posted
OK, I don't seem to have much luck with theology threads, so let me try a more neutral subject. I wonder if there is any practical upper limit to the intelligence that a real being, either organic, silicon-based, or both, can have?
This does require a bit of defining intelligence, of course. Let's just say that I am not talking about the ability to play chess. How about "The ability to process information and act upon it in such a way as to advance goals".
As an absolute upper limit, we can look to information theory. The speed of light, plus entropy, limits the amount of information that can be processed by a given mass in a given time. That really only applies to brute-force crucnhing of every possible outcome, though. Organic brains do not seem to work that way; instead they produce a response holistically. Not always the correct response, of course - that's where intelligence comes in.
A better limit might be the information available. No amount of information, for example, can give you a good decision on a moral dilemma. While that's not really a problem of intelligence as such, tehre are surely similar situations. Sometimes there are no good choices.
What do you think? Is there a maximum possible intelligence? Please do keep omnimax gods out of this, though. I want to discuss situations that are empirically testable, at least in principle.
Posts: 10645 | Registered: Jul 2004
| IP: Logged |
posted
I think if there is a max, it will be hard to reach. Maybe on a purely organic base, assuming that the only influence on the organic being is just evolutionary, then I think that there would be a max. It seems to me that lots of natural intelligence is based off of the experiences or stimulus the being faces. So when the being is exposed to more stimulus, and different kinds of stimulus, it seems like intelligence would naturally increase.
But if we talk about any kind of organic or silicon based being that can sort of...engineer its intelligence, then I think the max would be much higher. It would basically depend on the engineers ingenuity...the amount of data we can pack into hard drives, for example, is increasing drastically on the medium, and once we come up with new kinds of mediums, then we'll work at making that work best until we need to find a new way to work. I hope any of that made sense, my addled brain is running in circles right now...
Posts: 193 | Registered: Jul 2005
| IP: Logged |
posted
It's really hard to define intelligence well enough to make such an estimation, but if it's possible, I think info theory is the way to go.
It can put a hard upper limit on how much info is available in a string of bits. We can then use the speed of light, and the size of atoms, to say how close those bits can be together, and therefore how fast one can affect another.
This could then be used to get a theoretical maximum to the speed of solving certain well-understood mathematical problems (4 equations in 4 unknowns, say), assuming the problems are random and there is thus no sudden leap of intuition possible.
Moving it over to intelligence -- tough problem! Mostly because intelligence is so hard to define. No, I'll go further. Impossible to define, at least quantitatively and reliably.
Posts: 1877 | Registered: Apr 2005
| IP: Logged |
posted
That's our main problem with trying to come up with decent AI; we don't really know any empirical way to test intelligence. We don't really know how to quantify it.
We can keep advancing AI only by adding processing - the 'brute force' you mentioned. And eventually, we'll limit out on that. No one has been able to put forth any good analysis of how human beings are able to make intuitive leaps. These leaps, I believe, are what makes our intelligence an order of magnitude greater than other mammals that have similar, if less developed, faculties for language and tools.
Posts: 5462 | Registered: Apr 2005
| IP: Logged |
posted
Just as a sidebar, did anybody check out Animals in Translation? Mr. Card reviewd it a while back, and I checked it out, and I was just stunned by what the animals they talked about could do. They were muuuch more intelligent than we really give them credit for.
Posts: 193 | Registered: Jul 2005
| IP: Logged |
posted
KoM, that is a really excellent question. I'm not even sure if we understand what intelligence is. The most cynical definition I've come across is that intelligence is that property which is measured by intelligence tests.
Clearly, it is different than knowledge. Also, it seems as though it is more than one thing. Someone (or something) very clever at mathematical reasoning may be at a complete loss in a social setting, while another may be very clever at making people feel at ease while still being unable to learn a foreign language.
Perhaps intelligence may be defined at the potential to know. The greater the potential, the greater the intelligence. What is the limit to that potential?
It takes an intelligence greater than mine to quantify that.
Posts: 10397 | Registered: Jun 2005
| IP: Logged |
posted
It seems to me, though, that the "potential to know" definition runs into the problem of defining knowledge. A computer hard drive, presumably, could be made almost arbitrarily large, and already the one sitting on my desk could easily absorb a hundred Libraries of Alexandria. But it can hardly be said to "know" anything, it just has a lot of information.
Posts: 10645 | Registered: Jul 2004
| IP: Logged |
posted
Intelligence, to me, is the ability to take in new information, information outside your realm of experience, and react to it...correctly...intelligence is one of those words that's hard to describe without using the word itself.
Posts: 193 | Registered: Jul 2005
| IP: Logged |
posted
I read something once about intelligence being defined differently in different societies based on what aspects of it were most useful or advantageous. Like in primitive societies being good at hunting would be the most important factor of intelligence.
My short definition of practical intelligence is "pattern recognition." It covers the basic things like "The stove is hot, don't touch it again" to unravelling the more complex interactions between individuals. That being said, I think it's likely there would be an upper limit to intelligence, though I don't know how it could be determined what the limit is.
posted
My patched-together definition of intelligence, off the top of my head:
Intelligence: The ability to solve a problem one has never seen by drawing on a number of seemingly unrelated experiences.
I think our intuition has a lot more to do with what constitutes intelligence than mere memory or processing power. We can't program a computer to fit the description above because we don't know how we do it.
Posts: 5462 | Registered: Apr 2005
| IP: Logged |
quote:A better limit might be the information available.
Processing info requires energy and produces heat so that limitation must be a subset of the total info available.
Posts: 675 | Registered: Aug 2001
| IP: Logged |
posted
Yeah, the beings in the Transcend in Vernor Vinge’s A Fire Upon the Deep were seen as gods but understood to be basically very advanced intelligences.
I would say that our ability to conceive of intelligence (and have a conversation about it) is limited to our perceptions of what constitutes intelligence and what intelligence is useful for. We can only discuss artificial intelligence in terms of technology we’re presently aware of, and we are limited to our own intelligence when trying to conceive of something grander.
Posts: 5957 | Registered: Oct 2001
| IP: Logged |
If instead we were a mundane rearrangement of the fundamental elements of the universe, then it could be fairly argued that we ARE the universe. The part that is conscious, sentient, intelligent. If the universe is intelligent, then its ultimate limit would be the subset I referred to earlier.
Posts: 675 | Registered: Aug 2001
| IP: Logged |
posted
I just googled up an extreme concept: Matrioshka Brains.
quote:A Matrioshka Brain is a megascale structure constructed at atomic scale limits. It is essentially a Dyson Shell supercomputer, that uses all of the energy a star produces and all of the material in a solar system for "computronium". Because of their size, immense observational and computational abilities, Matrioshka Brains should have longevities at least as long as those of stars (~1014 years for smaller stars).
A Dyson Sphere is a shell enclosing a star in order to capture and use as much an energy output as possible. The writer at the link thinks it likely humans will be able make one of these mammoth brains in a century or so, which I find laughable. Even assuming the hardware could be built, who whould program it? We can't even debug million line programs.
Posts: 6316 | Registered: Jun 2003
| IP: Logged |
posted
To me intelligence is your ability to comprehend and understand. I don't believe there is a limit to intelligence, there is a limit to finding correct information that fills and amplifies your intelligence. If you admit for evolution, this is shown in Humans as well as in how it differentiates us from lesser intelligent beings such as animals, etc.
I think where most of life on this planet is a product of evolution. Humans are the only ones evolved enough (or intelligent enough) that can immunize ourself from it's effects by our intelligent understanding of nature.
Posts: 1533 | Registered: Sep 2004
| IP: Logged |
posted
Perhaps I didn't put that forcefully enough.
The universe is obviously intelligent as evidenced by little bits of it asking what the limits of intelligence are. The answer is everything minus the energy required to figure out everything.
Posts: 675 | Registered: Aug 2001
| IP: Logged |
Printer-friendly view of this topic
![[Razz]](tongue.gif)
