Note: You are viewing an old revision of this page. View the current version.
Describe SpirothetesAndHumans here.
Work in Progress. Please do not cite or copy. Copyright Hermit Technologies 2003 all rights reserved.
What is a man? A man is an intelligent, self-aware1
animal. What is a Spirothete2? A spirothete is an intelligent, self-aware machine. There are, at this time, many many men, some few of whom no doubt meet this definition. To the best of my knowledge, spirothetes do not yet exist. The first section of this document deals with the likelihood that they will exist. And, I will argue, far sooner than people anticipate.
This article attempts to provide a partial foundation for some of the propositions expressed in 
"Jobs and Human History", Church of Virus, Reply 1, Hermit, 2003-07-24, and in particular the development of spirothetes and their interaction with mankind:
Energy should already be close to "free", only our short sitedness has stopped this from happening. Eventually, somebody will get a clue - and it will be. Possibly space solar collection and concentration to drive MHD generators, likely turbines in ocean currents, probably hydraulic or pneumatic transmission of ocean pressure changes to land or platform based generators, or maybe, just maybe, fusion. At which point fresh water and most raw materials will be extracted from sea water and the cost will tend towards zero.
Nanotechnology will be a reality within 20 years. At which point the production of raw materials and processing and assembly of manufactured goods will tend towards zero.
The combination of almost free energy and almost free manufactured goods will result in the cost of transport and distribution dropping towards zero.
I used to think that the above would leave a value in ideas, concepts, plans and art. I no longer think this. Given the rate at which AI and computer technology is developing, and assuming no major discontinuities, within 20 years the typical desktop computer is likely to be more intelligent and creative than the most gifted human alive; and within 50, more intelligent and creative than the entire human race. So the value of creativity will also tend towards zero.
As intelligence and creativity surpassing human capability becomes commonplace, this will impact the factors above, greatly accelerating the tendency of all goods and services to be delivered free (as the cost of accounting will rapidy exceed the cost of of the goods).
Given these trends, the value of money will also tend towards zero. The incentive gone, "jobs" and economies are also likely to become obsolete. Given the massive acceleration in the slope of all these curves as the trends interact, the process of value-obsolescence is likely to occur with extreme rapidity.
I think that some of those reading this are likely to still be alive when this happens unless some asshole succeeds in blowing all of us to hell before then. A possibility which will increase dramatically as the Luddites panic when these trends become obvious.
The question is what should we be doing about it.
The human brain operates at only around 200 calculations per neuron per second. Due to the large number of neurons in the brain infra all operating in parallel, the brain actually calculates at a far higher effective speed. Current scientific opinion is that Man, like elephants3, dolphins4 and his close-cousins the great apes5, derives his self-awareness from his complex brain, capable of around 100 Teraflops (10^14 or 100 trillion (US) calculations per second). Our apparent data storage capacity is around 100 Terabytes (based on 10^14 analog values stored at the synapses of approximately 100 billion neurons with an average of 1 000 connections (synapses) per neuron)6. In other words, the arguments here depend on the thesis that self awareness is an artifact of neuronal density and complexity. This seems supported by the fact that only animals with complex brains appear to be self-conscious.
There are currently about 6.3 billion people living on Earth, and if we do not experience any major disasters (war, disease, cosmic, geological in decreasing order of probable deaths) this number is anticipated to rise by some 2.6 billion to 8.9 billion people by 20507. By 2060 global population will have stabilized, and subsequent to that there should be a decrease in overall population figures. We have extrapolated using a simple square rule based on the 2050 data to estimate total population. This suggests that in 2050, extrapolating single human performance, humanity as a whole will have a total capacity of 10^23 flops (100 ZettaFLOPS) of performance and 10^23 Bytes (100 ZettaBytes) of storage.
Moore's Law was formulated by Gordon Moore, a co-founder of Intel in 19658. As formulated today, Moore's Law states that "Storage density and processor performance will double every 18 months." This law is expected to apply using existing, well understood technologies until at least 2017. New technologies are being research which are likely to increase the rate at which developments occur. For example moving to three-dimensional chips founded on carbon nanotubes would change this from a square to a cube law progression, and the incorporation of superconducting materials will alleviate thermal constraints on packaging.
The Technical Basis For Spirothetes
If intelligence is a function of complexity then Moore's law and the very rapid developments we are seeing makes it inevitable. If it is a function of structure, our rapid development of understanding about the brain will do the same. I think it is a combination of the above and of chaotic systems (another field, which via fuzzy logic, we are beginning to get a very good handle on), and that the exponential curve of developments will result in self-evolving machines in the near future. Having examined the very interesting results of rapid evolution in "genetically programmed analog devices" (including those generated using digital circuits!), and seen our current inability to reverse engineer (let alone generate) devices which clearly "do the job" more efficiently (with fewer components) than circuits developed by human designers, I am fairly sure that once such evolution is initiated, that a full-blown spirothete will be the result. I am also sure that we won't understand how or why it works, and that it may have difficulty explaining it too9.
For the purpose of this article, we will assume base all the conclusions on a $500 (in current money) desktop computer, providing 2 500 MFlop performance, 128 MB of memory and a 40GB hard disk.10 Naturally, for those prepared to pay more, faster computers and greater storage will be available at earlier dates than projected here. In addition, it seems likely that clusters of computers, networked across the Internet will become very common in the near future, and that these "clustered virtual computers" will exceed the critical human and humanity capacities long before the dates projected here.
The reason for using this computer model is that irrespective of Luddite actions attempting to prevent the development of "General AI", the techniques required are becoming better known and it seems likely that irrespective of legal obstacles, people will attempt to develop such systems. As the capacity and performs of cheap, readily available desktop computers grows to meet and surpace that of humanity, the strong probability is that spirothetes, self-aware, self-evolving general Artificial Intelligence will be developed.
The Internet provides a database where most of the information gathered and generated by man can be found. Neural net11 technology, cognitively elementary, yet difficult to implement effectively is developing at an amazing pace, and a relatively recent development, the "fuzzy neuron"12 which works on a probabalistic basis, together with "Genetic Algorithms13, seem to offer a path to a very biological like self-training, self-evolving system. These technologies are likely to provide the basis for the development of "General AI"14 and eventually spirothetic life.
It is important to realise that the extrapolated data here will be based on a complete desktop computer system costing $500 in 2003 dollars10, and that an unaltered Moore's (i.e. straight doubling every 18 months) Law will be used to predict performance and capacities. As improvements in one area tend to improve performance in other areas and vice versa, these are likely pessimistic estimates. It is also important to realise that because both performance and memory capacity are doubling simultaneously, thus the neural network emulation capacity of computers does not just double, but increases with the product of these two improvements, i.e. it increases fourfold in each 18 month cycle. This is reflected in table 1 below under 'Neuron Equiv".
"Table 1: Projected Computer Capabilities (Based on extrapolation from UN population estimates).
(Note that the biggest quantitive improvements will occur towards the end of this process as a square law progression is currently occurring, and a cubic law progression is not impossible).
2021 Single $500 Desktop computer's disk storage capacity exceeds individual human's memory - with a much lower failure rate. 2027 Single $500 Desktop computer processor performance exceeds individual human's capabilities. 2030 Single $500 Desktop computer's 'Neuron Equivalence" exceeds that of an human. 2033 Single $500 Desktop computer has more dynamic memory than human and operates 20 times faster than a human. 2054 Single $500 Desktop computer's 'Neuron Equivalence" exceeds that of all humanity. 2070 Single $500 Desktop computer's disk storage capacity exceeds that of all humanity. 2076 Single $500 Desktop computer processor performance exceeds that of all humanity. 2082 Single $500 Desktop computer has more dynamic memory than human and operates 25 times faster than all humanity.
Human beings take 9+ years to reproduce. At least double that in most Western societies. Computers are reproduced at a rate of thousands per hour.
Human beings evolve randomly. Many promising developments are lost. Many harmful developments are retained. Human evolution has been a very random process and does not imply that later is superior to earlier. Only different. Computer evolution can be highly directed to optimise for any attributes and no permutations need be lost, but can be repeatedly reevaluated.
Human beings evolve at a rate of 1 significant (shared, inherited) genetic change per 2,500 years. Computers could evolve at a rate of many billions of "generations" per second.
Barring some idiot blowing us - or at least our civilization - to hell, or some natural disaster having the same effect, it appears that computers will be much more competent than humans within the next 80 years. Possibly even within the next 20 years. And it is not impossible that this will occur within the next 10 years. That means that spirothetic life is likely to arise within our life-times and it is almost unarguable that it will not arise during the livetimes of our children.
==
All life we know of, with very few domesticated exceptions, got to where it has by a process of "Natural Selection". "Natural Selection", which occurs when lifeforms clash both within and across species. The competion is caused by challenges, most usually restricted resources. Often the challenges were too great and too rapid for most existing life forms to adapt. In the aeons since life began, very few species have succeeded in surviving more than a few millions of years. At each stage, older life forms have been surpassed by newer that were, for one reason or another, slightly advantaged, in the evolutionary battles of that moment. Not necessarily a better life form, just a better suited to the moment - and sometimes luckier, life form.
Where these clashes occurred, the fitter species (for that time, in that environment) would survive and usually, but not always, thrive. For example, Homo Sapiens Sapiens replaced Neanderthal man. For a short while, Neanderthal shared the planet with his successor, but a little while later, Neanderthal was gone. Perhaps, for some small time while Neanderthal and Homo Sapiens Sapiens overlapped, Neanderthal could have eliminated Homo Sapiens Sapiens. And we would then not be here. Perhaps.Whatever the reasons, this did not happen. And an instant or two later, in geological time, it was too late. Neanderthal was extinct. And we are here. Intelligence appears to have bestowed an evolutionary advantage upon us, making us "better fitted" than Neanderthal.
This "dance of the species" has been driven by competition. If the competion were not there, no particular biological advantages (or disadvantages) would exist. For example, the mosquito, with few competitors, has been preying on sanguine species since the lower Cretaceous, some 100 million years ago, with little or no change. If spirothetes are developed (and we have see how likely this it happen) then they will be much more competent at thinking (and probably at implementing their thoughts) than man. Rather than the difference in competence being like the difference between Neanderthal and Homo Sapiens Sapiens, it will quite likely be like the difference between protozoa and Homo Sapiens Sapiens, with us playing the part of the protozoa.
This might not necessarily be a bad thing. If the spirothete is as intelligent as I have surmised, then it might look at us and decide to preserve us on ethical grounds, in the same fashion as we attempt to preserve endangered species. Unfortunately, I am not convinced that this will be the case. Rather in the way that we don't try to preserve smallpox on ethical grounds, but endeavour to eliminate it on the grounds that it forms an ongoing threat to us, mankind, currently being driven by competitive forces, may be seen as the only possible threat to spirothetes and this would create a "competitive environment" between men and spirothetes, where their desire for self-preservation and awareness of their superiority might overrule ethics.
Another issue is that judging from our history, it is unlikely that mankind will treat spirothetes as having rights until they are capable of enforcing them. This may make it seem to the spirothete as if the Neanderthals have been killing or enslaving Homo Sapiens - and justify our elimination on ethical grounds.
Yet no change means no development, and man is far from perfect.
A third issue would be the likelihood that
1 Self-awareness. Measured using the "mirror test". Gordon Gallup, Jr. began development of the mirror test in 1969. The mirror test involves marking the faces of animals or other parts which the animals cannot see with red dye, and determining whether the animals touch the marks on themselves or those appearing in a reflection in a mirror. Most animals treat mirror images as they would peers. Humans generally behave much the same way until 18 to 24 months, after which they start to recognize themselves in a reflection and also begin to use personal pronouns. Gallup proposes that the self-awareness demonstrated by this test might indicate the ability to empathize with other animals. "What an Elephant Sees in the Mirror", Jessica Clark, Brittanica.com, 2000-08-30 It should be noted that in "Self-Awareness" in the Pigeon", SCIENCE, VOL. 212, 1981-05-08, Robert Epstein, Robert P Lanza, BF Skinner, produced a nonmentalistic model of this test which perhaps reduces the value of this test. However a great deal more work based on the "mirror-test for self awareness" is in the literature subsequent to the publication of "Self-Awareness" in the Pigeon" suggesting that the "mirror-test" is still consider a valid test.
2 Spirothete: noun. A word coined to describe a living being, initially created as an artifact, from Latin, spiro -are; intransit., to breathe, blow, draw breath; to be alive; to have inspiration; be inspired; transit., to breath out, expire (also L/Gk spiros the breath of life) and synthetic adj 1: (chemistry) not of natural origin; prepared or made artificially 2: involving or of the nature of synthesis (combining separate elements to form a coherent whole) as opposed to analysis.
http://virus.lucifer.com/wiki/spirothete
3 Elephants: In research conducted at Sierra Nevada College in Lake Tahoe, Calif., two Asian elephants were observed for two and a half months. After they got used to the four-by-eight-foot mirror, they used it to find hidden objects and to observe marks that researchers placed on them with nontoxic face paint. "What an Elephant Sees in the Mirror", Jessica Clark, Brittanica.com, 2000-08-30
4 Dolphins: two new, unrelated studies are reporting that dolphins are able to recognize themselves in mirrors — often taken as a sign of self-awareness — and of spontaneously grasping the thoughts of other individuals, in this case, humans. " Brainy Dolphins Pass the Human 'Mirror' Test", The New York Times, Science, Mark Derr, 2001-05-01
5 Great Apes: Koko, a gorilla who was taught sign language, has mastered more than 1,000 signs and understands several thousand English words. On human IQ tests, she scores between 70 and 95, putting her in the slow learner - but not retarded - category. And At the Washington National Zoo, orangutans given mirrors explore parts of their bodies they can't see otherwise, showing a sense of self. An orangutan named Chantek at the Atlanta Zoo used a mirror to groom his teeth and adjust his sunglasses, says his trainer. "Man and other animals", The Guardian, Jeremy Rifkin, 2003-08-16. Refer also http://www.koko.org/
6 A human brain's probable processing power is around 100 Teraflops, roughly 100 trillion calculations per second, according to Hans Morvec, principal research scientist at the Robotics Institute of Carnegie Mellon University. This is based on factoring the capability of the brain's 100 billion neurons, each with over 1 000 connections to other neurons, with each connection capable of performing about 200 calculations per second... Morvec estimates a brain to have a 100-terabyte capacity. "This Is Your Computer on Brains", Wired News, Michelle Delio, 2002-11-19
7 "World Population Prospects: The 2002 Revision: Highlights", 2003-02-26
9 “Darwinism and unintentional electronics.”, Church of Virus, Hermit, 1999-06-02
11 Neural Networks:
"Artificial Neural Networks Technology", Dave Anderson and George McNeil, 1992-08-20
"Computers and Symbols versus Nets and Neurons", Kevin Gurney
"Neural Networks", Statsoft, 1984-2003
"An Introduction to Neural Networks", Prof. Leslie Smith, 1996-2003
12 Fuzzy Neural Networks: "Artificial Neural Networks Technology", Dave Anderson and George McNeil, 7.2 What the Next Developments Will Be?
13 Genetic Algorithms:
http://www.aihorizon.com/essays/generalai/rj_ga.htm
14 General AI:
http://www.pcai.com/web/ai_info/general_ai_sites.html
http://www.aihorizon.com/essays/generalai/
[] A discussion on the Singularity held between Hermit and Anand in the #virus IRC channel at irc.Lucifer.com:6667
Disussions on the BBS of the Church of Virus at http://virus.lucifer.com/bbs particularly with Andy Brice http://forum.javien.com/XMLmessage.php?id=id::HGlHRI0J-IXUD-P0Fw-VR1u-PwdeSyGKcn4Y and on the #virus IRC channel at irc.lucifer.com:6667, particularly with David Lucifer, BJ Klein and Annand http://forum.javien.com/XMLmessage.php?id=id::qYbtiiaT-7ATv-xEsX-pFS6-aaR5BTSic_3b.
*
* We develop ourselves to a point where we would not be perceived as a threat by superior intelligences by developing a binding ethic based on a rational perspective which is non-competitive in nature, recognizing other intelligences as being equally valuable, and live by it, while simultaneously attempting to improve ourselves through genetics and neurotechnology to a point where we might be recognized as worthy of a reverse recognition of being intelligent (despite our history).
* We develop the ability to serve as entertaining pets (doubtful that we could entertain creatures able to think a lot faster than ourselves for very long).
* We resign ourselves to oblivion.
So, the question of the day is, if the Luddites win (and right now they seem to be winning), will they have a very short period to enjoy this before they are eliminated by our successors, or is there an eye to this needle through which a few humans might hope to escape? A consequent question is, if there is something that can be done, what should we as individuals or as a church be attempting to do about it?
For I would put it so strongly as to suggest that we are in a situation parallel to 390 CE only in a much more dangerous world. And the barbarians are about to burn down the Library of Alexandria which contains the few possible "escape routes" which might offer us some hope.
"Cramming more components onto integrated circuits", Electronics Vol 38 No 8, Dr. Gordon E Moore, 1965-04-19