My NYT review of What Technology Wants

November 6, 2010 • 4:59 am

Up at the Sunday New York Times Book Review—it appears online a day early—is “Better all the time“, my review of Kevin Kelly’s new book, What Technology Wants.

It’s not a bad book.  In fact, parts of it are really interesting: the stuff on the history of technology, for instance, including Kelly’s stint with the technology-dubious Amish. But for me What Technology Wants was seriously marred by Kelly’s relentless progressivism, including his idea that evolution “strives” for certain outcomes (complexity, beauty, specialization, ubiquity, etc.), and that technology strives for exactly the same outcomes.  His view of evolution is unidirectional and teleological, and if that’s not the case, which it isn’t, then his idea that technology and evolution both follow universal “laws” (I think Kelly, a devout Christian, sees these laws as God given) breaks down.

An anecdote: shortly before the book was published, Kelly spoke about it at the New York Public Library. Radio journalist Robert Krulwich was there to moderate, and mentioned to Kelly that, contrary to his thesis, most “orthodox biologists” feel that evolution has no inherent direction.  Kelly blurted out, “Well, they’re wrong!”

The illustration, by Joon Mo Kang, is nice:

47 thoughts on “My NYT review of What Technology Wants

  1. Nice Review!

    I’m almost done with the book, and I’ve been hoping an evolutionary biologist like you would review it, because his sections on evolution set off alarm bells as I was reading it.

    Also, he seems to propagate a version of the “Environmentalist bureaucrats killed millions by banning DDT” canard, although he kind of backs off it a little the last time he mentions it.

    I also found the Amish over-praising a bit much at times. Although here again, to his credit, he does pull back in his final analysis and give a fair and critical assessment. In the end, I wouldn’t have minded except for the fact that he only mentions once, in a single aside, the fact that Amish women don’t get to do anything but play housewife. It’s maybe tangential to his focus, sure, but it’s something I think deserved more highlighting.

    I haven’t gotten to the last chapter yet, where apparently things get religious, so I was surprised to learn he’s a devout Christian. I didn’t find any hint of this in the parts I’ve read so far (although I can now see how his insistence on a direction for everything might fit well with this view).

    But on the biology, the thing that bothered me was that he kept saying, “most biologists think X, but a few charming rogues are starting to think Y” and would then proceed to make his argument as if Y were the unassailable truth. It’s a move he makes several times, and each one feels like cheating to me.

    1. “Environmentalist bureaucrats killed millions by banning DDT” canard,

      ??? But research says DDT cuts down on malaria so if it really was banned it would likely have killed many as malaria is a major killer AFAIU:

      “After correction for endogeneity, then spraying, insecticide treatment of nets, and education about malaria are all independently associated with reduction in incident cases of malaria or fever, while larviciding with temephos is not. The evidence suggests that although impregnated bed nets cannot entirely replace DDT spraying without substantial increase in incidence, their use permits reduced DDT spraying.”

      Is the canard that it was banned? “DDT was subsequently banned for agricultural use worldwide under the Stockholm Convention, but limited, controversial use in disease vector control continues.” [Wikipedia] But then I don’t get the connection to technology progress; the discussion of use being “controversial” would be more appropriate.

    2. Assuming means controversy on effectiveness, not environmental hazard. Ouch if the later, that would be immensely immoral. Then again, people never cease to amaze.

      1. He says it was effective but got banned because of environmental damage. But it’s a more complicated situation than that. It became less effective when used for agriculture instead of just to protect people.

        There’s a whole group of folks who use DDT as proof that environmentalists kill people, and when he first mentions it in his book, it seems like Kelly might be leaning this way. But in his summary of the subject, he pretty much gets it right. It looked to me like something he wrote and then learned more about and then clarified, but which wasn’t edited for maximum clarity.

        It was part of an overall discussion (which was interesting) about how it can be difficult to predict all the effects a technology might have on the world.

        1. Edward Tenner explores the unintended consequences motif in “Why Things Bite Back: Technology and the Revenge of Unintended Consequences”.

          I highly recommend this book, published in 1997, but not at all dated as he uses examples from agriculture, beach erosion and sports as well as technology.

        2. He says it was effective but got banned because of environmental damage.

          That sounds like part a fair description (that far), not the portrayed canard.

          It became less effective when used for agriculture instead of just to protect people.

          Yes, adaptation. That would make the Solomon result controversial on effectiveness.

          Ironically it still makes Kelly right if limitations on disease control followed from environmental concern, which may be the case. Certainly DDT seems useful in the current limited form (i.e. pure disease control).

          This seems to leave the question whether it is canard or not open.

          1. Basically I came away from the chapter with two different impressions. At first he seemed to be blaming bureaucrats a la the canard but his final word seemed more in line with the facts. The book has a number of these moments, where I’m left wondering where he stands or what he’s saying. Mostly thia is nuance, which is good, but too often its just lack of clarity, especially on subjecta like DDT which was an example to support a point rather than a central element of his argument.

  2. It’s His Noodliness! Evolution and technology must be directed by pasta.

    I like Kelly’s scientific argument for evolution directed by a sky fairy: Is so! Is so! Is so!

  3. “If there is a God,” Kelly writes, “the arc of the technium is aimed right at him.”

    So close, and yet….


    Clearly, there are no gods, no heavens. And yet, I will agree that the universe would be a much nicer place with heavens and gods — only, not the nasty ones like YHWH and Jesus.

    Which leaves us with with only one option: if we are to ever see Heaven, it will only be after we ourselves have created it and populated it with the gods we will have become.

    Building a Heaven from scratch is damned hard work, I’m sure you’ll agree. Those streets don’t just pave themselves with gold, after all. At least, not yet.

    …which leads me to the thing I find the most frustrating of all about religion. Those who profess the most interest in Heaven and are the most vocal about their desire to see it one day are also the ones doing their damnedest to hinder the rest of us in building it for them.

    As the saying goes: lead, follow, or get the Hell out of the way. Yet religion always seems to lead people into standing smack in the middle of door #3, scratching their asses as they keep wondering why their pitchforks are always getting stuck. I just don’t get it.



  4. I don’t really get why people always have to end up thinking that there is religious thinking behind seeing that evolution, in fact, have had a direction. The proof is all over the place but we should not confuse that with the idea that it knows where it’s going. I would like to hear proof that shows that things don’t increase in complexity and transcends but includes what came before.

    1. I’d like to hear proof that there is no invisible pink unicorn in my garage.

      Since you’re the one making the (disjointed to the point of incoherence) assertion, what say you provide some evidence for the teleological “evolution has a direction (and presumably a destination)” assertion.

      Should be a no brainer given that the “proof is all over the place”.

      1. I’d like to hear proof that there is no invisible pink unicorn in my garage.

        Well, of course She’s not in your garage. She’s right there, behind you, silently laughing over your shoulder.

        No, not that shoulder — your other shoulder. Erm…no, look, silly. It’s your other other shoulder, right over — oh, forget it. You’re hopeless.



      2. Steve, I think I made it clear that the there is no obvious destination, so why try to apply that to my post? If you can’t see the increasing complexity on earth or even the universe since it started I don’t know how to explain it to you. The problem is, as Jerry says in his review, we can’t really settle it because it is a matter of perspective. My concern is that we always try to look at the details to disproof the directionality but since the closer one gets to the details the more randomness appears. Evolution is a systematic affair, you need to back out to get the big picture which is not always part of the orthodox scientific practice.

        The smart and reflecting mind, that humans have today, would have happened (my believe) regardless if a meteor would have wiped out the dinosaures or not. It would not have been human but still self reflecting. Random selection and coincidences, yes, but still a direction. I cannot proof my theory in any other way then there are hints in our history, which Kevin brings up in his book. Teleological? No! You cannot separate mind and matter. It is one and the same process, in my view.

        1. You claim that the “proof is all over the place” and then you say “I cannot proof my theory in any other way then there are hints in our history”.

          Am I incorrectly seeing an inconsistency in what you claim ?

          And as for what Jerry Coyne said about increasing complexity in his review, it definitely is not a matter of perspective, there are perfectly adequate explanations that don’t personal belief.

        2. Magnus wrote:
          The smart and reflecting mind, that humans have today, would have happened (my believe) regardless if a meteor would have wiped out the dinosaures or not.

          What makes you believe that? The fact that (as far as we know) it only happened once, out of millions of opportunities for it to happen, would seem to argue the opposite. A one in a million fluke that it happened at all, meteor or no meteor.

          1. The fact that (as far as we know) it only happened once, out of millions of opportunities for it to happen, would seem to argue the opposite.

            Indeed. If we take (more or less, certainly sufficiently for a simplest model) independent attempts of “making life” (attempts in local environments over long time) and “human equivalents” by some definition (attempts in species, populations in local environments over long time) it is easy to see from Poisson processes exponential probability distributions:

            – Life is easy to make (stochastically and then likely deterministically, in the sense of rate of attempts et cetera).

            – “Humans” are hard to make.

            Life is so easy to make that the observer effect is erased. The ~ 1 Gy before we observe life (because of tectonics and what not) is precisely within a 95 % certainty interval for what we would expect on a 4.5 Gy old planet if it is an easy process.

            However, the observation of making “humans” are obviously *decided* by the observer effect as of yet, we effectively being the sole “humans” making that particular observation. So it is not only difficult, as opposed to abiogenesis data (shaky as it is), “humans” may well be unique.

            That “humans” would follow from evolution is the same as saying that my 6 on the dice had to follow from my throw. And the evidence proffered for that being the case is as rigorously “evident” as astrology outcome is.

            1. “the observation of making “humans” are obviously *decided* by the observer effect as of yet, we effectively being the sole “humans” making that particular observation.”

              Duh, unclear. I mean, when *we* (or any self-aware descendants) are several Gy old, the uncertainty would decrease towards the level of the abiogenesis observation.

              (Not that it will happen, the biosphere has a mere 0.5 – 2 Gy to go…)

          2. In fact, I just recently had to delve into this on the abiogenesis course I’m participating in. Hah, in fact it was 3 sigma certainty!

            Cutting & pasting:

            “Many texts describe how biogenesis could be a repeated process of attempts over time and locales. In the simplest model this belongs to the family of Poisson processes. Possible caveats is 1) it isn’t stationary after success (since existing life interrupts it) 2) it can succeed numerous times, in which case it is a Lévy process I believe. However for the purpose of this discussion it suffices.

            Such processes stacks most of their their distribution’s probability mass in the beginning. The pdf for the exponential distribution of a first attempt: .

            It is consistent with simplest possible stochastic process fitting the situation that we see a relatively short time. That was all I wanted to say at that time. Any clues why this is wrong is appreciated (but then you may also benefit from reading on).

            This time I was prompted to consider deeper:

            – We can test the model:

            In a normed distribution (observation time t = 1) for a one-sided 3 sigma test we want to have a set of distributions with at least 0.99 of the cdf: F(t,lambda) = 1 – exp(-t*lambda >= 0.99 -> t*lambda > 4.6. t = 1 -> lambda ~ 5, so waiting time T ~ 0.2. With actual time t* ~ 5 Gy, actual wait time T* ~ 1 Gy, which is observed.

            Thus a hypothesis that we see a stationary stochastic process with rate at least lambda = 5 passes a 3 sigma test on the available data.

            This can fail internally or by putting a more predictive hypothesis (since this is the simplest in its class). For the former, the caveats above first: 1) is not a fail. 2) is complicated, but 2.1) if it looks like a compound Poisson the model still applies and 2.2) post-event the phylogenetics looks like 1 success (a universal tree).

            “What if” the observation had been closer to observation time t? _Such_ test wouldn’t have been possible to do with such confidence. That doesn’t affect _this_ test that is to be scrutinized for its failure, not its success.

            – How is this “indicative of easy biogenesis”?

            In principle a stationary process means stationary mechanism in a stationary environment. This isn’t what happened at the start, since volatiles were collected, temperatures and pressures going down and tectonics started. But since it *looks* stationary it means it was close to stationary in some stochastic sense, even if it was frustrated in actuality. (For frustrated biogenesis, see Lineweaver et al papers.) In this sense the process was robust and the environment stable enough.

            Further smaller wait time, which could be even smaller if tectonics had allowed observation, means that there were many parallel attempts. Larger wait time considered above means deterministically either fewer attempts over long time for some reason or fewer successful for some other. Both is indicative of deterministic difficulty.

            – Is there a more predictive hypothesis? What would such look like, say for comparison’s sake if possible as stochastic process?

            Monod was suggested. I can envision the quoted material as a dynamic process; he seems to say that there was dotted i’s and crossed t’s so in some small volume of phase space biogenesis happened, and all (or enough) of a vast space was explored by the process and its environment before that volume was encountered. However I can’t see how that translates to a stochastic process, except that at the surface it looks like the exact opposite of “stationary”! Which doesn’t seem likely if given the above to compete with.”

            Yeah, I know, it sucks to have me as student.

    2. I don’t really get why people always have to end up thinking that there is religious thinking […] The proof is all over the place

      Oh yeah? Is that why TalkOrigins describe it as “a debate”? And how it is a “common misconception” “attributed … to social and religious attitude”:

      “One of the more common misconceptions, with a history long before Darwin, is that evolution is progressive; that things get more complex and perfect in some way. In fact, this view is attributed more to social and religious attitudes of 18th and 19th century European culture than to any evidence. It was a given that things are getting better and better, every way, every day. This persisted until long after Darwinism, until the middle of this century (e.g., Teilhard de Chardin). Even Darwin was ambiguous about it, talking on occasion about ‘perfection’ as a result of selection.

      At the time of the ‘modern synthesis’ [note 9] in the 1940s, the notion of progress was quietly dropped, with a few exceptions like Dobzhansky and Huxley within the synthesis, and Schindewolf and Goldschmidt outside it. Of course, heterodox writers (usually not biologists) like Teilhard and Koestler remained progressionists long after this. But by the 1970s, progress had been abandoned by working biologists.

      Recently, the issue has resurfaced, shorn of the mysticism of earlier debates. Biologist J.T. Bonner argued that there was a rise in complexity of organisms over the long term [1988], and others were arguing for a form of local progress under the terms ‘arms race’ [Dawkins and Krebs 1979] and ‘escalation’ [Vermeij 1987]. Gould [1989] felt so strongly about it he was moved to deny that, at least since the Cambrian explosion, there has been any progress at all.

      Much of the modern debate centres on what counts as ‘progress’.

      1. I never understood why anyone gave credence to Gould’s assertion that evolution is not progressive. Defining progressive as making use of new “ideas” or opportunities, it most certainly is. Each stage in the evolution of a structure makes possible the evolution of subsequent stages that would have been improbable without the earlier stage. Thus evolution is progressive in the same way that knowledge is progressive–each evolutionary innovation “stands on the shoulders” of earlier ones. Of course, this does not imply teleology or any hidden force or direction. It is purely mechanistic.

        1. From Coyne’s review: “some species, like fleas and tapeworms, evolved greater simplicity, losing important structures and organs.”

          Then to paraphrase Newton: “If I see so far, it is because I lose the important structures and organs of others.”

          Is that what you mean?

          1. It is possible that a somewhat complicated structure might evolve to a simpler form, but it may also evolve to a more complicated form that would have been improbable without the intermediate stage. The more complicated form makes possible yet more complicated forms, and so on. But U-turns are always possible, so I like your Newton paraphrase.

        1. Is it my fault that I always read that as “bleepity *bleep*” at first?

          Evolution is locally hill climbing at best, seeing that there is such mechanisms as drift and no outcome has been seen to be perfectly optimized.

          I’m trying to picture processes that are directed without having the process environment (boundary constraints) set up so, and I come up empty. Sure, processes like inflation which are breaking symmetries are “directed”, but without the symmetry they wouldn’t go anywhere. Same for entropy (time) direction and so on, without an expanding universe we would have no 2nd law.

          I assume seeing “directedness” in processes, then there are none known case nor reason to believe there is, is fully “common sense”/religious thinking.

          1. Actually, after more cogitating, I wouldn’t be surprised if there is a (deepity deep) connection between the known fact that laws are inherently CPT invariant and that processes at large seem to be non-directed.

            Unfortunately these are the types of connections that need precisely deep math to be explored in most cases.

  5. I would have thought that it would be easy work to draw an analogy between technology and biological evolution. Technology is a kind of meme, so it has a way to replicate, and is acted on by economic selection. And unlike the case for biological evolution, economic pressure does give a directionality toward the better, at least for technology, and at least measured by humans. From the review, it sounds like the book missed a simple argument like this in favor of magic.

        1. Technology does evolve by artificial selection, and by (lowercase) intelligent design. It has short-term directedness, but this leads in unexpected directions long-term. Alan Turing wrote of remote control of computers by telephone, but who could have foreseen the Internet before 1954 (the year of his death)? In 1967, a lift (US: elevator) engineer mocked the idea of computer control of lifts to me, because “computers can only deal in numbers”.

          One aspect of technological evolution that is underestimated is human inertia. I have a series of pictures of car door handles from the 1920s to the present. In retrospect, the succession towards flushness is obvious, but if a car had had flush handles in 1920, people would have rejected it as too outlandish.

    1. At least it is as easy to make the case for observed progressivism as much as in science. I don’t know why anyone would call observation “relentless”, aside from the fact that it itself need to be relentlessly pursued. 😀

      Even easier I think, as increased effectiveness in using environmental resources is explicit in technology, while (say) papers/scientist is a measure of implicit productivity.

      Technology is a kind of meme,

      Um, no, but techniques may be analogous to memes (I’m fairly sure you meant that). And then I assume by Dawkins (which I haven’t read on that), somewhat analogous to evolution.

  6. A mention of Robert Krulwich implores me to mention his and Jad Abumrad’s fantastic radio show Radiolab. Blows my mind every time. They kind of play off each other, Abumrad seems a bit more of a skeptic, it works wonderfully.

  7. Great review, Jerry–you are so satisfyingly readable; you & PZ both have the talent to write about big concepts so accessibly.

    As a potential reader with my biases (biology training, atheism), yours is exactly the sort of review of this work I’d be looking for; after reading it (the review), I’m still interested in the book but will feel much more comfortable about the reservations it will apparently elicit in me if/when I read it.

    If I were the author, though, I think I’d wish the Times had chosen someone else! Selecting reviewers is such a critical process, and the Times’ motives are often inscrutable. I love it when they get it right like this.

  8. I have an issue with one part of your review. You say that and I quote,
    the wheel is about as good as inventions get, but it didn’t appear in the New World or Australia until brought by Eurasians.

    The lack of wheel in the New World does not imply that the idea of building a wheel did not occur to the people in the new world. As far as I know, the lack of large domesticated animals made the wheel a very useless invention in the New World, until the Europeans arrived, bringing horses with them.

      1. Yes, yes, I knew this when I wrote the piece. I think I mentioned wheeled transport, though. They did have wheeled toys in mesoamerica, but, curiously, they never translated them into the full-sized thing. It would have taken too much space to discuss this all in a NYT book review!

      2. Fascinating link, tomh, esp. the discussion labeled “Why Not Wheels?”

        And it certainly does support JC’s contention above. (Though perhaps just a few adjectives–such as ‘load-bearing’–before the word “wheel” in the review might have sufficed to forestall these nitpicks? At any rate, I think the meaning is clear as it stands.)

  9. Magnus comments:
    I don’t really get why people always have to end up thinking that there is religious thinking behind seeing that evolution, in fact, have had a direction. The proof is all over the place but we should not confuse that with the idea that it knows where it’s going. I would like to hear proof that shows that things don’t increase in complexity and transcends but includes what came before.

    Absolutely, A fundamental aspect of scientific enquiry is the search for patterns arising in our universe and development of consistent interpretations that use minimal assumptions.

    The evolutionary patterns we observe right from stellar nucleosynthesis right through to the emergence of present technologies are so persistent and pervasive as not to be dismissed lightly.

    Just to look at the biological phase of evolution:

    The evolution of species is not a random process. It is driven by random events which produce mutations. Most importantly these mutations are then filtered by the prevailing environment. This is the process of natural selection which gives the development of life its direction, which, in a very limited sense, can be equated to “purpose”

    Problems with this arise because we easily can fall into the trap of anthropocentrism, whereby any phenomenon that exhibits what can be called “design” or “purpose” must involve a reflection of our own particular mental processes.

    As discussed further in my book “Unusual Perspectives” (Ch 10) this is a logical error of the “package deal” variety. Both the watch and the eye (or neither) can be considered to have design or purpose within this model. We consider ourselves to design such things as watches. This arrogance can only be justified in a very limited sense. In actuality, watches have evolved! Albeit by a non-genetic mechanism.

    They are products of nature and we merely the vehicles for their evolutionary progress.

    There is absolutely no need to invoke any kind of “creator” to explain the undeniable pattern that we observe. Nature being best viewed as simply an ongoing machine. No “initial conditions or “final goal” are required

    This and related interpretations are presented in my first book, “Unusual Perspectives”, which can be freely downloaded from the dedicated website

    However, some do find this a hard read and, if skimmed, it can be easily misinterpreted.

    My newly published work “The Goldilocks Effect”, is more concise and deals more specifically with this wider evolutionary model.

Leave a Reply