X-Message-Number: 14925 From: "Pat Clancy" <> Date: Wed, 15 Nov 2000 12:30:20 -0800 Subject: Re: Simulating People and Animals Lee Corbin in Message #14921 wrote > You raise two questions here; I meant only to ask about the > harder Turing machine case, not about other implementations, > such as neural nets. All computers, whether running a neural net program or anything else, or whether composed of one or a billion processors, are still just Turing machines, all logically equivalent - any Turing machine can do anything that can be done by any other Turing machine, it's just a matter of how long it takes. A machine implemented using a roll of toilet paper can accomplish the same thing as a billion-processor thousand-terahertz ultra-super- computer, it will just take a lot longer (and you'd probably need a really long roll). > > This leads to a host of very interesting questions, if you would be > so kind as to entertain them. I will assume that, speaking again of > a human body as a mechanism, we all disbelieve in the presence of > a non-physical host, or soul, that affects the actions taken by > a body; this is a fairly safe assumption on this list, because > neither Alcor nor CI is making any preparations whatsoever to > suspend souls. Okay, then > > >I hope that some other form of substrate can be developed that can > >support an artificial mind. But this will depend on our achieving > >some degree of understanding of how the mind works. Currently we > >have _no idea_ how the mind works. > > Well, evidently you know enough to pronounce it completely impossible > for computers to support artificial minds! Can you describe how you > know this, or at least why you feel it with such certainty, if you > admit that you have _no idea_ how the mind works? Well, one can make some judgements about what sorts of things are not explanations for the way something works, without knowing how it _does_ work. There's really no reason to think that the functioning of the mind is an _algorithm_, which is what is required to make it implementable as a computer program. In fact there are activities of the mind that would seem strongly not to be describle by any sort of finite, deterministic computation (which is the only kind you get from a computer). The detailed arguments, much better than I could present, can be found in various books ("What Computers Can't Do", Dreyfus; "Emporer's New Mind", "Shadows of the Mind", Penrose; "The Undiscovered Mind", Horgan; more on request). > > You recall that I have opened up the time domain in this discussion > to include billions of years, if necessary. So I guess that you > would not find it impossible for vast arrays of tiny computers to > successfully imitate, say, a cockroach. Is there any other point > short of humans in a sequence of animals leading up to man---say > ant, cockroach, snake, shark, mouse, cat, dog, chimpanzee, human--- > where you would be equally certain that simulation (as judged by > people) will never be possible? One of the arguments against the AI proponents is precisely that computer programs still are unable to simulate the most simple/primitive of animal activities - instead, they're good at things like Chess, which are really not "natural" activies of the brain, but which can be implemented by an algorithm. So, even the prey-recognition or locomotion capabilities of a "primitive" animal are still beyond the most sophisticated computer, whereas Chess grand-masters are in serious trouble. This is a big clue that computers aren't the right things to implement minds. > > What is it exactly that forbids the actions of humans or mammals > to forever resist simulation? I know that this is a very difficult > question, and will understand if you don't have any thoughts on it. > For example, will it be possible (presumably after many millions of > years) for computers to successfully imitate people who are asleep? Well, my feeling is that it doesn't matter how many millions of years of development take place, or what state of the brain is being simulated - it just won't happen with a Turing machine. > > On this score, I really am curious to know: do you think that it > will be possible to simulate people who are walking from one place > to another? How about people who are singing, or reciting poetry > with great affect? Ping-pong players? Politicians evading questions? > > I suppose that you will grant that at some point we will have some > sort of clunking robot that will take out the trash, beat any of us > at chess, convincingly talk like a psychiatrist (for a short while), > and probably do many other kinds of things, including running away > and shouting for help if anyone tries to turn it off. But what would > you guess is the key behavior that it will never be able to imitate? > I'm not saying that there will never be an artificial mind, just that it won't be based on a computer. The key test is the so-called Turing test - the artificial mind along with a set of real people are "behind a curtain" answering questions, and if you can't pick out the AM then it passes the test (i.e. it's "equivalent" to a real person). I just don't think a computer will ever pass that test. Pat Clancy Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=14925