This big, expensive book offers much food for thought. This review will be a reaction to the first editor's introduction, plus the clever reverse Turing test in Chapter 28 by Charles Platt with machines attempting to determine if humans have any intelligence. Basically, based on my sample of these two chapters, this book is a celebration of the coming extinction of the human race. I shall play the devil's advocate, and also take a meta perspective on the book, analyzing its significance as a social phenomenon instead of considering its contents.
Turing's famous paper on the imitation game (reprinted and annotated in this book), a remote conversation with a computer attempting to prove it is human, in addition to its intellectual fireworks, reflects the fact that Turing, as the French say, "felt uncomfortable in his skin," both as a male and as a human being. As this book indicates, this has now become part of the zeitgeist and a general social problem.
The general attitude I see here reminds me of remarks by Marvin Minsky I heard many years ago, when he called human beings "meat machines," and described the human race as a carbon-based life-form that was creating a silicon-based life-form that would replace it. At the time, his remarks seemed a bit mad, but now many people seem to feel that way.
Why is this? Well, our current society attempts to make people into machines, it behaves as if human beings were ants or bees. We are being forced to live in an anthill, beehive society. Obviously machines are better at being machines than we are, and humans feel ill-suited for anthill or beehive life. Human beings are made to feel obsolete, has-beens.
Robert Epstein's introduction argues that a super-human intelligence is inevitable and not far off in time, and that at best we shall be slaves or pets for the machine, at worse exterminated as annoying insects.
The authors are well aware of the amazing advances in computer technology that they believe make this possible, but perhaps they are less aware of the fact that the more we understand about organisms, the more molecular biology progresses, the more amazing living beings seem. The cells in the human body were originally autonomous living beings that have now banded together much like the citizens in a nation or the employees in a corporation. An individual cell is amazingly sophisticated, and, it seems to me, is best compared with a computer or even with an entire city.
So our artificial machines may not catch up with Nature's machines for a while. Can a century of human engineering compare with billions of years of evolution, essentially an immense parallel-processing molecular-level computation going on throughout the entire biosphere?
In a more optimistic scenario we are not exterminated, the machines will be our servants. Isaac Asimov thought that in the future human beings might live like ancient Greek aristocrats with robotic slaves.
Yes, machines can calculate better than we can, and remember things better than we can. Should we be very upset? Railroad trains go faster than a person can run, a steam-shovel can move earth quicker than a person, an airplane can fly. But human beings made those machines, and should be proud of it. Are we upset about the fact that we need to wear clothing in the winter? Not at all. People are not very fast, not very strong, they do not have fur or a tough hide, but they are extremely curious, clever, and imaginative, flexible and adaptable. Like the universal Turing machine, we are generalists, not specialists. We are not optimized for any particular little ecological niche.
It is also possible that eventually enhanced humans and humanized machines will become nearly indistinguishable, which doesn't sound too bad to me. It's much like wearing clothing or using a can-opener.
But maybe none of this will happen. Another possibility is that machine intelligences will remain unconscious zombies, monstrous golems lacking a divine spark, a human soul. For we are products of George Bernard Shaw's life-force, of Henri Bergson's "élan vital", and machines are not. This is of course not a fashionable view in our secular times, but let me try to give a contemporary version of this argument, one designed for modern sensibilities.
First of all, quantum mechanics, a branch of fundamental physics, has been telling us that the Schrödinger Psi function is real, more real than the particles it describes. Electrons in atoms are expressed as probability waves that interfere constructively and destructively. Atoms are like musical instruments.
Whatever the Psi function is, it is not material. It is more like an idea, and therefore gives support to those Platonic idealist philosophies that view spirit as more fundamental, more real, than matter. Of course, this is not a fashionable interpretation. Nonetheless Nature is giving us this hint loud and clear, even if we refuse to listen.
The latest version of quantum mechanics, now called quantum information theory, reformulates "classical" 1920s quantum mechanics in terms of qubits of information; information is certainly not matter. In my opinion quantum information theory is even less materialist than classical quantum mechanics.
Consciousness, quite mysterious at this time, is also more about information than about matter, I think. Could consciousness reflect some currently unknown level of physical reality? Could our current science be radically incomplete? Indeed, it may well be so. There may be many scientific mysteries yet to solve.
It is true that during the three-century plus history of modern science, each period thinks it has a nearly final answer, only to discover 25 or 50 years later some totally unexpected phenomenon that provokes a complete paradigm shift. Let me invoke a temporal rather than a spatial "Copernican principle." Why should our epoch be especially favored? Why should we have the final answers?
A simple linear extrapolation of the history of science suggests that a century from now things will look remarkably different. What did we know of quantum mechanics a century ago? Is it possible that, to use Wolfgang Pauli's trenchant phrase, our current scientific world-view "is not even wrong?" For our grand-children and great-grand children's sake I hope so. How boring if it should happen that there will be no fundamental changes in our scientific world view in the future. Why should Nature's imagination be as limited as ours?
So if our current scientific world view is not at all final, perhaps living beings do have something special that machines cannot attain, something that science will some day understand as well as we currently understand quantum mechanics, a scientific version, perhaps, of the soul or what the spiritual would refer to as a divine spark. How otherwise to understand cases of amazing human creativity? Pick your own favorite examples. I pick the composer Johann Sebastian Bach, and the mathematicians Leonhard Euler, Srinivasa Ramanujan and Georg Cantor. Can machines have that kind of creativity, that kind of inspiration? These men seem to have had a direct link to the source of new ideas.
Believers in Darwinian evolution by natural selection will argue that no vital spark, no élan vital, nothing at all divine is needed, just random mutations. I myself am a believer in Darwinian evolution. I am currently trying to develop a theory I optimistically have dubbed "metabiology." The purpose of metabiology is to prove mathematically that Darwinian evolution works. But I am open to the possibility that this may not be achievable. It would also be delightful to be able to prove that evolution by natural selection doesn't, cannot work. I would be happy either way, as long as I can prove it. Most likely my metabiological ideas will lead nowhere, but I feel my honor as a mathematician demands that I should give it a try.
And why have human beings become so defeatist? Is it more fun to work in a factory that produces robots than to conceive and raise one's own children? Or look at cars. I have been in remote corners of Argentina, where people seem almost completely divorced from the modern world economy and do everything themselves. They manage splendidly without cars, with horses and donkeys. These are self-reproducing cars, vegetarian cars, not ones that need petroleum.
No wonder that the contributors to this book have given up on human beings. People are ill-used in our modern society, and sensitive scientific intellectuals feel it. Scientists are now micro-managed. The refereeing and grant systems with everything decided by committees favors safe, conservative, incremental science. Can radical new ideas have a chance with our current "factory" science? I doubt it. Would Galileo, Newton, Maxwell, Darwin and Einstein be able to work in the current system? Would Euler, Ramanujan and Cantor? I think not.
As I said, human beings are not ants, they are not bees, they were not designed to be slaves. Let's look at particularly creative periods in human history, for example ancient Greece and the Italian Renaissance.
How come the ancient Greeks were so creative? I asked a Greek intellectual that once, in Mykonos, and he told me that the ancient Greeks discussed this, and noted that ancient Egypt was largely stable and un-innovative for millennia, the contrary of the ancient Greeks, because Greek city-states were small and separated by mountains or isolated on islands, and so imaginative individuals could be creative and affect things, while Egyptian geography permitted strong central, unified control of an empire, creativity was suppressed, and talented individuals could have little or no effect.
Similarly, the creativity of the Italian Renaissance probably had something to do with the fact that, even now, there is no Italian nation-state. Italians are first of all Tuscans or Sicilians, they are individualists, not Italians!
In both cases, ancient Greece and renaissance Italy, chaos and anarchy encouraged creativity, and kept it from being suppressing by the authorities.
What can we learn from this? That strong central control is bad for us. Immediate corollaries: The European Community was not a good idea. And the United States would be better off as fifty separate states. At least that's the case if you want to maximize creativity. I've already said what I think of the current refereeing and grant systems.
Let me wrap up my argument. People are not machines. It is time for people to stop trying to be like machines, because we have machines for that now. We should stop worshipping the machine, and instead unleash our creative, curious, passionate, inspired, intuitive, irrational individualistic humanity.
Gregory Chaitin
IBM T. J. Watson Research Center
http://www.umcs.maine.edu/~chaitin