From extropians-request@gnu.ai.mit.edu Sun Jun 27 23:29:03 1993 Return-Path: Received: from usc.edu by chaph.usc.edu (4.1/SMI-4.1+ucs-3.0) id AA03189; Sun, 27 Jun 93 23:29:01 PDT Errors-To: Extropians-Request@gnu.ai.mit.edu Received: from wookumz.gnu.ai.mit.edu by usc.edu (4.1/SMI-3.0DEV3-USC+3.1) id AA09311; Sun, 27 Jun 93 23:28:56 PDT Errors-To: Extropians-Request@gnu.ai.mit.edu Received: by wookumz.gnu.ai.mit.edu (5.65/4.0) id ; Mon, 28 Jun 93 02:13:42 -0400 Message-Id: <9306280613.AA21264@wookumz.gnu.ai.mit.edu> To: ExI-Daily@gnu.ai.mit.edu Date: Mon, 28 Jun 93 02:13:23 -0400 X-Original-Message-Id: <9306280613.AA21257@wookumz.gnu.ai.mit.edu> X-Original-To: Extropians@gnu.ai.mit.edu From: Extropians-Request@gnu.ai.mit.edu Subject: Extropians Digest V93 #0359 X-Extropian-Date: Remailed on June 28, 373 P.N.O. [06:13:41 UTC] Reply-To: Extropians@gnu.ai.mit.edu Errors-To: Extropians-Request@gnu.ai.mit.edu Status: OR Extropians Digest Mon, 28 Jun 93 Volume 93 : Issue 0359 Today's Topics: AIT VirtSem: The Game of "Go" as an Example [3 msgs] DIET: CoQ [1 msgs] DIET: any pure vegan culture? [5 msgs] DIET: any pure vegan cultures? [1 msgs] EVOLUTION/DIET: What proto-hominids ate [2 msgs] Fermat [1 msgs] I'm on vacation until 7/7 -- cul8r [1 msgs] Let it be known [1 msgs] MEDIA: B: _Jurassic_Park_ [2 msgs] Meta: List (was AIT VirtSem: The Game of "Go" as an Example [1 msgs] NANO:Critiques of Drexler [1 msgs] Administrivia: This is the digested version of the Extropian mailing list. Please remember that this list is private; messages must not be forwarded without their author's permission. To send mail to the list/digest, address your posts to: extropians@gnu.ai.mit.edu To send add/drop requests for this digest, address your post to: exi-daily-request@gnu.ai.mit.edu To make a formal complaint or an administrative request, address your posts to: extropians-request@gnu.ai.mit.edu If your mail reader is operating correctly, replies to this message will be automatically addressed to the entire list [extropians@gnu.ai.mit.edu] - please avoid long quotes! The Extropian mailing list is brought to you by the Extropy Institute, through hardware, generously provided, by the Free Software Foundation - neither is responsible for its content. Forward, Onward, Outward - Harry Shapiro (habs) List Administrator. Approximate Size: 51363 bytes. ---------------------------------------------------------------------- Date: Sun, 27 Jun 1993 17:14:38 -0600 (MDT) From: Stanton McCandlish Subject: EVOLUTION/DIET: What proto-hominids ate Quoth Eric S. Raymond, verily I saith unto thee: > > Stanton McCandlish says: > > > Counter-challenge. Find me any truly "pure vegan" culture. Period. [Perry Metzger:] > > The Jains. [Eric:] > Yup. Same example I'd have picked. Jains are a religion, and their members are made up of several of India's many cultures. This is like citing the Southern Baptists as a culture. -- Stanton McCandlish * Space Migration * Networking * ChaOrder * NO GOV'T. * anton@hydra.unm.edu * Intelligence Increase * Nano * Crypto * NO RELIGION * FidoNet: 1:301/2 * Life Extension * Ethics * VR * Now! * NO MORE LIES! * Noise in the Void BBS * +1-505-246-8515 (24hr, 1200-14400, v32bis, N-8-1) * ------------------------------ Date: Sat, 26 Jun 93 9:20:46 CDT From: extr@jido.b30.ingr.com (Craig Presson) Subject: I'm on vacation until 7/7 -- cul8r ------------------------------ Date: Sun, 27 Jun 1993 18:02:38 -0600 (MDT) From: Stanton McCandlish Subject: MEDIA: B: _Jurassic_Park_ Quoth Jay R. Freeman, verily I saith unto thee: > [re: the plant cloning] ok no argument there. My main gripe was that the FILM did not explain it, easy as it's been to show that it is possible. > > Would a veloceraptor attack a T.rex? > > That might depend on how strong the bonds were within the _Velociraptor_ > pack. Naturalist Adolph Murie, in _The_Wolves_of_Mount_McKinley_ (1944, > recently reprinted) has documented several instances of wolf packs > attacking grizzly bears when the latter came too close to the wolf dens. Ah, but the size difference is a whole other world...If there were examples of say, red foxes attacking bears I might be more inclined to go along with this. Also, the velociraptors were meant to be close to human intelligence in this film, and beligerent or otherwise, I think it highly unlikely the remain velociraptor would have done anything but hightailed it out of there (if she'd ever been dumb enough to get inside the building in the first place!) If we take the wolf example as valid, here's a counter example: hyenas and lions, who compete directly for food quite often. The films of confrontations I have seen (several, on Discovery Channel specials, and the old Nat'l. Geo. show on PBS) invariably have the hyenas heckling and annoying the lion, but being VERY wary of getting within attacking range, as the lion is considerably stronger, faster, larger and more dangerous. Much like the velociraptor:T.rex situation... Harly conclusive, but... > The _Velociraptor_ in question also looked as if it were undertaking a > rather effective attack; it looked to me as if it was trying to slash > at throat blood vessels. Considering that _T._rex_'s short "arms" > didn't reach the area where the _Velociraptor_ was, that might have > worked. Can't refute that, other than to say I didn't think it was very effective at all, and I cuckled when I saw it happening in the movie. Knowing WHERE to attack and making an EFFECTIVE attack are very different! > I think the biggest problem here is that the _Velociraptors_ had been > fed an enormous meal less than 24 hours before, and would not likely > have been hungry. Just so. > We know that one _Velociraptor_ was shot and likely > killed, at the start of the movie, but it seems unlikely that any of > the others had any reason to know what a gun was, or to think that > humans were dangerous. Well, unless this main character was a TERRIBLE shot, one of those 'raptors got a few slugs pumped into her before the silly domesticated primates got into the ceiling panels... > Furthermore, they had only just escaped from > captivity into the "human paddock" of the park, which likely contained > no other edible creatures much larger than an appetizer. If the > _Velociraptors_ had been hungry, it would make great sense for them > to go after the obvious large instances of dinner in residence. > It is also not clear that the insides of buildings were foreign to > them; they had to be raised somewhere before being put into the > paddock, what do you bet they were in inside enclosures. They might > even associate the interiors of buildings with security. Hmm...this may be so. I was under the impression that after hatching they were put in that pit, otherwise they'd be REAL dangerous. Remember the lab tech said they were quite deadly after only a few weeks. And it's not really the enclosures I'm thinking of so much as the foreign, noisy, metallic environment with bright lights, guns, etc. If they really WERE smart, I'd think they'd associate humans with not only the SOURCE of the food (don't bite the hand that feeds!), but also with the nasty electric fences, and the like, and would have run like hell and gone after food they had more of an instinct to hunt. > I agree about hunger, but as for success, they were doing pretty well! > Did everybody notice that during all of _Jurassic_Park_, no human > manages to kill a dinosaur? Yeah, other than the one that got blown away in the cage (that part was also silly. How many predators to you know that are SO bent on killing everything in sight, that they will continue attacking helpless prey while an even worse predator is shocking them to death and shooting them full of holes? I think not.) It kind of bummed me, I wanted them to toast at least one velociraptor! >;) > > > the predators we know tend to get away from humans [...] > > and are scared of fire and guns and loud noises [...] > Fear of noise may be instinctive. Fear of guns _per_se_ -- not merely > as noisemakers -- surely must be cultural Sure, I meant as noise makers, and as things that go bang, and wow suddenly you hurt real bad. Probably the SIGHT of a gun would do nothing to phase the dinos whatsoever, but the sound likely would, especially considering how skittish present day reptiles are. Anyway, this is part of the "Alien syndrome". No matter how dangerous those snacks are, the repitilian horrors go after them incessantly, and show NO intelligence whatsoever, EXCEPT where it helps them get to the noisy primate twinkies. This was especially silly in the Alien movies, since the aliens were supposed to be MORE intelligent than us, and had a hell of a spacecraft. > Fear of humans is evidently at least partly cultural, because it is > often not seen in animal populations with little contact with humans. > L. David Mech and Jim Brandenburg studied arctic wolves on Ellesmere > Island a few years ago, and found the wolves curious and not fearful; > after a few weeks they would approach within meters. This is *not* > typical of mainland wolf populations! Very true, and actually supports my point. Wolves are, comparatively speaking, quite intelligent, and are also vicious predators, yet they do not generally attack people on sight. To me the most believable dinosaurs in the film were the brachiosaurs (loved that sneeze scene ) and especially the poison-spitting dino. It was playful and curious, until it decided the fatboy was a suitable meal. > and also reported that one of the first brown hyenas they saw > initiated physical contact with them of its own free will on the > second occasion that they saw it. (Delia( Owens squatted down and the > animal -- a 100-pound carnivore -- came up and put its nose in her > hair.) Yikes, that's pretty brave...(on BOTH creature's part!) > As for fear of fire, Owens and Owens (_op._cit._) report that some of > the local lions occasionally would come and curl up on the opposite > side of the campfire from them, a few meters away from it, outside > their tent at their research site. Hmm this is interesting. Perhaps my idea that all the lights would scare them off is hogwash then... > > I won't even bring up the cage attack sequence at the start, that was > > just silly. > > There was a frightened animal in a small, strange place, being moved > around, with strange motions, noises and smells. It was indeed silly > that the cage was that poorly handled, but I find it plausible that > the critter within would be frantic and dangerous. See above. I don't fault that part; but the behaviour was not in accordance with what we know about animals. [feeding examples] Hmm well, I think the spitter did eat porky, but the T.rex looked just like it was killing, maybe I need to watch it again, so I won't press this one any further. > > Another thing is the wires. [...] [maybe they were just thin and carried a charge, were'nt for phyisical strength] Quite possibly, though that hardly explains how the 'raptors got up those 20'-looking walls... > > Would a T.rex really make THAT much noise and shaking-around as it walked? > > Could it really move that fast? > > There was an article in _Scientific_American_ on the last subject > within the last few years; the author suggested that a human running > at top speed might have been able to outrun a _T._rex_. You can try > it first ... No thanks, but I'd tend to agree with that guesstimate. Elephants CAN move pretty quick, but hell that T.rex was WAY bigger than several elephants put together, and had all weight on 2 legs, so it would be harder move. Related to this, I seriously doubt that it would have gone after humans anyway. Lions don't generally eat field mice as far as I know, and everything I've read on large carnivorous dinos (admittedly in my youth) suggests that they went after BIG, herbiverous dinos, like those cute brachiosaurs. > > there was no reason in hell to clone carnivorous dinosaurs. > > I'd pay more to see _T._rex_ than to see _J._random_veggiesaurus_. Sure, and disneyland would likely get a big boost if they had LIVE pirates with guns, (to borrow from the film itself) but it won't happen, since it's just too dangerous. As an aside, I remember a Judge Dredd comic from about 7 years ago or so, where the action took place in a theme park in Huston (capital of the Republic of Texas of course). I can't remember the name of the park, but all the attractions were VERY dangerous, and of course the park was packed with rednecks out to prove their manhood. It was quite funny, and most of them never got out alive... ANYWAY, like I say, it was a great movie, and I'll be sure to see it again, despite some goofiness here and there. -- Stanton McCandlish * Space Migration * Networking * ChaOrder * NO GOV'T. * anton@hydra.unm.edu * Intelligence Increase * Nano * Crypto * NO RELIGION * FidoNet: 1:301/2 * Life Extension * Ethics * VR * Now! * NO MORE LIES! * Noise in the Void BBS * +1-505-246-8515 (24hr, 1200-14400, v32bis, N-8-1) * ------------------------------ Date: Sun, 27 Jun 93 17:45:25 -0700 From: hfinney@shell.portal.com Subject: AIT VirtSem: The Game of "Go" as an Example Tim makes some good points about the evolution of complexity, but I keep coming back to the question: Is there really a complexity increase from the AIT point of view? It seems to me that you can take a program which includes (A) the rules of Go, (B) a bunch of agents with simple initial strategies, and (C) a genetic programming system to run tournaments among these agents, cross-fertilize among the winners to create new agents, give the fittest more descendents, etc. You then run this program for 10^12 cycles, or whatever, and you print the program of the most fit agent. The result will be a very good Go program, one which presumably encodes most of the complexity of the game. So, I have just described, in two sentences, a program which will derive a near-optimal Go strategy. This could be coded up in probably a few hundred thousand bits at most. That means that, by AIT terms, all the complexity in an optimal Go strategy is no more than this many bits. AIT doesn't see how much compute time is spent starting from this simple description to evolve that optimal program. Compute time is irrelevant in the measure of complexity used in AIT. Only program size matters. I think this shows that AIT's measure of complexity is not a good match to what we mean by complexity. AIT is, by this argument, blind to the increase in complexity caused by genetic algorithms, or equivalently by the theory of evolution. In AIT terms, evolution does not increase complexity, and a Go program with a random string (that plays terrible Go) is every bit as complex as one with an absolutely fantastic Go program. Where Tim said: > In a sense, a winning Go agent--one that has played a lot of games and does > well--has *no shorter description of its strategy* than just the strategy > (bit string) itself! I would say that the same thing is true of a losing Go agent which has a random program bit string. There is no shorter description of that string than the bit string itself. This is where AIT fails to recognize the complexity that we see. Now, I will add that IMO there may be one escape from this argument (which is basically that advanced by Tom Tolman). That escape is the possibility that the above-described procedure won't work, that it will not be possible to evolve a good Go program in this way. Two ways in which I could see it failing are, first, that the set of initial agents are all so bad that we evolve one which can always win against the others and their randomly-produced descendants, but it doesn't actually play very good Go. (It could be that this is true of people as well.) In Tim's "landscape" terms, we might say that it had found a local maximum which missed most of the complexity of the game. Or, it could fail because we need a lot of random bits to drive the GA properly, and a deterministic pseudo-random number generator isn't random enough. This is perhaps what Tolman would suggest. He emphasized the importance of non-determinism, and I believe he argued that the complexity that we see around us was dependent on quantum randomness, that evolution simply wouldn't happen in a deterministic universe. Applying that argument here, we'd say that the only way to drive a GA long enough to produce a billion-bit Go program whose AIT complexity was actually a billion bits would be if the starting point also had an AIT complexity of a billion bits. And this could be done if we took our 10^5-bit GA+Go program and added a starting random seed of 10^9 bits to drive the RNG. Intuitively it seems unlikely that the GA would fail to produce a good Go program. If a very good Go player could in fact be produced in this way, it would challenge the applicability of the AIT measure of complexity to evolutionary processes. Hal Finney hfinney@shell.portal.com ------------------------------ Date: Sun, 27 Jun 1993 20:48:21 -0400 From: "Perry E. Metzger" Subject: EVOLUTION/DIET: What proto-hominids ate X-Reposting-Policy: redistribute only with permission Stanton McCandlish says: > > > > Counter-challenge. Find me any truly "pure vegan" culture. Period. > > [Perry Metzger:] > > > The Jains. > [Eric:] > > Yup. Same example I'd have picked. > > Jains are a religion, and their members are made up of several of India's > many cultures. > > This is like citing the Southern Baptists as a culture. By your logic, Stanton, there is no such thing as a culture at all. Perry ------------------------------ Date: Sun, 27 Jun 1993 19:00:28 -0600 (MDT) From: Stanton McCandlish Subject: AIT VirtSem: The Game of "Go" as an Example Quoth Timothy C. May, verily I saith unto thee: > > [Sorry for sending this out a second time, if any of you have already > gotten it. I sent it out 24 hours ago and it has not appeared at my site. > Nor have I seen any comments on it. Very frustrating! Not knowing if it > made it out and all. As this has happened several times to me in the last > few weeks, I wonder if the either the new List software or the new List > server site--which I heard was being changed--is dropping messages?] > > > In this extended reply, I will use "Go" as a touchstone for discussing > string complexity, fitness functions, and the answer to the question of how > complexity can emerge from simplicity. [...] It DID reach my site the 1st time around. BTW, could you send another copy of the reading list? I didn't get THAT the first time, as far as I can tell... -- Stanton McCandlish * Space Migration * Networking * ChaOrder * NO GOV'T. * anton@hydra.unm.edu * Intelligence Increase * Nano * Crypto * NO RELIGION * FidoNet: 1:301/2 * Life Extension * Ethics * VR * Now! * NO MORE LIES! * Noise in the Void BBS * +1-505-246-8515 (24hr, 1200-14400, v32bis, N-8-1) * ------------------------------ Date: Sun, 27 Jun 93 18:11:43 PDT From: tcmay@netcom.com (Timothy C. May) Subject: MEDIA: B: _Jurassic_Park_ I had a question about the plants and animals in "Jurassic Park: How did they get so big in the several years that J.P. was supposed to have been in development? Even if cloning is possible, which I am prepared to grant, how could a 100-ton dinosaur reach full size in so few years? Or a triceratops? Or the others? (Elephants takes 15-20 years to reach maturity, I suspect other large animals are roughly the same...some calculations of energy budgets, food intake, muscle growth rates, etc., might be possible. Does the fossil record give any clues as to dinosaur maturity times?) And the plants and trees....I don't recall if there were any of the large Jurassic-era trees (ferns?), but they would certainly take many years--I would think--to reach large sizes. (But, like I said, I'm not sure any really large trees in the movie were supposed to be Jurassic-era trees.) Just wondering. -Tim -- .......................................................................... Timothy C. May | Crypto Anarchy: encryption, digital money, tcmay@netcom.com | anonymous networks, digital pseudonyms, zero 408-688-5409 | knowledge, reputations, information markets, W.A.S.T.E.: Aptos, CA | black markets, collapse of governments. Higher Power: 2^756839 | Public Key: PGP and MailSafe available. Note: I put time and money into writing this posting. I hope you enjoy it. ------------------------------ Date: Sun, 27 Jun 1993 19:32:02 -0600 (MDT) From: Stanton McCandlish Subject: DIET: any pure vegan culture? Quoth Perry E. Metzger, verily I saith unto thee: > > > Stanton McCandlish says: > > > > > Counter-challenge. Find me any truly "pure vegan" culture. Period. > > > > [Perry:] > > > > The Jains. > > > > Jains are a religion, and their members are made up of several of India's > > many cultures. > > This is like citing the Southern Baptists as a culture. > > By your logic, Stanton, there is no such thing as a culture at all. I beg your pardon? Please re-read the above: "and their members are made up of several of India's many CULTURES". Of course there are cultures. The Navajo are a culture. The Germans are a culture. The Japanese are a culture (not counting the Ainu, who are their own culture), etc. By anthropological definitions, a religion is not necessarily a culture, and a culture does not necessarily have a religion. Citing the Jains as a culture is also much like citing adherents to the Copenhagen interpretation as a culture. Just doesn't wash. Reviewing 2 of my sentences hardly gives you enough information(2) about 'my logic' to judge it like that, I think. [insert appropriate emoticon here]. -- Stanton McCandlish * Space Migration * Networking * ChaOrder * NO GOV'T. * anton@hydra.unm.edu * Intelligence Increase * Nano * Crypto * NO RELIGION * FidoNet: 1:301/2 * Life Extension * Ethics * VR * Now! * NO MORE LIES! * Noise in the Void BBS * +1-505-246-8515 (24hr, 1200-14400, v32bis, N-8-1) * ------------------------------ Date: Sun, 27 Jun 1993 22:49:14 -0400 (EDT) From: esr@snark.thyrsus.com (Eric S. Raymond) Subject: DIET: any pure vegan culture? > Citing the Jains as a culture is also much like citing adherents to the > Copenhagen interpretation as a culture. Just doesn't wash. The cases aren't parallel. Jainism isn't an intellectual abstraction, it's a total cosmological/ethical/religious system that conditions every aspect of a believer's life. In India, Jains tend to live in their own villages and form distinct communities, separate from the prevailing folk-Hindu/Islamic culture and perceived as separate by it. Much as the Amish do in the U.S. -- Eric S. Raymond ------------------------------ Date: Sun, 27 Jun 1993 22:58:47 -0400 (EDT) From: Harry Shapiro Subject: Meta: List (was AIT VirtSem: The Game of "Go" as an Example a conscious being, Timothy C. May wrote: > [Sorry for sending this out a second time, if any of you have already > gotten it. I sent it out 24 hours ago and it has not appeared at my site. > Nor have I seen any comments on it. Very frustrating! Not knowing if it > made it out and all. As this has happened several times to me in the last > few weeks, I wonder if the either the new List software or the new List > server site--which I heard was being changed--is dropping messages?] The software has yet to change. Delays are normally related to the abilities of the GNU Sendmail and/or network interuptions. I get up to 500 bounced messages EACH day. It is possible for a message to go out to the list and not make it to each list member. (FYI, our new software will send out a list of the entire day's messages, if you request it to, and missed posted can be requested.) ------------------------------ Date: Sun, 27 Jun 93 19:58:41 -0700 From: tcmay@netcom.com (Timothy C. May) Subject: AIT VirtSem: The Game of "Go" as an Example Hal Finney makes some thoughtful comments on my "Go" example, though I disagree with his conclusions. (But then the whole constellation of "information," "complexity," "entropy," and "randomness" ideas is notorious for producing varying reactions. This is seen in this recent debate, in debates in general about entropy and complexity, and in at least a dozen discussions I've seen over the years in various Usenet groups!) I'll try to limit myself to just a few of Hal's points: >Is there really a complexity increase from the AIT point of view? > >It seems to me that you can take a program which includes (A) the >rules of Go, (B) a bunch of agents with simple initial strategies, and >(C) a genetic programming system to run tournaments among these agents, >cross-fertilize among the winners to create new agents, give the fittest >more descendents, etc. I picked this Go example because it illustrates the "paradoxes" of information (which AIT and the related idea of "logical depth" do a better job of resolving than anything else I've seen). To my mind, Go clearly exemplifies "increased complexity." My shelf full of Go books from Ishi Press sure seems to have more than a few bits of information! (Experts, Go programs, difficulty...you get the picture.) The original Hilbert-Russell point of view, if updated to the modern computer-AIT era, would be that the Go only contains a few bits...the rest is just the logical extension, the inevitable outcome. (Indeed, a "Jupiter-sized brain" might take in the rules of Go and almost immediately see the "implications" and then play the game at a stupendous skill level. Note that I am not implying a mere Jupiter-sized brain can compute the 10^700 possible games of Go that I mentioned earlier.) >You then run this program for 10^12 cycles, or whatever, and you >print the program of the most fit agent. The result will be a very good >Go program, one which presumably encodes most of the complexity of the >game. > >So, I have just described, in two sentences, a program which will derive >a near-optimal Go strategy. This could be coded up in probably a few >hundred thousand bits at most. That means that, by AIT terms, all the >complexity in an optimal Go strategy is no more than this many bits. Perhaps. I understand a few groups are working on GA and GP approaches to Go, and we'll see how capable these programs (strings, state vectors, depending on your formalism) are. Also, be very careful when saying that a two sentence description is an actual algorithm--recall the various "recipes" for impossible dishes that start out, "First, catch a XXX." Or, the "description" of the "smallest number describable in six words."...I just described it, it must exist, but what is it? (This paradox had a lot to do with the Kolmogorov-Chaitin formutation of AIT.) >AIT doesn't see how much compute time is spent starting from this simple >description to evolve that optimal program. Compute time is irrelevant >in the measure of complexity used in AIT. Only program size matters. > >I think this shows that AIT's measure of complexity is not a good match >to what we mean by complexity. AIT is, by this argument, blind to the >increase in complexity caused by genetic algorithms, or equivalently by >the theory of evolution. In AIT terms, evolution does not increase complexity, >and a Go program with a random string (that plays terrible Go) is every >bit as complex as one with an absolutely fantastic Go program. "What we mean by complexity"? Speak for yourself, white man! Seriously, complexity is a confusing measure. No doubt about it. Kolmogorov, Chaitin, Solomonoff, and Martin-Lof came up with one measure: program length. This captures many important ideas about how to describe sequences or strings. Charles Bennett, also of IBM Yorktown Heights, came up with "logical depth" which captures the "running time" aspects. (Bennett's articles are available in several places, which I've mentioned several times on this List, such as Herken's "The Universal Turing Machine: A Half-Century Survey, and the PhysComp proceedings. A Martin Gardner book, whose name escapes me right now, has a guest article by Bennett which discusses logical depth and Chaitin's "Omega.") A Space-Time Diagram of Program Size and Program Running Time We on this List are severely hampered by our lack of pictures, fast feedback in front of a chalkboard or sheets of paper, and general lack of the normal tools for our types of discussions. ASCII just doesn't do the job. Here's a (painful, for me and perhaps for you) attempt to draw a picture that I can recycle for later use. The X-axis represents the length of a program, as in the Chaitin formalism, and the Y-axis represents the running time, or "logical depth," in Bennett's terms. ^ ^ | | | | program run time | | (time) | | "logical depth" | | (Bennett) | | |_ _ _ _ _ _ _ _ _ _ _ _ > program size---> (space) "algorithmic information content" (Chaitin) Now here's the same diagram with a few examples or cases thrown in: (BTW, no scales are given, but they may as well be seen as logarithmic.) ^ ^ | | | | C C = "life" (lots of cycles, program run time | lots of DNA complexity) | (time) | A A = Dawkins' "biomorphs" | (simple rules, long run "logical depth" | time to generate) | (Bennett) | B = complex string | D B D = simple string |_ _ _ _ _ _ _ _ _ _ _ _ > program size---> (space) "algorithmic information content" (Chaitin) I don't have a feel for where the "Go" example might fit on this...perhaps between A and C, in the sense that a good Go program may have a program size of hundreds of kilobytes and is "generated" by lots of cycles of evolution or whatever type of running is needed to "generate" it. (Needless to say, these diagrams are just that: diagrams. No great significance should be attached to specific locations.) And tradeoffs can be made, analogous to the space-time tradeoffs so familiar in complexity theory. Back to Hal's points: > >Where Tim said: >> In a sense, a winning Go agent--one that has played a lot of games and does >> well--has *no shorter description of its strategy* than just the strategy >> (bit string) itself! > >I would say that the same thing is true of a losing Go agent which has a >random program bit string. There is no shorter description of that string >than the bit string itself. This is where AIT fails to recognize the >complexity >that we see. Sure, in some sense even the "losing" strings have nearly the same complexity. That's expected, as the winners and losers are playing at nearly the same level! (The loser is only marginally weaker than the winner.) (It's like two heavyweight boxing contenders--they should be fairly evenly matched by the time they fight.) > > >Now, I will add that IMO there may be one escape from this argument >(which is basically that advanced by Tom Tolman). That escape is the >possibility that the above-described procedure won't work, that it will >not be possible to evolve a good Go program in this way. Two ways in >which I could see it failing are, first, that the set of initial agents are >all so bad that we evolve one which can always win against the others and >their randomly-produced descendants, but it doesn't actually play very good >Go. (It could be that this is true of people as well.) In Tim's "landscape" >terms, we might say that it had found a local maximum which missed most of >the complexity of the game. Like I said earlier, we'll see how well the GA/GP approach works for Go. Note that GA/GP has nothing per se to do with the arguments....I could have used neural nets as the example (in which case the "strings" would be reflecting the state vector, for example, of a neural net.) I won't comment further on this fellow Tolman, as his views sound far afield from what I'm talking about and his invocation of quantum mechanics seems unnecessary at best and actually misleading at worst. He was gone from the list by the time I joined a year ago, so I never saw his stuff directly. In conclusion, I maintain the Go example captures the ideas of AIT "in the real world." -Tim May -- Timothy C. May | Crypto Anarchy: encryption, digital money, tcmay@netcom.com | anonymous networks, digital pseudonyms, zero 408-688-5409 | knowledge, reputations, information markets, W.A.S.T.E.: Aptos, CA | black markets, collapse of governments. Higher Power: 2^756839 | Public Key: by arrangement Note: I put time and money into writing this posting. I hope you enjoy it. ------------------------------ Date: Sun, 27 Jun 1993 23:21:10 -0400 From: "Perry E. Metzger" Subject: DIET: any pure vegan culture? X-Reposting-Policy: redistribute only with permission Stanton McCandlish says: > > > Jains are a religion, and their members are made up of several of India's > > > many cultures. > > > This is like citing the Southern Baptists as a culture. > > > > By your logic, Stanton, there is no such thing as a culture at all. > > I beg your pardon? Please re-read the above: "and their members are made > up of several of India's many CULTURES". Of course there are > cultures. Yes, but your claim is that because the Jains are geographically distributed (they aren't, really, but never mind that) and a religion that they aren't a culture. Well, I'd argue forcefully that Southern Baptists ARE a culture, that we (that is, extropians) are a culture, and that without question the Jains are a culture. > Citing the Jains as a culture is also much like citing adherents to the > Copenhagen interpretation as a culture. Just doesn't wash. Quantum physicists are, however, a culture -- and so are computer people. (Having been at Usenix for the last week, I can tell you for certain that high-eschelon Unix hackers constitute a definate culture.) But as for the Jains, lets review: they marry other Jains, they live in villages together, they have very specific dietary and lifestyle constraints and they are pretty similar across India. I'd say they are a culture without any doubt. Perry ------------------------------ Date: Sun, 27 Jun 93 20:46:06 -0700 From: dasher@netcom.com (D. Anton Sherwood) Subject: Fermat Ray Cromwell writes: > So far we've been lucky. "Math is random", but we keep picking conjectures > that are decidable. Tim May: > I think most mathematicians felt that Fermat's Last Theorem was either > true or false (very probably true), and that one day we would know. Well, "cannot be disproven" would be equivalent to "true"; what would "cannot be proven" mean for Fermat's LT? *\\* Anton ------------------------------ Date: Sun, 27 Jun 1993 23:48:24 -0400 (EDT) From: esr@snark.thyrsus.com (Eric S. Raymond) Subject: DIET: any pure vegan culture? > But as for the Jains, lets review: they marry other Jains, they live > in villages together, they have very specific dietary and lifestyle > constraints and they are pretty similar across India. I'd say they are > a culture without any doubt. And I'd agree. -- Eric S. Raymond ------------------------------ Date: Sun, 27 Jun 1993 22:18:26 -0600 (MDT) From: Stanton McCandlish Subject: DIET: any pure vegan culture? Quoth Perry E. Metzger, verily I saith unto thee: > Yes, but your claim is that because the Jains are geographically > distributed I did not claim they were geographically distributed. You are misrepresenting me. > Well, I'd argue forcefully that Southern > Baptists ARE a culture, that we (that is, extropians) are a culture, > and that without question the Jains are a culture. [...] > Quantum physicists are, however, a culture -- and so are computer > people. (Having been at Usenix for the last week, I can tell you for > certain that high-eschelon Unix hackers constitute a definate > culture.) These are not cultures. They are subcultures. Please read some anthropology before lecturing on it. No anthropologist would accept these as cultures. If you wish to propose new definitions of "culture" then fine, but your use does not fit the conventional use at all. Maybe we need a culture(1) and culture(2)? -- Stanton McCandlish * Space Migration * Networking * ChaOrder * NO GOV'T. * anton@hydra.unm.edu * Intelligence Increase * Nano * Crypto * NO RELIGION * FidoNet: 1:301/2 * Life Extension * Ethics * VR * Now! * NO MORE LIES! * Noise in the Void BBS * +1-505-246-8515 (24hr, 1200-14400, v32bis, N-8-1) * ------------------------------ Date: Sun, 27 Jun 1993 23:59:07 -0400 (EDT) From: esr@snark.thyrsus.com (Eric S. Raymond) Subject: DIET: CoQ At the first meeting of the Malvern Extropian Nexus, Doug Platt held forth on the virtues of a food supplement called Coenzyme Q-10. Two things about his peroration caught my attention: 1) He said it's been really well-studied because something like 15% of the population of Japan is taking it under prescription --- so its good effects are well known and extensively verified, and 2) It's known to specifically effective against periodontal disease. The latter caught my attention, as I have excellent teeth but am woefully prone to gingivitis. My girlfriend and I bought some and have been taking it for about a week, and it's spooky --- we think we're already seeing and feeling the anti- gingivitis effect. This is purely anecdotal and uncontrolled, of course. But I pass it along for whatever it's worth. Thanks, Doug. -- >>eric>> ------------------------------ Date: Sunday, 27 June 1993 13:30:56 PST8 From: "James A. Donald" Subject: NANO:Critiques of Drexler In <9306241455.AA16822@wookumz.gnu.ai.mit.edu>, ashall@magnus.acs.ohio-state.edu (Andrew S Hall) wrote: > I am not saying Drexler is wrong, yet. It is just that some of his > statements ring a little warning bell in me built up from my days > of physical chemistry. Specifically, it seems to me that Drexler > dismisses the potentially harmful effects of random interactions > (as would occur if the nanomachine was treated as a reactive particle) > and then two pages later relies on thermal motion to achieve his > desired results. > > Before I sit down and do some calculations and grind through this with > the effort needed to confirm or deny, I would like to see some expert > criticism of Drexler. The only thing I have read was an old and fluffy > exchange in the WER. (from 1990?) If anyone can point me to some > technical critiques of Drexler, I would be grateful. When I have > a sufficient length of free time, I will do a search, of course, but > would appreciate anything that help out. Not an expert, but I think I have sufficient knowledge. The fundamental problem in constructing nanostructures is that it is extremely difficult to predict how the nano structure will behave, for example the unsolved, and perhaps insoluble, protein folding problem. Drexler is aware of this problem and rightly proposes to deal with this by creating nanostructures out of components whose consequences and effects are extremely isolated, so that a complex system can be treated as a large collection of separate subsystems, in the same way as we design computer programs and macroscopic machines. Unfortunately the resulting designs are violently chemically unstable. A random molecule banging about inside them will have the same effect as monkey wrench in your car gearbox, causing everything to randomly rearrange in a violently exothermic fashion. Drexler ignores the problem of friction. Friction becomes more and more severe as structures with moving parts get smaller and smaller. The only way to deal with this problem is the way that cells deal with it. All components are part of a gel or colloid, and motion is accommodated by the random thermal variations of the gel and its solvent. Structures are made out of jelly and fluid, rather than solid rigid components as in a Drexler design. Unfortunately it seems extremely difficult to construct large soft structures whose behavior within a gel or colloid is reasonably predictable, and Drexler does not attempt to do this. --------------------------------------------------------------------- | We have the right to defend ourselves and our James A. Donald | property, because of the kind of animals that we | are. True law derives from this right, not from jamesdon@infoserv.com | the arbitrary power of the omnipotent state. ------------------------------ Date: Sun, 27 Jun 1993 22:22:26 -0700 From: nobody@pmantis.berkeley.edu Subject: Let it be known > From: esr@snark.thyrsus.com (Eric S. Raymond) > > > But as for the Jains, lets review: they marry other Jains, they live > > in villages together, they have very specific dietary and lifestyle > > constraints and they are pretty similar across India. I'd say they are > > a culture without any doubt. > > And I'd agree. > -- > Eric S. Raymond Me too. -- No F. Content, Jr. Official Inspector of E-mail, Carrier of the Stamp of Approval, Department of the Useless Post. ------------------------------ Date: Mon, 28 Jun 1993 01:03:31 -0400 (EDT) From: esr@snark.thyrsus.com (Eric S. Raymond) Subject: DIET: any pure vegan cultures? Stanton, Perry: Much as I hate to break up a good fight :-), especially one I helped start, the argument over whether the Jains constitute a "culture" in the srict technical sense is irrelevant. Let's recall my original challenge to Perry: find me a pure-vegan culture living above 40 degrees north latitude in pre-industrial conditions. Since I had the Jains in mind as an example of a *tropical* vegan `culture' when I posed the challenge, I am certainly willing to accept a putative subculture analogous to the Jains as an answer to the challenge. Note 1: Perry, Stanton is right. Your definition of a `culture' is unusably loose for anthropologists. Cultures can and should be distinguished from subcultures by whether the population would be self-sufficient if the candidate `parent' culture were to vanish. The Amish were until recently and perhaps still are a culture; UNIX hackers are definitely a subculture. Note 2: I think the Jains *do* constitute a culture in the strict sense. Stanton, you need to study the Jains a bit if you want to argue this one. I have, because I was interested in their mysticism. Note 3: Anybody wanna make a side bet on the challenge, payable at the gala? -- >>eric>> ------------------------------ End of Extropians Digest V93 Issue #0359 ****************************************