machine can never really feel but they can learn how to respond in an emotional way, from our point of view thats basically the same thing. Too bad you can't just go to best buy and pick one up now though. In the show anyway the dude is basically "I don't care if you feel or not. I sure as hell do feel, let me be happy." It was just a cool idea so I thought I'd post about it.
@Reaper872: Go
Technically the robot assistants WERE the computers, and the guy wanted a robot assistant not just a computer (I've watched like 8 episodes).
Anyways, I actually thought about that idea and making a desktop helper (I'm a programmer). But it would require A LOT of effort for minimal functionality. Essentially you'd be creating an AI that can not only understand sentences (or at the very least commands), but could also speak back in a human-like tone and speech. All for the purpose of either: A - Just having a machine to converse with for the fun of it OR B - Doing things you could do with a few mouse clicks anyways.
That is unless you'd try and get the entity to understand complex commands, which would be even harder. But yea, interesting idea and it would be fun to play around with if it was ever made. However odds are the only chance of it being made is by a really dedicated independent team who for some reason or other, really want a desktop assistant.
As for the creepy bit, it really depends on how it's made vs who uses it. Though I doubt most people would be creeped out by it.
Hmmm lets see here.... parts needed: like 20 intel i7 990s, enough power to run 10 Vegas homes for a year, a team of really skilled programmers (the 1 person is 5 blizzard employees kind), quality speakers, sensors and motors galore. Thats pretty much all I can think of that would be needed.
edit: and enough of a hard drive to store the library of congress digitally.
Mostly I think it would be just kinda like the worlds coolest computer plus it would route phone calls, receive e-mail, and hopefully realize your upset and say nice things :)
facial recognition is already reality, a dude made a $7,000 robot for getting laid, the worlds first personal robot assistant platform has been created, o and hey go google "Honda Asimo." I wonder what would happen if you were to throw all of it into the same creation...
one thing has been bothering me though is that an emotion or a thought is really just a packet of chemicals that's read by a mind. I know this because my ELA teacher has us read "native son" and an article he gave us on the death penalty said that criminals have a chemical imbalance. I wonder if it would be possble to make a machine able to at least read the chemicals. I have slight AS which if servere (which I do not have) leaves you being really smart but having trouble reading faces or gestures and it leaves you socially awkward. Hey maybe we can make a robot that has a really really bad case of AS but at least it tries.
I was working on a language synthesizer thing for a while a couple years ago. It was fairly primitive, but basically I was breaking down English syntax mathematically. I didn't get it very far, but it could recognize a few things, mostly things like left, right, over, under, etc, and recognize some basic shapes. Like I would make an image in Photoshop with a circle above a triangle, and could type in "What's above the triangle?" and it would print out circle. Not very impressive, but it was kind of fun. I've been wanting to pick it back up again, but I've changed my mind on how I want to go about it. But I've been busy with other stuff that's more entertaining.
@Varine: Go
well it's better than nothing. It's too bad that you can't teach a machine the same way you can teach a 3 year old. Then again, I guess for most people having a real kid is time consuming enough. Man would I like to see a learning robot though.
What I want to get mine to eventually do is interpret input to work as a kind of advanced search and index algorithm. I don't much care about making it 'feel' as that would be ultimately pointless (I mean really, do you want your computer/helper robot to get in a pissy mood and be like "No, fuck you. I just ain't gonna do it!"?). The only real applicable thing I can see it being used in is security though, for indexing and searching through information, but there are already programs that do that. So I don't spend much time with it. I'm more interested in getting my game engine and getting alpha versions of a couple of projects done.
yeah well the idea has just been rolling around in my head. I'm not really interested in government application such as security or military. I'm thinking of more like a pet you can have instead of a cat if you care about your image. Searching though... that has potential. Reminds me of Ford's sync system except way better. If I made one it would be like a companion that you can trade in a tablet for. It takes notes, searches the web, tells you it's "opinion" based on the data it takes in, etc. etc. It would also kinda cure boredom at least until it no longer captivates you. It seems much more fun than just writing on a screen to me. People could also develop maturity through taking care of it too. Have it tell you it's tired and needs to be plugged in. Plus sometimes teaching someone else is the best way to learn on your own too. Imagine trying to tell a robot the meaning of a word or what an object is. It makes you really think about what it is yourself.
You CAN do that. It would likely be an extremely expensive project, but can probably be done relatively cheaply in terms of prototyping.
Basically you would need a camera and a large hard drive and extensive programming. Rudimentary things like left/right would need to be programmed in already, and then some kind of speech recognition and possibly synthesizing software (it would be easier if it was something like a dog that didn't talk. Then it just needs to make noise).
What I'm thinking of is kind of like an expanded model of this dog thing I had when I was little (it was a robot thing and it had some little games you could play. It got old really fast though, as you couldn't do much else). But anyway, you could teach it to recognize things through the camera (it would take a picture of something), so if you said, for instance, "Sit on the couch", you would need to show it the couch and repeat it a few times probably until it indexed the image with the phrase, and would probably need to walk around it so it knows what the back of the couch looks like as opposed to the front. If it walked around your house, it would take images, and eventually create something of a map and it can figure out where it is, etc. It would be a massive programming project (on scale with an operating system or so probably), as you probably know, but definitely doable with a decent amount of work.
Teaching it words would be different. Like attempting to teach it the difference between a lake and an ocean, for instance, as they are so extremely arbitrary (they're just symbols used for recognition, really; we attach far more complex meanings to them. Like you know that a book is a very broad category of things. Likewise, you know that Lord of the Rings is a book, specifically a novel, and then you can get into genres but that's irrelevant right now. You also know that something like a text book or diary does not fit into the category of novel, but it's still in the same broader category as Lord of the Rings, being a book).
I'm hoping no one said anything about terminator because that's just old. But I would like to point out Ghost in the Shell. Although it doesn't correspond to AI directly it relates to advancement of technology. If you want to waste some time watch that. If I'm ever bored I can waste hours thinking a bout ghost in the shell.
Artificial intelligence wouldn't be any more dangerous than you or I. You have a learning computer in your head, after all. There are "bad" humans and there will be "bad" AI assuming software tries to emulate the free-thinking nature of the most powerful computer in existence which makes your desktop look like an abacus; your brain.
I think the whole concept of trying to develop artificial intelligence is a little fruitless. At the end of the day, can an AI really surpass the human brain? And do we really even want that to happen?
@Eiviyn: Go
The question is are computers more efficient then the human brain? If computers can work better than the human brain as an AI then they'll be using 100% compared to us using 10% or whatever it was. But the AI would probably fry out quickly due to that.
Humans don't actually only use 10% of there brains that is a myth. However an AI could be much more efficient because it would be programmed intelligently and optimized for performance where as humans obviously are not (shown by the fact that despite having billions of neurons most of us still cant do 67*43 in our heads). However I don't see why you would want to make an A.I emotional as I certainly would not say emotion or even free will is necessarily for intelligence. In my opinion an A.G.I (Artificial General Intelligence) could be based on pure logic and still be able to comprehend abstract subjects but as no such A.I exists that is obviously just an opinion.
Intelligence is the ability to apply knowledge to a task. My calculator has the knowledge to work out 67*43, but no intelligence to apply it, making it as dumb as a rock. Don't confuse knowledge and intelligence.
I am not trying to imply that current programs have intelligence. You are clearly cleverer than your calculator and your computer and the most powerful supercomputer on Earth. My point is simply that intelligence does not require emotion or even self motivation nor are human minds optimized for performance (If you think about it theres no physical reason we should not be able to do 67 * 43 rapidly). If you had a machine which when asked how to say design an arbitrary object could come up with a logical plan to build that build that object based on the data available to it then it would be intelligent (imo). I think the idea that emotion is some kind of special human trait that cannot be mimicked is also just human arrogance. Although I can see no reason why you would want to build a machine that could "fall in love" I also don't see any reason why it would be impossible.
That all being said I think we should not focus our efforts on mimicking humans. We already have (nearly) seven billion humans on Earth. Instead we should focus our efforts on things humans are bad at (like tracking vast numbers of people or doing complex calculations) and not try to create "Self motivated" A.I.
@DirectorOfTheUED: Go
2881 with using only my head.
I don't really get what is so hard in that.
To explain how I got it with using my head.
67 * 43
60 * 40 = 2400
60 * 3 = 180
Aka it becomes 2580.
7 * 40 = 280
7 * 3 = 21
Aka it becomes 301.
2580 + 301 = 2881.
Sorry to budge in but i am raising a three year old (Reaper mentioned that exact age)
and it is fucking hard to teach them stuff that are not "monkey see monkey do" or other stuff that everyone would take for granted that a a machine could/would do (given the adequate prior instructions several times over)
feelings and emotions are the stuff that are really in question when talking about ai...
and i totally agree "we have a computer for a brain":
it's really the other way around .. every thing worth a damn has been reversed engineered from our bodies (or other living beings') or vegetation (the way they are, the way they work, grow, .. you name it..)
principal axiom to know for teaching kids: they speak before they know what they are saying.. (indeed before they realize that they are speaking :) )..
and they rely on emotion (theirs = completely mixed up from the start and their caretakers' emotions while using said words = completely whatever and yet overly active to the kid's future persona/character/etc )
they rely on emotion to teach them what the words mean (for the life of me, i can't understand how the asian's do it, since their "intonations" are part of the meanings of words..??? and i mean no disrespect! i'm genuinely puzzled by this) ..
the kids then go over and over (precisely like a cp would do) the words until they are linked to others enough that they get "dispassionate" with them.. and go on to others.. same for every thing really (everything accountable for, even fears or likes and dislikes)..
fascinating shite :)
as for my kid = he's awesome ;^) [insert a million smileys] and damn worth the effort
Well doing math in a matter of seconds is a bit unnecessary. Humans have the unique ability to invent ways to get around problems. Need to clear an enemy position with limited men? Invent a tank! Need to do quick research? Invent the internet! Humans don't necessarily need to know how to do everything on their own as we are an inventive species. Besides I could do math problems out myself but it takes longer than using a calculator. The sad thing is that eventually we will forget how to do things because machines are made to be able to do it. In the end we will forget how to make machines too. That's why a balance of both is needed.
As to the "if it's worth it" part, think about this for a second. It is much easier to take all people know and dump it into a machine, but the machine won't really understand what it has. A security robot has a hard time of determining the difference between friend and foe. People however, have a much easier time telling who the good guy vs the bad guy is. Also with all this technology that can do your match homework but no tech for socializing, it's an open market. Some people don't like being around others as much as some people. I for one am not a people person. I'm not a recluse but it's harder to be around people than computers and stuff. People are unpredictable, you can't always tell what someone is going to do. Current machines are boring. They are just boxes sitting on tables.
@houndofbaskerville: Go
exactly. Although I may not be a parent or necessarily want to be right now (I'm 17 ewww) i still would like to be able to educate something to it's advantage. Even if that thing isn't exactly alive it sure would be cool to teach it how to do things. It's just fascinating. Another idea I have been thinking of is how it would be possible to make a synthetic brain. With enough knowledge of how humans really think we could replicate it as best we could out of non-living materials.
Emotions in many ways could be described as "thinking shortcuts". A few billion years of evolution (or whatever you happen to believe created humans) have taught us the things that we emotionally consider disgusting cause disease. Even though we don't necessarily know why they are disgusted(The germ theory of disease is only a few hundred years old people have been disgusted for thousands of years) there is still a rational reason for it. This gives all humans a baseline of common sense that we don't have to learn (Even someone who was never taught not to they would be unlikely to walk into a fire or over a cliff). However I think this would not be particularly hard to mimic with programming. You would simply have to have a learning program in the machine as well as some human built programs which would do things like make sure it does not walk off a cliff.
The second role of emotions is to provide motivation. If you consider the world in a 100% logical way there is no reason to do anything. Even living and reproducing are "pointless". Therefor our main motivator are things like happiness and avoiding things like depression. However this would be unnecessary for a robot because presumably for it to be useful humans would be giving it its goals. For example you would not want a robotic car to be able to go Sunday driving because it feels like it you want it to go where you tell it to. This removes the need for self motivation.
As for the argument that humans are innovative I don't think there is any magical secret to innovation. Innovation is imo simply considering all the information you have and trying to use it in new ways. This is demonstrated by the fact that throughout history when the information necessarily to build it became available many things were invented by multiple people working independently almost simultaneously (eg the telephone). The fact that multiple people can invent the same thing at the same time shows that it is not some magical creative spark but simply that they happened to have all the information necessarily to invent the item in question and they happened to make the put it all together. With some very sophisticated programmatic I think that a robot which was capable of understanding arbitrary concepts in generalized logical ways would be able to innovate. (obviously no such machine exists and probably wont exist for allong time).
Finally about math. Although some humans are able to do that problem in our heads it takes us at best 10 seconds (if your really really good at it). A machine could do it in much much less than a second. The point of this is not to show that machines are smarter than us (they aren't) it is simply to point out that the process by which humans do arithmetic is very sub optimal therefor there is room for improvement if you were to attempt to build a machine with human like intelligence.