Date: Fri 26 Feb 1988 22:16-PST From: AIList Moderator Kenneth Laws Reply-To: AIList@KL.SRI.COM Us-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList V6 #41 - Supercomputers, Nonotechnology To: AIList@KL.SRI.COM Status: R AIList Digest Saturday, 27 Feb 1988 Volume 6 : Issue 41 Today's Topics: Reviews - Cray "Left Brains" & Spang Robinson Supercomputing, Opinion - Nanotechnology, Science Priorities ---------------------------------------------------------------------- Date: 25 Feb 88 06:29:18 GMT From: glacier!jbn@labrea.stanford.edu (John B. Nagle) Subject: Cray makes "left brains", says chairman Rollwagen ... Adding more and more parallel processors turns a computer into a new kind of animal, a thinking machine. Developers of thinking machines now are talking in terms not of 64 processors, but 64,000 interlinked processors. The linkages, not the processors, become the central part of the technology. Cray sees too much potential in its existing business to consider, at this point, going into the thinking machine business. "It's kind of like we make left brains. Even though they are getting more complex, it's still rational, linear, deterministic programs that we run. They [the thinking machine people] are trying to build the right brain, where the interconnections are as much of the machine as the processors. What you are going to do is just get them started and they're going to go off on their own and it will be fascinating to see how this works. But I think we're really talking about decades." (John A Rollwagen, chairman of the board of Cray Research, quoted in an article by George Melloan, in The Wall Street Journal of February 23, p. 29.) ------------------------------ Date: Sun, 21 Feb 88 22:36:39 CST From: smu!leff@uunet.UU.NET (Laurence Leff) Subject: Review - Spang Robinson Supercomputing Summary of Spang Robinson Report on Supercomputing and Parallel Processing Volume 2, No. 1, January 1988 Lead article is on Supercomputer Marketplace Cray Research 1985 - 100th machine shipped 1987 - 200th machine shipped Annual market - nine billion today, twenty billion in 1990 thirty billion in 1995 Academic Supercomputer Application Development Physical Sciences - 50% Geosciences - 15% Biosciences - 8% Social Science - 2% Math Sciences - 4% Engineering - 19% Multidisciplinary - 4% Worldwide Distribution of Supercomputers: U. S. 54% Japan 19% Other 8% UK, France, Germany, Canada less than 5% each Categorization of where the Supercomptuers are Research - 24% Universities 18% Defense 16% Aerospace 13% Petroleum 10% Environment 7% Nuclear Energy7% Service Bureau5% Automotive 5% Installed Base (1986): Supercomputers 228 Vector Augmented 190 Mainframes MInisupercompute 450 Superminicomputer 140,000 Workstations 110,000 The article also includes estimates for 1991 and beyond for most of the abov numbers as well as discussion. ______________________________ Article on Thinking Machines Connection Machine. As of May 1987, they delivered seven systems: They have raised a total of 31 million in equity capital with a great chunk untouched. ______________________________ The U. S. House of Representatives Subcommittee on Science, Research and Technology of the Committee on Science, Space and Technology plans a brief hearing in the February -early March time frame to review recommendations for futher support of Supercomputing. The Office of Science and Technology Policy has a report that says 1) U. S. government should development long range supercomputer support program because this is necessary for national security and the economy 2) joint research should be undertaken because of relationship with software and microelectronic 3) Network technology may become a barrier so further work should be done linking things up. 4) Long term support should be within available resources: They estimate $500 million on supercomputing which should be increased 70 percent. The National Network should be upgraded at a cost of 390 million. This includes upgrading existing networks to 1.5 MBit/second capacity, to be replaced with a 45 Mbit/sec and then to 45 Mbit/Sec. +_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_ Shorts: IBM has made a committment to Steve Chen's company that falls short of the $100 million needed to develop the product. Dana Computer has changed its name to Ardent Computer Technology. Cray Research will be selling a system to Construcciones Aeronauticas, S. A. Prime Computer Systems is announcing a system developed jointly with Cydronme. It is a $500,000 =- $600,000 machine using directed dataflow architecture and vector processors. A broker, Piper, Jaffray and Hopwood, recommended that investors accumulate Cray stocks. Sequent has announced an OEM agreement with Apricot computers of England. Westinghouse will be selling Scientific Computer System machines in a turnkey fashion for nuclear fuel customers. Encore made a profit of $114,000 for the year. Japan's MITI awarded Tandem its "Good Design Prize for Foreign Products" for its VLX mainframe and XL8 and V8 disk storage devices. Encore's President, James R. Pompa has resigned. Scientific Computers has released a version of System V Release 3 for its machines which includes UNICOS extensions. They also introduced a feature to increase the amount of main meory and to implement gather/scatter and compres instructions. University of Georgia had an ETA10-P shipped. BBN Advanced Systems has a low cost starter Butterfly system for $100,000 for universities. Topologix has introduced add-in boards for Suns containing four transputers a piece with up to eight boards per workstation. 'Convex has shipped beta site units of C2 which is targetted at 40 MFLOP's. Indiana University developed a system to interactively automate the conversion of programs for the Butterfly family of parallel processors. It is public domain. Ametek has announced a Series 2010, a hypercube based system, where messages pass through nodes without requiring processor time. Each node is a 68020 CPU plus 68881 floating point unit with 1 to eight MBYTES memory. It does not require expansions in powers of two and can grow to 1024 nodes. Dow Jones Informations has purchased a Connection Machine from Thinking Machines. The systems uses the Information retrieval technique of "relevant factor" in which the user evaluates what is returned from the first query to generate successive paths. A second 32,00 processor is on order for March delivery. One processor has 4.3% of the power of the entire world mainframe computer power. ------------------------------ Date: Sun, 21 Feb 88 17:06:54 CST From: smu!lewis@uunet.UU.NET (Eve Lewis) Subject: Nanotechnology, Science Priorities Urgent insert: On the morning of Friday, 12 February 1988, I heard the following NPR item:- The National Research Council has issued a report calling for a major shift in biology research. The report says the Federal Government should spend $200 million a year, for fifteen years, on a single genetic project. NPR's Science reporter, Laurie Garrett, has more: Garrett: For several years, some powerful scientists have been promoting the largest biological research effort ever attempted. They want to decipher the code of over three billion genetic mes- sages stored in human chromosomes. The National Research Council en- dorsed the project yesterday, but did not say which Federal agency should control it. Last night, Harvard University Nobel Laureate, David Baltimore, addressed the Annual Meeting of the American Association for the Advancement of Science. Despite his leading role in genetics and biotechnology research, Baltimore came out against "The Genome Project." Baltimore: When I walk around and ask people, "Do you feel in your own research, that what's holding you up, is a lack of sequence of the human genome?" I have yet to meet anybody who says, "Yes." Garrett: Baltimore says Federal money and research efforts would be better spent on a massive, co-ordinated attack on the AIDS virus. In Boston, I'm Laurie Garrett. End of this NPR transcript re: Baltimore's horrendous opinion. 1) Do researchers always know precisely what is "holding up" their research, or how said research would be spurred, stimulated, and aided by such an important store of data? 2) From Robert Kanigel's book, "Apprentice to Genius," which I just finished reading, discussing James Shannon of NIH, and I believe supporting my view that Baltimore's advice is a mistake: "Shannon's task was to align NIH's disease-oriented structure with the needs of basic research. The strategy he advanced all the years of his tenure as director was: Don't embark on a narrow search for disease cures at all." and, "`Knowledge of life processes and of phenomena underlying health and disease is still grossly inadequate,' he would write. Without such knowledge, it was a waste of time, money, and manpower to aim for the solution of a specific medical problem. He blamed the failure of polio vaccines back in the 1930s on lack of knowledge of the polio virus and techniques needed to culture it. He pulled the plug from an artificial heart program already approved because he didn't think cardiac functioning was well enough understood. "He didn't like the term basic research; he preferred calling it fundamental. But in the end it was the same. As he put it in an ar- ticle he coauthored for "Science" soon after becoming director, "The potential relevance of research to any disease category is [best] defined in terms of long-range possibilities and not in terms of work directed toward the quick solution of problems obviously and solely related to a given disease." Additionally, re: Ca. 1955, Dr. Seymour Kety, Director of Scientific Research for NIMH "In the long run, basic science would gain, leading to clinical ad- vances more abundant than if they'd been pursued directly. "Some years later, two clinical investigators, Julius Comroe and Robert Dripps, lent analytical force to Kety's intuition. The two undertook to examine the origins of the ten most important clinical advances in heart and lung medicine and surgery of the preceding thirty years. They tracked down 529 scientific articles that had, in retrospect, proven crucial to those clinical success stories. Of them, Comroe wrote, fully forty-one percent `reported work that, at the time it was done, had no relation whatever to the disease that it later helped to prevent, diagnose, treat, or alleviate.' Penicil- lin, the anticoagulant heparin, and the class of drugs known as beta-blockers were among them." Now, I refer to the above, because I believe that if David "reverse transcriptase" Baltimore's advice were followed, it would deprive the AI people and the neuroscience people (let's hope that some have a foot in each camp) of the BIOLOGICAL COMPILER IN THE NEURONS OF THE HUMAN BRAIN, to wit: the "reverse transcriptase" implicated in mental function. That is why it is too ironic for words, that Bal- timore should come out with such an opinion, in addition to its being pathetic that someone of his calibre should "think small." One wonders what interests have gotten to him. My article continues: On page A1 of the New York Times, 20 March 1987, there was a report of a meeting called by the American Physical Society. I tell you, these people were chortling in a state of absolute mania, in regard to the discoveries in "superconductivity." The head: "Discoveries Bring a `Woodstock' for Physics." Byline: By James Gleick. Was I jealous? Was I envious? I tell you that I was totally sick, to the max. But I also tell you that less than one year later, I am beginning to feel rather good, about AI (artificial intelligence, as we all know) and about NI (natural intelligence). Interdigitation between these two disciplines will forge an ab- solutely unbreakable bond, and after a struggle of two and a half millenia, WE will chortle at our own "Woodstock." I can see The New York Times article's banner headline in my "mind's eye," now: "DISCOVERIES BRING A `WOODSTOCK' FOR NEUROSCIENCE AND ARTIFICIAL INTELLIGENCE." (We will also be able to define in neurophysiological and molecular biological terms, just what is that "mind's eye.") Now, re: the Drexler concept: Godden in an article, "Intelligent Nanocomputers," dated Friday, 15 January 1988 @ 09:46 EST, discusses K. Eric Drexler's "Engines of Creation." What he explores particularly in his review is the chapter on AI and nanocomputers. "Drexler makes the fascinating claim (no doubt many will vehemently disagree) that to create a true artificial intel- ligence it is not necessary to first understand intelligence. All one has to do is simulate the brain, which can be done given nanotechnology. He suggests that a complete hardware simulation of the brain can be done, synapse-for-synapse and dendrite-for-dendrite, in the space of one cubic centimeter (this figure is backed up in the notes)." However, the real "Engine of Creation" is, in actuality, a NANO-NANOTECHNOLOGY. It would be most resourceful, and productive to access the BIOLOGICAL NANO-NANOTECHNOLOGY, and with this inspiration," with this wealth of clues, create an AI NANO-NANOTECHNOLOGY. Biological Nano-Nanotechnology IS Molecular Biology. But take heart, because there is no conflict between Biological Nanotechnology and Biological Nano-Nanotechnology. Indeed, the latter is the raison d'etre, the vis a tergo, the provider of templates, the sine qua non, for the former, which provides the model, and the source of reference for Drexler's "Engine of Creation," which is - one judges from Godden's review - an isomorph of the human brain. I only recommend that AI nanotechnologists avail themselves of the wealth of experimentation and information in Biological Nano-Nanotechnology. Every structure in the human body, not ex- cepting, for sure, the human brain, drags with it a phylogenetic residue, from further back in time than perhaps we care to remember, a DNA riddled with karma, "sins," choices made at forks in the evolutionary road, even choices not made at such forks, and all, rattling like the chains of Marley's ghost. We may be scared, in- hibited by built-in protective mechanims, which is why we have not yet solved "the biggie." I here refer to precisely what is the human mind, in neurophysiological and molecular biological terms. The structural genes of the human genome determine the morphology of the human brain, not to neglect the neurotransmitters, receptors, etc. In my view, the human genome is the engineer that masterminded the "Engine of Creation." There is no antithesis implicated; indeed, after swallowing the theories of Galileo, Darwin and Freud, the species (ours) had best prepare itself for another humongous gulp, when the genetic constraints on human thought are revealed and sub- stantiated. Godden's review of the Drexler book apparently elicited some reac- tion. I refer specifically to the article dated 01 Feb 88 (11:53 PST), by John McCarthy , and that from Wednesday, 03 Feb 1988 (01:25 EST), by Marvin L. Minsky . Dr. Minsky is optimistic in the extreme, as I am: "Progress in this direction certainly seems faster than almost everyone would have expected. I will make a prediction: In the next few years, various projects will request and obtain large budgets for the "human genome sequencing" enterprise. In the meantime, some- one will succeed in stretching single strands of protein, DNA, or RNA across crystalline surfaces, and sequence them, using the STM method. Eventually, it should become feasible to do such sequencing at multi-kilocycle rates, so that an entire chromosome could be logged in a few days." That is why Dr. Baltimore's advice is so alarming. With his prestige and influence, whomelse will he recruit for that bandwagon? John McCarthy (01 Feb 88) describes two extreme approaches to AI, the "Logic Approach," which is his preference: "Understand the common sense world well enough to express in a suit- able logical language the facts known to a person. Also express the reasoning methods as some kind of generalized logical inference." and the "Instrumental Approach," which McCarthy finds less useful: "Using nano-technology to make an instrumented person. (This ap- proach was suggested by Drexler's book and by Eve Lewis's commentary in AILIST. It may even be what she is suggesting)." McCarthy points out certain problems with the "Instrumental Approach," which on a Nanotechnological (Drexler) level would be a recruitment of a functional isomorph of the human brain, and on a more intricate, biologically more "basic," Nano-Nanotechnological level, would involve a functioning human genomime ["genomime" is a word I coined for the genomic equivalent of the brain isomorph]. He says "Sequence the human genome. Which one? Mostly they're the same, but let the researcher sequence his own." Dr. McCarthy is overlooking, in the most anti-serendipitous manner, what neuroscience, embryology, and most particularly, which he did not mention, PHYLOGENY, have to offer. Just for starters, the evolu- tion of the genome, pari passu with the phylogenesis of the pineal organ and the phylogenesis of the inner ear, as traced throughout the vertebrate phylum, would give artificial intelligence such a wealth of data, such an embarrassment of riches that it would hardly begin to know what to do with. This article is fast becoming a "gontzeh megillah," and I really must wind it up, but when McCarthy refers to "the human genome," and says, "which one?" that is enough to get me started again. As far as feeding "culturgens" (E.O. Wilson's term) into the super-duper parallel computer, does McCarthy really believe that there are no racial differences, no sexual differences, not to mention individual differences, in human brains? Does he think that it's "cultural pressure" that accounts for the Oriental reverence for the dragon, and for the Occidental St. George, slayer of same? "Evolution keeps going, and even if we don't do anything artificial, we won't be the same in ten million years." - Marvin L. Minsky Meanwhile, before McCarthy abandons the "Engine of Creation," en- tirely, I suggest that he check out the "Sexually Dimorphic Nucleus of the Hypothalamus," S.D.N., for short - just for starters. McCarthy states: "However, experience since the 1950s shows that AI is a difficult problem, and it is very likely that fully understand- ing intelligence may take of the order of a hundred years. Therefore, the winning approach is likely to be tens of years ahead of the also-rans." In fact, we are not referring just to the experience of the last few decades, but that of the last two and a half millenia. If you would read the "Works of Plato," specifically the Socratic dialogues in "Phaedrus" and "Theatetus" and compare the Dialogues to Minsky's "Society of Mind," or to Lopate's interview with him, or to Sir Francis Crick's seminar on "The Impact of Biochemistry on Neurobiology," at Cornell University on 6 May 1986, or to Jonathan Winson's "Brain and Psyche," or Michael S. Gazzaniga's "The Social Brain," or J.Z. Young's "Programs of the Brain," or the other per- ceptive contributions in this area, then you would have to ack- nowledge that as far as the struggle to comprehend memory, or in- ternal representation, or vision, or "How We Know Universals," is concerned, we would have to paraphrase the Queen in "Through the Looking-Glass," and admit that it's taken all the running we can do, just to keep in the same place! Nonetheless, solving this problem is do-able, and we are better off with the information about the trillion neurons, with the 10,000 connections each, and the plethora of neurotransmitters and recep- tors. And we will be better off yet with sequencing the human genome with the three billion base pairs, and the introns and the exons, and the promoters and the enhancers and the repressors and the even the "junk" DNA. Maybe, especially the "junk" DNA. Surely, we will have come up with a robust theory, or theories, to interpret all that data, but let it be there to interpret, just as we have to make sense out of all the neural pathways and connections and peptides and receptors. We don't need any "bottom-line," pessimistic, "applied research" types shoving us off that track, or plunging us back into the dark ages. So much for Baltimore's suggestion. Keep the faith! ------------------------------ End of AIList Digest ********************