X-Message-Number: 25538 From: Date: Fri, 14 Jan 2005 09:52:44 EST Subject: AI again Peter Merel argues that the problem of general-purpose artificial intelligence is intractable so far, with nothing much further in sight. He recognizes that the problems of algorithmic approaches do not speak to the possibility of other approaches or hybrid approaches, but perhaps he gives less than due recognition to the implications of biology. If a tiny insect or mite can negotiate terrain, how hard can the problem be, in the longer view? If it has been done, it can be done--the Precedent Principle. Richard B.R. argues that both humans and general-purpose machines will always be inferior to special-purpose machines. He also tries to tie this to evolution--that we evolved to reproduce, and everything else is a side effect. Looking at that last, it is irrelevant. Humans of 50,000 years ago were genetically nearly identical to modern humans, and capable of designing and flying space ships, even though they had not yet invented agriculture. We can sometimes get clues to what we are by looking at origins, but what counts in any case is what we are, not how we became what we are. Richard also dismissed my point about idiot-savants by emphasizing the "idiot" part. That's not relevant. There are also people with superior math abilities who are otherwise normal, and little if any reason to think that savant status necessarily sacrifices other normal abilities, even though that sometimes happens to be the case. My main point was that "small" biological differences can result in enormous practical differences or consequences. Again, raising the mean IQ just a little can raise the number of geniuses by a lot, and they in turn can move civilizations. In a vaguely similar way, a "small" advance in AI architecture could result in major practical improvements. RBR also said that Einstein's life work could be encapsulated in a few lines, and said: >This is great for the >human species but is pathetic in absolute terms. Well, everybody's entitled to a few blunders, but the first and second parts of the above sentence seem a bit inconsistent--and surely the first part is the more important, even if the second part has any real meaning. Robert Ettinger Content-Type: text/html; charset="US-ASCII" [ AUTOMATICALLY SKIPPING HTML ENCODING! ] Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=25538