X-Message-Number: 25560 From: Date: Sun, 16 Jan 2005 11:27:05 EST Subject: improving on nature Peter Merel has noted that it may have taken Nature an immense amount of space and time to produce DNA and its consequences, with the suggestion that we cannot count on quick solution of AI problems merely because Nature has produced intelligence. He does agree that we might speed things up, but perhaps does not carry this far enough. I think there are lots of hints that we probably can speed things up by important factors, as many others have also pointed out. 1. There are many things Nature never tried, that we in our brief tenure have already used successfully. For example, as far as I recall, the wheel does not exist in living things--but wheels and gears are important, successful devices in artifacts. And we have "birds" and "fishes" made of metal that fly higher and faster, or swim faster, than anything made of meat. That we have already outdone Nature in some areas does not in itself prove that we can do it in all areas. But it's a powerful hint. And remember that active enterprise and large resources for these endeavors are very recent--an eyeblink in history. 2. Evolutionary advances worked under at least one very great handicap--that improvements or new versions had to be reasonably consistent with previous models. As a crude example, if a mutation had a much larger head and brain but was otherwise the same, it could not survive--the system was not self-compatible. Also, in the case of sexual reproduction, a single mutant, carrying a dominant gene, had to be able to mate successfully with a non-mutant in order to produce mutant offspring. (For example, a mutant child with a large head would kill a mother with a small pelvis.) So successful and lasting improvements were few and far between. We don't have those handicaps. We can make leaps instead of creeps. 3. In some instances Nature has simply blundered in obvious ways. For example, the human eye has a blind spot because the anatomy is wrong, and the vermiform appendix is both useless and potentially dangerous. Human engineers would not have made those blunders, and human engineers could improve on Nature in some cases by the mere expedient of doing something similar but without the blunders. 4. We have many resources not available to Nature, or Nature near the surface of a planet. For example, cryogenics and hyperbarics and vacuum and use of rare materials. Evolution had to depend almost entirely on wastefully generous amounts of raw materials and long-lasting dominant conditions. 5. Motivation. Just a reminder--Nature doesn't give a damn, and we do, and that makes a difference. Part of Peter's argument concerns the "burden of proof"--we know that such-and-such is a hard problem, and no one has solved it yet, so we cannot make an optimistic presumption. But the correct application of probability theory in such cases is to take a step back and look at the larger picture. Those interested may look at my essay on probability on the CI web site. 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=25560