Customer Reviews

Darwin Among the Machines: The Evolution of Global Intelligence (Helix Books)

5 star:		(10)
4 star:		(8)
3 star:		(4)
2 star:		(1)
1 star:		(1)

 

 

George Dyson has the rare skill of being able to put flesh on ideas. He is particularly good at Samuel Butler(evoked in the title essay) and a few Darwins: Erasmus (a great character and, we learn here, Mary Shelly's inspiration for Dr. Frankenstein), his grandson Charles (Origin of Species), and brief mention of Charles' grandson Sir Charles Darwin (who headed the National Physical Laboratory (NPL) which employed Alan Turing, but was unable to gain support for Turing's project to build an "Automatic Computing Engine" in 1945). Selected against.

The Chapter on Butler is worth the price of the book. Readers will also encounter many obscure names brought alive with interesting detail and then fit into the evolution of a familiar technology. For example, Dyson explains how wooden tally sticks, used as a primitive, secure means of record keeping in the English (twelfth century) pre-history of banking, both facilitated the establishment of a banking system and served as an early precursor and model for encryption keys.

Familiar, iconographic names, Charles Babbage and John Von Neuman, to name just two examples, are shown in somewhat different, and more human, light than they are usually presented. Babbage, for example, was a prophet of telecommunications whose early ideas for what we now call packet switching revolutionized the British mail system. Babbage analyzed the operations of the British postal system and found that its costs were governed more by switching than by distance. His recommendaton of a flat rate service was introduced in 1840 as the penny post. Von Neuman's influence is described in detail in many places, for his contributions to mathematics, game theory, computing, the Cold War defense system, and the Institute for Advanced Studies at Princeton.

Students looking for a concise description of the history of "distributed communication" (most familiarly now the Internet) will also find a great and amusing chapter in this book. Dyson has written a remarkably compact description of how the issues and concerns of the defense establishment encouraged the creation of what we now know as the Internet.

The boundlessness of the book, its avoidance of the shelter of one or a few strict disciplines, is among its greatest attractions. If anyone ever asks you what a liberal arts education is, point them to this book. There is no better book on how ideas live and grow across generations.

Darwin Among the Machines is science writing, intellectual history, personal essay, and more.

Several have criticized Dyson's philosophical and historical treatise "Darwin Among the Machines" for not articulating exactly how a global intelligence might emerge from today's synthetic biological and computational networks. But as Dyson says in the preface, the past is where we find answers, and the future merely a fog of questions "to which the answers are up to us." In the next 200 pages, Dyson explores the history of an idea: that man will someday create a form of artificial life, with intelligence that may match or exceed our own.

It may astound some readers to know that these ideas date much farther back than Alan Turing's "Turing Test," or Vannevar Bush's influential essay "As We May Think." Consider the following quote from Thomas Hobbes (1651): "Nature is by the Art of man, as in many other things, so in this also imitated, that it can make an Artificial Animal." Or consider this excerpt from Samuel Butler's 1859 essay, which serves as Dyson's main theoretical foundation: "As the vegetable kingdom was slowly developed from the mineral, and as in like manner the animal supervened upon the vegetable, so now in these last few ages an entirely new kindgom has sprung up ... It appears to us that we are ourselves creating our own successors."

Careful to acknowledge his predecessors, Dyson profiles the lives of some of the most prescient Enlightenment- and modern-era thinkers in captivating detail. In so doing, he traces the evolution of the "Artificial Animal" from its earliest incorporeal appearances - as merely an idea - to its current computational incarnation in neural networks. But Dyson doesn't stop there.

In fact, he goes on to argue that the global telecommunications network (primarily the internet) may provide the appropriate architecture for a kind of global, distributed intelligence to evolve. Here Dyson borrows from Leibniz, who noted that the "soul" may be "born when the machine is organized to receive it, as organ-pipes are adjusted to receive the general wind."

To further support this claim, Dyson draws parallels between the development of increasingly efficient machines and the processes of biological evolution. In fact, this is one of the most interesting parts of the book, in part because the language in which Dyson details the principles of evolution might be considered dangerous today, in the midst of the raging Intelligent Design debate. For example, Dyson suggests that evolution itself may embody a kind of intelligence, though we frequently perceive it as merely a shallow process, highly dependent on chance and randomness.

As Dyson points out, this perception gets to a fundamental semantic confusion surrounding "intelligence," a phenomenon well known to AI researchers in which problems once thought to require intelligence are then seen as trivial after an algorithm is designed to solve them. As Dyson points out, intelligence may simply be a word we use to describe behavior that corresponds to our view of how humans behave. Not believing in "'the existence of an intelligence behind the achievements in biological evolution may prove to be one of the most spectacular examples of the kind of misunderstandings which may arise before two alien forms of intelligence become aware of one another.' Likewise, to conclude from the failure of individual machines to act intelligently that machines are not intelligent may represent a spectacular misunderstanding of the nature of intelligence among machines."

Ultimately, whether you agree with Dyson's perspective is besides the point. This is not a scientific book; many of the ideas are purely philosophical, and the logic used to support Dyson's assertions frequently rests on historical anecdote and analogy. These should not be considered weaknesses, however. The real, lasting value of "Darwin Among the Machines" is Dysons's imaginative and graceful writing, his impeccable historical research, and the conceptual ease with which he integrates ideas from ballistics, biology, hydrodynamics, set theory, Cybernetics, and uncountably more esoteric subjects.

Though I won't dispute that many of these exciting ideas are far-fetched, Dyson has found powerful allies for his assertions, from Hobbes and Leibniz to Goedel and Von Neumann. So if you find yourself believing - or simply wanting to believe - in these groundbreaking ideas, then you're in fine company.

While I understand and (to a certain extent) agree with all the positive comments from the reviewers on this page, I find myself unable to share the sentiments.

The book is well written in the sense that Dyson provides a rich series of anecdotes and historical facts to back up the connection between the evolution of man and machine that he posits as his central thesis. And while I really appreciated those anecdotes, I didn't find that he really earned my belief. I often found that he made leaps of logic in the way he lay the thing out that were expected to stand more on the charmingness of the stories and/or the pithiness of the quotations that preface each chapter than on a real well-constructed argument. I often didn't follow how he got from point C to point D and frankly I found many of the connections that he made rather streched in appropriateness.

This said, Dyson is clearly a very smart guy with a lot of interesting things to say on the topic, so despite my concerns/disbelief, the book is a worthwhile read.

I bought this book in the hope of reading some intelligent speculations by the author about evolution, machines, and AI, which is what the title suggested I would find. However, it turned out to be a history of the evolution of computers with old speculations from the computer pioneers concerning the evolution of computers injected along the way. To be fair, the author does have an overarching thesis that he tries to weave into the historical narrative whenever some past speculation seems to lend it some support. It is that the World Wide Web - that well known network of millions of computers - may some day, at a certain critical size and running who knows what software (certainly not the author) will become intelligent in some way (also not specified by the author). Come to think of it, I think the author has used the historical angle of the book - the similar speculations of the computer pioneers of the past - as a device to lend credence to his thesis - a kind of proof by consensus. I remain unconvinced, however. His arguments (where there were any; it was hard to tell his arguments from narrative) were very weak and unconvincing. To his credit, the author did a tremendous job of scholarship for the historical side of the book. However, he left the speculative side undeveloped (at the most weakly developed) and, therefore, the book was unappetizing to me.

Wow! What a bunch of excellent reviews! They all provide excellent information. That won't stop me from entering my own review, but I will present it in the form of a self-test.

Answer questions for the intended reader (IR) of the book.

1. Does the IR have one or more books relating to Science with the word "fuzzy" in the title?

2. Does the IR describe time sitting in front of the computer as "fun"?

3. Does the IR cite Gödel's theorem at recreational gatherings?

4. Does the IR have both a cellular phone and a Personal Digital Assistant?

5. Does the IR cite evolutionary theory as an explanation for everyday phenomena?

6. Is the IR a member of the Sierra Club or Greenpeace and a regular reader of one of the following: Scientific American, Nature, Science, or Technology Review?

7. Does the IR regularly read Discover magazine and is a college graduate (or equivalent autodidact)?

I think you get the idea.

If "yes" is the answer to one or more of the above, this book should be a hit. If you said "yes" to three or more, this book is a must have!

This book was really extraordinary to read. Dyson correctly approaches digital evolution and delivers expansive and well thought theories. Really thought provoking and well laid out. While the prose is difficult sometimes the ideas burn through it and communicate some very powerful ideas that still bounce around my head months after reading it. The exploration of macro-consiousness in the economy and the world-wide network were really enlightening and almost unsettling.

How does the development of "artificial" intelligence fit into biological evolution? George Dyson suggests that the fit is seamless. This profound investigation of the history of thinking machines and evolutionary theory is brilliant and engaging. It offers a far more palatable look at the human-machine future than the misanthropic vision of Ray Kurzweil in THE AGE OF SPIRITUAL MACHINES (Viking, 1999 ). (Palatability is no assurance of accuracy, of course, but it sure feels better going down.) Where Kurzweil sees machine intelligence as better than human and confidently predicts that we will upload ourselves and abandon our bodies, -- drawn initially by the superiority of cyber sex -- Dyson envisions a spreading macro intelligence that will involve humans in wholly new ways -- and suggests it is already emerging. He predicts that much like other life forms which share their eco-niche, we are apt to become symbionts with the machines, each doing what we do best and benefitting the other. But perhaps I have gotten ahead of myself here. Are there living, thinking, artificial minds on this planet already? Dyson asserts that it depends entirely on one's definition of "living," "thinking," "artificial," and "mind." There are self-replicating cybernetic entities evolving within computer networks. Self awareness is not yet evident, but it is not entirely clear that it doesn't exist. How will we know when and if it emerges? We don't have a clear definition of consciousness as it applies to our own condition, which makes consideration of the whole issue iffy at best. Tracing the history of evolutionary theory, following the startling course of mathematics in the past century, noting the lightning fast advances in silicon technology, and all underlain by a trenchant sense of human development, Dyson weaves a fascinating tale. He pulls in the Darwin family, Darwin's critics (then and now), oddball tinkerers, mainstream theorists, nuclear physicists, Turing, Godel, and von Neuman, and the science fiction of Olaf Stapledon in the telling. A highly rewarding if sometimes difficult read. Dyson ends with a quote from Thoreau, suggestive of his own open-ended view of our future: "In wildness is the preservation of the world." Exactly so.

The title implied that this book was about Artificial Life - evolution as applied to machines. Rather it is more of a collection of biographys of scientists involved with this field.

I found the book very difficult to read, with very little relevant details for modern researchers in Artificial Life and complex systems. Even as a 'historical' book it is long winded, and does not bring the people being described alive. Rather dissaponting for a book with such a nice title.

Those of us who have spent more than a few years with computers have read a great many books about the history of computing, and they seem, for the most part to be of the "kings and battles" school of history writing: Long lists of men and machines that convey the image of computing history as strictly an additive process, in which advancement is due to improvments in the low-level technology of the hardware. Thus the 1940s and 50s are summed up as the era of tubve computing, the 1980s become the era of the microprocessor and so on. We're also given a few names to spice up the narrative. Babbage makes the obligatory appearance, as do Turing, Von Neumann and Grace Murray Hopper.

But Dyson presents a very different approach. Rather than concentrating on the machines that are the exemplars of computing technology at any given time, he chooses to concentrate on the philisophical assumptions that underly the philosophy of computing at a given time, as well as how the availible technology and economics dictated how that philosophy was to be realized. At the same time, he gives us a story about the evolution of life on earth, and the interplay between theories of life, and theories of computing. Not an easy trick to pull off, and yet Dyson's narrative is both rigorous in its science, and compelling as a story.

Most histories of computing begin with the abacus, or some other tallying device, the natural consequence of seeing computers as simply larger and faster arithmatic engines. Even those who begin the story with Babbage tend to focus on the notion of Babbage's machine as a device for calculation, despite the obligatory quote from Lady Ada Lovelace about the potential of the caclulating engine for purely symbolic computing. But Dyson begins his narrative in an unexpected place- Hobbe's Leviathan. For Dyson, Leviathan is the first theory of the emergent properties of complex systems, as well as the first theory of how reliable systems can be built from unrelaible componants- a theory as applicable to biological evolution as it is to the problems of vaccum tube based computers of the 1950s or to the studies and simulations of complex systems that gained so much momentum in the 1980s and 90s. The questions raised are traced both through their impact on computing and on biology. Dyson introduces us to the brilliant and mostoly forgotton work of Baricelli in creating a numerically based "artificial life", reasearch that was going on at the Institute for Advanced Study while Dyson was a child growing up on the grounds.

Dyson also brings a new clarity to the evolutionary debate. As he explains it, the question is not between a theory of creationism relying on faith and evolution relying on evidence; it's much more subtle and complex. We actaully have three possibilities: Design from without, design from within, and selection from random processes. The former is nominally the creationist position, but it's also the position of such people as Rupert Sheldrake. The selection from random processes model of Darwin does have flows, as creationists like to point out, but those flaws do not automatically have to put us in the Creationist camp.

This brief discussion barely touches the surface of this complex and wonderful book. It's certainly one of most stimulating and thoughtful books of the last few years to touch upon the topics of complexity, self-organizing properties of complex systems and related topics, and may prove to be one of the most influential works in influenceing the direction of work in these areas in the next few years.

Darwin Among the Machines is a darn good book. Academically rigorous, but (thankfully) not academically styled. It is an important contribution to AI not in a technical sense but certainly in terms of improving our awareness and understanding-our preparedness-for what so many of our greatest thinkers have concluded is inevitable. I hope and expect we have not heard the last from George Dyson. He says that "Everything that human beings are doing to make it easier to operate computer networks is at the same time, but for different reasons, making it easier for computer networks to operate human beings." There may be no more important job in the world, than to tell people that. I have written a somewhat lengthy summary and commentary of the book at http://ai.miningco.com/library/weekly/aa060897.htm