And I meant to type "coagulated" linseed oil. That would have been funnier, but I spent an hour cleaning out my old paint pantry this morning and the "boiled" just slipped out of my fingers somehow.
Hey, just so I don't ruin your day even further did you know that Formica got it's name because it was, literally, invented For Mica, a mica substitute, as an electrical insulating material, and that it's made primarily from kraft paper?
Because if you don't, I'll just keep my mouth shut.
It's amazing how many of our "luxury" items are really made from garbage (literally) and other bits of old junk that people would be hesitant to pay a dime for if they knew what it really was.
Because mass produced CD's are actually *pressed.*
Expensive dies have to be made.
Look at it this way. If you want a model of a boat, just *one* model, it'll cost you several thousand dollars to have a professional modeler make it for you. That's an expensive model. Boat, boat models in plastic only cost pennies each to produce.
Yes, but the *mold* and injector equipment cost tens of millions.
For one boat this is doofey.
For one CD it's cheaper to spend $.25 on a blank and pay someone $5 to do the burning ( and if you're selling for $15 you turn a profit).
For a million CD's it's cheaper to make a mold, buy expensive machines to crank blanks through the in less than a second each with essentially no expensive human labor involved.
It's just like any other One Off vs. Mass Produced economy of scale problem.
Oddly enough this is exactly the same reasoning I use for removing figure skating as part of the Olympics(tm, please submit your bribes early and often).
I like figure skating, but believe it's an artform, at least once the compulsories have been gotten past.
"I don't know art, but I know what I like" Which makes all art awards little more than a popularity contest among the particular group of judges.
As opposed to, say, the marathon, where the guy who crosses the line first wins.
Some actors say, "To be nominated is honor enough." I think those are the actors with their heads screwed on straight.
just the opposite. The Acadamy largely refuses to recognize good films that aren't popular. Mr. Redford founded the Sundance Film Festival for just this reason.
I think you're confusing a bias against popular films (like, say, Gone With the Wind, Oscar winner, most popular of all time, and really bit of a piece of crap) with a bias against Burt Reynolds.
If I ask a humorist directions, and he gives them to me, where is my cause for dissatisfaction? ( Unless he gives me "joke" directions off a pier or something).
Well there ya go. I *wasn't* talking about chess. I was talking about AI.
As for the white to win I've already posted a stipulation to an AC that I just pulled it out of my ass. It makes no difference to any AI argument I've made. So what the hell.
You have missed the point entirely that I have just demonstrated a game that can be played on a 3x4 board (since godzillion was left undefined) which has an infinite number of possible moves ( a number much, MUCH larger than the number of atoms in the universe, in fact, infinately larger) and thus can't be solved by a computer at all by your reasoning.
Yet a five year old can solve it in a matter of a few seconds.
Yes, this directly relates to the problem of AI in playing chess. If you have enough RI to see it.
Therefore it's impossible to play a game of chess.
Think about it.
Humans routinely solve problems with computational trees larger than the number of atoms in the universe. It's *easy* for humans. It's impossible for computers.
Not to mention solving problems that aren't computable by *any* number of discrete steps.
By the way, who told the computer to ignore certain lines? It didn't figure it out for itself.
The question isn't whether they do, it's whether they *have* to.
You *can* learn to play an excellent game of chess by being told nothing but the rules. And many have. A few have even learned the rules just by observation.
The fact that many ( and I might argue most) people are idiotic robots would be beside the point if it weren't for the fact that even chimpanzees have shown greater problem solving abilities in certain areas than any computer has, or can be shown to be capable of.
Look, I'm not arguing that AI isn't possible. I don't see any inate reason why people can't build some sort of machine that can "think."
I'm simply stating that showing it can add two beads to two more beads and then hold them up for review ain't it. A Z80 isn't smarter than a human because it can recalculate a spreadsheet faster than a whole room full of accountants.
The accountant knows what the spreadsheet *means* (whatever that means).
Deductive and inductive reasoning are the key to intelligence and that has to be the proper goal of AI, and a computer that can't do it ain't smart.
*A* human was asked to determine whether an object contained the specified weight in gold, and the result was the law of displacment.
I've only asked for the Towers of Hanoi, which I was able to solve without hints. Is it really too much to ask a computer to demonstrate the same skill before I aknowledge it as my equal?
Now how about a computer that *concieves* the Towers of Hanoi.
More specifically it means "under" (sub) "judicial review" (judice).
In other words rules of behaviour in public that could effect the case apply. This includes making statements to the public that could influence a judge or jury. Or for that matter doing *anything* that could blow your case, like, ohhhhh, confessing to a buddy or harassing a witness.
Under some jurisdictions violating sub judice can actually bring charges of contempt of court. I don't know if Columbia is one of these.
Sequestering jurors is based on the principles of sub judice.
Or, as it is more commonly explained by lawyers to their clients:
"Look, just stay home and keep your damned mouth shut. Ok?"
By definition any game solvable by a human is solvable by a computer programed by a human.
That isn't an example of AI. That's an example of solving an equation. A very complex equation, perhaps, but in essence it's no different than than a computer adding 2 and 2 and getting four, and no more proof of intelligence.
Here's something for you to try. Write a computer program for Tic Tac Toe that only knows the *rules* and *deduces* on its own the perfect strategy.
Yes, there is, but I haven't been able to find it on the web ( which it predates).
It either hasn't been posted or I havn't "solved" my search parameters.
Here's a simple logical proof that a "perfect" chess playing computer can't be made though ( perfect being defined as it wins every game whether it plays black or white).
If it played itself, one of it could only draw at best, since, by the rules, there cannot be two winners.
The number of "moves" possible in Go is so large that it hasn't even been determined how many there just might be.
You're falling into the same trap as the "chess guy." Mathmatically the number of possible moves does not necessarily have anything to do with the complexity of the game as most of the moves may well be trivial moves a human player would never make.
The *rules* are critical to the complexity though.
To the extent that a game follows mathmatical law Godel's Theorem applies and one could construct a game with true, but unprovable, theorms of solution.
The number of possible *discrete* states is completely irrelevant.
Most of them make no sense. To an intelligent being.
The question isn't how many states, the question is the complexity of the mathmatical algorithm required to solve it.
It would be perfectly possible to create a game ( a rather pointless one) with 10 times the number of possible states of chess but solvable quite easily by any freshman in mathematics. Or a child.
For instance, take a "chess" board a godzillion+1 x godzillion in size. Each player gets one token. The players place their tokens on diagonal opposites. Each player can move their token one space in any direction. First player to reach the other side wins.
The winning strategy is both simple and obvious, and first player to move wins every game. Despite the large number of possible states.
You are confusing big and impressive numbers with complexity.
The most trivial ( and for AI purposes useless) example is the single fair toss of a fair coin.
Parchisi is a more advanced version. It is not resolvable to a guarunteed set of moves to ensure victory ( a draw being provably impossible).
This is why games involving chance are so popular with the, ummmmm, populace. A five year old can beat an 80 year old who has been playing parchisi all of his life. Whereas the 80 year old will *never* lose a game of tic tac toe to the five year old, because it is not only solvable, it has been solved.
Various form of games of cooperation are also unsolvable because it isn't even possible to define in advance what the goal is, let alone a stratagy to achieve it.
Your point of view ( and that of many AI researchers) relies on the concept that universe is a predictable machine if one simply knows all the parameters, whereas it is now known that even if all the parameters are known results may be fundamentally unpredictable.
By using the mathmatical ( which includes the logical) algebra.
Some games are reducable to such analysis. Chess is one of those games. While certain theorems regarding chess are as yet unsolved ( like what the perfect sequence of moves is) some of them are.
Among the theorems that have been proven are that there is at least one perfect stratagy and that if white plays it, white will always win.
This is one of the reasons a perfect chess playing computer is being sought. It is *known* that it is possible.
The search continues for that stratagy, either by empirical example or mathmatical analysis.
The point being that chess is a, theoretically, *solvable* game. The precise solution isn't known, although we have a good deal of empirical data regarding possible solutions. (Although white to win has been proven)
The chess computers rely on this empirical data, not on thinking. They *compute.* Big deal.
To really demonstrate a machine that has something of the sort that could be truely called AI it will have to compete with a human player on at least a near even level at a complex and *unsolvable* game.
Chess is the beginners level of game playing computers, and they're just about "getting there." Go is the Holy Grail, and they ain't even close. To date no one has made a Go playing program that can reasonably hold it's own against even a relative novice.
Slashdot Mirror
And I meant to type "coagulated" linseed oil. That would have been funnier, but I spent an hour cleaning out my old paint pantry this morning and the "boiled" just slipped out of my fingers somehow.
Hey, just so I don't ruin your day even further did you know that Formica got it's name because it was, literally, invented For Mica, a mica substitute, as an electrical insulating material, and that it's made primarily from kraft paper?
Because if you don't, I'll just keep my mouth shut.
It's amazing how many of our "luxury" items are really made from garbage (literally) and other bits of old junk that people would be hesitant to pay a dime for if they knew what it really was.
KFG
came to own the rights to the Folkways catalog?
Moses Asche gave it to them. It was a donation.
This could stand as a good model for titles that have been removed from the catalog.
Plus, you could even turn a profit. The Smithsonian is a *non profit*, donations are tax deductable.
Art collectors take advantage of this fact all the time. Why shouldn't the music industry?
KFG
Because mass produced CD's are actually *pressed.*
Expensive dies have to be made.
Look at it this way. If you want a model of a boat, just *one* model, it'll cost you several thousand dollars to have a professional modeler make it for you. That's an expensive model. Boat, boat models in plastic only cost pennies each to produce.
Yes, but the *mold* and injector equipment cost tens of millions.
For one boat this is doofey.
For one CD it's cheaper to spend $.25 on a blank and pay someone $5 to do the burning ( and if you're selling for $15 you turn a profit).
For a million CD's it's cheaper to make a mold, buy expensive machines to crank blanks through the in less than a second each with essentially no expensive human labor involved.
It's just like any other One Off vs. Mass Produced economy of scale problem.
KFG
Oddly enough this is exactly the same reasoning I use for removing figure skating as part of the Olympics(tm, please submit your bribes early and often).
I like figure skating, but believe it's an artform, at least once the compulsories have been gotten past.
"I don't know art, but I know what I like" Which makes all art awards little more than a popularity contest among the particular group of judges.
As opposed to, say, the marathon, where the guy who crosses the line first wins.
Some actors say, "To be nominated is honor enough." I think those are the actors with their heads screwed on straight.
KFG
just the opposite. The Acadamy largely refuses to recognize good films that aren't popular. Mr. Redford founded the Sundance Film Festival for just this reason.
I think you're confusing a bias against popular films (like, say, Gone With the Wind, Oscar winner, most popular of all time, and really bit of a piece of crap) with a bias against Burt Reynolds.
KFG
who said any of them were supposed to be funny?
If I ask a humorist directions, and he gives them to me, where is my cause for dissatisfaction? ( Unless he gives me "joke" directions off a pier or something).
KFG
If you don't see the humor in using burlap soaked in boiled linseed oil as quality luxury flooring, well, I guess you just aren't looking hard enough.
KFG
Well there ya go. I *wasn't* talking about chess. I was talking about AI.
As for the white to win I've already posted a stipulation to an AC that I just pulled it out of my ass. It makes no difference to any AI argument I've made. So what the hell.
KFG
I just pulled that out of my ass.
It certainly makes no difference to any of my arguments.
KFG
When CompUSA comes up on /. as a topic it's usually trying to figure out which one of your possible selections the staff belong to.
Go figure.
KFG
You have missed the point entirely that I have just demonstrated a game that can be played on a 3x4 board (since godzillion was left undefined) which has an infinite number of possible moves ( a number much, MUCH larger than the number of atoms in the universe, in fact, infinately larger) and thus can't be solved by a computer at all by your reasoning.
Yet a five year old can solve it in a matter of a few seconds.
Yes, this directly relates to the problem of AI in playing chess. If you have enough RI to see it.
KFG
Therefore it's impossible to play a game of chess.
Think about it.
Humans routinely solve problems with computational trees larger than the number of atoms in the universe. It's *easy* for humans. It's impossible for computers.
Not to mention solving problems that aren't computable by *any* number of discrete steps.
By the way, who told the computer to ignore certain lines? It didn't figure it out for itself.
Think about that too.
KFG
The question isn't whether they do, it's whether they *have* to.
You *can* learn to play an excellent game of chess by being told nothing but the rules. And many have. A few have even learned the rules just by observation.
The fact that many ( and I might argue most) people are idiotic robots would be beside the point if it weren't for the fact that even chimpanzees have shown greater problem solving abilities in certain areas than any computer has, or can be shown to be capable of.
Look, I'm not arguing that AI isn't possible. I don't see any inate reason why people can't build some sort of machine that can "think."
I'm simply stating that showing it can add two beads to two more beads and then hold them up for review ain't it. A Z80 isn't smarter than a human because it can recalculate a spreadsheet faster than a whole room full of accountants.
The accountant knows what the spreadsheet *means* (whatever that means).
Deductive and inductive reasoning are the key to intelligence and that has to be the proper goal of AI, and a computer that can't do it ain't smart.
*A* human was asked to determine whether an object contained the specified weight in gold, and the result was the law of displacment.
I've only asked for the Towers of Hanoi, which I was able to solve without hints. Is it really too much to ask a computer to demonstrate the same skill before I aknowledge it as my equal?
Now how about a computer that *concieves* the Towers of Hanoi.
KFG
More specifically it means "under" (sub) "judicial review" (judice).
In other words rules of behaviour in public that could effect the case apply. This includes making statements to the public that could influence a judge or jury. Or for that matter doing *anything* that could blow your case, like, ohhhhh, confessing to a buddy or harassing a witness.
Under some jurisdictions violating sub judice can actually bring charges of contempt of court. I don't know if Columbia is one of these.
Sequestering jurors is based on the principles of sub judice.
Or, as it is more commonly explained by lawyers to their clients:
"Look, just stay home and keep your damned mouth shut. Ok?"
KFG
No, you misunderstand. See the guy above who posted about the halting problem.
Certain problems can be demonstrated to be solvable, but not reducable to a mathmatical solution.
Chess, for instance, has been shown to be so reducable, even though we don't know what that solution *is.*
No one who knows what they're talking about has *ever* claimed that a computer being able to solve an algorithm is a sign of intelligence.
What you find interesting is simply how many things that aren't obviously mathmatically predictable that turn out to be so.
KFG
By definition any game solvable by a human is solvable by a computer programed by a human.
That isn't an example of AI. That's an example of solving an equation. A very complex equation, perhaps, but in essence it's no different than than a computer adding 2 and 2 and getting four, and no more proof of intelligence.
Here's something for you to try. Write a computer program for Tic Tac Toe that only knows the *rules* and *deduces* on its own the perfect strategy.
Not so easy.
Now do it again with the Towers of Hanoi.
KFG
Yes, there is, but I haven't been able to find it on the web ( which it predates).
It either hasn't been posted or I havn't "solved" my search parameters.
Here's a simple logical proof that a "perfect" chess playing computer can't be made though ( perfect being defined as it wins every game whether it plays black or white).
If it played itself, one of it could only draw at best, since, by the rules, there cannot be two winners.
KFG
The number of "moves" possible in Go is so large that it hasn't even been determined how many there just might be.
You're falling into the same trap as the "chess guy." Mathmatically the number of possible moves does not necessarily have anything to do with the complexity of the game as most of the moves may well be trivial moves a human player would never make.
The *rules* are critical to the complexity though.
To the extent that a game follows mathmatical law Godel's Theorem applies and one could construct a game with true, but unprovable, theorms of solution.
Even one with a fairly limited number of moves.
KFG
The number of possible *discrete* states is completely irrelevant.
Most of them make no sense. To an intelligent being.
The question isn't how many states, the question is the complexity of the mathmatical algorithm required to solve it.
It would be perfectly possible to create a game ( a rather pointless one) with 10 times the number of possible states of chess but solvable quite easily by any freshman in mathematics. Or a child.
For instance, take a "chess" board a godzillion+1 x godzillion in size. Each player gets one token. The players place their tokens on diagonal opposites. Each player can move their token one space in any direction. First player to reach the other side wins.
The winning strategy is both simple and obvious, and first player to move wins every game. Despite the large number of possible states.
You are confusing big and impressive numbers with complexity.
KFG
Chaos theory can be used to prove it.
The most trivial ( and for AI purposes useless) example is the single fair toss of a fair coin.
Parchisi is a more advanced version. It is not resolvable to a guarunteed set of moves to ensure victory ( a draw being provably impossible).
This is why games involving chance are so popular with the, ummmmm, populace. A five year old can beat an 80 year old who has been playing parchisi all of his life. Whereas the 80 year old will *never* lose a game of tic tac toe to the five year old, because it is not only solvable, it has been solved.
Various form of games of cooperation are also unsolvable because it isn't even possible to define in advance what the goal is, let alone a stratagy to achieve it.
Your point of view ( and that of many AI researchers) relies on the concept that universe is a predictable machine if one simply knows all the parameters, whereas it is now known that even if all the parameters are known results may be fundamentally unpredictable.
KFG
By using the mathmatical ( which includes the logical) algebra.
Some games are reducable to such analysis. Chess is one of those games. While certain theorems regarding chess are as yet unsolved ( like what the perfect sequence of moves is) some of them are.
Among the theorems that have been proven are that there is at least one perfect stratagy and that if white plays it, white will always win.
This is one of the reasons a perfect chess playing computer is being sought. It is *known* that it is possible.
The search continues for that stratagy, either by empirical example or mathmatical analysis.
KFG
The point being that chess is a, theoretically, *solvable* game. The precise solution isn't known, although we have a good deal of empirical data regarding possible solutions. (Although white to win has been proven)
The chess computers rely on this empirical data, not on thinking. They *compute.* Big deal.
To really demonstrate a machine that has something of the sort that could be truely called AI it will have to compete with a human player on at least a near even level at a complex and *unsolvable* game.
Chess is the beginners level of game playing computers, and they're just about "getting there." Go is the Holy Grail, and they ain't even close. To date no one has made a Go playing program that can reasonably hold it's own against even a relative novice.
KFG
Bury the buggers in Japanese copiers, that'll show 'em.
KFG
I'm having fun watching how many people seem to have *totally* missed the point of your post, which was rather amusing itself.
Good show.
I'm a cat person (no, not THAT kind of cat person) so I'll have to reward you with Meow Mix.
KFG
comes around behind the lines and pops me from the back.
Go figure.
KFG