X-Message-Number: 302 From att!Venus.YCC.Yale.Edu!LEVY%JANE Mon Apr 15 14:08 EST 1991 Date: Mon, 15 Apr 91 14:08 EST From: LEVY% Subject: Eric Klien's prediction about reanimation possibility To: Message-id: <> X-Envelope-to: [ Re: msg #293 - Eric Klien's prediction about reanimation possibility ] It's ludicrous. We could have subatomic sized computers, and unless we knew how to program them, they'd be of no more use to us than a corkscrew. The problem these days is software, not hardware. Machines don't know anything, and we still don't know what to tell them. [ Simon, good point, but not the last point. I recall reading (in sci.nanotech?) that the availability of greatly increased brute force computing power can free us to use algorithms that are both conceptually simpler and more robust, although much less efficient. For example, one may use a simple two-dimensional array rather than a complex tree structure (for representing a robot's field of view) and base one's pattern-matching algorithms on that representation. Current AI programs may thus become simpler and problems currently too complex to easily program may become more conceptually tractable. Also, when we have moderately sophisticated nanomachines to help us analyze a human's brain structure we should quickly get good ideas on how to build smart machines ourselves. (One possibility is to just copy the entire brain structure - even though we do not entirely understand what makes its organization intelligent - simulate it, based upon our current understanding of what aspects of brain operation are important, and see what happens.) Thus, while we certainly have a lot of basic science to learn and complex design issues to solve to enable reanimation, greatly improved hardware can greatly simplify our software problem. Sometimes technology does simplify our lives. - KQB ] Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=302