The history of life's development on Earth has apparently been an increasingly faster emergence of computational complexity (or modeling intelligence') within a special subset of locally emergent forms. Curiously, these new forms are often much more resource efficient (per physical or computational output), denser, or miniaturized, so that they continually avoid resource limits to their accelerating growth. Historians have long noted that significant cultural advances (neolithic tool kits, architectures, language, civil society, law, science) emerge at an accelerating rate in human history. Many scholars (Jared Diamond, James Burke, Robert Wright) consider such factors as increasing population density, technological diffusion, and communication rates to be key drivers of these sociotechnological transformations.

Over the last millennium, rates of planetary technological innovation and diffusion have broadly accelerated as a whole, with ever-briefer pauses between new phases of acceleration. This accelerating trend in what may be called the "average distributed complexity" of our socio-technical systems has been apparent even as wars, local catastrophes, and revolutions have caused discontinuities within specific civilizations. In other words, while catastrophes continually occur in specific cases, some type of general immunity, resiliency, or social learning continually emerges in our most successful physical systems (wther civilizations, economies, cultures, or technologies) on a distributed and redundant basis. Like the human immune system, we are beginning to discover that all computational networks encode their own forms of immunity, keeping them on an accelerating growth curve for long spans of time.

As perhaps the most dramatic example of global acceleration, recent data show that our modern computer technology, when considered as one broadly distributed planetary system or "substrate," has been smoothly and continuously doubling in average complexity for the entire twentieth century. Ray Kurzweil's data propose that performance/price ratios in purchasable computing systems were originally doubling every three years in our 1890 mechanical computing systems, and are now doubling every 12-18 months.

Most curiously, this acceleration has been highly immune to the fortunes and catastrophies of individual technology companies – even of major social, political, or economic crises, such as World Wars, the Great Depression, and our current recession. It has been maintained through at least five dramatically different computer engineering and manufacturing paradigms: mechanical, relay, vacuum tube, transistor, and integrated circuit information processing machines.

Today we are creating a panoply of successively more miniaturized and resource-efficient computing architectures, several of which are growing measurably more autonomous (evolutionary, biologically-inspired, self-directing, self-monitoring, self-provisioning, self-repairing, and partially self-replicating) with each new computer generation. An impressive array of new commercial applications for these semi-autonomous systems (e.g., Google's cluster architecture, electronic design automation software, reverse compilers, self-diagnosing and semi-autonomic systems, pattern recognizing neural networks and genetic algorithms, innovative machine learning paradigms such as support vector machines) have further increased our breathtaking pace of technological change.

Where does this continual acceleration phenomenon come from, where is it going, and what does it mean for the near future of humanity? AC2004 is the place where today's leading thinkers explore science, technology, business, and humanist dialogs in accelerating change.

Come join us in Palo Alto this November as we investigate some of the most fascinating and important issues of the modern era.

For Further Study

One of ASF's long-term goals is to encourage the development of multidisciplinary educational programs exploring the drivers of accelerating change at the graduate level. We see this as helpful within two broad domains: 1. Acceleration Studies (a semi-quantitative, predictive, policy, and applications-oriented program of study) and 2. Evolutionary Development and Phase Transition Studies (a technical, theory-oriented research program).

1. Acceleration Studies includes such subjects as forecasting, systems theory, science and technology studies and roadmapping (infotech, physics, nanotech, biotech, neuro and cognitive science), technology assessment and policy, history of science and technology, cybernetics, sociology and economics, information science, productivity metrics, engineering and operations research, future studies, and forecasting, including trend extrapolation and analysis. This program would focus on the benefits, choices, and risks of a range of accelerating systems of change, and would necessarily also consider the emerging sociopolitical and ethical issues of our apparently imminent transition to machine intelligence. Today's technology policy graduate programs offer a weak start toward this kind of curriculum, but have a long way to go before they become broadly acceleration-aware.

2. Evolutionary Development and Phase Transition Studies includes complex systems research, singularity theory, nonlinear mathematics, evolutionary developmental biology, systems and astrobiology, physics and astrophysics, theory of computation, information and autonomy theory, philosopy of science and technology, and other disciplines relevant to modeling accelerating physical domains of change. This program would focus on dynamical models of change in complex systems, including the universe as a complex system, and would necessarily also consider philosophical and teleological issues of the meaning and purpose of universal change in relation to current scientific theory. Again, today's complex systems graduate programs provide a tentative start toward this kind of curriculum, but have many shortcomings with regard to broadly modeling accelerating change.

We believe a broad understanding of science and technology is vital to understanding and modeling accelerating change, and should be a necessary prerequisite to graduating methodologically-sound technology consultants and "futurists," in the ASF definition. In skimming through much of the loosely-informed future studies work since the 1970's, it becomes clear that many futurists in recent decades have been both forecasting-challenged and science and technology-unaware. There is much to learn before one should engage in falsifiable extrapolation about the future.

Harold Linstone, editor of the journal Technological Forecasting and Social Change, is one of a handful of scientifically-grounded futurists who presently champions this perspective. When you require predictive validity as a basis for your efforts, you rapidly come to understand that only a subset of future events are particularly easily predicted, making them uniquely important to model and understand, from a policy perspective.

Most centrally, many varieties of accelerating technological change are surprisingly predictable/ forecastable meta-trends. They don't revert periodically to baseline, as do so many cyclical or pendular social changes (market bull/bear cycles, political centralization/decentralization cycles, etc.), but instead continue to accelerate relentlessly, irrespective of culture.

We suggest this kind of change is thus something both mainstream futurists and the general public really needs to understand better, in order to substantially improve our collective decisionmaking.