X-Message-Number: 24046 Date: Tue, 04 May 2004 22:28:48 -0700 From: Mike Perry <> Subject: Back to Cryonet; brains and computers I've been away from Cryonet for over a week now due to the press of other things, but have now read the intervening messages. I belatedly wish Mike Darwin a happy birthday, and acknowledge his many positive contributions to cryonics (while not overlooking the fact that nobody is perfect, including him). On the personal level, I'll credit Mike with influencing some of my own attitudes about cryonics as well as personal identity and survival, which in turn is reflected in my book, *Forever for All.* On another topic, I'm very sorry about the cremation of a CI member, as are others, and hope people will take a lesson from this. And I wanted to comment on Thomas Donaldson's remark from #23984: >Yes, there are right now computer parts which >work by eliminating certain circuit paths within them; when we have such >parts that can also grow circuit paths within themselves the prospects >for making a machine working like a human brain will look much better. >(And note one basic problem: even on a small scale, the total number >of possible paths is factorial of N, where N is the number of nearby >neurons, at least 1000 if not more. It's not practical to implement >such a part simply by including all possible paths and opening some >of them to substitute for growth I think by "path" Thomas means that each of N components, "neurons," is included exactly once, that is, you take a tour. (If the components are thought of as cities, you have a "traveling salesman's" tour.) With N components, there are N-factorial possible paths in this sense, quite a lot if N is in the neighborhood of 1000 or more. I'm not sure what is meant by "including all possible paths" but it's easy to see that one would not have to have N-factorial separate structures or objects to realize all possible paths without having growth of connections. We can imagine each component is equipped with a 1-way connection to every other component; each connection can be in one of two states, open or closed. Open means a message can be transmitted from the one component to the other, closed means it cannot. Then clearly we can realize any path among the N components by opening the appropriate connections, keeping the others closed (assuming we also single out a first or starting component as the beginning of the path). This is N*(N-1) total connections, not N-factorial. The bottom line seems to be that it may not be so difficult to reasonably duplicate the brain's function in non-growing but programmable hardware. It would, of course, be more advanced hardware than we have today, but still recognizably computerlike. Mike Perry Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=24046