Slashdot Mirror
Turing Machine Implemented in Life
PixelJuice writes "Paul Rendell has implemented a Turing Machine in Life here. Perverted, but still kind of impressive. The site also contains a few useful links to Turing Machine principles and Life components." Normally I save this sort of stuff for the quickies, but this is to out there. I can't believe this works... but wow. (CT:Link seems to have gond thud. But thanks for the hate mail reminding me not to forget the letter v. I never knew a single letter deserved so many 4 letter words. Makes me love this job oh so much)
To understand the halting problem, you first have to understand what a Turing Machine is. As mentioned in posts above, a Turing machine is an infinitely long "tape" (magnetic, hole-punch, whatever) which can be filled with "symbols" (instructions, data, whatever). In addition, there is a "head" which can move back and forth over the tape and read a symbol or write to it. The Turing Machine "runs" by reading some starting symbol on the tape carrying out the instruction that symbol represents. As you can imagine from the structure of the Turing Machine, there aren't too many instructions that are capable of being performed - move left or right n units, or basic increment & decrement operations to symbols themselves.
In a nutshell (I'm glossing over details a little here), everything you consider to be a digital computer is just a form of Turing Machine, and all Turing Machines are equivalent in capability (assuming they have infinite memory), as long as you don't care about how fast they are. In other words, you could emulate a Linux box using the Life Turing Machine if you had enough memory. It would be very slow.
Anyway, as far as the halting problem is concerned, the main thing that interests people studying Turing Machines/programs is whether the head moving back & forth over the tape ever stops - in other words, will it get to the 'end' of the program. This isn't the same thing as 'failing' as you would normally consider. This is why a compiler can't determine whether a program halts. For example, if we consider C, the following program neither halts nor fails:
main()
{
while (1);
}
Whereas this program 'fails' and halts:
main()
{
int foo = 0;
int bar = 10 / foo;
}
and this program doesn't 'fail', but it also halts:
main()
{
int foo = 2;
int bar = 10 / foo;
}
All these programs compile, but only the last 2 halt. Although a person can look at a simple program and tell if it halts, the Halting Problem proof (which I can't remember) shows us that is not algorithmically possible to determine if a program halts without actually running it, and even then, we only can show that a program halts, but not that it will never halt (since we'd have to run it an infinitely long time to demonstrate that it doesn't halt).
This proof actually has very broad applications in mathematics beyond computer science. It has also been used to show that given a set of primitive 'axioms' (like the basic axioms of Euclidean geometry), it is possible to construct truths which cannot be proven by the same set of axioms. It's been about 10 years, but if you're really interested, read the book "The Emperor's New Mind" by Roger Penrose, an extremely insteresting & informative survey of 20th century math & science. It's somewhat flawed by Penrose's personal intellectual biases (for example, he refuses to accept that randomness has any important role in the way that intelligence/conciousness works), but its packed with nerd food.
In that case, it was being used in an entirely question-begging sense.
-- the most controversial site on the Web
Not quite. You're forgetting the eensy detail about having an infinite Life grid (or Turing tape) to work with. But for practical purposes (except that there aren't many practical uses for Turing emulation) yes, digital computers do the same stuff. That's a big reason why people still learn about Turing machines.
Hmm... if you could create a physical Life grid (not necessarily infinte, but it would have to be really big) it might be able to solve NP-complete problems quickly due to massive parallelism (each cell being a very simple ALU & 1 bit memory).
And YOU are not intelligent. It's actually a bundle of neurons deep within your cerebrum. You are a meatspace cellular automaton.
Pure Heaven! We had it much rougher! Back in my day, Turing hadn't even come up with the Turing Machine yet. We had to think up the idea ourselves, implement it with watery dirt from the bottom of a lake, work 28 hours in the mill, and then forget how we did it, every day!
Someone you trust is one of us.
No, because it just happened to be. You're really going need to be clear about the meaning of a "necessary condition" if we're going to avoid talking at cross purposes.
When you were a kid, you didn't know shit about the world. The teacher passed out pointed out green leaves and you, in your infantile mind, saw no necessity for it to be so. Now having learnt about chlorophyll, you do. So were you unintelligent before? After?
No, you're not getting it. Your concept of "chlorophyll" is necessarily connected to chlorophyll because of a causal chain of events between the set of objects which are chlorophyll and the idea of it in your brain. And it has to be that way. The lookup table's entry for chlorophyll might or might not be causally related to the set of objects which are chlorophyll, but certainly don't have to be in any systematic way. Therefore, the lookup table is just a lookup table, and its set of ordered pairs cannot be interpreted as meaningful utterances.
There are more promising arguments for artificial intelligence, but surely there must be something intuitively wrong to you about a criterion of intelligence that doesn't distinguish between human beings and lookup tables?
-- the most controversial site on the Web
And since both of *those* are implemented in the laws of physics, then shouldn't physics be used as the base model of computation?
No.
Why? Because the model of computation that we already have has more power to explain computation. Just because you can build a computer out of Life patterns does not mean that those patterns are inherently computational. You could build a computer out of Legos if you wanted to, or photons, or anything else. You could argue that the "most simple" material used to construct your computer was the "most basic" and therefore "most true" model for a computer. But that would not give you any more insight into what a computer is.
What does give you this insight is not the knowledge of what the computer is made out of, but how the components function and work together to do the computing process. The "count-the-neighbors, add births, remove deaths" rules for life give none of this information.
That a Turing Machine can be implemented in Life is very interesting, but it must be understood that Turing Machines in Life universes are engineered objects, and that this engineering requires quite a bit of additional knowledge than the simple rules for the Life universe.
You see? You see? Your stupid minds! Stupid! Stupid!
You mean I have no recourse to answering this question apart from metaphysics? Gimme a break! I can give it food to see how it responds, shine light on it, etc. And you think these things don't mean a thing? That is possible for a sea slug to be unresponsive, but yet capable of abstract mathematical thought?
No serious researcher gets bothers to defend metaphysics anymore. You are fighting a battle lost a hundred years ago.
MSK
Now, implement Life in BrainFuck and run it on your Life BrainFuck engine...
Sorry, the tape is not infinite, but merely unbounded.
"Rub her feet." -- L.L.
Correction: It is well known that it was theoretically possible to simulate a Turing Machine in the Life universe. A Turing Machine had never been fully implemented before, but various Life patterns had been discovered or engineered which could carry out the functions of various components of a Turing Machine. But the entire thing had never been put together, it being a monstrously complex and unbelievably huge pattern.
You see? You see? Your stupid minds! Stupid! Stupid!
Yes, well. It's not an easy line to draw with geeks, is it? www.unixsex.com
- undoware.ca
Now surely, somebody has used some kind of diagonalisation to implement the game of life on a Turing machine. How powerful does the hardware need to be to run a single Life->TM->Life stack at faster than glacial speed?
I in no way believe that your table-method is a feasible mechanism for creating an intelligence. However, if it was implemented, and was completely 100% indistinguishable using any and all conversational techniques from a real intelligent person, then it should be deemed intelligent.
But, like I said, I am almost 100% sure that your proposed method for producing such intelligence is fundamentally flawed and impossible.
Responses which do not mean anything are not evidence of intelligence.Understand that what your describing does not satisfy the conditions of the Turing test. You are arguing that because it cannot always come up with an intelligent response, it is not intelligent. Turing would completely agree with you, as would I. So far this is completely valid according to the Turing test.
So now, exclusively consider if you can't tell the difference. No more "It's not feasible", or "You would be able to tell the difference" statement. If to all the tests you apply and that can be applied, it appears to be intelligent, is it? If not, why?
Then you should probably avoid cheap Chinese restaurants.Invalid argument. I said "cannot be distinguished", and I suspect a simple analysis (DNA or other chemical analysis) would determine what I was eating. Therefore, I would be able to distinguish your "unknown" thing from a duck unless it was a duck.
Several years back I read a chapter from "Brainchildren" by Daniel C. Dennett. According to Dennett, Turing's test was meant to be a show stopper for those endless philosophical discussion about what intelligence is. Turing's idea works just like auditing for an orchestra. In the case of orchestra, the candidate should only be judged based on his/her musical performance rather than his/her appearance. So if you want to judge whether a machine is intelligence, we can reuse the same idea: put that "thing" into a black box and talk to it. In the case of Turing's test, the ability to perform intelligence conversation is used as a judging criteria for intelligence.
As you probably figure out now, Turing's test is not meant to be "fool-proof". Rather, it is a test that is meant to be good enough. If a machine passes the Turing's test, then it is reasonable to consider it as intelligence. But not all intelligence machines will necessary pass the Turing's test. Also, keep in mind that the test is meant to be executed a lot of times by different judges asking questions which could be simply "off-the-wall" (i.e. out of the machine's areas of expertise). Passing the "pure" Turing's test is not as easy as it seems.
Turing is no fool. At least, he already accounts for inhuman-like intelligence in his infamous test.
It is this offhand dismissal of this large body of work that deserves to criticized. Go ask streetlawyer about that. He appears to know lots, having spoken to so many researchers.
Not wanting the Enemy to bomb you is the mother of lots of fun stuff... :)
-c
I have discovered a truly remarkable proof which this margin is too small to contain.
IMHO, the only way to adequately refute the Turing Test as a measure of intelligence is to refute Turing's theory on what intelligence is. Turing defined intelligence recursivly, an intelligent thing is one which can convince another intelligent thing of it's intelligence. If yout take "human" as the benchmark of what is intelligent, then any computer that can convice a significant cross section of humanity of its intelligence is intelligent. The Turing Test is intimatley linked to Turing's theory on what intelligence is, and is the perfect test for what Turing defined as intelligent. To discredit the Turing test you first have to disprove his definition. The trick is that in order to really do that you have to have an alternative definition. So just come up with an encompassing definition of intelligence, then you can build a test for it. Of course to be accepted, your definition will have to be considered intelligent by those who study intelligence. So for all intents and purposes your new defintion of intelligence will have to pass a Turings Test for intelligence, in order to be accepted as intelligent... Which kinda argues in the guy's favor.
I don't need a million points of light, just two points of multi-mode fiber and a 10 Gig-E router.
Remebering is simply an ability to store information. If you call that intelligence, then a card index is intelligent.
As for your duck example, what if I were to then demonstrate that it was NOT a duck? Would it still be a duck? Would it just cease to be a duck? Or would it just prove that it was never a duck in the first place?
Remebering is simply an ability to store information. If you call that intelligence, then a card index is intelligent.
As for your duck example, what if I were to then demonstrate that it was NOT a duck? Would it still be a duck? Would it just cease to be a duck? Or would it just prove that it was never a duck in the first place?
I said nothing about being human.
Philosophical questions regarding what it is to be human are irrelevent. The question is, ONLY how to determine if something is intelligent.
Do not equate being human and intelligence. I didn't.
Turing said that "if a machine could convince an interrogator that it was human, then the machine should be considered intelligent". But you didn't even capture the same logic in your counter statement, so you haven't proved Turing is a fool, or anything else. Had you said Does this mean that If I were to convince an interrogator that I was Chinese I should be considered to be intelligent? you argument would have meant something... assuming you're not Chinese... I would say you are intelligent.
You two are arguing what amounts to information theory's equivalent to the abortion debate. The argument cannot be resolved or 'won' with logic, since the disagreement between the two sides lies in the most basic assumptions that each party holds about the nature of conciousness & intelligence. BMazurek is arguing the 'strong AI' position (one that I hold), streetlawyer is arguing the 'weak AI' position (held by Roger Penrose, among others).
I imagine that BMazurek probably also believes that what we call 'conciousness' is also an illusion, an aritifact of the emergent behavior we call intelligence (as I do). Some people won't except that idea, and hold to the 'weak AI' position, which more or less says that what is commonly considered conciousness or intelligence is something beyond what can currently be defined by science, and that an algorithmic computer cannot ever be intelligent, since it is impossible to model intelligence algorithmically (it is called 'weak AI' because they do believe that algorithmic computers can be used to achieve certain subsets of behavior that can be useful (expert systems, vision, etc.), but cannot achieve 'conciousness'). People like Penrose use the Halting Problem and work by Euler to justify this position 'mathematically', but only with an extreme amount of handwaving (in my opinion).
The problem I have with Turing Test opponents is that they have no proper means of determining if something is 'intelligent', or if they do have such means, they invariably would exclude 99% of the Earth's current human population by setting impossibly high standards. In they end it boils down to faith, in that they say, "Humans are intelligent because I know I'm intelligent, and since I'm a human, all other humans are intelligent, too", which is certainly one way of 'proving' humans are intelligent (to an individual), but useless for determining if non-humans (living or otherwise) are also intelligent.
Rob (or Hemos), get back to that terminal and fix that!
"Ancillary does not mean you get to rule the world." --U.S. Circuit Judge Harry Edwards, speaking to the FCC's lawyer
Actually Conway himself published this result in 1983 For a reference see: Winning Ways for Your Mathematical Plays (1983 Berlekamp,ER Conway,JH Guy,RK, Academic Press ,New York)
I mean, can a human being tell it apart from another game of Life?
The Turing Test Page
... "Could you imagine a Beowulf Cluster of these things?"
*goes back to trying to get QMail up*
The Church-Turing principle only claims that there is no computer that can solve more problems than a Turing Machine. Clearly there are computers that can run faster than the fastest TM, but that's irrelevant. I'm not sure why this guy is implying otherwise; maybe he's just caught up in the "quantum computing" hype.
MSK
Dangit, meant to change the subject!:
u k.co/gol/tm.htm+life+turing+machine&hl=en
The link appears to be dead, but not in Google:
http://www.google.com/search?q=cache:www.rendell.
Or did you have to work at it? Remembering is a small part of learning. "Learning" by rote is just remembering. Nobody has tried to teach that way for almost a century, because it doesn't work. Learning means you know how to apply knowledge too.
It's a little after noon EST and this link appears to be broken. Did the /.ing of this page crash the server? Anyone got a mirror on this?
I need a funny sig
I now have the most pounding headache just looking at the initial pattern, let alone reading the description. How the hell did the guy figure out the components of a Turing machine with discrete Life patterns? How the hell did he get them to work together like that? How the hell do you manage to convert something as esotoric as a "Stack" into a Life pattern?
This just scares me....
Syllable : It's an Operating System
Just wish I could find a functioning link to this one. Anyway, given Turing's insight, and the period the paper would almost have to be novel and original. Why its still classified, who knows? Maybe the offical secrets office or whatever they call themselves. What strkes me about WWII is that almost everything we now do with signals and communication comes out of that one time. Well okay, maybe not satellite and digital, but even so, the foundations were laid then. Lot of innovation in a very short time.
get a life... (or a brainfuck or a aw shucks now you got me all confusified)
Secondly, no one ever takes what he says about AI as gospel. He is an early pioneer into the science of cognitive awareness. That is like saying that we take everything that Newton said as the utter fact without striving for more. There is so much more out there, he was just one famous name that did a lot of ground breaking work.
Also, The turing machine is a mathematical model of a computer that can do anything, but because it involves an infinite in the formula it is more or less impossible to tangibly create.
And also, who are you to declare what would have come into existence? Are you omnipotent? I sure hope not, because if omnipotence is handed out to those as closed-minded and ill informed as yourself than I'm surprised the universe hasn't folded into a singularity yet.
Dacels Jewelers can't be trusted.
The link in the slashdot story is incomplete. The page on that server does not exist. Nobody on this board even has a clue where the page is
I think you can do it with some kind of white noise generator, for starters. I think that the Intel 815 chip set has a hardware random number generator in it..
Don't search for extraterrestrial intelligence - create them yourselves!
Think about it, it works.
MSK
The entire question of being and sentience is a metaphysical question at its very heart, one which Dennet, Descartes, and others have been quabbling over for centuries. Your views on "intelligence" and "sentience" will essentially come down to which broad metaphysical category your beleifs fall into (i.e. materialism, dualism, etc.).
Trying to dismiss metaphysics as being "irrelevant" is to essentially throw out the window these hundreds of years of logical proofs and discourses by some incredibly intelligent people.
I think you're trying to fight a battle that you haven't really done any research into.
---------
"Cut word lines. Cut music lines. Smash the control images. Smash the control machine." - William S. Burroughs
This is the reason I endure weeks of half-way interesting /.-articles; to read comments like this.
I couldn't stop laughing. Thanks!
I respectfully disagree; I particularly disagree with your definition of 'insight'. Insight is that spark of brilliance that cuts through a problem and either reduces it down to its essense or, ultimately, solves it. This is not a communal process. It's something that takes genius. True, unadulterated genius.
If insight was indeed a communal process, frogleaps in science would be commonplace. You wouldnot *have* a list of brilliant scientists to list, like you did. The fact that you do proves that someone (Einstein, Archimedes...) made that frogleap. That was the insight. The people around them set the stage, defined the problem, then poured over their solutions, validating them and ultimately making them famous; but they didnot provide any insight...
We used a TM simulator in my "Philosophy and Computers" class. It had an unnecessarily clever name, and an unnecessarily clever interface. It also ran only on MacOS. Anybody know the one I'm talking about? It's right on the tip of my tongue.
Anyway, since I found the above program a little too constraining, I just put one together in MS Access in about fifteen minutes. It's really pretty trivial, especially if you understand the significance of TMs to begin with.
If you're lazy, a simple Google search turns up many other prospects.
MSK
Interesting... I just did a Google search for "entymology," and it returned a mix of pages about etymology and entomology.
Bugrit! Millenium hand and shrimp!
You're not confident in convincing other people?
Tough luck...
In principle, it all *CAN* be faked. Whether or not it can in practice is another matter entirely. Depending on your metaphysical position (materialism, dualism, etc.) one can argue, with alot of logical backing that the operational definition of intelligence is either totally valid, or totally invalid.
-----
"Cut word lines. Cut music lines. Smash the control images. Smash the control machine." - William S. Burroughs
we're just all hoping that artificial intelligence will make up for a lack of natural intelligence.
How did they implement the infinite tape?
Just wondering B*)
SLOGEN [ http://ungdomshus.nu : Sebastian cover music]
Uhh, I'm kinda confused about the 'Halting Problem' or 'Halting Program'. One of the descriptions of the problem is that no program could be made to determine if another program will fail... Based on my __LIMITED__ understanding of this concept, I would argue that statement. Wouldn't a COMPILER be a program that determines if another program will fail? I'm expecting to be proven wrong _very quickly_ , (so don't let me down), because this seems too simple
I said "done properly". By your definition of pass, those computers that fooled 50% of the people have a 50% chance of being intelligent. If it passes one person, it is intelligent. I don't think that is the case. So, in real terms, what is a Turing test? How many people? How long for each person? What percentage have to believe in the intelligence? Theoretically, it seems like it would have to be infinite people/infinite time. But no human could pass that. They'd either stop answering due to boredom, or forgetting that, die well before an infinite test was complete.
yeah, .co is Colombia for fsck's sake. What's this guy trying to pull?
Pope
Freedom is Slavery! Ignorance is Strength! Monopolies offer Choice!
It doesn't mean much now, it's built for the future.
I don't see how you, being human, is not a necessary condition for there being a spinal cord. And having qualified by statement, restricting attention to those with functional spinal cords and functional knees and legs. Then it is necessarily true that such knees would jerk. It is not "just happens" to be true. The chain of causation is very strong in such a case.
Here's a link to a pretty cool applet that shows exactly how CA works in real time.
This applet will even let you choose the size of the matrix you want to play with, set your own start pattern, and count generations as it goes. (It also shows the famous r-pentonimo we've all come to know and love)
Group 2 can be discarded. Group 1 (CLASS = Programs) can be further subdivided into:
Programs that fail
Programs that do not fail
--
--
Pay no attention to the errors in my post. I am the great and powerful Oz.
I was referring to the lack of logical meaning, when the subset of the lambda-calculus is mapped to the intuitionistic logic. Church, Curry, and many others doubted the logical meaning of such lambda-expressions. Sure, you could define it to do something, as you stated, but you can do that with any formal mathematical construction. Therefore, it might mean something in a formalism sense, but in a logical sense... it cannot be mapped to a logic proof, like many other lambda-expressions, and therefore, its meaning on a intuitive mathematical level, comes into question.
I particularly disagree with your definition of 'insight'
This being slashdot, I was not being precise. I suspect that we actually are in close agreement. I fully agree that the actual "spark of brilliance" is almost always an intensely individual event. How many great works of art were done by committee? I also agree that great insight typically (but not always) requires "unadulterated genius." (occasionally perfectly pedestrian minds stumble across import insights)
My original point is that history shows that the individuals who produce the great insights usually do so in the context of stimulating peer interaction. The thought is wholly their own, but without the context of informed, questioning peers the spark of genius rarely ignites. And it is this that makes me pessimistic about the coming Wolfram contribution. Stephen has _completely_ cut himself off from feedback, constructive or otherwise. He may pull it off, but this approach is at odds with the historical record.
"one treats others with courtesy not because they are gentlemen or gentlewomen, but because you are" --G. Henrichs
Ah, but you'd think that being able to fake intelligence would be much harder to do than merely being intelligent.
;)
Therefore, anyone who could fake it would be quite intelligent indeed.
Heck, just watch the moderation on my posts sometime!
---
pb Reply or e-mail; don't vaguely moderate.
pb Reply or e-mail; don't vaguely moderate.
There was a story on Slashdot recently about a Turing Machine implementation in Minesweeper. The paper is here (in yucky pdf format).
His proof is very similar to the Life proof -- which makes sense, because when you think about it, Minesweeper is a lot like Life. (That's 'Life' with a capital 'L'... I'm not trying be profound here.)
MSK
If I understand Gregory Chaitin's interpretation of this sort of thing correctly, he might say that this is simply a case of an expression that is "randomly" true, i.e. true for no particular reason, thus unprovable, having no meaning, but nevertheless being fundamental to the existence of the formalism to which it belongs. In that sense, its usefulness in the absence of meaning itself has meaning, or at least an explanation, so I can still sleep easy tonight! ;^)
The machine can theoretically have an infinite tape, it is just the creation of the infinite tape which is impossible. However it might be possible to build a pattern which generates tape faster that it is used. In fact I'm confident that such a pattern will be possible but I don't think the necessary bits have been found yet. One slightly underhand way round this is to ask the observer to add the tape as needed. This means stopping time and adding a bit to the pattern then starting time. The machine will act the same if this bit was there before it started or added later and the procedure can be repeated until the observer gets bored. Paul Rendell P.S. The website HAS been \.ed out of existence. The server administration are trying to sort something out but that probability means waiting until this discussion has finished.
It can be shown that a Turing machine on a grid (with move up+down along with left+right) can be simulated by a machine with merely a tape. From there, it can be seen that a Turing machine can simulate a life grid, and so the two are equivalent.
Calmacil
I can't seem to face up to the facts, I'm tense and nervous and I can't relax... --Talking Heads
In a recent submission I posted a review of a new Doctor Who book which just came out. It has been written by Paul Leonard and features as one of the main characters Alan Turing.
It is an interesting book, featuring also Graham Greene the novelist as a character.
It is all about a mysterious encrypted transmission coming from Germany in the 40's. One of the locations in the book is wartime Malta.
SCIREV.NET - fanfics,reviews & more
More ironic that AI people need a book to defend their own field. Go check out McCarthy's book Defending AI Research: A Collection of Essays and Reviews .
I can see we're going to have to start calling you Sherlock. My assertion is that "the Turing test is useless because it could be passed by a look-up table, and look-up tables can't be intelligent".
To "test" whether something is intelligent is a category-mistake in any case. Whether something is sentient (which is implied by the interesting sense of "intelligent") is a metaphysical question.
-- the most controversial site on the Web
Because once something gets classified, it takes about a billion dollars of FOIA lawsuits to change that....
/* This post not warrantied for mission critical applications. */
The problem you note in your second paragraph is indeed a tricky one; however, I have to say that the universe is under no obligation to provide us with any test at all for whether something is intelligent.
-- the most controversial site on the Web
OT question: Does anybody know how the hell the Bomb got its name? Seems a rather odd way to describe a computer...
Bugrit! Millenium hand and shrimp!
You clearly do, and so did my interlocutor. The matter at issue at that point was what "serious researchers" did or did not defend. They do defend the role of metaphysics and do not defend the Turing Test. I called someone on a point of bullshit.
Turing made a good step by suggesting that we define intelligence as a behavioral trait
He made a step. It wasn't a "good step", as it suggested as a criterion of intelligence a property which could be staisfied by a canonically non-intelligent entity (a look-up table). There are numerous behaviourists (going back to Hobbes) who don't agree with Turing Tests in any case, so this couldn't have been his original contribution. In any case, Locke anticipated the Turing Test by about two hundred years, when he discussed whether a talking animal would have to be accepted as human.
Noone has really had any better way of defining it...at least, you haven't suggested one yet
The universe is not under any obligation to provide definitions. The fact that the only definitions suggested have been utterly unsatisfactory may mean that no definition is possible, or that the question is wrongly posed or illusory. It's not necessarily a good thing to do to go running around proliferating bad solutions.
-- the most controversial site on the Web
Therefore, it is not my position that machines can always fake intelligence. But I can accept the operational definition, becuase I know that sometimes we don't make judgements based on the full data. If I can make a quick judgement that someone is intelligent based upon meagre evidence, then a machine could satisfy it.
You may say I am not looking hard enough and you will be correct. Then my question is, which you you contend: that we can always look hard enough past fakery, or fakery will always win?
In fact, I don't see the necesity of believing either statement at all.
I know this is kinda redundant, but...
When talking about Turing machines, no assumptions are made about efficiency. Just because any real, physical computer has restrictions on the number of operations it can perform in a given period of time doesn't mean that you have to place the same restriction on a theoretical Turing machine. Yes, quantum mechanical Turing machines can do some things in fewer steps than their classical equivalents, but the number of steps is usually not a consideration. For the purpose of theoretical discussion, classical Turing machines do just fine. While quantum computing is an extremely important development for the people talking about how fast you can calculate something, they're pretty much irrelevant for the people talking about whether or not you can calculate something.
Bugrit! Millenium hand and shrimp!
rofl
omg, good one.
I think it flew over the moderators heads though.
Sigs are awesome huh?
One could fix it by replacing the tape (stack) with one that extends itself and does not require a pre-existing pattern. It is theoretically possible to build such a tape (using a universal constructor if nothing else), but no efficient method of doing it is known.
People replying to my sig annoy me, that's why I change it all the time.
And by addressing in your sig the issue of why people talking about your sig annoys you, you are just asking to be annoyed. Therefore I feel it is my duty to annoy you.
Long signatures suck.
Wow, a Turing machine implemented in game-of-Life. You could program it to do anything. You could even make a spell checker run in Life! :)
I too remember that begin old news many years ago too
"John von Neumann proved years ago that a universal Turing machine could be realized in two dimensions, and Conway constructed a universal Turing machine in his two-dimensional Life world" (D.C.Dennett "Fast Thinking" The Intentional Stance, 1989).
DCD went on to support 2D intelligence along with a couple of arguments on how speed is necessary to be intelligent.
BTW, to refute the parent post, Life is an interesting simulation BECAUSE of its predictability. From any one state, one can determine ANY of the future states. It is NOT backwards predictable (as you might imagine). This has meaning for us because we live in a world that isn't future predictable but IS past predictable (we know our history but not our future).
---
---
I'm just an ordinary man with nothing to lose.
It wasn't called the bomb, rather it was the bombe. Can't help with the entymology(sp?), though.
Even more OT question. Anyone know of an online entymology(sp?) resource?
this is nothing new.
The cellular autonoma game "Life" was invented by the mathematician John Conway in order to create a Turing maching. Conway wanted to show that cellular autonoma could be used to create Turing-equivalent computations.
Once again, you miss the most point by taking things out of context. The word remember was being used to describe the ability to adapt and derive an intelligent response to a given situation. I will avoid the obvious derision that could be made here.
As for your duck example, what if I were to then demonstrate that it was NOT a duck? Would it still be a duck?Obviously not. It has failed the test for "duckiness".
Would it just cease to be a duck?No, it never was a duck. There may have been a perception that it was a duck, however, your theoretical test, if correct and validated, would show that it never was a duck.
Or would it just prove that it was never a duck in the first place?Precisely.
I don't understand your point.
It's both a play on "Alpha-Bits" cereal and on the notion of the infinite - sorry, unbounded - tape. I think that makes about as much sense as the delightful "frog and mouse" allegory that you linked.
I thought Aleph-Nought is the cardinality of the natural numbers. What do you think it is?
Implementing a Turing machine in Life isn't such a conceptual leap if you remember that Conway invented the game by simplifying the designs for cellular automata that John Von Neumann invented as a conceptual tool for designing computing machines. (Von Neumann's automata had many possible values for each cell, and his goal was to create a system in which the operating instructions were protected by the structure itself from corruption. He also imagined a system that would fully reproduce itself on demand.)
Well, if you're being lazy, you could simply tile your house with a binary encoded version of DeCSS. Then have a webcam point at your floor - any machine around the world could simply decode the b/w tiles.
No Life snapshot required.
-- perl -e'print pack"H*","6e656d6f406d38792e6f7267"'
What he means is that given, say, a specific configuration at step 20495, you can't always determine what the configuration at step 0 is.
Also, you probably coudln't predict, say, if any given block will stay white whilst the counter goes to infinity (such prediction ability plus the above computability result would enable one to solve the halting problem).
John
John_Chalisque
Consider this: if you're determined to prove that machines are fundamentally different from humans in some way that makes them ineligable for intelligence, then you have to also prove that humans are not themselves simply elaborate finite state machines. I don't think you can do this.
As I stated in this comment, humans can decide whether an algorithm/loop/procedure will terminate (or not) depending on inputs. Machines cannot. I'll look up the exact theorem/law/axiom when I get home tonight.
Not having notes or a textbook in front of me, I'm making the assertion (or assumption) that a Turing machine can simulate all finite state machines. In my response to the original post
...then you have to also prove that humans are not themselves simply elaborate finite state machines. I don't think you can do this.
If, according to the halting problem, a Turing machine ( or its equivalent) cannot determine whether a given program halts or not, and a human can, that human cannot be a finite state machine, no matter how "fiendishly complex."
So if you take this program
Begin
i=0
While ( i is less than 10 )
i=i+1
End While
End
and you can figure out that this program terminates then congratulations, you're not a machine. I figured out that it halts, so I cannot be a Turing machine nor can I be a "fiendishly complex" finite state machine. Make the program as complex as you wish, a human will still be able to determine whether it terminates or not. "Likely not correct" and mathematically proven (the Halting problem) are quite different.
and now the fucking link won't even work, it's 404
At this moment in time it's a '404'. If I go to the root of rendell, it comes up with the default Apache page.
Something screwy is going on. Anyone got a mirror?
So here's a key question:
Can every (or any) HUMAN pass the test, done properly?
You're trolling, right? Or are you simply unaware that his highly proprietary, (practically) solo effort Mathematica is one of the most amazing computer programs ever produced?
--
whuppy enjoys smelling like diesel fuel
size of memory isn't usually considered when determining number of states. usually just the controller.
If a machine is clever enough to fake intelligence indefinitely and in all situations, doesn't that make it intelligent?
The meta-Turing test counts a thing as intelligent if it seeks to devise and apply Turing tests to objects of its own creation.
-- Lew Mammel, Jr.
I purchased Mathematica 4.0 100% legally. It is a student version, but it is fully functional. When I first purchased it wolfram demanded that I enter an ID generated by the program, as well as a CD identifier key into their website, which then spit out a serial so i could use it. So far so good. Until I purchased a new Hard drive. apparently, the cd generates a unique key based on your hard drive, no unique key, program no work. Sure, I probably could call up wolfram and ask for a new serial, but do i really need to tell them when i've purchased a new hard drive? Oh, and the kicker is the program will not run off the CD except as a reader. I dont know how they do it, i can set the date back to when i purchased it and it still wont work off the CD, even though after i first bought it it had a 2 week "register or else" fully functional version on it. My advice, dont buy any software that wolfram makes.
Kinda interesting that this discussion on artificial intellegence has so many people calling each other idiots.
No, I don't trust in god. He'll have to pay up front, like everybody else.
It should be stressed that the equivalence, stated above, is a language equivalence. In real world software development, language equivalence has become inadequate because of our highly interactive computational systems. For this, people like Robin Milner have been researching interactive calculi and heirarchies of equivalence relations, such as behavioural equivalence and strong equivalence in the pi-calculus.
Personally, I don't think that it is stressed enough, in undergrad University courses in theoretical computer science, how language equivalence is a very weak equivalence for computational systems. As we have just seen, language equivalence equates two machines, one which is far faster than the other!
For example, two computational machines (not the ones mentioned above) are (language) equivalent, but one of the computational machines can sort a list in constant time, O(1), while the other can only sort a list as fast as O(n*log n). One reason that many people are so excited about quantum computing, is that while it is computationaly (language) equivalent to other general purpose computers, its massive parrallel structure allows it to do things asymptotically faster than the general purpose computers of today.
I think you're really being petty.
How often do you change your hard disk?
I moved my licence from one machine to another, and Wolfram gave me a new code almost instantly.
I've just bought another faster machine, and I'm going to migrate the licence to that now.
It takes 10 times longer to install a new hard disk than it does to type in a new code. Get a perspective, man!
FP.
Also FatPhil on SoylentNews, id 863
No. I'm not. I'm arguing that it can never come up with an intelligent response, because none of its responses are meaningful, because they do not have the correct "word/world link". Even if it could, always, in every case (per impossibile, of course, if I am allowed to ask questions about real numbers) come back with something that appeared to be an intelligent response, I would still say that in no case did it ever come up with something that was an intelligent response.
No more "It's not feasible", or "You would be able to tell the difference"
Since I have never made any such claim, I don't know who you're addressing.
. If to all the tests you apply and that can be applied, it appears to be intelligent, is it?
No.
If not, why?
Because it is a look-up table and therefore there are no necessary connections between the words in its responses and their referents.
I suspect a simple analysis (DNA or other chemical analysis) would determine what I was eating.
You are more wrong than you can possibly imagine being. For a start "duck" is not something which is given to us by nature -- it's a taxonomic term. To give a Turing-style operational definition of "duck" in terms of DNA, you would need (1) a canonical duck DNA sequence, by reference to which all other ducks would be judged (2) a stipulation about how different from the canonical duck a given DNA sequence could be, and still be a duck and (3) a measure of what "different" meant in this context; which similarities and differences mattered.
You mmight be able to get away with an arbitary stipulation in this case, though it would almost certainly defeat you to define "duck" in the Chinese chef's sense (as opposed to the naturalist's). But it's ludicrous to expect to be allowed to get away with a similar move in the case of sentience, which everyone has direct experience of, and which, unlike ducks, is an irreducible concept given to us by nature.
-- the most controversial site on the Web
By "the box" I meant the man+book+room combination. The man doesn't know Chinese, and (I guess I misspoke) the box doesn't know Chinese, but the man in a box knows Chinese just fine. I don't think that your interpretation of Searle's counterpoint applies in this case.
An artificial intelligence binary program written to a hard drive is not intelligent, and the computer running it is not intelligent, but the artificial intelligence program+computer combination is artificially intelligent.
I read Searle's article a while ago and got annoyed and stopped before finishing. So I guess I know the first thing about artificial intelligence, just not the last thing.
--
There are no trails. There are no trees out here.
As I stated in this comment...
Stating doesn't make it so. If you're talking about the Halting Problem, it is correct that a certain class of machines (those with computing power equivalent to a Turing (how's that for irony?) machine) cannot decide whether or not, given a certain input, a machine of this class will halt. However, your contention that humans can accurately make such a decision is likely not correct.
If you want to argue about whether or not the human mind is a fiendishly complex state machine, you can't start by assuming your conclusion.
Floating face-down in a river of regret...and thoughts of you...
How do you construct such a thing? By building up abstractions. You construct "wires" from gliders and mirrors and define basic logic gates. For storage, delay lines (composed of gliders and movable mirrors) are one choice. To put everything together, writing a program that compiles logic into an initial state is probably a good idea.
I have read the Forbes article. Initially, I was intrigued. As Stephen revealed bits of his work, I kept recalling the famous G.H. Hardy quote
Stephen has left the open, peer reviewed world of academia to pursue his highly proprietary , solo research effort. We know how well this works for software, we shall see how that works for math...
"one treats others with courtesy not because they are gentlemen or gentlewomen, but because you are" --G. Henrichs
I wrote some code to run life on 3d surfaces. http://www.speakeasy.org/~morse/life
How were you hoping to find the answer to a metaphysical question?
No serious researcher gets bothers to defend metaphysics anymore
Let's play a game. You name one serious philosopher that you've had a long conversation on this subject with, then I'll name two serious cognitive researchers that I've had a long conversation on this subject with. First one to run out of names loses. You go first.
-- the most controversial site on the Web
Cool, so now Permutation City can be real?
For those who don't know, it is a novel by Greg Egan about an alternate reality created simply by encoding the rules of its existence in cellular automata. It's a great read if you're into this sort of stuff.
--
Bush's assertion: there ought to be limits to freedom
Let me rephrase that. A TM can simulate any computer program that can be written for todays computers. TM can simulate more than mere algorithms, since an algorithm is something that terminates on all inputs. TM's may not terminate, and may run forever, but that would be useless.
Any program you think of will eventually be converted into machine language, that has op codes, registers, and memory. It's not very hard to write a TM to simulate the op-codes of a CPU, and include cache (as a tape) as well as registers, and memory.(It may be tedious, but not hard). Then any program that is compiled to that CPU can be simulated by that TM.
If you replace your video card and monitor with a tape, you get the same thing as a TM. If you made a video card that could read a tape of a TM (hypothetically of course) then you could actually run Quake 3. So if your computer took in a TM input, then you can still run any program you can today.
Steven Rostedt
Steven Rostedt
-- Nevermind
This may have appeared recently on /. when I wasn't paying attention, but just in case...
Check out this article from Forbes on Life; Turing Machines aren't the only things that can come out of Life programs.
That makes sense. The cardinality of the set of integer points on a grid is the same as integer points on a line. Therefore you can map the grid positions to tape positions indefinitely. Good point.
Link has been /.ed, what the hell is a turing machine?
Dirty Pirate Hooker
Is this link working for anybody? Seems to be a page that redirects to itself after 0 seconds, and then finally I get "Connection closed." Or has it been /.ed?
You're spot on about Penrose's book. I had some trouble with his personal biases, but there is a ton of interesting stuff.
Didn't Gödel's incompleteness theorem precede the halting proof, though? It's not the set of axioms that is significant, in any case, it is the power of the formal system that is used to reason from them.
Floating face-down in a river of regret...and thoughts of you...
Yeah. Google.
Sig, meet "end user."
He might have, but I came in after the researcher stuff. As the one who would know, I can honestly say I don't think it matters.
The universe is not under any obligation to provide definitions.
Things behave the way they behave, and it's us who label them as intelligent or not intelligent. So of course it would be up to us, and not the universe, to define it.
That said, I don't think we need or want a completely formal definition. But we should have some sort of criterion for telling when something is, indeed, intelligent - just to avoid the whole thing that happened when they found the new world.
The idea that intelligence can be evaluated on behavior is a pretty good one; I've successfully used it to conclude that my friends are, and sea slugs aren't. Maybe Turing's test formalized that a little too much, and so broke it somewhat, but the base concept is important.
As for Hobbes, well, I haven't read any of his works, so I can't speak on him. But a lot of great philosophers do make stuff up as they go. ;)
This is a bit late for a reply on this thread, but I couldn't help but take the first sentence as a little bit ad hominem meum...
However, this was never proved by Turing. In fact, it has been shown, for certain notional machines, that this Thesis is false! Though, my personal belief is that these notional systems are not even mathematically sound, let alone do they obey the laws of nature.
Another personal rant of mine, is that the Turing machines are horrible ways to invision computation. They give no intuition for a software developer. Other foundations for computer science are much better.
jacking
||| I still can't believe Parkay's not butter.
I still can't access the page
www.droppingdimes.com
That's ok for one page. But the hyperlinks within the one page typically don't work, one would have to go to the trouble of manually putting them into another google request.
We need google to change the hyperlinks in it's cached pages to refer to the other google-cached pages.
Of course no matter how convenient, it'll never happen. Legally troublesome. (Well, at least for 36-100 more years :)
Where did they manage to find infinite tape? The last time I went to Walmart they were out. :-)
"Computer Science is no more about computers than astronomy is about telescopes." - E.W. Dijkstra
Argh!
http://www.rendell.co.uk/gol/tm.htm
-- Pete. (Must...use...preview!)
Monochrome - Probably the UK's largest internet BBS
A few minutes of looking hasn't turned up proof of the original intentions either way, but here is a quote from an article posted in 1991:
So it appears that there have long been proofs that Life could host a universal Turing machine, but there had been no explicit construction.
This still doesn't answer the question of the purpose of Life. At this point, all I can say is that from what I vaguely recall reading somewhere, I am right and you are wrong. But I don't think a google search on "the purpose of Life" will turn up the answer.
Prime numbers are exactly what Alan Greenspan says they are -S. Minsky
And don't forget Neal Stephenson's Cryptonomnicon, in which Turing pops up occasionally. The Diamond Age, or a young lady's illustrated primer, also by Stephenson, has info on Turing machines as well.
But you are right in that Conway was also looking for the properties you describe.
At any rate, it appears that consturctions of UTMs in Life have been around for a while. One was published in 1982 in a book by Berelekamp, Conway and Guys Winning Ways for your Mathematical Plays .
Prime numbers are exactly what Alan Greenspan says they are -S. Minsky
Quantum computing hasn't really changed anything: anything that can be determined by any sort of algorithm, can be determined by a Turing machine.
But more important is the idea of proof: it doesn't make any sense. In order to do that, you would have to formalize computability, and that is what a Turing machine is.
What lends credence to the idea is that there are many other formalizations that end up exactly the same. Just like ZF set theory (which is either incomplete or inconsistent, impossible to prove which, as a side note).
I almost couldn't tell that it wasn't a REAL broken link ...
Oh hmm....
---
Rob Flynn
---
Rob Flynn
Pidgin
...be sure to read The Code Book by Simon Sihgn (sp?). It documents his activities at Blectchly Park and provides an interesting account about the logic and thought process he used to eventually creating the Turing machine which would play a crucial role in breaking the German enigma.
Then I must be misunderstanding what you are saying. As soon as you say "none of it's responses are meaningful", I read it as "none of it's responses are intelligent". The difference seems to be hinging here, and perhaps I do not understand your argument as a result.
Because it is a look-up table and therefore there are no necessary connections between the words in its responses and their referents.Let me understand what you're saying. Please correct me if I'm wrong, I don't want to put words in your mouth. Are you saying that because encoding context in such a table mechanism is difficult it can't be intelligent?
Let's go back to the beginning, and start very simply. A person sits down at a machine, using what appears to be something like the UNIX talk program. That person begins to have a conversation, which proceeds as long as they want. Let us assume that after they walk away from that computer, they are convinced they were talking to a real human being. If that person were wrong, and they were conversing with a machine, is the machine intelligent?
Personally, I would say yes. If you disagree, please state, as simply as possible, why.
OK, I'm officially confused. What does the Turing Machine actually do?
-Cliff
It's called Turing's World, and it was written by John Barwise and John Etchemendy, philosophers at Stanford.
I never could have done the homework for my computability and logic class without it. Debugging turing machines on paper is a bitch!
--
Back in the early '80's the Washington University (wustl) Computer Science Department used a neat Turing Machine emulator in its introductory CS classes. One of those things that makes absolutely no sense when you are doing it, but three years later you wake up and say "_that's_ what that was for!!!".
Does anyone know if this emulator still exists? If so, has it been ported to Windows or Linux (or MS-DOS for that matter - it originally ran under the UCSD p-System!)?
Just a bit of off-topic curosity.
sPh
A look-up table can't be considered as intelligent because it is a look-up table, not for any other reason.
The answer to your third question is that the machine might or might not be intelligent, but the test you suggest proves nothing either way.
-- the most controversial site on the Web
What's the real link?!
He was stating the conclusion without giving an argument (unfortunately, a common practice). The usual argument is to describe a TM, and then describe a program which is halting but which the TM could not possibly prove to be halting, and the conclusion usually drawn from this is that the human mind must have more 'power' or something than a TM.
I actually remember seeing this a few months ago, linked from UserFriendly, but it's still a very cool link.
I wonder how long it would take to run a program of any decent length. Wasn't it like, 7,000 generations for one step, or something?
--
Matthew Walker
My DNA is Y2K compliant
Matthew Walker
http://www.tweeterdiet.com/ - My Diet Tracking Tool
Can you convince him? If so, how?
And also, I don't think you can compare convincing someone to be of a certain nationality to convincing someone that you are an intelligent being.
Turing made a good step by suggesting that we define intelligence as a behavioral trait (which is really metaphysics saying that metaphysics isn't needed). Noone has really had any better way of defining it...at least, you haven't suggested one yet.
"You made a Turing machine out of cereal?"
"You made a Turing machine out of a popular magazine?"
"You made a Turing machine out of the quality that distinguishes a vital and functional being from a dead body?"
--
--
Mod up a post Rob doesn't like and you'll never mod again
If I remember correctly from what the prof said, when the game was invented they set up the rules in such a way so that it was hard to predict the outcome of a given configuration. There could be other rules than an organism living only if it is next to one or two of its own kind. That rule was chosen simply because it makes life hard to predict.
From there somebody showed that you could make wires and gates using these rules and we know you can make a turing machine from (infinite number of) wires and gates. So an infinite life board can be used as a turing machine.
Am I the only one that thinks that this guy may be a little biased? His email address is an untypable lambda-expression, which has questionable logical meaning ;-)
Ok, actually, I agree with you, Tom. Turing machines are clumsy ways to view computation. The lambda calculus is nice, but it is awkward, whenever you are thinking about heterarchical computational problems (like graph algorithms). However, for heirarchical computational problems, the lambda-calculus works like a charm.
May I suggest an even better formalization of computation: The Fusion Calculus, which is related to Robin Milner's pi-calculus.
The pi-calculus is the result of Robin Milner's work, which he started after initially completing ML back in the late 70s and early 80s. Back then, Milner could see how the lambda-calculus was inadequate, as the foudation of a general purpose programming language.
The most interesting aspect of the lambda-calculus is the Curry-Howard isomorphism, which equated a program with a proof in intuitionistic logic (Brouwer Logic). To make a long story short, the lambda-calculus is sequential in computational nature, and in a way, the Brouwer Logic is a sequential form of proof. Here is the best part!!! Variations of the pi-calculus are showing a similar isomorphism with a newer kind of logic: Girard's Classical Linear Logic. Classical Linear Logic is a parrallel, constructive logic, which can encode Brouwer logic proofs. So we are begining to see that the pi-calculus and Linear Logic are generalizations of the lambda-calculus and Brouwer Logic.
Tom, and anyone else, if you want to learn more about the pi-calculus, then I suggest Robin Milner's new book
Anyway, I understand that one of Saint Turing of Computing's original papers written just before or during WWII is *still* classified. I believe it had to do with information theory and detecting 'signal from noise' if you will, or perhaps dealing with stego matters, or how to detect information others are trying to hide.
I wonder sometimes what its contents could possibly be, and if it was truly novel, or whether it is something known but its implications have been overlooked
==================
By the time you have reached perfection, there's nobody around you to share it with.
==================
By the time you have reached perfection, there's nobody around you to share it with.
So, I don't know if it makes the point that Life was actually designed to be yet another minimal system that could implement any Turing Machine.
Prime numbers are exactly what Alan Greenspan says they are -S. Minsky
It's amazing how people (in all aspects of life) assume that someone else is wrong, just because someone else's suggestion does not match the person's preconceived notion of what they should be hearing -- especially without even bothering to check of they are right or not.
rendell.uk.co contains the correct page (as evinced by Googol, even though it is currently slashdotted). "rendell.co.uk" has no nameserver lookup, and "www.rendell.co.uk" is a completely unrelated site and does not mention Turing machines at all.
The domain "rendell.uk.co" is registered to Paul Rendell, as of 10 July 2000, and the domain "rendell.co.uk" to Webhound Ltd., as of 16 Sep 1999.
To use a cliche, "People hear what they want to hear"
This is from my cache [can't post front page - just an image]:
The Turing Machine Program
The program I chose is one that duplicates a pattern of 1's. With 2 1's on the tape to the right of the reading position it takes 16 cycles to stop with 4 1's on the tape. This takes over one hour on my computer.
The state transition table for this program is as:
State
Input Symbol 0
Input Symbol 1
Input Symbol 2
0 Find next v=1
D:R, V:0, NS:2
D:R, V:2, NS:1
D:L, V:2, NS:0
1 Write 2*v=2
D:L, V:2, NS:0
-
D:R, V:2, NS:1
2 Convert v=2 to v=1
Halt
-
D:R, V:1, NS:2
Where:
D = Direction
V = Value of symbol to write
NS = next state
A P240 gun has been placed to insert the instruction "D:R, V:0, NS:0" when the blocking glider is deleted, which it is in the initial pattern above. This starts up the Turing Machine.
Back to Turing Machine Main Page
Free Anne Tomlinson!!
He developed a machine that helped boost automata and state logic to develop the computer you are using to complain about one of its creators. Alan Turing was brilliant and years ahead of his time. Show some respect.
--
Good quote, too many chars. Seriously, the slashdot 120 char limit sucks!
Church and Turing never said anything about computers in general. They made claims about variations of logical computing machines (LCMs) with computation defined as anything that can be described by "rules of thumb". Now, depending on your definitions of: "rules of thumb", data, program, machine, and computer... you are making wildly different claims about computation.
Church and Turing NEVER endorsed what you said, that "The Church-Turing principle only claims that there is no computer that can solve more problems than a Turing Machine. "
If you don't believe me, then read Church and Turings original papers on the subject, or check this discussion of the myths surrounding the Church-Turing thesis.
This incorrect interpretation that you have of the Church-Turing thesis discourages interest in alternative forms of computing.
So how would you convincingly demonstrate that something is intelligent in a way that would distinguish a human from a machine that could communicate with you in such a way that it is indistinguishable from communication with a human? Are there non communication-based ways of judging whether something is intelligent?
But if you convinced a Chinese interrogator that you were chinese, then perhaps you might as well be Chinese.
Sort of like the man in the Chinese box. Of course the man doesn't know Chinese. The box knows Chinese.
--
There are no trails. There are no trees out here.
That should be http://www.rendell.co.uk/gol/tm.htm...
A computer can do things that a Turing machine cannot. (e.g. the C=64 had a nondeterministic random number generator)
How do you know that *I* am a human and not a bot?
Since a turing machine can be implemented in Life, this means Life is equivalent to a turing machine. Since Life is simpler than a TM, doesn't this actually mean Life should be used as the base model of computation, rather than a TM?
Yes, he was a good mathemetician. But just because of that we shouldn't take everything he says about AI as Gospel.
Besides, the modern PC would have come into existence with or without his help.
Quoting from your link:I also never claimed that the C-T Thesis has been proven; in fact, I strongly suspect that it is unprovable. Like you said, it's a "rule of thumb," but I am not aware of any violations of it.
MSK
While it is true that the peer review process is, as you say, the validator and not the initiator of insight, the vast major of advances in insight come from interaction with peers. Contrary to the myth, there d*mn few examples of the "lone genius." Einstein, Liebniz, Darwin, Newton, Galileo, Watson & Crick, Gauss, Archimedes....all don't qualify. Mendel is about the only one I can think of who does.
The "stereotypical mad scientist" is so popular because it is an incredibly useful device for narrative -- it spares having to explain to the audience how the "McGuffin" (to use Hitchcock's term) came about. Also, the "mad scientist" or "lone inventor" is a comforting myth for non-technical types, so it is in the self-interest of technical innovators to nurture the myth, even if it is utterly untrue."one treats others with courtesy not because they are gentlemen or gentlewomen, but because you are" --G. Henrichs
Google's cached copy of the explination of how the machine works
This makes my brain hurt, but wow, I find it amazing that someone took the time to create this. Bravo!
-inq
Here's a dump...it's bad formatting tho.
The Alan Turing Internet Scrapbook
Computable Numbers, 1936
and the Turing Machine
maintained by
Andrew Hodges Alan Turing
home page Scrapbook index Short Biography Bibliography My Books
-----
Mathematical Logic
In 1935 a course by the Cambridge mathematician M. H. A. (Max) Newman introduced Alan Turing to the frontier of research in mathematical logic.
Logic is not well represented on the Web, and unfortunately the Gödel home page doesn't tell you anything about Kurt Gödel's 1931 work that completely rewrote the agenda in the foundations of mathematics. This is just mentioned at the end of a worthwhile MacTutor summary of the Beginnings of Set Theory.
You can also read the famous 1900 speech by the German mathematician David Hilbert which did much to set the agenda for twentieth century mathematical research. Hilbert's later question about the 'decidability' of mathematical assertions set the stage for Turing's work.
This Encyclopaedia Britannica article on Logic discusses the background to decidability in mathematical logic.
I strongly recommend Martin Davis's new book The Universal Computer, the road from Leibniz to Turing as a complement to my own work.
Turing Machines and Computability
Responding to Hilbert's question about 'decidability' in mathematics, until then unanswered, Turing had the idea now called a Turing machine as his formalization of a what had informally been described as a 'method,' or in Turing's favourite expression, a 'rule of thumb.'
The Turing machine concept involves specifying a very restricted set of logical operations which are, however, sufficient to encompass anything that in modern terms would be called an algorithm.
The American Mathematical Society has a page explaining the Turing machine concept.
Turing argued that his formalism was sufficiently general to encompass anything that a human being could do when carrying out a definite method.
The Universal Machine
He had the further idea of the Universal Turing Machine, capable of simulating the operation of any Turing machine.
A Universal machine is a Turing machine with the property of being able to read the description of any other Turing machine, and to carry out what that other Turing machine would have done. Turing gave an exact description of such a UTM in his paper (though with a few bugs).
Another one was given by Roger Penrose in his book The Emperor's New Mind, and you can see this on Roman Verostko's page.
After 1945 Turing was able to embody the idea of the universal machine in his plan for an electronic computer: this is described on another Scrapbook Page.
Turing Machines Today
In my book I described the concept of the Turing machine in terms of the ideas which existed in 1935. But in fact it's now almost impossible not to think of a Turing machine as a computer program, and the Universal Turing Machine as the computer on which different programs can be run.
We are now so familiar with the idea of the computer as a fixed piece of hardware, requiring only fresh software to make it do entirely different things, that it is hard to imagine the world without it.
But Turing imagined the Universal Turing Machine ten years before it could be implemented in electronics.
Now, by a twist of history, the computer itself can be used to simulate the working of a Turing machine, and one can actually see on the screen what in 1936 was only possible in Turing's imagination.
Go to another Scrapbook page for
a Turing machine simulated in Java.
You can make it run a sequence of steps while on-line.
You can also copy the Java code and adapt it yourself.
Java Computability Toolkit - a 1998 release
A new Web resource is available from SUNY Institute of Technology at Rome, NY. It is a freely downloadable Turing machine simulator in Java. The writers say: 'It is built with collaboration and user-friendliness in mind and will always be free!' Go to the JCT site.
Turing's World
For an older full-scale Turing machine simulator, with a mass of documentation, there is Turing's World software.
This page by Kari Coleman develops a serious Turing's World algorithm for a decision problem in first-order logic, and thus exhibits the use of the Turing machine within the context of mathematical logic to which Turing originally applied it. The coded algorithm is downloadable.
Other Turing Machine Descriptions and Simulations On-Line
For a good description of Turing machines (but with outdated links) see this page by David Matuszek
Amother interesting description, including a Java simulator, is given on this page by Andreas Ehrencrona
There are other websites with information on (downloadable) Turing machine simulators.
These are maintained by:
Suzanne Skinner (another Applet)
David Matz (for Microsoft Windows)
Cristian Cheran (for Microsoft Windows)
Stefan Milius (in German. Java to download, not to run on-line.)
David Woodruff (in C).
SUNY Binghamton (Java simulation with documentation)
Turing Machines in DNA?
Alan Turing's definition of a Turing machine was not intended as a blueprint for how one would actually build practical computing machinery. The very primitive actions of reading and writing and moving one step at a time are like atoms of computation, and the atomic level is too time-consuming for what is needed in practice. However it appears that there is one modern field in which this atomic level of simplicity may be just what is needed. This is explored by Ehud Shapiro in a June 1999 paper on using the Turing machine model for a DNA computer. See his page for further information, press reports and links.
Turing Machines in Real Life
Paul Rendell has a page on building Turing machines in Conway's Game of Life.
The Uncomputable
Turing's definition of computability entailed the fact that uncomputable numbers and functions can be exhibited explicitly. The most famous uncomputable function, which Turing defined himself in 1936, is one that distinguishes between halting and non-halting Turing machines. Turing used this to answer Hilbert's question in the negative: there can be no one definite method that can decide all mathematical questions.
A version of the halting problem is given on a page by Mike Yates, which explains Turing's development of Cantor's diagonal method, and gives a proof of the essential result.
Mike Yates has a special connection with this problem. He was the first research student of Robin Gandy, who was in turn Alan Turing's first. Mike Yates was also greatly stimulated by Max Newman's knowledge of mathematical logic, and found him a great encourager just as Alan Turing did. He became Robin Gandy's collaborator, and is now the editor of the remaining volume of Turing's Collected Works, in which an annotated edition of Turing's 1936 paper On Computable Numbers will appear.
Another uncomputable function arises from the Busy Beaver problem, which is fully described with many links to other work on computability on this page by Michael Somos.
Computability, Complexity...
Turing machines, in providing a sort of atomic structure for the concept of computation, have led to new mathematical investigations. One development of the last 25 years, which Turing did not himself foresee, is that of classifying different problems in terms of their complexity, defined in terms of Turing machines.
A Nottingham University undergraduate course on computability and complexity (16 lectures) is now available on-line thanks to Dr A. N. Walker.
.. and quantum computing
Turing machines, regarded as the foundation of 'classical' computing, also provided the model in the 1980s for the new theory of quantum computing.
Computability and the Philosophy of Mind
Alan Turing described his concept of the Turing machine in terms of 'states of mind', and his work has important implications for the philosophy of Mind.
This Rutgers University course by Charles F. Schmidt has an extensive discussion of computability and artificial intelligence. It also has an excerpt from Turing's original 1936/7 paper on this page.
As indicated on the Scrapbook page on Mind and Matter, it is not surprising that Turing immediately drew this connection in his original 1936-7 paper. Turing's interest in the nature of Mind preceded his knowledge of mathematical logic, and had a powerful emotional base.
It is also notable that the Turing machine picture of computability has a definite physical sense to it, being based on what people actually do. This also reflects Turing's prior interest in physics, as well as his do-it-yourself engineering sense.
After the Second World War, Turing took a strong line that computers would be able to perform anything that people do in thinking (see this Scrapbook Page.) In my book I took the view that in taking this line Turing was simply developing to its full extent the idea of the Turing machine imitating 'states of mind'; and this is not only the generally accepted view, but the view that Turing himself argued in his post-war writings.
Accordingly it seemed to me that by 1936 Turing had rejected his youthful ideas about free will and the role of quantum-mechanical physics. As I put it, Christopher Morcom had died a second death, as Turing set out to explore the world of computability.
However I now think the development of Turing's ideas was more subtle. Although he certainly became fascinated by the role of computation after 1936, I suggest that until about 1941 Turing left open the idea that the uncomputable might play a role in human thought. Then he changed his mind. My reasons for this shift of judgment are set out in a new short text:
My own new text on Alan Turing as a philosopher of Mind appeared in November 1997. It is Turing, no. 3 of a series The Great Philosophers published by Phoenix (London) and Routledge (New York). It includes a substantial amount of Turing's original writing, and in particular big chunks of On computable numbers. My commentary explains how the Turing machine concept is related to Turing's philosophy of Mind, relating Turing's thought to Roger Penrose's ideas about computability.
More details, an extract from the text,
translations, and reviews.
Amazon page with information and review.
Church's Thesis and Turing's Thesis
A new Scrapbook Page will be prepared to link to items now on the Web which address the significance of computability. For the moment, note the article by B. J. Copeland in the Stanford Encyclopaedia of Philosophy. This has worthwhile criticism of the many loose statements to be found in present-day literature of what Turing achieved and claimed. However in my view Copeland's analysis is itself skewed by his 'super-Turing-machines' agenda (see the following section), and this article could well give a highly misleading impression of what Turing had to say about the scope of computability.
Logical Consequences for Alan Turing
Turing spent most of the next two years at Princeton University, based in the powerful research group in mathematical logic headed by Alonzo Church.
The work he did in 1937-8, his most difficult and most abstruse, charted new territory in trying to bring uncomputable numbers into some kind of order.
This page by Barry Cooper, University of Leeds, describes some modern research on the lines that Turing started.
To do this, Turing extended his concept of the Turing machine with abstract constructions he called 'oracles.' These would perform uncomputable operations. Turing explicitly wrote:
We shall not go any further into the nature of this oracle apart from saying that it cannot be a machine.
This has not stopped the philosopher B. J. Copeland from advancing the claim that Turing would have supported a project to 'construct' such oracle-machines, which he calls 'super-Turing-machines.' He holds out the prospect of 'the biggest revolution in computing since 1948.' See this Scrapbook Page for my comment on this remarkable announcement, and this page for my discussion of Copeland's claim that Turing was leaving room for such a possibility in his 1950 paper.
Alan Turing had the chance to stay at Princeton in 1938, but he returned to Britain and at about the time of the Munich agreement began helping the British government with the problem of deciphering German communications.
Turing's work in logic had in fact stimulated an interest in ciphers, as well as in actual physical machinery.
No-one could have guessed where this would lead, not even Ludwig Wittgenstein with whom Turing argued about the philosophy of mathematics. See Wittgenstein's Lectures on the Foundations of Mathematics, Cambridge, 1939 for a transcript.
Turing and Wittgenstein did not discuss the philosophy of Mind, then or later. Many people have wondered what they would have said to each other. John Casti has written an imaginary conversation, The Cambridge Quintet, involving such a dialogue; see also a page of comment by Chris Mitchell.
More mathematics, real and imaginary
Turing's work at Princeton, as described in my book, also involved work on complex analysis and the Riemann zeta function. Its wide-ranging mixture of topics has inspired a passage in the novel Cryptonomicon by the science-fiction writer Neal Stephenson, which you can read in in this excerpt.
The extended dialogue written for Turing there is rather more thoughtful in content than anything usually found in fiction. It certainly outdoes the feeble 'I'm researching Riemann' statement attributed to Turing in Robert Harris's thriller novel Enigma. (soon to appear as a film Enigma with screenplay by Tom Stoppard who I hope will do rather better.)
The real Alan Turing in late August 1939, sailing at Bosham, Sussex. Behind him is Fred Clayton, another young Fellow of King's College, and between them the two refugee boys Bob and Karl from Austria whom he and Fred helped to get asylum in Britain.
While they were there the pact between Hitler and Stalin was signed and war became inevitable.
The Loss of Logic
The coming of war meant that Turing never again concentrated on mathematical logic, and he did not follow up the ideas he had in 1937-38 on 'ordinal logics.'
The war was to take Alan Turing to the heart of the world's affairs, and soon he was combining his logical ideas of computability with the leading edge of practical technology. He grasped this chance with great enthusiasm.
But the war also exiled him from the opportunity to develop his pure-mathematical ideas at the height of his powers.
Last updated 10 September 2000.
I am always grateful for feedback and suggestions for new links: andrew@synth.co.uk
--
That's it. I can safely say I did not understand what a Turing Machine is after reading it tho:)
I don't. You could be a dumb program designed by a skilled expert manipulator to predict everything I could possibly say in response to you. That would make you a big lookup table. A big lookup table is not intelligent.
So you *are* equating being human and being intelligent, as far as I can see.
I believe posters are recognized by their sig. So I made one.
Perhaps you already know this, but I can't sit idly by while you cast aspersions on a most interesting function, the fixed-point combinator:
It returns the fixed point of any function expressible in lambda calculus. If that doesn't have an important logical meaning, I don't know what does!
right Here, but if you have to ask how much....
try { do() || do_not(); } catch (JediException err) { yoda(err); }
'Course, when I were a lad, we never 'ad any of this game o' life nonsense, no, we'd be hand coding turing machines with orange peel and lumps of coal. And for backups we used to 'ave to brand the machine state on our own arms, and then our dad would hack 'em off at the shoulder before rubbing salt into the wound and laughing like a madman. And if we so much as complained, we'd be making punch cards out of our own saliva for a week.
And you tell the youth of today - they won't believe you.
Turing's "simulation game" (nowadays known as the "Turing test") avoids the issues of a descriptive definition by focussing on an operational definition.
If you consider his operational definition to be the wrong way to define intelligence, then perhaps you can share your profound insights into the nature of intelligence with the slashdot readers by providing your definition for us to criticise.
It's a exact copy of the URL in the article - mistakes and all!
If it has an intelligent and coherent response to any possible thing you might say, why is it not intelligent?
Arguments like "but it can't learn" don't hold water. If it wasn't able to learn from prior experience and/or context, then it has not been able to construct an intelligent and coherent response to you.
The Turing Test is an extremely simple and revolutionary way of describing what it means to be intelligent.
Can you convince him? If so, how?
By speaking Chinese (I'm assuming we're still using teletypes as in the Turing test), and introducing myself with a Chinese name. A few deliberate mistakes in english word ordering mighnt help give the impression too.
And also, I don't think you can compare convincing someone to be of a certain nationality to convincing someone that you are an intelligent being.
It would be a lot easier to fake my nationality. Intelligence can be faked too.
That depends. If you consider clubbing someone/something/me over the head and looking if he/she/it/I hit(s) back a form of communication, then no, there aren't...
People replying to my sig annoy me. That's why I change it all the time.
It was always difficult to keep those little stick people from falling out of the cars too.
Arm yourself with knowledge.
It wouldn't have to learn. It would just have to remember. This could be hard coded too. It wouldn't be solving anything that a human had already solved.
Are there non-communication based ways of judging anything?
Just because you have to communicate with something to deem it intelligent doesn't mean the turing test is a fair measure of machine intelligence.
Forgeting the whole philosphical thing about weak and strong AIs I think the turing test is pretty good. Having said that I think a machine should have to convince *everyone* that it is human before being deemed intelligent to some level.
BBC's Tomorrows World programme did some big Turing test and 50 odd percent of people thought the computer was a person. Yeah, 50% of non-techy people who will ask the predictable questions. A lot of people could be fooled by Eliza if they don't have the knowledge to guess were the computer might trip up (jokes, emotions, questions about my previous comments, dodgy context etc). Doesn't mean Eliza on my Amstrad CPC 6128 was intelligent.
One of the first programs I compiled to run on my first Linux PC back in '94 was 'xlife'. This was a superb implementation of Conway's life, and came with some quite complex patterns.
The most memorable was the prime number seive. This consisted of two trains heading away from a central point in perpendicular directions, leaving a trail of mirrors. A third train produced a glider which would bounce back and forth between the mirrors.
A slow-period glider gun at the origin fires a stream of gliders diagonally between the two rows of mirrors. For any non prime number, the new glider will hit one of the bouncing gliders and be destroyed, leaving the bouncing glider intact. The result is that only prime numbered gliders from the central gun can escape.
There was also a cute pseudo-random sequence generator.
Operational definitions are fine for "placeholder" terms of science, but intelligence isn't one of them.
By the way, no serious AI researcher bothers to defend the Turing test these days. You're fighting a battle which was lost twenty years ago.
-- the most controversial site on the Web
the stupid thing isn't a turing machine, its a turing tape. all these posts were made without any clue as to what the guy actually did. the point of doing it using the game of life is that it is a simulation not a freaking turing machine. something can go on infinitely in such a simulation because guns can be set up to shoot infinite gliders. it's not a turing machine!
Oh yeah, and cahnces are we would have had the theory of relativity without Einstein too, maybe it would have just taken a bit longer... The point is, he was there to help. Another example : chances are we'd have had the protestant church without Martin Luther (no, not King, the original), someone else would have noticed how corrupt the roman catholic church was, and broken away like Martin Luther did. BUT, your mistake is assuming that all that is inevitable. It isn't. Turing was there, and he did his thing. YOU CANNOT SAY what would be happening now if he hadn't. All you can say is that Turing's work ( and Church's - but that's another story) DID actually lay the foundation for modern computing. I don't see you inventing phased matter transportation or whatever right now, so just keep quiet about belittling other's achievements. And anyway, without Turing, chances are we'd all be speaking German now here in England. His theories were instrumental in breaking the German codes in WWII.
So he designed something that he didn't think would have much Real Life application. Welcome to math. Boole thought he was designing the most 'pure' math possible, something which would have no conceivable use whatsoever... and now Boolean Logic is the basis for most of the world's digital computing devices.
As for the Chinese bit... well, according to the latest census forms, your 'race' depends not on your ancestry, but on which 'culture' you claim. You could very well be Chinese, as far as the government's concerned, if you think you are....
---
Good judgment comes from experience.
Experience comes from bad judgment.
Ludicrous. Looking up call-response pairs in a huge table is not intelligence. "How can you distinguish" is a ludicrous defence; if your interlocutor is a computer, you *know* whether it is doing this, or whether it is generating responses.
The point is that ordered pairs in a look-up table have no necessary connection to one another, so they cannot be taken as referring to one another. Therefore, the computer's replies in this case have no reference, therefore no content, therefore no meaning. Responses which do not mean anything are not evidence of intelligence.
If it smells like a duck, tastes like a duck, looks like a duck, and cannot be distinguished from a duck in any way, is it a duck? I would say yes.
Then you should probably avoid cheap Chinese restaurants.
-- the most controversial site on the Web
Turing was attempting to advance the argument about intelligence with an operational definition. Whether you agree with the definition of or not, it appears that you would rather take on faith the idea that "it can all be faked".
If you really believed that, then it appears to me that you believe in ineluctable truths. You may as well believe that there's a pink dragon in your room and nobody can convince you otherwise.
This means that since you can do a turing machine, you can do my favorite language Brainfuck with Life. I am still waiting for M$ to bring out Visual Brainfuck (VB for you folks), to make the interface stuff easier... This might be a breakthrough.
And how do you distinguish between this "remembering" and "true intelligence"?
If you can't distinguish between the two, either the definition or criteria your are using are insufficient, or the things are functionally equivalent.
Therefore, if you can't distinguish between someone who is intelligent and a machine that is claimed to be intelligent, what's the difference? As far as can be ascertained, they are both intelligent.
If it smells like a duck, tastes like a duck, looks like a duck, and cannot be distinguished from a duck in any way, is it a duck? I would say yes.
Not that I'm discounting this guy's work - it is amazing. But it's not impossible or miraculous. An infinite Life board *is* Turing-complete after all, it's already been proven that it can solve any computationally-solvable problem. And, therefore, it's simply an isomorphism (if you could really apply such a term to a process) of this Turing machine - or, more accurately - a Life board is a superset of all non-working and working Turing machine configurations. Anyway, just my two cents...
...but it's being eaten...by some...Linux or something...
I wouldn't exactly say "fool", though, although I agree with you on the asinine nature of his writings on machine intelligence. He did crack Enigma, afterall.
-- the most controversial site on the Web
Friends, I urge you to check out the lambda calculus or combinatory logic . Both of these are simpler (IMO), complete, and are a good complement to turing machines for certain kinds of problems. Here's combinatory logic, for instance:
K a b => a
S a b c => a c (b c)
That's it. All computable functions can be built out of Ss and Ks defined this way... holy mindfuck, Batman!
So "being intelligent" equals "being able to sound like a human"? I think not. There may be lots of higher intelligences out there, and none of them will be able to fake being a human, because they aren't, and they can't answer all kinds of questions about our emotions, the way a hangover feels, etc. It's a test of faking humanity, not of intelligence.
I believe posters are recognized by their sig. So I made one.
Why and why not?
They may well be functionally equivelent. However, that hardly makes them the same. If you wish to label that painted goose as a duck, feel free. But I submit that you just aren't looking close enough.
- The Racksasha
OK, this is really sick, I know...
:( ...
So you have a Turing machine running in a life grid. It'll need a bit more memory, but the hard part is done.
Next you port Bochs to the Turing machine.
Then you run DeCSS under Bochs.
Finally, you get a contract to tile some large area, and tile it with black and white tiles corresponding to some snapshot of the Life matrix.
I don't want to know how big the matrix would have to be though
Another thought - seeing that Conway's rules seem to be the simplest possible set which allows the formation of complex dynamic structures, howabout etching life patterns into deep space probes?
It's interesting to note that having claimed that humans are distinct from computers, you acknowledge now that there is a circumstance in which you can't tell them apart.
So the question to you now is this: Is shoving bits around one of those circumstances? Why or why not?
Please answer the question.
As you can see, you haven't been able to convince me that you know the first fucking thing about artificial intelligence, therefore by your own test you don't.
-- the most controversial site on the Web
No, it is not necessary, as I have a severed spinal cord. (not really). But this isn't going to get you anywhere. Jerked knees don't refer to anything and knee jerking isn't meaningful.
-- the most controversial site on the Web
The link appears to be pointing to a clean default installation of Apache. Paul has outdone himself doing a Turing port soo quick.
To Terminate, or not to Terminate, that's the question - SCSIROB
So how would you convincingly demonstrate that something is intelligent in a way that would distinguish a human from a machine that could communicate with you in such a way that it is indistinguishable from communication with a human? Are there non communication-based ways of judging whether something is intelligent?
If I recall correctly from my automata theory class, it is impossible for a machine to determine whether a given algorithm/function/procedure/loop terminates or not. So it would seem that you could figure out whether it was a machine (or at the very least a stupid human) by asking a statistically significant amount of "does this loop terminate or not?" questions.
Most people, if you challenge them to define intelligence (and ask them if a computer is intelligent) will immediately start to try to find a definition that is based on the difference between humans and machines. You seem to be doing this.
Consider this: if you're determined to prove that machines are fundamentally different from humans in some way that makes them ineligable for intelligence, then you have to also prove that humans are not themselves simply elaborate finite state machines. I don't think you can do this.
...has implemented a Turing Machine in Life here
That link points to whatever.uk.co. Should it be co.uk like so?
Connah
Connah
"Your mouse has moved. Windows NT must be restarted for this change to take effect."
It is fun to implement stuff in Cellular Automata, but since it is proven that CA are computationally equivalent to Turing Machine, what is the value of this work? The original post sounds as if it is sort of technological/scientific breakthrough. Dr. Dymov from Lomonosov Moscow State University used very little CA (fewer states and smaller local neighbourhood for a cell -- like 3 neighbours only instead of 9 in Life as long as I recall) to model arbitrary boolean circuits. That looked fun. Would be more challenging if the folks tried to use a CA simpler than Life for their TM modelling.
Maybe you are a form of artificial ignorance.
p.s. Just kidding. I couldn't resist.
I'm a REAL nerd now, I actually understood that
----------
Technoli
I wish I could see the web page, but:
Microsoft OLE DB Provider for ODBC Drivers error '80040e37'
[Microsoft][ODBC Microsoft Access Driver] The Microsoft Jet database engine cannot find the input table or query 'TA_Errors'. Make sure it exists and that its name is spelled correctly.
/Error_URLNotFound.asp, line 29
Oh, well, I guess I'll check back later......
Is there a necessary condition now?
When you were a kid, you didn't know shit about the world. The teacher passed out pointed out green leaves and you, in your infantile mind, saw no necessity for it to be so. Now having learnt about chlorophyll, you do.
So were you unintelligent before? After?
A computer too can discover and learn about such things, in rather limited and specific instances. It matters not whether you want to call it intelligence or not. This is already reality.
You are just lucky to be living in a time when this distinction can still be maintained, largely.
Therein lies the heart of our disagreement. This is precisely the Turing test. I think it is a very simple yet powerful rationale for describing when something is intelligent.
If you have an alternative, I would genuinely love to hear it.
A look-up table can't be considered as intelligent because it is a look-up table, not for any other reason.So if it were implemented using another technique, you might consider it intelligent? Dismissing a program that appears to be intelligent because of it's implementation does not seem right to me.
what are you talking about? That link is to the homepage of Professor Gregory M Chaitin, who I hope most of the readers here will have heard of. You clearly haven't. Sucks to be you. I was doing the "you're just trolling" troll before you were born, laddie.
-- the most controversial site on the Web