A True or False Prophet?
by BradfordWill Wright's 5 Prophecies about Artificial Intelligence has a few points to make including an interesting take on intelligence:
MACHINES WILL NEVER ACHIEVE HUMAN INTELLIGENCE
According to Wright, one of the main benefits of the quest for AI is a better definition of human intelligence. "Intelligence is whatever we can do that computers can't," says Wright. One key difference between human and machine intelligence, is that while computers are excellent at problem solving within narrow parameters, "humans can automatically scale the solution to the problem." The problem is one of creativity, which machines are unlikely to ever develop. "Computers aren't great at fundamentally recreating intelligence," says Wright. "But they're great at harvesting intelligence." By harvesting both human intelligence and human creativity, machines can begin the long journey towards their own kind of sentience.
It would be most ironic if undirected processes were able to surpass what advanced intelligence could not intentionally design would it not?
October 28th, 2008 at 11:09 am
This is why I like to tie in Turing machines. I think that algorithms which require class 4 features of a class 4 machine to implement are the limit which define what "creativity" can do which "non-creativity" cannot. For those of you interested, see my short intro video on the subject. I'm currently in the process of writing a paper for the BSG on it, but that may take some time. However, I have an early draft, if anyone is interested. My email is jonathan at bartlettpublishing dot com.
Comment by johnnyb — October 28, 2008 @ 11:09 am
October 28th, 2008 at 6:38 pm
Bradford:
Not if you allow for the possibility that advanced intelligence can be the result of undirected processes. Paraphrasing Rock: evolution can be seen as a form of machine learning.
Author of article:
Paraphrasing John von Neumann: if you tell me exactly what creativity is, I can build a machine that can do just that.
The brain is a machine made of meat.
Comment by Raevmo — October 28, 2008 @ 6:38 pm
October 28th, 2008 at 6:57 pm
"It would be most ironic if undirected processes were able to surpass what advanced intelligence could not intentionally design would it not?"
I think it would be more ironic that intelligent superbeings, or omniscient god used a sloppy slovenly process waiting billions of years for live to 'emerge' from a brutal and bloody process, who were so benevolent in their designs they thought giving children diseases and mental deformities a wondrous thing.
Two can play the irony game. It's more ironic that the universe is half-assedly put together, then any so called smearing of artificial intelligence, or some ancient flawed idea that the mind is not physically bound to the substrate of space and time.
Most people in the ID community are profoundly ignorant, they need to go read some books on what cognitive science has discovered in the last 30 years:
Philosophy in the Flesh : The Embodied Mind and Its Challenge to Western Thought by George Lakoff and Mark Johnson (Paperback – Dec 1, 1999)
From Molecule to Metaphor: A Neural Theory of Language (Bradford Books) by Jerome A. Feldman (Paperback – Mar 31, 2008)
Descartes' Error: Emotion, Reason, and the Human Brain by Antonio Damasio (Paperback – Sep 27, 2005)
I can't take anyone seriously who doesn't keep up with cognitive science and harkens back to primitive ideas of mind body dualism, or that the brain and the mind are separate. They are the most fallacious and pernicious ideas to ever grace the face of existence, and it comes right out of a misinterpretation of the bible and christian theology and other failed notions such as descartian mind body dualism.
Comment by XDragon — October 28, 2008 @ 6:57 pm
October 28th, 2008 at 10:32 pm
It would be most ironic if undirected processes were able to surpass what advanced intelligence could not intentionally design would it not?
That's the predictable response but it does not address the point. The manner in which intelligence is arrived at would not impute advantages to a non-telic process compared to one in which an advanced intelligence is able to analyze and problem solve purposeful solutions. Intelligence does not forego random advantages. Rather it is able to direct a process so as to avoid dead end sidetracks. Intelligence would not confer disadvantages compared to an undirected process. So the irony remains.
Comment by Bradford — October 28, 2008 @ 10:32 pm
October 29th, 2008 at 6:47 am
irony, incongruity between the actual result of a sequence of events and the normal or expected result.
Then it wouldn't be incongruous. That's a reasonable point. Why doesn't Bradford agree.
That doesn't address Raevmo's point. Accepting that intelligence has evolved from non-intelligent processes, then it is not incongruous to expect it to do so again. Then it's just a matter of degree and kind.
I think you're saying that intelligent beings can emulate evolutionary processes. And they do. But that's the whole point. Humans can create an artificial evolutionary process that works much faster than a biological evolutionary process, and so it might quickly surpass the original, even without additional input from the designer.
It would be ironic like Frankenstein is ironic—because the maker doesn't recognize the implications of his own creation.
But the author's point, an expert in machine intelligence, is that machines will boot-strap their way to sentience through various forms of feedback and reprogramming. The evolutionary process, once started, could easily exceed the original creator's capabilities.
Irony is in the eye of the beholder.
And these 'minds' will be unexpectedly alien. That's the irony.
Comment by Zachriel — October 29, 2008 @ 6:47 am
October 29th, 2008 at 7:10 am
Zach:
Do you consider this a testable scientific prediction? Would its opposite be one as well?
Peace
Comment by fifth monarchy man — October 29, 2008 @ 7:10 am
October 29th, 2008 at 7:30 am
In that case the following is valid:
Comment by ID guy — October 29, 2008 @ 7:30 am
October 29th, 2008 at 7:39 am
Futurology, the study of the future of human history, straddles the line between social science and the humanities.
Machines already use feedback and self-programming mechanisms. We would expect this trend to continue. But like with most such extrapolations, we can't accurately project the distant trajectory of such trends.
Even the common notion of 'sentience' may be an inappropriate metaphor for a highly developed machine intelligence.
Comment by Zachriel — October 29, 2008 @ 7:39 am
October 29th, 2008 at 7:50 am
I thought the Solar System was formed by the gravitational collapse of a nebula.
Comment by Zachriel — October 29, 2008 @ 7:50 am
October 29th, 2008 at 8:11 am
Don't think. Also the nebula hypothesis is just that and perhaps not even worthy of a hypothesis status.
Comment by ID guy — October 29, 2008 @ 8:11 am
October 29th, 2008 at 11:32 am
Sounds like constraints to me. In any case unless machines are able to self-replicate how is this truly an evolutionary process? Human input is not separated from reprogramming.
Comment by Bradford — October 29, 2008 @ 11:32 am
October 29th, 2008 at 6:17 pm
ID guy:
How does CS Lewis' thoughtless claim follow from my response?
CS Lewis:
In a sense it's true that the evolution of humans was an accident. If a meteorite hadn't wiped out the dinosaurs we probably wouldn't be here. Similarly, given that we did evolve, your existence is an accident (no comment on whether a lucky one) contingent on your parents meeting at the right time, having sex at the right time, etc. Does this series of accidents render your beliefs automatically suspect? I think not. Like I said before, it makes a lot of adaptive sense for brains to build reliable approximate models of their environment. Thus, we have good reasons to trust our senses and our brains' "simulations" to estimate plausibilities of future events.
Comment by Raevmo — October 29, 2008 @ 6:17 pm
October 29th, 2008 at 7:41 pm
That's contrary to the opinions of the vast majority of scientists in the field. Most not only consider the model a valid hypothesis, but consider the evidence convincing (including that the planetary orbits are on a plane, and that we can directly observe nebula in various stages of collapse). Is there a reason why you think they may be wrong on this?
Comment by Zachriel — October 29, 2008 @ 7:41 pm
October 29th, 2008 at 8:17 pm
Yes, that's one reason why the author suggests that machine intelligence will be inherently alien, while still plausibly being far more powerful.
It wouldn't be biological evolution, but it can still evolve in that each generation varies from its ancestor.
It can be.
Comment by Zachriel — October 29, 2008 @ 8:17 pm
October 29th, 2008 at 8:35 pm
The question is can it be separated and still generate human-like intelligence? Incidentally, if the answer is yes then we still have an intelligently designed end-product. The intelligent design is at point of origin. Sound familiar?
Comment by Bradford — October 29, 2008 @ 8:35 pm
October 29th, 2008 at 8:48 pm
The author's answer is "no". To do so would require a full understanding of the human mind, something not yet available, and an explicit effort to build a simulator. However, an even more powerful alien intelligence might evolve along its own path.
Intelligence would be at the point of origin regardless of the answer. But evolution could result in something completely beyond the kin of its creator.
Comment by Zachriel — October 29, 2008 @ 8:48 pm
October 30th, 2008 at 7:16 am
Zach:
My question is could we recognize such a devise as intelligence even if it was very alien?
How would we make this determination? Would we look at artifacts?
The implications for ID are profound.
Bradford:
I agree instead of being an argument for MET such a result would be a powerful demonstration of frontloading
Raevmo
According to the author it’s “automatically scaling the solution to the problem”
Build a machine that can do that why don’t ya
Peace
Comment by fifth monarchy man — October 30, 2008 @ 7:16 am
October 30th, 2008 at 7:38 am
As intelligent? More than likely, by most reasonable definitions of intelligence, such as the ability to solve problems, to plan, etc.
By testing for intelligence. I think you mean would we recognize it was intelligently designed. We may or may not immediately recognize its origin, depending on how closely it resembles similar human manufactured devices or evolved organic beings. And yes, we would look for evidence of its historical evolution, including its origin. Of course we would.
As science fiction, ID may be worth pondering. As science, it is contrary to what we know, and completely lacking in supporting evidence.
Comment by Zachriel — October 30, 2008 @ 7:38 am
October 30th, 2008 at 8:12 am
Building machines that evolve no more shows that life was designed to evolve than building a machine that flies shows that birds were designed to fly.
Comment by Zachriel — October 30, 2008 @ 8:12 am
October 30th, 2008 at 8:41 am
XDragon:
I first encountered this argument courtesy of Ken Ham and Answers in Genesis. Strange bedfellows.
I'm sure the world is holding its breath about what you take seriously.
Written by a true expert on the Bible.
Comment by Bradford — October 30, 2008 @ 8:41 am
October 30th, 2008 at 4:58 pm
Zach:
So we don't have to know about a designer except that it solves problems and plans inorder it call it an Intellegent designer very interesting.
No the fact that it was intelligently designed would be obvious at least to me. I meant would we recognize it as an intelligent designer? And how would we make this determination? If we could do it for one “alien” designer could we do it for others.
How would we test for intelligence in a "very alien" entity? Especially one we cannot understand?
It seems to me that we would have to look at the product of this machine it's artifacts whether digital or otherwise.
That is what ID is all about determining the intelligence of a designer by looking at it's artifacts.
Peace
Comment by fifth monarchy man — October 30, 2008 @ 4:58 pm
October 30th, 2008 at 7:00 pm
Obvious to you is not scientific evidence, of course. The Earth is obviously fixed. We'll obviously never know what stars are made of. Matter is obviously solid.
You seemed to have inadvertently slipped from intelligent to intelligently designed to intelligent designer. Please try to be a bit clearer in your statements.
Intelligence is a notoriously hard concept to pin down precisely, but it usually entails learning and planning. And with an alien entity, it may be even more difficult. Notice that definitions of intelligence typically refer to behavior or capability.
Well, the methodology is obviously faulty as it keeps generating false positives.
Comment by Zachriel — October 30, 2008 @ 7:00 pm
October 30th, 2008 at 7:41 pm
Zach
I never said it was. Science is not the only road to knowledge I would question the sanity or truthfulness of anyone who thought that a computer was not a designed thing and would question the motivation of anyone who demanded scientific evidence before he would agree to that fact.
Do we have to pin it down precisely before we can decide if a machine is intellegent?
I know you feel this subject is hard but everything worthwhile in life is
How would we know if an alien entity/computer was learning or planning?
Is not an artifact nothing more than a tangbile record of behavior?
No slippage on my part I've been interested in computers as intelligent designers from the start of this thread. You seemed to have missed it because you are only interested in mocking others. Perhaps you should be a little more openminded in your appraisals of others staments
Computers that boot-straped their way to sentience through various forms of feedback and reprogramming. would be "designing" the next model would they not?
In light of that fact the question I want to now is "Would they be intelligent designers?"
I was unaware there was a standard methodology. In fact a methodology is what I'm after.
Please compare the standard methodology to the method you would use in the case of a bootstrapping designer so I can see its flaws. How would we know if your methd had produced a false positive?
Peace
Comment by fifth monarchy man — October 30, 2008 @ 7:41 pm
October 30th, 2008 at 8:55 pm
That's true. However, your unenlightened opinion on empirical matters is not an argument.
I am more than happy to accept that an electronic computer is designed.
You are confusing the reasonable acceptance of well-known facts with whether something is empirically demonstrable. When someone who should be familiar with computers refuses to admit basic facts, i.e. they're handwaving, it may mean they have ulterior motives. Or it may just mean they haven't the foggiest idea what constitutes a valid argument.
But if someone were unfamiliar with computers, a Jovian perhaps, I would hope you would be willing to teach them about the subject. To be willing to demonstrate how they are made, and why. That you could support your claim.
You have to have some idea of what you mean by the term. It used to be considered 'obvious' that only human minds could play chess, that playing chess represented the epitome of intellectual achievement. It turns out that walking across the street to buy a loaf of bread is much, much harder for computers than playing chess.
By observing its behavior, e.g. rats and mazes. But that typically requires some knowledge of the entity being tested. In other words, observing and interacting with the subject.
If it's an artifact, you have already drawn your conclusion.
Computers meet several measures of intelligence, and they can create new designs. This is non-controversial.
The standard ID methodology is called the Explanatory Filter (Dembski), but the method is specious.
Comment by Zachriel — October 30, 2008 @ 8:55 pm
November 1st, 2008 at 1:53 am
By definition, a creative machine is an oxymoron.
Creativity is the ability to bring something into existence, something novel. This is opposed to simply imitating something.
Merriam-Webster definition of creative:
"2. having the quality of something created rather than imitated"
Merriam-Wesbter definition of machine:
"1. an assemblage of parts that transmit forces, motion, and energy one to another in a predetermined manner"
Anything that a machine does is simply an imitation or implementation of an idea that its inventor had. How can a machine create something in a predetermined manner? If you believe this is possible, then you disagree w/ the scholarly-accepted definitions of the words creative and/or machine.
Then either you disagree with the definitions of creative and/or machine, or you truly believe that humans don't possess the ability to be truly creative.
If it is the first option, then the only way your opinion would be of any value is if you are a well-studied, respected lexicographer, since people who devote their lives to the definitions of words write those. If it is the second, then you simply deny everything we experience.
Or perhaps you could create another option?
Comment by uoflcard — November 1, 2008 @ 1:53 am
November 1st, 2008 at 8:53 am
There is more than one definition of machine. But to prevent a semantic diversion, I would suggest you read "artificial intelligence" in place of "machine intelligence".
Computers are capable of generating novel solutions to problems, original chess strategies, even novel works of art.
Comment by Zachriel — November 1, 2008 @ 8:53 am