Brief History of Intellectual Discussion of Accelerating Change
Accelerating Universal Phases of Physical-Computational Change
© 1999-2008, John Smart. Reproduction, review and quotation encouraged with attribution.
Origins The Idea of Progressive Change
To consider the origins of the idea of accelerating change, we should briefly go back to a much earlier one, that of progress itself. As Historian J.D. Bury reminds us (The Idea of Progress, 1920), the idea of progress in any human domain other than spiritual (e.g., social, intellectual, technical), versus stasis, moral decline, or cyclic fluctuation, has been a quite recent emergence in human history. We see no evidence for it at the start of human civilization in Mesopotamia with the Sumerians, circa 3,500 B.C.. Surprisingly, it was missed entirely by Greek civilization during its "Golden Age" of imperial democracy and scientific flowering, 500-300 B.C. Even the rise of the Roman empire was not explicitly (e.g., in the written record) associated with progress! Consider the historical context. Great empires had a long history of rising and falling. Most intelligent folk simply could not believe in the idea of continual progress when the pay for the Roman soldier was a fixed number of denari for the last three centuries of Roman rule (e.g., 100-400 A.D). The idea became further untenable in the West as Rome itself collapsed, as city sizes shrunk, and as Europe entered long political and ideological eras of escalating warfare and repression.
As we have argued elsewhere (see this technology chart for the European "Dark Ages" of 400-1400 A.D., for example, or Lynn White's superb Medieval Technology and Social Change, 1966), a millenium of post-Roman regression in the scale of organization of human social systems had surprisingly little impact on the continuing acceleration of technological progress, the steady advance of artifice, machines, and tools that made cottage industries into factories and irreversibly eroded feudalism. Unlike human social systems, which oscillate on a pendular dynamic between phases of centralization and decentralization, technological evolutionary development has been on a much more smoothly accelerating, and increasingly self-catalyzing trajectory. Technology is moving toward its own true autonomy.
As social systems decentralized in the post-Roman era, the scale of technological innovation, transfer, and diffusion simply readjusted to the more localized social systems of the era. After the fall of Rome, no massively centralized government remained to maintain elaborate aqueducts, roads and cities. Nevertheless a number of other more scale-appropriate technologies accelerated in complexity and diversity. Even religious institutions, for all their repressive activities, were swept up in a desire for practical technological innovation. Consider the water wheel, which spread rapidly across Europe after 700 AD, one of many artifacts built by Benedictine monk-engineers. This desire for useful innovation became steadily more institutionalized, and by 1200 A.D. Franciscan monks even spread the doctrine that they were getting closer to God by practicing the "useful arts."
During the Middle Ages, the great rich-poor divides of the Roman era were also flattened back to more equitable social structures. Consider the similarity here to the way that modern globalizationmulti-local and decentralized versus previously centralized economic developmentwill rationalize First and Third World divides over the remainder of this century, lifting the developing world far more than the developed one. To many of us in the First World, afflicted with such excesses of success as obesity, endless entertainment, and a decreasing desire to sacrifice, the next few decades may seem like a time of relative "stagnation", backsliding, or incremental improvements by comparision to previous periods of dramatic advance. Our accelerating technology can easily provide more benefits than the finite and information-overloaded human mind can absorb.
But like the European Dark Ages, should such unfortunate socioeconomic events come to pass, don't let that fool you into concluding that technology is also subject to the whims of social fluctuations. It's advances march forward to an increasingly independent beat every year. To the emerging nations and the planet as a whole, the accelerating impacts and benefits of technology in coming years will be conveyed far more broadly and rapidly than the ones that gave those of us in the First World our enviably high standard of living. Accelerating change never slows down, it just moves to new substrates (e.g., technology vs. biology) and dives "under the hood," (e.g. Third World vs. First World advances), becoming less obvious from certain human perspectives.
As Bury notes, the idea of progress in the material realm was missed, amazingly, even for most of the European Renaissance (1300's to 1600's, 14th-17th century). Only by the 1650’s, near the end of this cultural explosion, did the idea of an unstoppable force of progress finally begin to emerge as a possibility to the average literate mind. This is well captured in the synthetic works of such pioneers as Jean Bodin (Method, 1566 and Colloquium, 1588) and Francis Bacon (e.g., Novum Organum, 1620, The New Atlantis, 1626), each considered forerunners of the scientific method.
The idea finally infected the public consciousness during the European Enlightenment (mid 17th and 18th centuries), beginning with such thinkers as Renes Descartes (Discourse on Method, 1637), Blaise Pascal (Pensees, 1660), and Jean-Jacques Rousseau. A century later, American philosopher-statesmen such as Thomas Jefferson, James Madison, and Benjamin Franklin would greatly strengthen and extend Enlightenment ideals and ideas in the political domain. Technological progress in particular was promoted by such scholars as Anne-R-J Turgot, Reflections on Formation and Distribution of Wealth, 1766, who noted the "inevitable" march of technological progress that had occurred even during Medieval Europe. Peripheral observations on the inexorable quickening of technology also appeared in Adam Smith's writings (1723-1790). In An Enquiry Concerning Political Justice, 1793, the transcendentalist William Godwin, possibly the earliest transhumanist philosopher, predicted that advancing knowledge and information dissemination must lead to the inevitable ascendancy of mind over matter, including a shrinking of the importance of the state relative to the individual, an eventual “total extirpation of the infirmities of our nature,” and extension of human life “beyond any limits which we are able to assign.” Curiously, he also predicted the decline of biological procreation as part of this transition, a phenomenon that has been observed in all first world countries in recent decades.
In the 19th century, the historian August Comte also considered the topic of systematic progressive change when he formulated his sociological "law of progress" in the 1830's. As the developmental progress meme spread, scholars such as Karl Marx (1818-1883), and Herbert Spencer (1820-1903) began crafting their own elaborate theories of human progress toward what they conceived as socially desired ends. This era also saw the Russian futurist and father of transhumanism, Nicolai Fyodorov (1827-1903), who theorized about the eventual perfection of human bodies and society, advocated for space and ocean colonization, and whose posthumously-published two-volume work, Philosophy of the Common Task (1906) argued that the highest common purpose of humanity, and a way reduce violence in human culture, must be scientific research on radical life extension, discovery of methods of physical immortality (presaging modern ideas of chemopreservation and uploading), and even efforts to resurrect the dead (the ultimate aims of modern history and anthropology, one might argue). While these early social progress models had shortcomings, they showed vastly greater subtlety and maturity than earlier utopian writings, such as those of Plato (427-347 BC) and Thomas More (1478-1535 AD).
All developmentalist progress models assume a trajectory, a hierarchical emergence that must unfold over many years and stages to achieve a long-term, developmentally guided end, just as in biological development. It is our position that the best of these models are what we may call evolutionary developmental, or "evo devo." They recognize that only the general outlines or broad properties of each developmental emergence can be statistically determined. The bulk of the particulars in evolutionary developmental emergences are always evolutionary, by which we mean unpredictable, creative, and locally unique events. We see this in the unique, selectionist, adapted particulars within any developing organism. Ttwo identical twins, for example, are unpredictably unique in molecular organization, tissue-architecture, fingerprints, brain wiring, etc. So it is likely to be on any two Earth-like planets in an Evo Devo Universe (Smart 2008). Mostly unique local evolutionary events, but also with a few highly-similar developmental emergences as well.
Earth's First Singularity Theorist Henry Adams
All of this set the stage for a historian by the name of Henry Adams, who in the 1890's began documenting the rapid development of science and technology at the turn of the century. It was Adams, observing the profound new forces of the dynamo, the internal combustion engine, and the railroad, who was apparently the first in the written record to explore the idea of the inevitable acceleration of progress leading to a coming global "phase change" (commendably, he even used this physical analogy) in environmental dynamics. A century later, the mathematician and science fiction author Vernor Vinge would aptly term this phase transition a "technological singularity."
Adams was the great-grandson of U.S. President John Adams (the second President of the U.S.) and grandson of John Quincy Adams (sixth President of the U.S.), so he had daunting shoes to fill. He rose to this challenge by becoming one of the most thoughtful historians of technology that we have yet had. Adams', The Education of Henry Adams, 1918, is considered one the greatest autobiographies in U.S. history.
Adams technological singularity insights first appear in rough form in "A Law of Acceleration", an essay written in 1904 in which he first surmises the existence of "A law of acceleration, definite and constant as any law of mechanics, [which] cannot be supposed to relax its energy to suit the convenience of man."
He further develops his ideas in, "A Rule of Phase Applied to History," 1909 (contained in Degradation of the Democratic Dogma, 1919), a treatise which proposes that the world may now be engaged in an inexorable acceleration toward a coming phase change in the relationship between technology and humanity, some time between 1921 and 2025. Adams used the phase change concept in the same vein as Josiah W. Gibbs (of Gibbs Free Energy and Gibbs Phase Rule fame), the brilliant but little-known Amercian physicist who described systems thermodynamics and equilibria changes in terms of energy and entropy.
In "A Rule of Phase," Adams speculates that a simple "Law of Squares" much like Newtonian inverse square laws, determines (we might assume only statistically, not absolutely) the average duration of each new phase in a developmental process of local computational complexity. He envisioned a 90,000 year Religious Phase (what we might today call the Age of Modern Humans, Jared Diamond's "Great Leap Forward" of complex linguistic and cultural innovation which began circa 100,000 years ago in Africa, and led to the behaviorally modern Cro-Magnon invasion of Europe 40,000 years ago), followed by a 300 year Mechanical Phase (e.g., Industrial Information and Computer Ages), followed by a 17 year Electrical Phase (e.g., the Symbiotic Age), followed by a 4 year Ethereal Phase (e.g., Autonomy Age), which would subsequently "bring Thought [from the human perspective] to the limit of its possibilities." Given the difficulty of timing the start of each phase, he suggested that the asymptote (the phase change singularity) might occur anywhere between 1921 and 2025. Rather amazing foresight for 1909!
Some other quotations from this important work: "A law of acceleration, definite and constant as any law of mechanics, cannot be supposed to relax its energy to suit the convenience of man." He also wryly notes that "Fifty years ago [e.g., the 1850's], science took for granted that the rate of acceleration could not last." In an interesting side note, Adams considered his conventional education to be "defective," because it did not equip him to either understand or live properly in a world of accelerating, transformational science and technology. Almost 100 years later, I'd have to agreefor the most part, U.S. education and culture remain painfully acceleration-unaware, something I hope will start to change in coming decades.
As important as Adam's insights were, they were not yet enough create a critical mass of new thinking on the subject of change. When this work came out in 1909, at the dawn of the Age of Automation, we seemed almost ready to collectively entertain the idea of intrinsically accelerating progress. But then came mechanized warfare (WW I, 1914-18), followed by large scale communist oppression (eventually totaling more than 60 million deaths). All told, politically connected deaths of 170+ million in the 20th century showed the strong limitations of human-engineered accelerating progress models, of human-centric utopias like those of More, Marx, and Spencer. Today the idea of accelerating progress remains in the cultural minority, even in the "developed" world. It is viewed with mild interest but also deep suspicion by a populace that has been traumatized by technological extremes, global divides, and economic fluctuation.
A Global Phase Change in our Understanding of Universal Change The 1930's
In my analysis to date, the most dramatic "phase change" in our collective understanding of the nature of change appears to have began in the 1930's, a transition that was directly related to the major advances in physical theory, via Albert Einstein in relativity and Niels Bohr, Werner Hiesenberg and Erwin Schroedinger in quantum mechanics, which occurred at this time. These advances, most notably the discovery of hidden continua and quantizations of matter, energy, space, and time, gave scientists worldwide a new confidence that our entire physical environment could be accurately modeled with quantitative and qualitative tools, and launched a quest for understanding and unification which has burned brighter, faster, and stronger every year since.
We must also acknowledge a debt to speculative fiction for developing the collective vision of the scientific, technical, and lay communities beginning around this time. Perhaps the first popular confrontation of the technological singularity occurred in a science fiction short story by the legendary John W. Campbell, The Last Evolution (in Amazing Stories, August 1932, and The Best of John W. Campbell, 1976). Here Campbell contemplates the ever-accelerating implications of computers gaining the ability to design even more powerful copies of themselves. Often directly motivated by Campbell and his successors, a long chain of scholars from different disciplines have confronted this fascinating idea in subsequent years. [But unfortunately, as Judith Berman notes ("Science Fiction Without the Future," 2001), modern science fiction increasingly avoids the growing challenge of thinking about accelerating technological change.]
During the 1930's, the pioneers of modern information and computer science also began asking a series of questions that are fundamentally related to the issue of accelerating change. Is computation a framework that can describe all physical interactions? Can we show that computing done on one physical system is equivalent to that done on another? Are there questions a computing system can ask that can never be answered by it? And in the boldest question, is there evidence that the universe itself is a unitary system for computation? If so, we can examine a variety of separate processes of physical acceleration, so called "exponential and asymptotic domains of physics" such as those that produce black holes (cosmological singularities), and for the first time, understand them all as related forms of universal computation.
In 1931, one of the most important mathematical insights to date was made by Kurt Gödel, in the form of the Incompleteness Theorem. This concept allowed us to understand that every finite computational, formal logical system must have areas of intrinsic uncertainty, undecidable questions that can neither be proven nor disproven from within the system. Presumably, the persistence of these uncertainties and their interaction with the physical environment spurs the creation of increasingly more sophisticated physical-computational systems over time, in a process of hierarchical emergence over universal history.
The idea of a universal symbolism has very long and rich history in philosophy and logic. It was perhaps first clearly advocated by Gottfried Wilhelm Liebniz, co-inventor of the calculus, back in the 1680's (for more, see the excellent History of Philosophy: Descartes to Liebniz, 1976). But a key successor idea, the hypothesis that all physical systems can be described in comparable computational terms, even when done by very different entities (molecules, human minds, or technological machines) took the "phase change" of the 1930's physical and logical insights before it could arrive.
It was first formalized by Alonzo Church, Alan Turing, and Emil Post in a series of working hypotheses beginning in 1936. They proposed the existence of a type of "universal computation", through logically equivalent "finite state machines" (later called Turing machines), and this formalization came to be known as the Church-Turing thesis in computer science. The C-T thesis was a major breakthrough in our suspicion that computation underlies and unifies physical reality.
In 1938, Harvard poet and polymath R. Buckminster Fuller published Nine Chains to the Moon, a creative rant on the nature of world systems. Chapter 38, "Ephemeralization," posited that in nature, "all progressions are from material to abstract" (ephemeralization, a form of digitization), and "every one of the ephemeralization trends.. eventually hits the electrical stage" such that even "efficiency (doing more with less) ephemeralizes". This idea would eventually mature into a leading candidate for a physical driver of sustained acceleration in technological systems. In 1939, an obscure Physical Review paper by J. Robert Oppenheimer described how a star might theoretically collapse into an object so dense not even light could escape its gravitational clutchesthe first modern discussion of the accelerating physical phenomena that were later to be called black holes.
The Roots of Applied Computer Science and Military "Big Iron" Computing The 1940's
In the 1940's, during a very dramatic chapter of human history, the necessities of WWII gave rise to our founders of modern computing such as Konrad Zuse, Alan Turing, and John Von Neumann. Building on the prehistory of mechanical computing, as first explored by such pioneers as Charles Babbage and his "Difference Engine" (1871) these individuals were the first to create truly sophisticated and functional large scale digital computational devices. Not unexpectedly, they were also early explorers of both artificial intelligence and accelerating technological change as topics of serious inquiry.
In particular, both Turing (see The Universal Computer, 2000), and Von Neumann (John Von Neumann and the Origins of Modern Computing, 1990) were important early explorers of the idea of ongoing computational acceleration, and of the computational nature of both human and universal intelligence. It appears that Von Neumann may have been the first, some time in the late 1940's or early 1950's, to use the mathematical-physical term "singularity" to describe his vision of a coming "runaway" progression in computational events (see Vernor Vinge's "The Coming Technological Singularity").
At the same time, the prospect of humans designing friendly, non-monstrous artificial intelligences was first seriously explored by such science fiction greats as Isaac Asimov, and stories utilizing his "Laws of Robotics" (co-formulated with John W. Campbell) began appearing in 1940. In 1947, at the same time John Von Neumann was having his singularity insights, the French researcher François Meyer wrote "L'acceleration evolutive. Essai sur le rythme evolutif et son interpretation quantique". This paper may have been the first simple log-periodic acceleration model applied to evolutionary history. If so, Meyer's model should include a finite time singularity at some particular future date, but I have not been able to access and translate the original as of yet. (See also Meyer's 1954 essay, "Problematique de l'evolution."). Also in 1948, French historian Daniel Halévy published "Essai sur l'acceleration de l'histoire (Essay on the Acceleration of History) which has been called "penetrating" by some social historians (Alain Silvera, 1966).
Also in 1948, Claude Shannon wrote A Mathematical Theory of Communication, thereby creating a new field, information theory, and a new model of the world in which the currency of all computers and communication channels could be quantitatively unified under the idea of binary digits, or bits. Thereafter, the double exponential growth in speed, capacity, and power that occurs in these unique physical systems would become readily apparent to all who would choose to measure it.
Early Digital Physics and the First Commercial Big Iron The 1950's
In the 1950's, Norbert Weiner (beginning with Cybernetics, 1948), Von Neumann, and a number other pioneers of electronic computing began to apply the theorems of computer science to all the laws and systems of the natural world, sometimes poorly, sometimes indiscriminately, but occasionally with real success. One still unknown innovator of these ideas at that time was Ed Fredkin (and Tommaso Toffoli somewhat later), who introduced and greatly refined the idea of treating the physical universe as a unitary computing system, a paradigm which came to be known as digital physics in subsequent decades. A history of this profoundly important and still little-discussed development, including the concept of the universe as a type of cellular automaton, can be found at Fredkin's excellent digital philosophy website.
The idea of universal computing also achieved perhaps its first major popularization at this time, from Isaac Asimov in a famous 1956 short story, The Last Question. In this story, the universe is proposed to be a self-organizing computational system that ultimately defeats entropy by recreating itself in a recursive manner. This speculative idea, universal rebirth as the only way around "the entropy problem," represents the most credible solution to the question of the long term survival of universal intelligence that has yet been proposed. The influence of this little story an its successors on the Western scientific heritage has grown steadily since, and is often underestimated.
Another major information processing modelling success of this decade was Frank Rosenblatt's powerful perceptron model of the human neuron. First published in 1957, it spurred the early development of neural networks. This same year, the launch of Sputnik propelled the US-Soviet cold war in the direction of a sustained thirty year space race, one that peaked in intensity with the first moon landing in 1969. The post-Sputnik generation saw the last broadly-sustained promotion of science and technology education and application in our great nation, which has grown far too complacent since. [People do their best with a coach and a few common, well-defined goals. I have proposed elsewhere that the development of cheap, ubiquitous broadband connectivity and a conversational user interface, both locally and globally, should be today's primary technology goal, the galvanizing U.S. "Moon Shot" of the early 21st century.]
Ongoing informal discussions by visionaries in the scientific community at this time led to such contributions as Richard Feynman's ("There's Plenty of Room at the Bottom" 1959), a revolutionary article that implicitly assumed accelerating change would continue in a very important new domain, miniaturization.
The Great Goal of Science and Technology and The IBM System/360 The 1960's
The 1960's were driven by a can-do, space race optimism. John F. Kennedy's rousing speech in 1961, met with tremendous public support for this very costly venture, and sci-tech development became our national Great Goal. This engendered a new popularization of futures thinking, and new understanding of accelerating change in the scientific and engineering communities.
In 1964, Gordon Moore, co-founder of Intel, observing manufacturing trends in the new medium of integrated circuits, made his now famous observation that circuit densities were doubling every 18 to 24 months, and would continue to do so for the forseeable future. At first, Moore's law simply defined the economic and engineering environment specific to computer hardware development, and gradually an entire industry of scientists and engineers came to not only intellectually appreciate, but to tangibly experience the local effects of continuous accelerating change. Throughout the 1960's and 1970's a growing number of insightful analysts, engineers, systems theorists and cyberneticians developed models of reality which began to incorporate trend curves of accelerating computational change, and these strong-growth-assuming models made their way into various industry and generalist publications, most prominently, Scientific American. Also in 1964, the first affordable integrated computer system emerged, IBM's radically innovative System/360.
Following the excitement of the New York City World's Fair (1964-1965) and it's many optimistic, future-oriented exhibits (e.g., Bell Telephone's PicturePhone, G.M.'s Futurama: Cities of the Future), and riding the rising enthusiasm of the pre-Apollo era, serious futurist thinking beyond the science fiction authors and academics first came into its own. Yet both of these groups still had much to contribute. Of particular note, Stanislaw Lem, published (in Polish only, unfortunately) his nonfiction human-machine convergence masterwork Summa Technologiae, 1964, and I.J. Good ("Speculations Concerning the First Ultraintelligent Machine," 1965), published in a professional publication (Advances in Computers) what was perhaps the first clear conceptualization of the coming technological singularity, moving the topic another step closer to legitimacy. In 1966, the great historian Alfred Toynbee, wrote a chapter, "Acceleration in Human History," in Change and Habit (1966) which, like Henry Adams, charts a series of rapidly accelerating phases of biological and, with the emergence of humans "1Mya," technological acceleration in Earth's history. Toynbee charted the same pattern to a lesser degree in religious development, while noting it's ambiguity in political forms, and its apparent absence in such domains as art. He also briefly speculates that human consciousnes and culture form the start of a "metabiological" phase of planetary development.
Leading the popularization of futures scanning, scenario, and trend analysis in this fertile time was Edward Cornish's World Future Society, and the bi-monthly publication of The Futurist magazine, beginning in 1967, as well as Olaf Helmer's Institute for the Future, a research-oriented think tank started in 1968. By the close of the 1960's no one in computer science really knew how long curves of technological acceleration might last (excepting a very limited number of visionaries like Von Neumann and Lem, as mentioned), or how relevant they might be to the larger world of human affairs. But the gate was open, and the concept of the singularity was finally a viable inquiry (though one with professional risk to discuss too openly or forcefully, particularly in conservative academic circles).
It was also during this decade that the physicist John A. Wheeler coined the term "black hole" to describe the final accelerating developmental stage of sufficiently massive stars, still-theoretical entities whose subsequent structure would become impenetrable from our universal vantage point.
Increasing Social Awareness of Acceleration and the Minicomputer The 1970's
In the 1970's, the idea of accelerating change as a permanent feature of modern life entered broadly into the public consciousness with Alvin Toffler and his revolutionary Future Shock, 1970. The first and second chapters of Shock, "The 800th Lifetime" and "The Accelerative Thrust," remain as engaging as they day they were written, an eloquent overview of the profound speedup that technology has brought to modern life. Shortly afterward, Scientific American editor and polymath Gerard Piel wrote a less well-known but equally prescient work, The Acceleration of History, 1972, which considered the ongoing acceleration of science and techology, and the many social implications of continued exponential growth. Throughout this period the implications of Moore's law as a signifier of accelerating computationally-driven scientific and social change became increasingly understood by systems thinkers.
Also beginning in 1970, F.M. Esfandiary (later, FM-2030) published a series of short, accessible, and passionate works of technology optimism that assumed and began to explore the integrative, efficient, and self-balancing features apparently emerging within technological systems. His three major works, Optimism One, 1970; Up-Wingers: A Futurist Manifesto, 1973 and Tele-Spheres, 1977 were eventually republished in accessible paperback form in 1977 and 1978, becoming known as the "Transhumanist Trilogy," and setting the stage for the Extropian and early transhumanist social movements of the 1980's.
At this time Carl Sagan published The Dragons of Eden, 1977, a seminal book that brought the "Cosmic Calendar" metaphor to mass public attention, dramatically demonstrating that important emergences of universal complexity appear to have occurred in an ever-accelerating manner through at least the last six billion years of universal developmental history.
Where did the Cosmic Calendar metaphor originate? We are not yet clear on this point. My colleage Ted Kaehler reports that when he took ninth grade biology from "Mr. Peterson" in Palo Alto, CA in 1965, he used the calendar analogy. He recalls that one calendar year was mapped to the age of the Earth, and students were consistently amazed at how recent homo sapiens was. He suggests checking high school geology and biology texts from the 1960's to discover where the calendar metaphor was first used. What is clear is that after Dragons of Eden (1977), and Sagan's Cosmos television series (1980) a very curious phenomenon, our apparently universal record of accelerating development, had finally entered the mass consciousness.
The roboticist Hans Moravec also emerged on the public scene in this decade. Moravec is arguably the most important single pioneer and advocate of deep thinking on accelerating computational change in the 20th century. He began writing about exponential growth in computer power beginning in the mid 1970's while working at the Stanford Artificial Intelligence Laboratory (SAIL). His fascinating June 1974 essay "Locomotion, Vision, and Intelligence" led to a 1976 essay, "The Role of Raw Power in Intelligence" and an even bolder piece, Sept 1977's "Intelligent Machines: How to get there from here and What to do afterwards" (incorporating 1974's "Locomotion"). These essays were widely xeroxed and commented on in the SAIL, MIT-AI, and Carnegie Mellon University-AI communities at this time, as well as migrating to the computer science departments at other major universities, such as the University of California. If I recall correctly, a graph of Moravec's accelerating computer power curves even made it into Ted Nelson's visionary computer futures work, Computer Lib / Dream Machines, 1975.
In Feb 1979, after two years of editorial delay, Moravec's ideas finally reached the general public through Analog: Science Fiction and Fact, in an essay titled "Today's Computers, Intelligent Machines, and Our Future." The last section of this groundbreaking essay "considers the implications of the emergence of intelligent machines, and concludes that they are the final step in a revolution in the nature of life. Classical evolution based on DNA, random mutations and natural selection may be completely replaced by the much faster process of intelligence mediated cultural and technological evolution." Considering the future of computer-human coevolution, Moravec concludes we are rapidly headed for a post-biological form for all local, living intelligence: "In the long run the sheer physical inability of humans to keep up with these rapidly evolving progeny of our minds will ensure that the ratio of people to machines approaches zero, and that a direct descendant of our culture, but not our genes, inherits the universe."
This may turn out to be a fundamental insight into the future. At the same time, it is one so simple and elegant that I also arrived at it privately in 1972, as a young high school student contemplating the nature and purpose of human existence. I'm sure many others did as well.
'Cambrian Explosion' of Complexity Science and The Personal Computer The 1980's
The 1980's saw a modern version of the "Cambrian Explosion" in our scientific understanding of accelerating computational change within various specialties, with simultaneous breakthroughs in the study of complexity, artificial life, neural networks, connectionist and parallel computation, and several other fields best left for a "Longer History." Sagan's enormously successful public television series Cosmos, 1980 also introduced the Cosmic Calendar to an even wider audience, and futurists such as John Platt ("The Acceleration of Evolution," The Futurist, 1981) and others chipped in with their take on the central acceleration.
In 1981, in Critical Path, Buckminster Fuller published an expanded version of his concept of ephemeralization, the apparent driver of accelerating change, "the invisible chemical, metallurgical, and electronic production of ever-more-efficient and satisfyingly effective performance with the investment of ever-less weight and volume of materials per unit function formed or performed". In Synergetics 2, 1979, he defined ephemeralization as "the principle of doing ever more with ever less weight, time and energy per each given level of functional performance"(italics mine). Such statements provide what may be among the first published descriptions of what we may call the STEM efficiency/density or generically, "compression" of computation, independently derived by myself in the 1970's, and first published at this website in 1999. STEM efficiency is the idea that the leading edge of local complex systems always discover how to do more computing with less space, time, energy, and matter per salient computation (however measured), due to special preexisting universal structure. They naturally run down a STEM efficiency and/or STEM density gradient, built into the physics of the macrocosm, microcosm, nanocosm, and eventually, femtocosm. The idea of STEM compression / ephemeralization seems the current leading candidate for the central driver of the accelerating technological change we see today.
In 1982, Patrick Gunkel, philosopher of "ideonomy" (a nascent attempt at a philology for the human 'meme space'), coined the word "metamachines" to refer to "mechanical minds or beings whose intellectual, psychic, behavioral, or ontic level resembles or even surpasses man's." At this time he also attempted a broad categorization of possible transitions to greater-than-human machine intelligence but did not complete the manuscript.
In January 1983, revisiting and extending Von Neumann's insights, an up-and-coming science fiction writer by the name of Vernor Vinge, writing in the First Word column of Omni Magazine, introduced the idea that the ever-accelerating evolution of computer intelligence itself might soon produce a kind of technological singularity, "an intellectual transition as impenetrable as the knotted space-time at the center of a black hole, and the world will pass far beyond our understanding." Vinge concluded in this prescient piece, as I did independently in 1972, that inevitable technological singularities in intelligent civilizations represented the most logical explanation for the "vast silence" of the cosmos. This silence is commonly known as Fermi's Paradox, after Enrico Fermi, who first popularized it in 1950.
By the mid 80's, consideration of accelerating change from a systems perspective finally became broadly accessible to the general public, through groundbreaking popular works by such authors as Marvin Minsky (Society of Mind, 1985), Erik Drexler (Engines of Creation, 1986), and Hans Moravec (Mind Children, 1988). These three books respectively represented a theory of mind as an emergent collective computational system, a framework for applying computation and embodied environmental interaction on as-yet-undreamed scales of miniaturization, and the first coherent projection of the meaning of robotics and the new computer and cognitive sciences for the future of mind in the larger universe. These early authors risked professional reputation in their own fields to advance their unique ideas, and each deserves special recognition for their courage, conviction, and clarity of vision.
During this same decade Stephen Wolfram, a key thinker in digital physics, and one of the creators of the Mathematica modelling software, became a leading investigator into the life- and physics-simulating properties of cellular automata (CA). His recent synthetic work, A New Kind of Science, 2002, presents further evidence of the value of the CA paradigm, and is well worth investigating. In 1987, Pierre Grou, a pioneering French economist and systems theorist, noted in L' Aventure Economique that economic evolutionary development since the neolithic period can be described in terms of a hierarchical emergence of "dominating economies" (e.g., Egypt, Greece, Rome, Byzantium, Southern Europe, Netherlands, Great Britain, United States, China?), all driven forward in an accelerating crisis/no-crisis pattern. He would develop this fascinating insight into more formal singularity models in later years.
This productive decade also culminated with a provocative presentation in 1989 by John A. Wheeler (of "black hole" fame) entitled "It From Bit." In this famous speech Wheeler advanced the idea that every "it" (every particle, every field of force, even the fabric of the spacetime continuum) derives its function, meaning, and existence from a series of binary choices, or "bits" available to it. Physical reality, in this perspective, arises from the posing of countless numbers of these binary, yes/no questions among the its, as they construct their interrelationships in an inexorable computational process. As another refinement, evolutionary developmental ("evo-devo") biologists (e.g., Rudolf Raff, The Shape of Life, 1996), began to carefully explore the interaction between evolution and development in biological systems over long spans of time. Evo-devo biologists made it possible for systems theorists to begin to consider that complexity emergence may proceed via both evolutionary and developmental processes over universal history, and to admit that we are only beginning to understand the fundamental difference between the two.
As yet, few scholars in the domain of digital physics have considered the developmental implications of our planet's accelerating history of local computing (e.g., Moore's law, Vinge's Technological Singularity, Kurzweil's law of Accelerating Returns, Smart's law of Locally Asymptotic Computation) when considered within both a universal and multiversal framework. I have made such a proposal in my developmental singularity hypothesis, and I expect significant new work in this area will eventually be forthcoming.
Early Singularity-Awareness and The GUI-based Internet The 1990's
In the 1990's, a flood of publications that were implicitly singularity-aware, and a bold few that were explicitly singularity-aware, such as those by John Brockman (ed., The Third Culture, 1995), Damien Broderick (The Spike, 1997) and Richard Coren (The Evolutionary Trajectory, 1998) became available to the general public. Broderick's work is noteworthy in that it is quite accessible, and the first lay publication on the topic of the singularity. Coren's text is pioneering, and quite useful in this early era of singularity studies, in that he applies a logistic model to events in cosmology, biology, civilization, and sociology. His analysis proposes a global phase change singularity, an "instantaneous" rate of meaningful physical change (at least as seen from our slowspace perspective) circa 2140 ± 10 years (2130-2150).
What must happen after this point? Coren cites the famous systems theorist Derek De Solla Price, who suggests a "Sorites Paradox" solution. In the Sorites Paradox, the Greeks noted that often in life, an increasing quantity or capacity of some object eventually enables a qualitative change in the nature of the object. The engineer and futurist Brad Holtz states something similar, as a general rule of thumb: "Three orders of magnitude allows a paradigm shift."
When do you have enough grains of sand to call a sandpile? Enough fax machines to form a fax network? Enough local computation for autonomous machine intelligence to emerge? At the phase change, the process being examined changes its nature in such a way that the quantity being measured no longer captures some of the key qualities or dynamics of the emergent entity. A singularity has occurred.
Also of note in the 1990's, the transhumanist philosopher Max More, the artist Natasha Vita-More (Extropy.org), the economist Robin Hansen, the transhumanist philosopher Nick Bostrom (Transhumanism.org), again Hans Moravec (homepage) and most centrally, mathematician and science fiction author Vernor Vinge, in his seminal "The Coming Technological Singularity," 1993, brought many of these ideas to serious critical attention with their excellent discussions of the singularity meme and its variants using the mass medium of the new millennium, the Web.
But perhaps most visibly, and most importantly for the public consideration of these ideas, inventor and artificial intelligence pioneer Ray Kurzweil published two seminal books in this decade, Age of Intelligent Machines, in 1990, and Age of Spiritual Machines, in 1999. The former book makes a case for the astounding growth in computational complexity in recent decades, and the reality that A.I. must play a central role in 21st century life. The latter book takes these ideas much further, and is our recommended introductory primer to singularity ideas for the lay public.
In this same decade, Eliezer Yudkowsky also developed prolific writings in the technological singularity, a community of A.I. investigators and "singularity advocates", and a nonprofit A.I. venture with transhumanists Brian and Sabine Atkins. Finally, in 1999 I started Acceleration Watch.com as a generalist website to educate lay thinkers on issues of accelerating computational and technological change.
Small Steps Toward Singularity Studies and a CI-based PlanetNet The 2000's
In 2000, Laurent Nottale (an astrophysicist), Jean Chaline (a paleontologist), and Pierre Grou (an economist) published an admirably interdisciplinary paper, "On the Fractal Structure of Evolutionary Trees," which applies log-periodic analysis to the main crises of evolutionary civilizations. They followed this up with a groundbreaking book, Les Arbres de l'Evolution (Trees of Evolution), 2000, which models universal, life, and economic development all on a fractal, log-periodic acceleration. This book has several similarities to Richard Coren's, but makes a case for a more rapid acceleration on a planetary scale. Grou and the others note that hierarchical events emerge in an accelerating alternation of crisis and plateau, punctuation and equilibrium. Their acceleration model reaches a macro-scale singularity, a global time critical, at 2080 ± 30 years (2050-2110). The trio continue to publish (note this 2002 essay) on their fractal model for acceleration, and with luck their ideas will gain wider critical consideration in coming years.
We can note here that the crises (whether galactic supernova, extinction event, or economic crash) they use as chart points are locally debilitating but they never slow down the log-periodic acceleration of the network. What I call the average distributed complexity (intelligence, interdependence, and immunity) of the network always accelerates, quite smoothly when viewed from the perspective of the network as a whole. Catastrophes are locally disruptive but nonlocally educational and catalytic. This is a heartening and enlightening realization.
In 2001, complex systems scholar Didier Sornette and physicist Anders Johansen published a paper, "Significance of log-periodic precursors to financial crashes." They noted that hierarchical emergence to new regimes often involve an accelerating approach to a finite-time singularity, followed by a phase transition, which may or may not be locally "catastrophic," as in a financial crash. They started to believe that this pattern could be used to predict some stock market crashes months before they actually happen. This led Sornette to publish a fascinating work, Why Stock Markets Crash, 2003. This book gives a tour of the theory of critical phenomena, and then applys a log-periodic model to historical economic crashes. Sornette and Johansen's model predicts a critical time for global phase change at 2050 ± 10 years (2040-2060), and they offer three scenarios for the meaning of this change: 1) economic collapse, 2) a transition to economic sustainability, or most interestingly, 3) superhumanity.
I find the first scenario unsupportable from systems theory (e.g., a global singularity could be locally catastrophic, but would be catalytic to the global network). The second scenario is best seen as a subset of the third (intelligent technology, far more resource efficient and inner-space oriented with each new generation, can be expected to be an exemplar of immobile sustainability from the human perspective). The third scenario seems to be the essence of the present dynamic, from my perspective. Successively accelerating waves of automation cause the economic shocks we experience. If the model is correct, these wave fronts are likely to arrive at increasingly rapid rates and yet be progressively briefer and more locally confined in their "catastrophic effect," while being globally catalytic in the process.
Sornette, Anderson, and colleagues have developed models for anticipating both local economic crashes, and for interpreting the log-periodic data of the entire economic history of civilization. While their macrohistorical analysis might be quite insightful, their attempt to decipher the local signatures of finite-time singularities in either individual or global economic systems may or may not turn out to be predictively valid (they predicted a 1990's Nikkei crash in advance, but have few data points at present).
If the dynamics of hierarchical emergence are fractal, and we see finite-time singularities leading to phase transitions everywhere, this still doesn't guarantee us that we can see local patterns as easily as we can see the Big Picture ones. Indeed, I'd argue the opposite. In an evolutionary developmental universe, the vast majority of local change appears to be driven by deterministic chaos, by strange attractors, and the noise of pseudorandomness might easily obliterate signal on a local level. My intuition tells me a predictive model could work very well at the network level, but would be much less reliable describing dynamics of individual nodes. We shall see.
At present, in 2005, only a few tens of thousands of individuals have been exposed to the singularity meme, mostly through the web. But soon, the general concept of the singularity is likley to become at least a bit more broadly known, though still not yet mainstream.
It is periodically confronted by networking think tanks such as the nanotechnology and future scenario leader Foresight Institute and has been increasingly explored by great journalists, inventor visionaries, and social theorists such as Ed Regis, (Mambo Chicken, 1990; Nano, 1995), Douglas Rushkoff (Cyberia, 1994), Danny Hillis, "Close to the Singularity" 1995, Peter Russell (Waking up in Time, 1998), James Glieck (Faster: The Acceleration of Just About Everything, 1999), and again, Hans Moravec (Mind Children, 1988; Robot, 1999). The French historian Pierre Nora is also writing eloquently about the "acceleration of history," echoing his colleague Francois Meyer's work a half-century earlier.
Promisingly, Ray Kurzweil has published an excellent precis "The Law of Accelerating Returns," (2001) of his forthcoming book The Singularity is Near, 2005. In 2000 he also created a prominent website showcasing these future-relevant ideas, KurzweilAI, headed by transhumanist editor Amara Angelica. Kurzweil's next work will go a long way toward making the general, professional, and scientific communities aware of the implications and inevitability of continuous accelerating change, and is the latest in a long tradition of careful popular analysis of these topics.
The physicist Seth Lloyd, in two bold, insightful, and well publicized digital physics papers ("Computational capacity of the universe," Physical Review Letters, 2002, and "Ultimate physical limits to computation," Nature, 2000, has also advanced simple and powerful conceptual models of the intrinsic limits of universal computation. Such models may yield significant singularity-relevant insights into the developmental future of our universe, as I argue in my own forthcoming work.
Additional important authors from this rich history of increasing awareness of accelerating computational change may be seen on the Speculative Topics page of this site. Happy reading!
The New Accelerators: Computation, Engineering, and Science
Recalling the initial A.I. optimism of the 1960's, both conservative and many embarassingly ambitious [1, 2, 3, 4, 5] grass roots, boutique, and startup A.I. projects are again flourishing in software and hardware. But this time around there are countless effective proprietary implementations as well (ie, HNC's neural networks, Xilinx's FPGAs, Chameleon's RCPs or Google's use of Beowulf clusters) which have solved important human problems and made fortunes for their inventors. Nanotechnology (both computational and general) makes its own bold strides, through pioneers like Jim Von Ehr at Zyvex , as well as broadly within MEMS, microphotonics, microfluidics, materials science, solid state physics, and many other domains. The autocatalytic loop is well underway, following myriad pathways toward greater systems complexity.
When you're ready to investigate the field of A.I., consider supporting the following three organizations. First, for all of computer science, the Association for Computing Machinery (ACM). You may join at $38 or $185/year (student/regular rates), to subsidize the computer science field and gain full access to their extensive "Digital Library". A lite version of this resource is available for free. Second, more specific to A.I., explore the American Association for Artificial Intelligence (AAAI), and their informative "A.I. Topics" resource. Consider joining at $50/year to help their important educational work. Third, to encourage the development of evolutionary computation, arguably the most long term promising of many important areas under the A.I. umbrella, consider joining the International Society for Genetic and Evolutionary Computation (ISGEC) at $50 or $120/year (student/regular rate).
To get an overview of topics and players in A.I., Stuart Russell has a nice introductory portal, A.I. on the Web. At some point, if you plan to contribute to the development of this field, you will need to decide which of several useful intelligence-building strategies seem most important and relevant to your own personal path.
From my own generalist's perspective, there are three strategies that seem particularly important: 1. Evolutionary computation (evolvable hardware and software) and knowledge representation (classic A.I.), 2. The cognitive and neurosciences (decoding the essential structure of the human brain using molecular, cellular, organismic, and social sciences), and 3. Electrical engineering and general I.T. (processor, computer, network, and systems design and manufacture). I currently think they will be necessary to the task in that priority order, and consider these the three most interesting and robust current research paradigms where continued breakthroughs are likely to catalyze the development of sophisticated A.I. I may be misconstruing or leaving something out entirely here, of coursethat's part of the excitement of the quest, for the full story of emergent A.I. has yet to unfold.
To begin your survey of the leading edge, consider perusing general and academic portals on genetic and evolutionary computation, on evolvable hardware, on neural nets, and computational neuroscience [1, 2]. The annual Genetic and Evolutionary Computation Conference (GECCO) is a showcase for this newly emerging paradigm, and a must-attend event for those looking to understand accelerating trends in computational autonomy. This is only a beginningthere are other valuable A.I. conferences to explore, far too many to list here.
Have fun and change the world!
Apologies to Those Who Have Been Overlooked
Many, many other institutions, publications, technological systems, and individuals could also have been mentioned in this "Brief History"this essay showcases only the more historically relevant or more singularity-aware individuals who are helping us to gain a conscious understanding of the accelerating manner in which we are co-creating our complex future. If any reader thinks I've made mistakes, misrepresentations, or omissions, please let me know. Most of the credit, as always, should go the vast majority of unsung individuals and institutions who work daily to improve, speed up and balance the myriad technological systems of our lives, perhaps without conscious recognition that they also play a role in some deep universal developmental process.
The Near Future: Gradually Increasing Social Awareness of the Singularity
Even though we futurists observe continously accelerating (and ever more "self-catalyzed") computational and technological change, we should not expect a continuously accelerating recognition of this within human society. On the contrary, human social systems in recent decades have shown evidence of "saturation," or a definite maximization in their rate of comfortable absorption of new products and ideas, leading us to such healthy countertrends as voluntary simplicity. Over the next decade we can expect increasing articles on the singularity in popular magazines, and as this meme becomes familiar to the general public, many more groups will weigh in with their own unique perspective on the universal reasons for and likely future directions of inexorable accelerating change.
But it is all too probable that most of humanity will contine to pervasively ignore this phenomenon until it finally becomes "surprisingly apparent," some time between 2020 and 2060, depending on your analysis. Such seems to be human nature, so our greatest challenge for the present may be an educational one. As Glen Hiemstra of Futurist.com likes to say, once we understand a paradigm, can start talking about the kind of preferred futures we would like to see, but the broad understanding and acknowledgement must come first. Our choice of the path we take now, in the context of accelerating computational change, is always the issue with greatest current relevance.
As acceleration watcher Robert Trask eloquently stated: "It is tough to try to explain to someone what all this is about without coming across as a fanatical moron. Oh well... [we] just need to learn more, think more, and get better at it. I look forward to your newsletter in the hope that I can better grasp this beautiful yet somehow horrific inevitability."
It is our still unrealized hope that we may grow to understand, to simply and intelligently explain, and to scientifically demonstrate that the accelerating transition to computational complexity occurring in the technological infrastructure all around us, while we humans often strive for greater simplicity and peace, on our own scale, is neither horrifying or dehumanizing, but actually quite natural, balanced, self-protecting, and integral to the inherent design of the universe.
The universe appears to be a system that produces complex, self-examining subsystems which develop exponentially greater local 'modeling intelligence' over time. Occasional catastrophes are an integral part of this creative process, but they act as selectional and catalytic events at all levels, apparently never threatening the informational content of the great bulk of extant systems at any substrate scale. Furthermore, the more computationally complex any local systems become, the more limited and constrained, and the less subjectively violent the catastrophies they appear to endure.
Such observations evoke real optimism for the future, and at the same time they inform our current personal and social choices. The insights we gain in studying the evolutionary development of complex systems, at all substrate levels, may deeply advise us of better vs. worse paths we may take on a daily basis, as we move ever closer to the apparently unavoidable universal attractor of human-surpassing technological intelligence within the forseeable future.
For your part, you can promote the development of an evolutionary developmental understanding of accelerating change whenever it seems appropriate. For more, see Accelerating.org. If you've got a few spare dollars (a typical annual donation to a small nonprofit is $25), we welcome you to contribute to our nonprofit with your tax-deductible donations. Thanks for reading.