The Foresight Institute's Founding Vision
page 1 | page 2 | page 3 | page 4
Profiles of the Possible
In the following chapters Drexler explores what will become possible after the advent of atomically precise manufacturing, which he at that time named "The Assembler Breakthrough".
Engines of Abundance
The first chapter in this section lays out why and how assemblers would be able to replicate themselves in less than 15 minutes, and build anything else the same size just as fast. By the 1980s engineers were already considering large self-replicating robot factories. Eventually, it was thought, robots would not only do all the work assembling robots and other equipment, but also make the needed parts, run the mines, etc. Such a sprawling system of robot factories was thought to be feasible, although cumbersome. Such factories would have to make and test parts, as well as assemble them. Drexler points out that molecular replicators would have an important advantage: their parts are atoms, which come ready-made, identical, and perfect.
Artificial molecular replicators would also have an important advantage compared to evolved biological replicators. Once evolution selected a basic pattern of making cells from soft molecules like DNA, RNA, and protein, there was no easy way to change the basic machinery. Engineers, however, are not limited to building from the same soft, moist components biology evolved. No breeder would be able to guide the evolution of a horse into an automobile, but engineers build automobiles. Using assemblers, in his example, engineers will be able to build mechanical nanocomputers such that an entire computer is smaller than a single neurological synapse, and a million times faster. Further, the mechanical arms of an assembler will be about fifty million times shorter than a human arm, and thus will work about fifty million times faster. Drexler estimates that an assembler of a billion atoms could copy itself in a bit more than 15 minutes. At this rate a single assembler, if it could be supplied with sufficient materials and energy, would produce a ton of assemblers in less than a day, and in less than two days would outweigh the Earth.
As an example of how a vat of assemblers could be programmed to build large objects, Drexler describes the construction of a large rocket engine, in less than a day, from atomically precise materials resembling diamond and sapphire that would have over 90 percent less mass than a modern engine of metal. More advanced designs could include a vascular system and assembler and disassembler systems programmed to mend worn or broken parts, or even to remodel itself for changing requirements. Drexler suggests that such systems could make disassemblers to break down rock for raw materials and solar collectors to provide energy so that 'forests' of assemblers could grow space ships from soil, air, and sunlight.
Drexler later realized that macroscopic desktop machines would be much more practical molecular manufacturing systems, and much safer, than would microscopic replicating machines. In Chapter 14 of Nanosystems he refers to such a system as an "exemplar manufacturing system", and in later work as a "nanofactory" or as a "productive nanosystem" [86.1 MB movie]. One of Foresight's major projects from 2005 through 2007 was to work with Drexler and Battelle to develop a Technology Roadmap for Productive Nanosystems. Nevertheless, his conclusions here on the awesome manufacturing capabilities of systems of assemblers remain true for nanofactories and productive nanosystems:
Assemblers will be able to make virtually anything from common materials without labor, replacing smoking factories with systems as clean as forests. They will transform technology and the economy at their roots, opening a new world of possibilities. They will indeed be engines of abundance.
—K. Eric Drexler Engines of Creation
Thinking Machines
In the next chapter Drexler turns his attention to another emerging technolog of vast transformative potential—artificial intelligence, or AI. He describes how trial and error behaviors in simple organisms can lead in more complex organisms to brains with internal models of how the world works. Genes can thus provide for both evolving behaviors and evolving mental models, which in turn can evolve during the course of a single life as individuals learn what works and what does not. Imitation and speech allow mental models to spread through populations and to new generations. Mental models lead to judgments of what knowledge is useful and what is not, and good results lead to goals. Clear thinking and accurate models become goals. Drexler thus describes how the evolution of goals produces both science and ethics, and cites Marvin Minsky's views on the mind as an evolving system (a society) of both simple and complex agents that communicate, cooperate, and compete.
The interactions of mental agents reflect the evolution of memes. Thus minds evolve their mental models of the world through varying idea patterns and selecting among them, guided by knowledge learned from others. The fact that intelligence in humans can be seen as a natural process, makes intelligence in machines less surprising.
Drexler sketches the evolution of artificial intelligence from the early mechanical calculator of Babbage to the early expert systems of the 1980s that achieved startling successes in very narrow domains of knowledge. The definition of artificial intelligence, however, seems to be a moving target so that once the expert systems succeeded, they were no longer considered true artificial intelligence.
Is self-awareness the true test of AI? In discussing the Turing test of whether a computer is intelligent, Drexler notes that self-awareness (consciousness) evolved in people because our mental models of the world must include other people, their abilities and inclinations, and how they relate to us in order to make plans that involve them and ourselves, so that self-awareness is no more than a special part of interacting patterns of thought. Though there seems to be nothing magical about self-awareness, Drexler notes that no computer can now (1986) pass the Turing test.
Drexler proposes that passing the Turing test, however, is not the most useful goal for AI research. He distinguishes two types of AI. Technical AI deals with the physical world and leads toward automated engineering and scientific inquiry. Social AI deals with human minds and leads toward passing the Turing test.
Drexler describes contemporary efforts in automated engineering, specifically Douglas Lenat's EURISKO, designed to explore new areas of knowledge guided by heuristics, and the successes it had achieved by the mid-1980s.
He argues that technical AI systems adapted to automate scientific inquiry and engineering design will both speed the race to develop assemblers (nanofactories) and increase the impact of atomically precise manufacturing once it is developed. While describing the immense impact of AI, Drexler also claims that it differs from atomically precise manufacturing in that developing it will likely require new science as well as engineering multiple generations of new tools. A Foresight Background essay "Dimensions of Progress" published in 1987 explores the effect of developing atomically precise manufacturing on the development of AI, and the effect of developing AI on developing APM. Developing institutions to deal with such a complex fabric of opportunities and threats is another part of Foresight's mission.
The World Beyond Earth
The next chapter explores the promise of the space frontier—abundant resources sufficient "to build the practical equivalent of a thousand new Earths"—artificial space habitats of continental scale. Drexler argues the case for habitats in free space over settling Mars in an article originally published in the L5 News in 1984 and reprinted with his permission by Foresight: "Space Development: The Case Against Mars". A description of space vehicles 90% lighter than similar conventional vehicles, and of a space suit that feels as comfortable as wearing nothing but "will keep you comfortable, breathing, and well fed almost anywhere in the inner solar system" illustrate the leaps in system performance that APM will bring. The trio of replicating assemblers (atomically precise manufacturing), automated engineering, and access to space resources point to a future of very low costs and abundant resources.
Drexler paints a picture of immense abundance including O'Neill cylinder-like space habitats made of strong carbon-like materials that will contain not only continent-scale land masses, but seas "wider and deeper than the Mediterranean". By his estimates, the space near Earth could harbor a total land area more than a million times Earth's area. Nor would we be forever confined to the solar system. Large banks of lasers orbiting the sun could push spacecraft attached to lightsails to the stars at a substantial fraction of the speed of light, reducing travel times from millennia to less than decades.
Drexler's hope is that the prospect of such abundance will encourage the cooperation that humanity will need to prevent these technological advances from being turned into weapons to destroy humanity. This hope that the opportunities afforded by emerging transformative technologies will inspire cooperation to use these tools wisely inspires Foresight's efforts to educate the public and policy leaders on both the perils and potentials of these technologies.
Engines of Healing
The next three chapters elaborate the application to medicine and life extension of complex molecular machinery, described by Drexler as "cell repair machines", made possible by atomically precise manufacturing and automated engineering. The first of these chapters begins with the observation that the most delicate surgery is still coarse from the viewpoint of individual cells, and the observation that drug molecules "are too simple to sense, plan, and act." Drexler foresees molecular machines controlled by nanocomputers that will examine and repair biological molecular machines within cells, bringing "surgical control to the molecular domain." Drexler then describes how molecular knowledge derived from modern biomedical technology will be combined with atomically precise manufacturing and technical AI systems to build cell repair machines to enter cells and modify their structures to bring them into compliance with molecular-level descriptions of healthy tissue. With a vivid description of how small molecular machinery would be compared to a typical human cell, Drexler makes clear that many small repair machines (perhaps thousands), each with a small computer to control it, and a "large" central nanocomputer to coordinate their activities, would together occupy only a small portion of the volume of the cell. A cell would have a thousand times the volume of the central nanocomputer and a million times the volume of single repair device. Such machines could identify any of the myriad molecules in a human cell and record the type and position of each large molecule.
Cell repair machines will be able to selectively "destroy the dangerous replicators, whether they are bacteria, cancer cells, viruses, or worms." They will be able to make whatever molecular or cellular repairs are necessary, and even to regenerate fresh tissue when needed. Using the example of damage to the brain to define the limits on what cell repair machines could do, regenerating brain tissue could not restore lost memories or skills if the structures in the brain encoding those memories and skills had been lost. "Loss of information through obliteration of structure imposes the most important, fundamental limit to the repair of tissue."
A final capability that Drexler foresees is the induction of reversible biostasis by stopping metabolism and holding cell structures firmly in place. Comparing the reversibility of biostasis with the irreversibility of severe brain damage illuminates the revolution that cell repair machines will bring to medicine. Today physicians must try to preserve the function of tissue because today's medicine can only help tissues to heal themselves. Cell repair machines will change the central requirement from preserving function to preserving structure. Once the molecular and cellular structures of healthy tissue have been described, cell repair machines will be able to restore health by restoring healthy structures.
Robert Freitas has continued Drexler's study of cell repair machines with his four-volume Nanomedicine book series, two volumes of which have been published as of 2016. From 1998 through 2006 Freitas curated the Foresight Institute's Nanomedicine section and Nanomedicine Art Gallery, in which he presented his research and design studies, and related designs, news, and images related to cell repair machines and similar nanomedical robots.
Long Life in an Open World
In the next chapter Drexler considers the consequences of long, healthy lives achieved by inexpensive means. He gives particular attention to how nanotechnology (atomically precise manufacturing) will make obsolete pro-death memes concerning over population, pollution, and resource depletion. Paralleling the development of cell repair machines built from molecular machinery to heal the body, cleaning machines built from molecular machinery will heal the planet by rearranging the atoms of dangerous molecules or by isolating dangerous atoms. Molecular machines will make cheap solar power a reality and enable removal of excess carbon dioxide from the atmosphere, preventing drastic climate changes. Inexpensive and sophisticated robots will clear and restore lands damaged by twentieth century civilization. With improved spacecraft built from molecular machinery, the solar system will be opened to human settlement.
Drexler points out that despite these advantages of advanced nanotechnology, even the cleanest molecular manufacturing machines will produce waste heat, and the solar system is finite, while the stars are distant. So long-term limits still exist. Less clear are the effects of long life on culture and political conflict. On the other hand, the hope of long life and abundance may increase hope and motivation to solve such problems as exist or arise.
Writing in 1999 in Nanomedicine, Volume I: Basic Capabilities, Robert A. Freitas Jr. estimates that the finite ability of the Earth to handle waste heat from human activities without disastrous effects on climate limits the amount of "active continuously-functioning nanorobotry" to ~10 kg per person (6.5.7 Global Hypsithermal Limit).
Writing in 1986, Drexler offers no timeline for the development of cell repair machines and the other benefits of what we now term atomically precise manufacturing, but he does speculate that a 30-year-old could, with the help of incremental progress in medicine and biotechnology, survive to the 2030s and enjoy the beginnings of cell repair technology, and thus make it into the 2040s and 2050s to benefit from advanced cell repair technology and indefinite life-span. He concludes by asking if today's medical technology might be able to stop biological processes in a way that tomorrow's cell repair machines will be able to reverse?
A Door to the Future
This question leads to the next chapter on biostasis and cryonics. Drexler first establishes that the essential requirement for any biostasis is to preserve the brain structures that encode personality and long-term memories since cell repair machines could repair or replace any other damaged or missing tissue. He then presents evidence that long-term memory is encoded in structures that are reasonably robust and likely to remain intact for some brief period after clinical death. He cites fixation with glutaraldehyde, cryoprotection and vitrification with propylene glycol, ethylene glycol, or dimethyl sulfoxide, and freezing in liquid nitrogen as credible methods of initiating biostasis with current technology that should be reversible with the future technology of cell repair machines.
Having presented clear arguments for the future feasibility of cell repair machines and the present availability of biostasis technology that could be reversed in the future with such cell repair machines, Drexler addresses why cryonics is not widely used. To counter the argument that cell repair technology is not available, he cites several current examples where powerful technologies have developed over several decades from very modest beginnings once feasibility was understood. He cites examples where seemingly dull, unimportant facts were found to have dramatic consequences, including great medical advances. He hopes that, as more people become aware of the benefits that atomically precise manufacturing, technical AI systems, and space resources will bring, and of the option to open a door to the future with biostasis, a more lively interest in the future will develop, and this will spur efforts to prepare for the dangers that advanced nanotechnology will bring.
The Limits to Growth
After devoting six chapters to exploring the capabilities that nanotechnology will bring, Drexler devotes the next chapter to exploring the limits to growth. The most basic limits are set by natural law. New scientific understanding does not always open new possibilities for technology. For example, Einstein's special relativity set a firm limit to how fast anything could travel through space. In a historical recounting of the surprises and revolutions in physics during the preceding century, Drexler builds a case that it should have been clear to 19th century physicists that their understanding of matter was incomplete, as indeed it was. He then builds the case that our understanding of the world of atoms and molecules, planets and stars is now complete, and that the frontiers of physics now lie beyond the world of atoms, with particles that are too unstable and energies that are too high to be relevant to building machines and technologies that would be useful to beings living in the world of atoms. It appears that atoms will remain the sole building blocks of stable hardware so that engineering will remain a game played with pieces and rules that are already known. At most, unexpected new particles would add pieces, not eliminate rules. Drexler develops this point by explaining why molecular machinery cannot affect the nature of the atomic nucleus ("more futile than trying to flatten a steel ball bearing by waving a ball of cotton candy at it"), and further, why nuclear-scale (femtometer-scale) machinery does not seem possible (ferocious electrostatic repulsion, requirements for pressures found only inside collapsing stars, etc.).
Because genuine limits exist to properties of matter, like strength and heat resistance, etc., the point will eventually be reached with an increasing number of technologies where our capabilities will cease growing. "We may misjudge the limits today, but wherever the real limits lie, there they will remain."
Having established that true limits to progress do exist, Drexler debunks the assertions of certain popular writers in the early 1980s that the accumulation of waste heat and entropy (disorder) on earth will bring a halt to human progress. In fact, these limits apply not to the Earth, but to the universe as a whole. There will in fact be a limit to the amount of industrial activity that can occur on Earth, but the vastness of space, or at least of the solar system, is available for industrial activity. Entropy will indeed limit the ability of the universe to support life, but not until an immense, cosmological span of time has passed.
Drexler makes the effort to debunk obviously bogus ideas because they can easily spread in our society due to inadequate "mental immune systems". The spread of bogus ideas distracts scarce attention from the real problems that will accompany accelerating technical revolutions. Before considering, in the third and final section of his book, these problems and how to improve our "mental immune systems" to deal with them, Drexler turns his attention to the real limits imposed by finite resources and the speed of light.
Drexler explains that emerging technologies will make available clean resources on the Earth that are vast compared to the resources we now use; that the solar systems contains resources that are vast compared to the resources available on the Earth; that the galaxy contains resources that are vast compared to the resources available in the solar system; that the visible universe contains resources that are vast compared to the resources available in the galaxy. Nevertheless, Malthus was essentially correct. In the long term any population of replicators, like humans, will expand at an exponential rate characteristic of its most rapid replicators. Once the resources of the solar system are saturated, expansion can only occur at the frontier and will be limited by the speed of light. This expansion will result in cubic growth of resources. Mathematics shows that exponential growth will soon overtake cubic growth, probably within one to two thousand years. Thus, even with continued expansion into space, unlimited exponential growth remains impossible.
If other civilizations already own the resources of the universe, then exponential growth will be further limited. Drexler argues, however, that contrary to popular suppositions, we may be alone in the visible universe. His primary argument is that astronomical observations of other galaxies and of other sections of our galaxy reveal an enormous waste of resources (dust, starlight) that the laws of evolution require any advanced civilization in the vicinity would have already exploited.
Drexler paints a picture of a future of great but limited growth, in which growth will eventually become limited everywhere except at the frontiers, which will over time become more and more distant. In most areas growth will be limited to fields like art and mathematics, the "world of mind and pattern", in which growth does not demand ever increasing supplies of matter and energy.
Against this background of realistic limits, Drexler addresses why "obviously bankrupt" models of the future that posit no new technological breakthroughs during the coming century remain popular. He believes this popularity rests in the desire to avoid the challenges of understanding and controlling the technology race. Warnings of bogus limits thus distract from dealing with true limits, from identifying goals worth pursuing, and from planning effective strategies for dealing with coming problems.
page 1 | page 2 | page 3 | page 4
|