Mind Children

This original article by Dan Polansky describes book Mind Children: The Future of Robot and Human Intelligence by Hans Moravec, 1988. The book is mentioned in the Encyclopedia Britannica article on the author, making it somewhat notable. The book can perhaps be best described as science-fiction without a plot, packed with more interesting ideas than most science-fiction novels, albeit presented as a description of actual future.

Page references in this article are for the paperback edition, ISBN 0-674-57618-7.

Destruction of living things and humans

The preface suggests that robots will be developed that will be comfortable with destroying humans and perhaps other living things as well (page 1).[1] The language is not wholly explicit; the wording used is that "They [the genes] have produced a weapon so powerful it will vanquish the losers and winners alike", and "the human race has been swept away by the tide of cultural change, usurped by its own artificial progeny". Here, "artificial progeny" refers to robots. A possible consolation is that some of these robots will be emulating individual human minds, thereby ensuring their continuing quasi-existence; see section Mind uploading.

The mind children metaphor

The metaphor of "mind children" has powerful rhetorical effect. It implies that since humans are fine to pass the future to their biological children, they should be equally fine to pass it to the children of the mind, the superintelligent thinking machines or robots. Marvin Minsky is happy to pass the future to them, and refers to Hans Moravec and his mind children metaphor.

Artificial intelligence predictions

The book predicts human-level intelligence for supercomputers by 2010 and for personal computers by 2030 (pages 64 and 68).

Self-improving thinking machines

The book says: "Sooner or later our machines will become knowledgeable enough to handle their own maintenance, reproduction and self-improvement without help. When this happens, the new genetic takeover will be complete. Our culture will then be able to evolve independently of human biology and its limitations, passing instead directly from generation to generation of ever more capable intelligent machinery." (page 4).[2]

And, "A postbiological world dominated by self-improving thinking machines will be as different from our own world of living things as this world is different from the lifeless chemistry that preceded it. A population consisting of unfettered mind children is quite unimaginable. We are going to try to imagine some of the consequences anyway." (page 5)

Further, "Eventually humans [...] will become unnecessary [...] as the scientific and technical discoveries of self-reproducing superintelligent mechanisms are applied to making themselves smarter still." (page 102)

Mind uploading

The book predicts mind uploading, a process by which a simulation of a human mind is transferred to a human-like robot (pages 108-112).[3] The patterns of activity of the person's real biological brain are scanned and based on the observed patterns, a faithful simulation is created. This would enable extreme longevity by allowing the mind to transfer between a succession of robots, each new robot replacing a worn-out old one. The result would be not a true immortality but quasi-immortality. The book calls this "transmigration".

One consequence the book predicts is a 1000fold increase of mind speed since the emulating hardware would not need to keep the emulation as slow as the real biological phenomenon (page 112).[4]

A related term is "whole brain emulation".

In a 1986 article, Moravec mentions that "Vernor Vinge devised a particularly slow and gentle [mind] transfer method in True Names, his novel of the near future." True Names is from 1981.

Mind uploading via "brain surgeon" is already described in Moravec (1979).

Pattern-identity position

Related to mind uploading is Moravec pattern-identity position, which he contrasts to body-identity position. He starts by quoting someone's objection: "Regardless of how the copying is done, the end result will be a new person." (page 116)[5] This is in reference to mind uploading, called "transmigration". The book goes on to argue that living bodies keep on renewing themselves, old cells dying and new cells being created, and therefore, what creates their identity across time are the continuing patterns. This line of argument leads him to claim that a perfect simulation of what was previously human mind in a biological body taking place in a human-like robot is pattern-identical to the original and thus identical. The idea is briefly mentioned in Moravec article online, with the "pattern Identity" spelling. Bamfordand and Danaher mention Moravec and find similar treatment in Kurzweil (2000), Koene (2011) and Goertzel and Ikle (2012).

Performance boost

As per above, the book predicts is a 1000fold increase of mind speed since the emulating hardware would not need to keep the emulation as slow as the real biological phenomenon (page 112).[6]

A similar boost appears in Moravec (1977): 'Advantages become apparent as soon as the process is complete. Somewhere in your machine is a control labelled "speed". It was initially set to "slow", to enable the simulations to remain synchronized with the rest of your old brain, but now the setting is changed to "fast". You can communicate, react and think at a thousand times your former rate. But this is only a minor first step.'.

Space colonization

Moravec envisions space colonization of other galaxies, going beyond our galaxy, the Milky Way. Since, as per Moravec, "Sooner or later an unstoppable virus deadly to humans will evolve, or a major asteroid will collide with the earth, or the sun will expand, or we will be invaded from the stars, or a black hole will swallow the galaxy. The bigger, more diverse and competent a culture is, the better it can detect and deal with external dangers". (Similar passage is found in Moravec 1977 article.) The passage is intended to show that technological progress is necessary to avoid dangers, and since one of the dangers is a black hole swallowing the galaxy, the remedy must be to get beyond the galaxy. Further, "Resurrecting one small planet should be child's play long before our civilization has colonized even its first galaxy" (p. 124)."

"Eventually humans [...] will become unnecessary [...] as the scientific and technical discoveries of self-reproducing superintelligent are applied to making themselves smarter still. The new creations [...] will explode into the universe, leaving us behind in a cloud of dust." (page 102) The vague phrasing "explode into the universe" suggests the superintelligent machines will expand far beyond the Solar System.

Further evidence is this: "A spectrum of scales will come to exist--from [...] to star-spanning superminds for big problems" (page 125). It is not clear what "star-spanning" refers to, though: a supermind that is part of a star? Or a supermind that is spread across multiple star systems?

Further, "The human race will expand into the solar system before long, and human-occupied space colonies will be part of that expansion." (page 101) And further, "Imagine the immensely lucrative robot factories that could be built in the asteroids".

Development beyond imagination

Moravec seems to envision a world of thinking machines that will be beyond human imagination: "A postbiological world dominated by self-improving thinking machines will be as different from our own world of living things as this world is different from the lifeless chemistry that preceded it. A population consisting of unfettered mind children is quite unimaginable. We are going to try to imagine some of the consequences anyway." (page 5)

Resurrection

The notion that faithful simulation of a human mind in a machine is as good as the original leads the book to conceptualize resurrection. Thus, "Wholesale resurrection may be possible through the use of immense simulators." (page 123) And, "It might be fun to resurrect all the past inhabitants of the earth this way and to give them the opportunity to share with us in the (ephemeral) immortality of transplanted minds." (page 124)

The above will be enabled by feats of simulation, for which see the next section.

The book's treatment of resurrection is covered in The New Atlantis article.

Simulation

The book predicts incredible feats of simulation. In mind uploading, the whole brain of a human can be simulated in a human-like robot, faithful enough that the human can consider the result to be identical to themselves. But single human is not enough: "Now, imagine an immense simulator (I imagine it made out of a superdense neutron star) that can model the whole surface of the earth on an atomic scale and can run time forward and back and produce different plausible outcomes by making different random choices at key points in its calculation. Because of the great detail, this simulator models living things, including humans, in their full complexity." (page 123, quoted in the "No Loyalty to DNA" review below)

Unstopability of technological development

Moravec argues that technological development of thinking machines is unlikely to be stopped, as a result of friendly or unfriendly competition between nations, and since these machines are key for very long term survival. Since "If the United States were to unilaterally halt technological development (an occasionally fashionable idea), it would soon succumb either to the military might of unfriendly nations or to the economic success of its trading partners. Either way, the social ideals that lead to the decision would become unimportant on the world scale." (page 101) And further, "If, by some unlikely pact, the whole human race decided to eschew progress, the long-term result would be almost certain extinction. The universe is one random event after another. Sooner or later an unstoppable virus deadly to humans will evolve, or a major asteroid will collide with the earth, or the sun will expand, or we will be invaded from the stars, or a black hole will swallow the galaxy. The bigger, more diverse and competent a culture is, the better it can detect and deal with external dangers".

Nuclear war

As per the book, "Engaged for billions of years in a relentless, spiraling arms race with one another, our genes have finally outsmarted themselves. They have produced a weapon so powerful it will vanquish the losers and winners alike. This device is not the hydrogen bomb--widespread use of nuclear weapons would merely delay the immensely more interesting demise that has been engineered."

What the metaphors describe is the biological evolution of humans, who then are about to create a thinking machine, which will then obliterate its makers. What is remarkable is the idea that this will happen even in the case of a nuclear war, as if the resulting technological and social setback were just a minor step back to be overcome again given geological time. There is no idea of a limited window of opportunity.

Nanotechnology

The book assumes remarkable development in nanotechnology: "Before long, conventional technologies, miniaturized down to the atomic scale, and biotechnology, its molecular interactions understood in detailed mechanical terms, will have merged into a seamless array of techniques encompassing all materials, sizes and complexities. Robots will then be made of a mix of fabulous substances, including, where appropriate, living biological materials." Apparently, no arbitrary combination, mixing and achievement is off limits.

Immortality via accelerating time

The extreme longevity achieved via mind uploading is apparently not enough. The book wants more and thinks it can be achieved. The book speculates that "subjective infinity" could be achieved by increasingly accelerating time at the end of the universe (pages 147-149). Indeed, in exponentially accelerating subjective time, a subjective sense of infinite life can be achieved in finite amount of time; think of Achilles and Tortoise. The book refers the ideas back to Freeman Dyson's Infinite in All Directions (1988) and Barrow and Tipler's The Anthropic Cosmological Principle. Similar ideas were further developed by Tipler in his later The Physics of Immortality, not referenced by Moravec. Similar ideas were developed as part of Dyson's eternal intelligence concept in 1979 in Dyson's article Wikidata:Time without end: Physics and biology in an open universe.

Culture

Moravec uses the word "culture" to cover not only human phenomena but also those of the superintelligent thinking machines. For instance, "the human race has been swept away by the tide of cultural change, usurped by its own artificial progeny". The result is a curious rhetorical effect: without Moravec context, one would not use the phrase "cultural change" to refer to replacement of biologically embodied humans with artificial thinking machinery.

Collaborators

The book was edited with the help of Harvard Press. Reviews of drafts were written by Vernon Vinge, a singularitarian. Referees who provided comments included Rod Brooks, Richard Dawkins, Kee Dewdney, Bruce Donald, John Dowling, Bob Forward, John McCarthy, Pamela McCorduck and others.

Criticism

Original criticism of the book follows:

• Generally, the book belongs to the field of science-fiction even if it does not present itself as such.
• Development of human-level intelligence in silicon may be impossible; we do not know. There are limits of what can be done in physics, and the limits of what can be done in carbon, hydrogen, oxygen and other elements of living things via growth from tiny biological cells may differ from the limits of what can be done in silicon via manufacturing and the differences in the limits may be unfavorable to silicon. If it is not so, the book should analyze why not, but it does not do that. The book does nothing to show possibility. Extrapolation from temporary exponential growths of computing capability into many decades into future is no solid science; if experience teaches us anything, all apparent exponential growths in nature are in fact logistic growths, S-curve growths.
• In retrospect, the prediction of human-level intelligence for supercomputers by 2010 already failed. The prediction was made by extrapolation of technology growth data points. This provides tangible evidence for how problematic such extrapolation is.
• Mind uploading and emulation in silicon may be impossible. The book does nothing to show possibility.
• Space colonization, including Mars, may be impossible. The book does nothing to show possibility.
• Simulation of the surface of the whole Earth including all living things in a neutron star sees to be a pure unfettered fantasy.
• Immortality via accelerating time approaches pure speculation, whether done by Moravec or Tipler.
• Humans may dislike being uploaded into machines as mere simulacra and may take steps to prevent silicon-based machinery from taking over the biology.
• No mention is made of the need of the machines to preserve the Earth's biosphere in its actual form to sustain the planetary homeostasis (which simulated analogs cannot do), to ensure sufficient energy sources to last for ages, to avoid hugely disruptive changes to planetary environments, and to ensure material recycling of machine bodies or run the risk of running out of mined raw materials, in contrast to the marvelous recycling of matter achieved by the actual living things. All these problems are either supposed to not exist or to be soluble with the use of the developed machine superintelligence, none of which is necessarily true. The biosphere and the living things are largely left without mention, except when they appear in the word "postbiological", which implies their absence or irrelevance.
• Dangerous technological progress can be stopped as a result of international treaties, provided enough powerful countries agree and are ready to enforce their will to stop the lethal threat to genuine biological humans (as opposed to simulacra) against other countries that disagree. The requirement of developing thinking machines to get the minds out of this galaxy (since, beware of the black hole) is very unrealistic given current knowledge.

Reception

Joseph Weizenbaum warned that "Mind Children" was as dangerous as "Mein Kampf."

Moravec mind uploading was covered in detail, including a full quote, in Jerry Mander's In the Absence of the Sacred, 1991, a book critical of the uncritical acceptance of all development of new technology by the Western industrial civilization.

Further reading

• Book info for Mind Children: The Future of Robot and Human Intelligence, Hans Moravec, 1988, frc.ri.cmu.edu, archived at archive.org
• Book Reviews of Mind Children: the future of robot and human intelligence, frc.ri.cmu.edu, archived at archive.org
  • ANYTHING WE CAN DO THEY CAN DO BETTER, frc.ri.cmu.edu, archived at archive.org
  • THE SOULS OF THE NEW MACHINES, frc.ri.cmu.edu, archived at archive.org
  • No Loyalty to DNA, The New Yorker, 1989, frc.ri.cmu.edu, archived at archive.org -- has multiple quotes
• W:Hans Moravec#Mind Children
• W:Mind uploading
• Strolling up the garden path by Igor Aleksander, 1988, nature.com, a review of Mind Children
• Superhumanism, 1995, wired.com
• Matter Over Mind (a review of Mind Children) by Roger Penrose, 1990, nybooks.com
• Virtual Futures: Cyberotics, Technology and Post-human Pragmatism, edited by Joan Broadhurst Dixon, Eric Cassidy, 1998, monoskop.org
• The Universal Robot in Vision-21 Interdisciplinary Science and Engineering in the Era of Cyberspace, by Hans Moravec, 1993, nasa.gov -- presents some of the ideas from Mind Children