X-Message-Number: 25533 From: Date: Thu, 13 Jan 2005 12:44:44 EST Subject: tail of the curve Richard B.R. writes in part: >The development of so-called general-purpose AI will be great for >creating believable computer characters. However, it won't solve >any significant problems whatsoever, and it will have only a >slightly broader range of applications than humans themselves I think he is missing something here, or let's call it two things: 1. There is only a relatively tiny difference between apes and humans, yet humans have immensely greater effective abilities than apes. There is also a biologically tiny difference between geniuses and idiots among humans, but the geniuses can accomplish immensely more. The normal curve of distribution of intelligence means that a small displacement of the mean results in a huge difference, percentagewise, in the number of geniuses or idiots. 2. RBR says, roughly, that humans are pretty bad at everything (compared to special purpose machines) because they are general-purpose. But some "idiot-savants" can do amazing machine-like cognition, which proves that their degree of mastery would be relatively easy for everyone to attain eventually. If some have already done it, how hard can it be? In other words, RBR says that general AI machines would have "only a slightly broader range of applications than humans themselves." If that were true, it would imply that any one human could have at most only a slightly broader range of abilities than any other, or than any ape for that matter. But we know that is not the case. Even "small" changes can make big differences. 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=25533