X-Message-Number: 8083 From: (Thomas Donaldson) Subject: CRYONICS Re: CryoNet #7993 - #7997 Date: Wed, 16 Apr 1997 00:35:39 -0700 (PDT) Hi Guys! 1. To Mike Perry: You seem to assume that without mathematics a civilization must NECESSARILY be behind us. I never said that nor implied it. I am saying that they may be so far in advance of us that they have forgotten all of our present science and mathematics as the prattling of infants. My test for an intelligent civilization (and when you come down to it, this bears a lot on what I think of the Turing test, too) is not what they can say to us but what they can DO. As for whether or not we will eventually forget all of present mathematics, I will say that this is certain to happen. But also I expect it to take many thousands of years. Mathematics IS fundamentalto the way WE think about the world NOW. That is why it may last longer than our physics. As for forgetting, how many readers of Cryonet remember how to work a slide rule? 2. To John Clark: We may simply be unable to connect, but as you may remember, I provided my OWN quite explicit test for intelligence and awareness in a robot.You are welcome to disagree with it and explain why, but by ignoring it you fail to convince me that you have coped with the issues I raised. You may wish to read my answer to Mike which sets out my ideas also on the Turing test. Yes, some people may decide that the existence of Chess playing robots now means that playing chess does not use human intelligence. As I understand this issue, however, no one involved in programming such machines, even the strongest, makes any pretense that they are trying to imitate the way (say) a Grandmaster plays chess. They are writing a computer program which evaluates all the possibilities out to a certain distance, has facilities to test whether to go further along one or another tree, etc etc. This is very good programming, and I admire their abilities. But they themselves would laugh at the notion that they were creating anything like a human capability --- certainly, a different capability, but not human. As for implementing an accurate model of a neuron in a "modern mainframe computer", forget it. Neurons are very highly parallel, themselves. That is how their biochemistry works. You see, all those different volumes (and surfaces) with molecules floating in them or stuck to them are actually doing parallel computations: there are enough that the enzymes, the cofactors, and the substrates will quickly come into contact with one another and carry out their reactions, producing another set of biochemicals to be go through THEIR specific reactions. Sure, it is a variety of parallel computer that we do not yet know how to implement ourselves. To say that we just have bags of chemicals totally misses what's happening. (As a matter of fact, biochemists have begun to borrow ideas from electronics as a means to describe how this goes on and all fits together). Most important, it does not consist of a few reactions among a small number of molecules, but many closely related reactions among many molecules. There simply is no comparison. I think a suitable emulation can be done --- I would hardly claim it cannot. But no present computer is really equal to the task. Maybe a large computer built of nanotechnological parts may do it --- or maybe we will discover that these denigrated bags of chemicals are really one important direction to go if we want to improve our computing. We will see what the future will bring. That's all for now! Long long life, Thomas Donaldson Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=8083