Transcript of Episode 138 – Brian Arthur on the Nature of Technology

The following is a rough transcript which has not been revised by The Jim Rutt Show or by Brian Arthur. Please check with us before using any quotations from this transcript. Thank you.

Jim: Today’s guest is Brian Arthur. Brian’s a leading economist and complexity thinker. He’s best known for his pioneering work on positive feedbacks and increasing returns in the economy, I.e., what happens when products gain market share, find it easier to gain further market share, and the role of these network effects in locking the markets into the domination of a single player. Things like Microsoft Windows, Facebook, et cetera. Brian was one of the dudes that first saw that as a phenomenon, and explain it from an economics perspective. Brian’s also one of the pioneers of the science of complexity, science about patterns and structures self organize and emerge. He’s a founding member of the Founders Society of the Santa Fe Institute. In 1988, ran it’s first research program. I didn’t know that Brian, that was interesting when I looked up your background. He helped found the Santa Fe Institute’s work on complexity economics. Welcome Brian.

Arthur: Thank you. I’m delighted to be here. You’re a very old friend Jim, I should say. Thank you.

Jim: Yeah, indeed. When I first came out to the Santa Fe Institute as a researcher back in 2002, you and I were actually office mates for a while.

Arthur: Really? That’s great. Yeah.

Jim: You were traveling most of the time, so I mostly got an office to myself which was handy. Sometimes we both be in the office at the same time. We had a few conversations from time to time. Yeah, we go way back. It’s great to have you on the show.

Arthur: Thank you, delighted.

Jim: Well, Brian works in lots of interesting areas today. We’re going to talk mostly about his book, The Nature of Technology: What It Is and How It Evolves. Now, this is a book that’s much talked about, but perhaps not read as much as it should be. I’m going to encourage my listeners, if you find today’s talk interesting, get the book. It’s very readable, full of lots of good stories, as well as some very deep ideas. First, one of the things you discuss early in the book is that, oddly enough, despite the importance of technology in our world, it isn’t really a rich academic field. Theories of technology are not a rich academic field unlike say, the history and philosophy of science, where there’s lots of good mud pie fights going on all the time. Why do you think that is?

Arthur: It’s the oldest thing. I think there’s just a few vague ideas about technology. One, is that somewhere across the way, maybe on campus in the engineering department, people are thinking about technology deeply. Turns out, mostly they’re not. Engineers, I was told by other engineers, are not interested that much in thinking about the foundations of technology. There’s another reason as well, and that’s technology is viewed as kind of ugly sister, a lesser sibling to science. Science is glory. Science creates this science, managers that science, puts men on the moon, not engineers, et cetera. I think that technology has suffered a lot of bad rap. We blame technology for all kinds of things, and we glorify science equally for all kinds of things. I was trained as an engineer and I thought I wanted to really take a good, hard look of technology, hence the book.

Jim: Yeah, it’s really very, very interesting. You’re right, I looked around a little, did some side research, there’s still not all that much about real theories of technology and how it has come to be. We did recently have on our show, Matt Ridley, talked about his new book, How Innovation Works. This hit on some of the things, but frankly it was more a pop history of innovation, et cetera, with a little bit of theory. One thing I would like to ask you about is, he did make the distinction, same one that’s often made at Santa Fe Institute, the distinction between invention and innovation. Invention is something, a primary new thing. Say for instance, discovering that moving magnetic fields can generate electricity. Versus an innovation, the application of something like that kind of discovery to the development of electric dynamos for producing electricity in bulk. You don’t use that exact distinction, any particular reason why?

Arthur: I find that when I got into seriously looking at technology, there were a lot of concepts, invention, innovation, things like that, words that were bandied about, each with a kind of penumbra of meanings and huge amount of vagueness. I got a bit tired of trying to wrestle with those. I just decided I’d strip everything down, only mentioned innovation if it was needed. That distinction you mentioned by the way, it goes back to Schumpeter, Joseph Schumpeter, about a hundred years ago or more. He was fascinated by invention, but he thought the real news was innovation, when an invention enters the economy and gets used. I decided I wasn’t going to write a pop book about innovation. I wasn’t going to worship geniuses and people upstairs in attic, inventing things. I wanted to look at reality. I read a lot of lab books and I talked to people. I did a lot of homework on this. I was trying to look at reality and bring back whatever news I could from, what’s still a very much a frontier field.

Jim: Very interesting. At least what I took away, what you substituted instead, is actually something very interesting. You start at the base with phenomena, that the capture of phenomena, meaning sort of things from the world of physics is your base. It’s interesting. I’ve recently been doing some research on where hydrogen fits in the carbon neutral economy. During the same time, I was rereading your book for this podcast. As it turns out, one of the core phenomena that’s involved in hydrogen as part of the carbon neutral economy, is all how electrolysis is done, how water is split into hydrogen and oxygen to capture the hydrogen. Now turns out, there’s four different ways that this catalysis can be done and each one harnesses a different physical phenomena. Your use of phenomena, as sort of the base to build from, frankly I probably wouldn’t have thought of it quite that way if I hadn’t been reading the book at the same time. Maybe talk us through a little bit about what you mean by phenomena, and what it means to sort of capture, encapsulate it.

Arthur: Sure. I must say I had a real blast writing this book. I thought it would be two years to write. I started, oh, around 1997 and the book actually took 12 years to write on. That was because as I started to look at individual technologies, I discovered I knew quite a bit. I had a top class engineering degree. I knew a lot about technology, but I wanted to know very well about two dozen different technologies. I mean, being reasonably expert in at least a dozen of those and the history of them and who had done what. As I started to look into different technologies, not just casually, but very, very deeply and read and read and read several books and plunge into them and look at research notebooks and stuff, I began to realize that they all had this common base. They’re all based upon one phenomenon or more usually several.

Arthur: I got fascinated by things like dendro chronology, basically the art of dating things, dating pieces of wood like lintels, archeologically, by looking at the tree rings. Tree rings have a pattern year by year. If it’s a wet year, the pattern is different from if it’s a dry year, as the rings go out. You can nail down when a tree fell, to be used in some archeological site, maybe 2000 years ago, to the exact year, pretty much. I got fascinated by this. I discovered that every technology I used, was based on some phenomena and usually funny enough, many phenomena. I poetically began to describe technologies as orchestrations of phenomena, or programmings of phenomena. It’s as if you’ve got all these different phenomena in your toolkits, electrical ones, ones that are more physical, like in construction systems. You’re reaching into that toolkit to put several of these together, to use for a specific purpose.

Jim: Yeah. I like the fact that you manage the ground map all the way back, one of mankind’s earliest technologies, the hand ax. Basically, a rock with a sharp edge on it. What’s the phenomenon, it’s two, right? Transfer of momentum and the solid state physics of the materials, right?

Arthur: Exactly.

Jim: Okay. Your theory works. You can take it all the way back to the hand ax.

Arthur: One of the real kicks I had was reading history. I tried to read up close individual accounts. I started to read about Roentgen, who had discovered x-rays, as you know, accidentally if I recall, around 1895. He started to experiment with these rays using Crookes tubes, now we call them cathode ray tubes, by putting wood in between the photographic plates and the emitter. Putting lead in between, he got no image. Then, he persuaded his wife to put her hand in between. He got this perfect skeletal picture of her hand showing her wedding rings and presto, there is a phenomenon now by which you could look at the internal structure of limb and see bones and see them very clearly.

Arthur: Within about three months of publicizing that, suddenly people were trying to do that themselves and radiology or x-ray systems were born with that one thing. I conjectured one time to an audience that his wife must have died of cancer. It was kind of a, maybe a rather a dark joke. I was just, “Oh, she must’ve died of cancer.” Before I finished the sentence, almost a hand goes up and says, “Yes, she did.” There’s a human story behind all of these inventions.

Jim: I recently watched the movie about the Curries. They both died from radiation related phenomena.

Arthur: Yep, as did many people in Los Alamos. I remember our mutual friend, George Cowan, who started the Santa Fe Institute. George told me at about age 80 with a great chuckle, he says, “I’m the only one left alive from the Manhattan project.” He said, “I should have been dead 10 times over with all the radiation I got, but I survived it.”

Jim: Yeah. That’s a probabilistic thing. Let’s move on to the next part, which I think is an interesting insight that you had, which is the relationship to an economy and its technologies. You kind of reversed in some sense, the usual conversation.

Arthur: First of all, digging back a little here, trying to get at these ideas. I was taught economics at a PhD level actually in Berkeley quite a long time ago, but I don’t think subject has changed. We were basically told that an economy was a system of exchange, usually with well-formed markets and prices and so on, a modern economy. The economy created every sort of technology. You’d have little factories and firms, possibly making steel. The next thing would be, maybe 1860 and then somebody comes along with a better process. Bessemer, for example, now there’s Bessemer steel. We can produce better steel in bulk. The emphasis was always on the economy as a kind of container. From time to time, it would create new technologies. You could sort of slide these out of the machine, like a unit in a fancy digital machine, and slide in the new technology and the economy would be better off.

Arthur: That’s the story we were given. I don’t think it’s wrong, but as I did more and more research, I decided that something different was going on, not quite independently. A fuller picture was this, it’s not that the economy creates technologies, certainly it does, but that’s not the full story. A much rounder story, is that technologies create the economy. Here I would go back to really good economists before the 1870s, like John Stuart Mill, Karl Marx, at least as a good observer of the economy. I’m not that keen on his ideological stance. These people observed the economy and realized that an economy came out of its means of production, that was the phrase used. In other words, an economy is built around its technologies.

Arthur: If you’ve got water wheels and grain mills, later on if you’ve got railways, the economy forms around the technologies that are in use. They’re the skeleton of the economy if you like. The brains might lie in the market. The sinews might be the actions human beings take, et cetera. It’s the technologies that give you the economy. They also bring along with them the eras. We have an era of railroads and the era of electrical machines and the era of canals before the railroads, and so on.

Jim: I was thinking about that as I was reading it. I decided, let me play a thought experiment. Supposed we go all the way back, right, not all the way back, but a long way back to one of the greatest transitions in human history, actually pre-history, was the development of settled agriculture, right? I go, “What’s the phenomenon that was there?” I said, “Basically, it was a kind of a folk genetics, right?” They didn’t call it genetics. They didn’t obviously know about genes. They didn’t know about DNA, but they somehow or other, stumbled into a low tech technology of selective breeding of plants and animals to allow for high enough yields to make it worthwhile to settle, rather than to continue to forage. I go, “That seems to work.” You talk about cultural eras. Settled agriculture, we’re still in the epoch of settled agriculture, interestingly enough, right? Then I was thinking about some others, the automobile turned into the suburban era, which we’re still living in here in the west. Those arguments made a lot of sense to me.

Arthur: Yeah. There’s a wonderful economic historian called Carlota Perez.

Jim: Yes, I’ve actually had a guest on who was a student of hers. We talked about her work a fair amount.

Arthur: I’m a big admirer. She points out that these technologies, or I would call them bodies of technology or domains, these big collections are clusters of technology that bring you electrical systems or chemical systems, or later on electronics or railroads. These are clusters of technology. They define new areas and they over time, the time would be decades, they actually redefine society. We’ve had these clusters coming in, all the way from the start of the industrial revolution, which would be around the 1760s. We had canals coming in. We had waterwheels and then by about the 1820s, steam engines. Then canals were ousted in the 1840s, 1850s, by railroads and so on. Each of these, like you’re pointing out about suburbia, transform society.

Arthur: It’s remarkable that in the horse and buggy era, you know where I’m standing here in Silicon Valley. I’m in Palo Alto, it was a string of little towns 120 years ago, along one railway line. Menlo Park, Palo Alto, Cupertino, et cetera. Now, we have a conglomeration of those with the automobile. Suddenly the game changed, we could get from a town to what became the suburbs rapidly and quickly. Gasoline was relatively cheap. We didn’t have to get on a horse. We didn’t need a chauffer. We could do it ourselves. These bodies of technology or clusters of technology, not only do they create the economy, but when new ones come along, they recreate the economy. In doing so, they create a new way of being, a new form of society. I think Perez got all this absolutely right. I was greatly influenced by her work. If I recall it’s, Technological Revolutions and Financial Capital, if I remember the title well. Brilliant book, I recommend it very highly to people.

Arthur: I think what I would like to emphasize, bit of a long answer here, but is that the causality goes from some need. We need to transport goods more easy, then some cluster of technologies, railway, locomotives, rails, better steel, and so on, giving us a cluster railroad technology that is used by industries. The industries get transformed by the new cluster of technologies, and that in turn transform society. Notice that at the back of all of this, technology is driving. It’s in the driver’s seat. This is what economists talked about and notice before 1850, 1860, sad to say that when analytical economics came along, it liked to keep everything constant and static. Technology went out of fashion for theorists, and wasn’t much talked about.

Jim: I’ll insert this here. I have it later in my notes, but that’s a perfect time to get to the Schumpeterian perspective and contrast it with the equilibrium economics, that somehow became all the fashion in the economics profession. Just for the audience, if you could actually describe the equilibrium economics that is still taught an awful lot. I took a couple of courses and three economics courses when I was an undergraduate, including one from Paul Samuelson himself, and a small class of 20 students. It’s all about equilibrium and yet, it’s not.

Arthur: Yeah, that’s an, “Oh my God,” right there, that you were in the same classroom as the great Samuelson. If you’ve gone to Harvard, you might have encountered Schumpeter. They weren’t that far apart. There’s an odd thing happened in economics. Let me see if I can encapsulate this fairly quickly. People looking at the economy, Adam Smith, David Ricardo, Mill, Marx, line up the usual theorists, they had a very keen idea of the entire interrelation of markets and prices appearing in those markets, that gave the players incentives, the purchasers and the consumers, to save and to produce and all this sort of thing. They were also aware that none of this was static, that the economy was changing all the time. Around 1870, algebra came in as a technique or a tool and calculus. Top economists, really good ones, Jevons, Alfred Marshall and others, started to look at the economy mathematically. It was difficult to do that with having new things coming along. If you have X, Y, and Z, a year later, you could have X, Y, Z, K, L, and M. This didn’t suit algebra.

Arthur: … the KL of n. This didn’t suit algebra at all. So economists started to imagine; let there be n different types of goods, n is fixed, and there might be an improvement on making n, but we’re not going to talk about really new things. We’re not going to talk about changes in structure, we’re not going to talk about the economy in non-equilibrium. We’re going to nail it to a board like it would with a butterfly and try to figure out how it flies by staring at its wings. I’m being a little bit facetious here. But there was a drive away from looking at the economy as changing and evolving. Let me pause here and just bring this to Schumpeter. Joseph Schumpeter had grown up in Austria, mostly in Vienna, was fascinated by how an economy formed and he drank in this equilibrium static framework, he even wrote a book about it in 1906 and presented this with great aplomb to Walras who was the chief equilibrium theorist at the time. He was getting old.

Arthur: So the young Schumpeter travels to Switzerland, knocks on the door of the great Walras, who’s now quite old, and unannounced and Schumpeter says, “Here I am. I am the person who sent you the book about economic equilibrium.” The great Walras at the door and he says, “Oh”; Schumpeter was 26 at the time I think. Walras looks at him and thinks, “This guy’s very young,” and he says, “Thank you very much for sending me your father’s book, it was magnificent.” Schumpeter says, “No, no. I wrote that book.” Walras says “Yes, it was a wonderful book and congratulate your father.”

Arthur: Anyway, Schumpeter goes back to Vienna, totally pissed. That was Schumpeter finished with equilibrium, and he spent the remaining decades of his life showing that the economy was always changing, always discovering new structures, new technologies, and it was nowhere near equilibrium. It was a platonic ideal, this idea of equilibrium, but it was adopted for convenience. When Schumpeter tried to say otherwise, at least in the English speaking world, this was not a popular thing.

Jim: Yet that’s so strikingly at odds with history. Some other work that I’m doing; we go back and track the trajectory of social evolution since 1700. In 1700, people forget how amazingly primitive our world was. Most people lived in dirt huts without any glass in the windows, maybe some animal hides or something like that. We basically had only some of the old technologies, things like gravity for water systems if we were lucky. The Romans had better water systems than the big cities had at 1700. You had a bit of water power; we knew about siphons, simple machines like the wheel, folk genetics like agriculture, but the scope was pretty small.

Jim: Today we can fool around with DNA using CRISPR, we’ve harnessed quantum effects, and things like hard drive heads and increasingly in chip design, and now, of course, quantum computing. Things like general special relativity are used in products like GPS. It’s been all in 320 years, utterly amazing. It’s kind of astounding that economists being right in the middle of all that somehow didn’t notice it or didn’t think it was of the essence.

Arthur: I remember when I was very young economist, I was sent to Nepal. It was 1975 and Westerners hadn’t been there that much, they hadn’t been allowed in. It was essentially medieval, and this shocked me. It was summer in the Himalayas obviously, and Kathmandu… It’s quite some altitude and not that warm, and yet their wasn’t glass in windows, as you mentioned. For me, this was a sense of wonder, it was like going back to the year 1400 or so.

Arthur: I had similar experiences in the Middle East at the time, and Syria and Jordan. Now things have caught up a lot, obviously both in Nepal and there, but it’s a marvel. What has happened is totally marvelous about technology, and I think it’s created our world more than anything else.

Arthur: One thing I do want to say in fairness to modern economics, it doesn’t say there’s no technology. It just says, we’re going to take a series of snapshots. We will look at the economy in equilibrium before the Bessemer process. [inaudible 00:28:19] we got that snapshot, everything’s static, and so on. Bessemer comes along, the equilibrium shifts. It’s like watching a very slow motion movie. Is it wrong? I don’t know, but it’s not to my taste. What Schumpeter brought along was to say; everything’s in motion, it’s all in motion, things are changing. There’s new things made out of the old and those new things come along and there’s no way where you could halt the process very plausibly, and say, “This is static.” It might be static for three days. But as you know, in fast moving markets say, for oil products, that they’re not even [inaudible 00:29:05] right there. It’s an analytical convenience. And I support it as… you’ve got to start somewhere, but it doesn’t mirror reality when you really do dig into technology.

Jim: From that we move from Shumpeter towards your own theories. I think you guys both share the view that every technology has to come from already existing components. Maybe talk a little bit about that, how that works.

Arthur: It’s an odd thing. I was working on increasing returns or what’s now come to be called network effects in the early 1980s, and I began to read a lot about technologies and that kind of resurrected my engineering background. I began to notice… I started to ask myself, then I looked at radar, I looked at continuous wave radio. I was trained as an electrical engineer, but I looked at other technologies coming with penicillin as a therapeutic technology, things like that. I began to notice that all of these technologies had something in common. They were all put together, or constructed… think of a radar set or a television set. They were put together from components that already existed. Those components that already existed were also technologies, means to purposes and this enthralled me.

Arthur: Any technology… say take a sophisticated one, like the F35 fighter jet, which is current, you’d find that it’s put together from parts and assemblages and [inaudible 00:31:09], that are themselves technologies. Those technologies… sub-technologies are put together from other parts and components and sub-assemblies; they’ll have their own purposes. I began to realize that jet engines were not made of jet engines all the way down, but technologies were combinations of other sub-technologies and sub-assemblies that were combinations of further technologies all the way down, maybe 6, 7, 8, 9 levels. I’ve found that very satisfying and really interesting. I began to research this idea that technologies were combinations; that’s usually associated with Schumpeter, but Schumpeter wasn’t really into technology. He was into what he called Means of Production. His uncle, I think, owned a factory. It was a guy called Robert Thurston in 1880, writing about steam engines that mentions that anything like a steam engine is a combination of other technologies that already exists.

Arthur: It’s an odd thing. If you think of a steam engine, you can imagine a big lumbering one in those days in the 1880s, it’s combination of boilers, linkages, at least one system where the piston cooling equipment as something that will regulate the speed at which it’s turning over, et cetera. It’s a combination of many technologies, each of which as other sub-parts. So I find this fascinating. It’s an idea that occurs and occurs. To me this was an original insight, but I quickly discovered it goes way back. I think that it’s something that gets rediscovered every 20 years or so; that technologies are basically combinations. This got me thinking about evolution. Let me pause here. If you want to ask about that.

Jim: No. Why don’t you continue? And I guess one of things I would steer you towards is kind of evolution happening at multiple levels.

Arthur: I began to realize that novel technologies are created out of parts that already exist. They have to be created from somewhere. It’s a bit like saying some new computer program, some new application, if it’s written, say in Python or Linux or something, is created from software parts that already exist. You put them together in some logical combination, they do something. The kicker in this whole thing that nobody had seemed to point out before, is that when you create a novel technology, it can be used later as well. So it becomes a building block for yet further technologies to be made up out of. So when we began to create things like a laser, round about 1960, there wasn’t much of a use and envisaged for it. Now lasers are components and all kinds of things all the way from computers to smart phones, to GPS devices, et cetera, they’re indispensable.

Arthur: Those devices themselves are components. GPS is a component in the larger navigational system, say onboard a freighter ship, et cetera. I began to realize, and this was the key breakthrough for me, that the whole collection of technology, the whole shooting match, the whole numerous thingies we call technologies, built out… in a way built themselves, builds itself from itself. Think of it this way; there’s a big collection of technologies. The year is whatever, 1800, and novel technologies; we have cog wheels, we have levers, we’ve pulleys, we’ve ropes, we have all kinds of stuff and we can put all those together and make a steam engine. Now with a steam engine, new things come possible, like steam locomotives with the railway system, yet newer things become possible. I realized that technology was more like a chemistry. The collection that we call technology, in general, is more like a chemistry.

Arthur: There’s many different types of molecules. There’s many reactions possible, and occasionally we get new compounds formed and we can toss those back in the chemical system and those become new components for yet further technologies. It’s as if you had this Lego set, it’s a metaphor that isn’t particularly perfect, but your Lego set consists of a few small yellow blocks that can… and there’s a few green ones of different shape, and there’s a few red ones. You discover in the course of using these, that certain combinations are useful and they repeat, and you get tired of putting these together and you say, I’m going to call this particular combination, keeps repeating, I’ll call it some new name. I’ll throw it in the oven, I’ll fuse it together, I’ll make multiple copies of this. Then you throw this new combination back in the Lego set.

Arthur: So the Lego set keeps expanding. Or if you like technology, the whole collection creates itself out of itself with an awful lot of human help, I should add. You can go from not having too many pieces, say in the Egyptian era, two, three, 4,000 years ago to having many technologies now and considerably complicated ones.

Arthur: I began to realize that’s how technology evolves. Darwin’s mechanism doesn’t give you properly, a theory of evolution for technology. If you want a theory of evolution, you’re looking to say something like; The technologies of today descend in some plausible way from the technologies of a thousand years ago, and so on. A lot of people have tried that as a theory of evolution for technology, they said that there might be small variations in different technologies and every so often that would give you something startlingly new that could be useful.

Arthur: This is very much Darwin applied to technology. It works to some degree, you can get different forms of bows, historically different shapes. And maybe some of those were cooked up by randomly or by luck. Then you suddenly discover you have something better. Then that becomes a new thing for use in further applications.

Arthur: But it left too many gaps, and I wasn’t happy with that. I began to think, where did the jet engine come from? Around 1927 to 1936, both in Germany and in England, it wasn’t from varying air piston engines, I can tell you. Similarly, about the same time radar didn’t come from varying radio circuits and magically getting something different. The jet engine works on a totally different principle than air piston engines, radar works on a different principle from standard radio. They came out of looking at how particular needs could be satisfied. In the case of the jet engine; How do you fly in thinner air? Maybe it’s 35,000 feet instead of 4,000 feet. You can’t get a piston engine to work that well, it doesn’t breathe well at that altitude. So we needed something different and Frank Whittle and Hans von Ohain invented, pretty well simultaneously, the jet engine.

Arthur: Similarly with radar, people in Britain and in other countries started to worry; if somebody is menacing us with armors in five years, it was around 1936, how can we detect fleets of bombers coming, say across the channel, you could do that acoustically with people with very sharp hearing, listening with big concrete air trumpets, so to speak, but that wasn’t good enough and radar got invented. It’s very much a technology creating its new parts and new pieces and adding those to the collection of building blocks for the creation, in turn, of yet further technologies.

Jim: I think I’d like to highlight, one of the key distinctions between the evolution of the technology and let’s say, biological or Darwinian evolution is that evolution of technology while there is some random variation, look at the different manufacturers of cars, almost [inaudible 00:41:37] in some, in short term sometimes seems almost random, but especially to these bigger examples, these transitions they’re driven by a human purpose or, our human need. So when you focus on; I need an airplane that can fly above 35,000 feet, and if it’s a little faster, that would be even better. Then focusing on that need causes human ingenuity, engineering skills, et cetera, to look fairly far afield, and as you point out, in the history of the jet engine, while it was for purposes of aerial flight, radical, it used some technologies that already existed such as compressors, et cetera.

Jim: I’m sure it used some existing metallurgy though it may have had to invent a little bit along the way. It’s this concept of purpose or human need that molds the trajectory and makes it what it is, and of course, makes it much more efficient in the time sense than Darwinian evolution what takes billions of years to accomplish stuff.

Arthur: I completely agree with all you’ve said and I think you’ve put it very succinctly. One thing I want to mention here, in passing, is the whole idea of invention. We typically think of a new technology, take radar as being invented, or maybe the jet engine being invented. That brings up romantic pictures of people in attics, experimenting and being alone and maybe a semi outcast from society, kind of like Isaac Newton fiddling around with alchemical apparatus and things like that, the genius inventor. All that you said is absolutely on target, but I began to wonder exactly where these novel, radically novel technologies come from. They don’t come from Darwinian variation. They come, as you said, from felt needs. I began to realize there was really no theory of invention, at least on all the research I did over 10 years, I couldn’t find any that I found very satisfactory.

Arthur: [Muslin 00:43:55] said it’s from thinking intensely and then a creative act happens. But that begs the question; what’s the creative act? I was struck when I read Darwin’s book backwards and forwards. Naturally I realized Darwin was not that interested in evolution itself. The descent of species from previous versions of species, he felt that to be well established, he was interested in where new species came from. So it was the origin of species that interested him and through all these little incremental changes that might happen, say snails on two sides of a cliff system in an island in Hawaii, they might diverge a bit until, after a long time, there might not be able to interbreed and you’d have a new species. It’s all very gradual. I began to realize that for a theory of evolution of technology, I needed a theory of invention.

Arthur: So I read, and I began to see, very clearly, that in all cases of invention I looked at, it was basically not a matter of genius, not a matter of being an outcast, or whatever, it was always a matter of problem solving. I began to realize it wasn’t that mysterious. It is wonderful and you can admire the ideas people come up with, and as an engineer slash mathematician, I’ve had my share of insights like anybody in science. So you could admire that, but I began to realize that it’s past the awe and wonder, it’s not that mysterious or magical. I began to realize that invention in the cases I…

Arthur: To realize that invention in the cases I was looking at was a matter of having some problem or need. How can we detect enemy aircraft? You begin to look for some principle, maybe a team does, maybe some smart people do, maybe several teams, maybe several different countries this is happening. And you think that there might be this principle or that principle. In the case of radar, one principal’s stood out, and that is that people already knew that if you had high frequency, radio bings pointed at anything metallic, say 20 miles away even over the horizon and you’d get some reflection or distortion of those beams, and if you could pick up an echo of the distortion, you’d be able to say there is something there that’s metal likely in the aircraft. So you begin to see some principle, but it’s not that easy to make that real.

Arthur: And you think, well, how can I make this principle work? I could do it using these components that are sitting in my toolbox. However, that brings up subproblems and then you have to solve the subproblems, which might have further subproblems. So you’re going back and forth. There’s nothing much more mysterious than this. I began to realize if you were sitting in Palo Alto where I work and you worked in the city, San Francisco, and normally you might drive in, but one day your car’s in the shop getting repaired. How do you get to San Francisco? Now principals would correspond to sort of thinking, “Wow! Maybe I could get my spouse to drop me off at these railway station, and then I can take Caltrain into the city. Maybe I could get up early and get a ride with a friend, on the other hand, that would be inconvenient for him. Maybe I could do this. Maybe I could do that.

Arthur: So what you’re looking for is an overall principle, and then you’re looking for a concrete way to realize the principles, to make it real. And then you’re saying, “Well, okay. But if my wife were to drop me at the station, that would inconvenience her and she’d have to get up a bit earlier.” So you’re trying to solve all these problems. Essentially, invention as far as I can see works like that. Sometimes you’d have a brilliant insight into a phenomenon. “Oh yeah, we could use x-rays in such and such way. But I would, we build a proper machine, proper technology out of that to look at bones.” I began to realize also like another metaphor that maybe you’re standing in the Himalayas and maybe it’s around 1950 and 1960, and you’re wondering how to climb Everest or K2, say, “Oh Carrie, we could go up the Western call and we can make this approach. Some people have pioneered that and we could then cross over to such and such a ridge and then go from there.

Arthur: So you’re putting together an overall route, but each of those overall routes consists of sub paths or sub routes. Each of those has its own subproblem. There might be ice falls, there might be rockfalls, there might be cliffs along the way that you have to wonder how can I do this bit? So in all the cases of invention I was looking at, turned out that you’re trying to solve a problem. Finding an overarching idea, or principle, will bounce high-frequency radio waves off something out there and pick up the distortions. But then how do I make that real with real components? And that’s where combination comes in.

Arthur: And then that brings subproblems. If you bounce a high frequency radio waves off something 20 miles where they’re going to be large and loud, how can you switch that off to hear the echo and time the original waves or drown out any echo? So you’ve got all these subproblems and that’s why it takes a lot of expertise, not so much genius, but familiarity. And going from there, some people are very good at this sort of problem solving. Bit like doing Sodoku. I can do this if I could do that. I could do that, if I could do the… You’re looking problems up and down the hierarchy.

Jim: Absolutely. In fact, in my own business career, they were… We did, I think, two actual inventions along the way, and one in particular. And it had just that attribute, okay? We had this problem, which nobody had ever solved. It was not unsolvable, we didn’t think. So we’d go down this road. Well, that doesn’t… Oh, there’s this problem. We didn’t realize. This broadcast data satellite network that we’re going to use has a higher error. It doesn’t have a higher error rate, the error rate has a statistical characteristic that the vendor didn’t tell us about. Either errors were clustered and that would screw up our applications. That way I had to figure out an error detection and correction scheme, and then we had to figure out how to do it in a cost economic fashion, and then we had to work it into a bunch of other stuff.

Jim: So it was like, okay, you go down this, you stymie it, you backtrack, you go up there, and then of course in the business case, it all had to work for the price we could afford to pay for that sub system. So there was this outer constraint. If I can’t do it for… I turned out $200 a month per customer, it made no sense in the context we were doing it, so we threw out some solutions that would work, but they couldn’t have done it for that price. So, it’s this trying this, trying that, backtracking, throwing it away, discovering problems nobody even knew existed because they’d never tried to do this before. Very much like doing Sudoku. That’s a perfect example. One of my favorite examples, and this will then lead to my other favorite topics, is the Wright brothers, right? Because in some degree, they were those cranks you were talking about, right?

Jim: Two guys, right? Neither of them college educated. Though smart boys with lots of practical knowledge of how to do things, worked more or less secretly for a number of years and solved one problem after the other including ones people didn’t even know existed. Like whoops! We have to be able to steer this airplane, not just get it off the ground. And they tried one thing, they tried another, and they eventually got it to work.

Arthur: I particularly like the Wright brothers as an example. It’s often said the Wright brothers invented modern aircraft. Not quite. The whole idea of having a kite flight contraption was, Arthur, before the Wright brothers and people like Langley and others had been looking at this a decade or so before them. What hadn’t been solved was the subproblems that you were talking about. Nobody knew. Everybody said, “Okay, you could have fixed wings, you could have something that might resemble a fuselage.” But number one, how do you control that? Secondly, how would you power that if you wanted to keep somebody up there for more than 30 seconds? You could use an internal combustion engine, but they were heavy. So how would you do that with an aluminum engine. How would you… What sort of wing surfaces would work? What sort of propellors would work and control systems?

Arthur: So what the Wright brothers did was they solved the key subproblems and they did it by… Not so much by theory, but by trial and error. They were smart as a whip, as you said. Minds like steel traps. They weren’t afraid of experimenting, they were bicycle mechanics at heart. They did construct a lightweight engine to put on their machine and so… So they solved the four or five subproblems, and at least once those sub problems resolved, then you had a primitive aircraft around about 1903 and a few years after that. So I think they’re a very good example. I wouldn’t say they’re the inventors of the aircraft. They’re the people who made it possible though. More than anyone else.

Jim: Yeah, and I’m going to jump in on one point because this is where I thought I wanted to transition to next. That’s why I love the Wright brothers as this transitionary epoch. The light motor was important and the internal combustion engine, wasn’t very old at that point, right? They’d been designed for tractors first and then cars. But cars were big and bulky, and weighed wait an extra hundred pounds, didn’t matter that much, but it made all the difference in the world to get off the ground. But I’d like to focus more on the lift and control surfaces.

Jim: Interestingly, there was very little science about this at the time. But, Wright brothers did write to hundreds of people all around the world who’d been trying various experiments and very crude stuff surprisingly. And so, getting man off the ground, was actually done more or less old school tinkering with relatively little insight from science. So I thought that this would be a great place to open it up to you to talk about the relationship between science and technology. And again, we know for instance that the steam engine classically, was developed before there was really any knowledge about heat engines and in fact, it was the invention of the steam engine that focused the scientific community on understanding the dynamics of heat engines, which led us to the whole field of thermodynamics. So I’ll open it up to you to talk about the intricate and sometimes surprising relationship between science and technology.

Arthur: It’s a topic close to my heart because I’m sure like you, Jim, I absolutely adore science. I love science. At heart, I think I’m an engineer and so I’m fascinated by this connection between science and technology. What I’d say is this, that I read a lot. Obviously, I read and read because it took me a dozen or so years to write this book or at least to do the research for it. And I began to see definitions of technology as… In fact, there was a very good engineer, late engineer called [Truckcell 00:57:06] who defined technology as… He said, technology pure and simply is applied science. Now, I love technology, I love science, but that didn’t ring true to me. In fact, a lot of people think science comes first and then technology. Actually, it’s the other way around. I began… As I got into this more and more deeply, and I hope your listeners are not offended, but I wanted to look at the truth.

Arthur: We’ve had technology, as you mentioned for 11,000 years. In fact, probably for 300,000 years, if you take human language, or fire, maybe 80,000 ago, they used fire. Things like that. We’ve had technology a long, long time as human beings. Science is relatively new. You can find Greek ideas of course, but modern science dates roughly from the 1600s or so. And I don’t quite know how I would want to define science, but just as a approximate definition, I would say, it’s the close observation of nature and study of nature’s phenomena. Something like that. But if you look at that definition, what is this close observation done with? Turns out it’s done with technologies or we call them instruments in science. So when Galileo, in January 1610 starts to… He gets the idea for telescope, he makes a zone. He goes up into this room and he starts to look at Jupiter, partly for astrological purposes. Jupiter was moving in retrograde at the time and he wanted to see if you could measure that. He noticed four fixed stars behind Jupiter that nobody had seen because he had better magnification and need good eyesight.

Arthur: And lo and behold, the following day, the stars have moved a little bit and it moved again the next day, and he came to this, ‘Oh my God’, or ‘holy shit’ moment is, “oh my God! These things are revolving around Jupiter. They are moons. Oh my God!” So with the coming of the telescope, we got a new era in astronomy. With the coming of the microscope a bit later, we got a new era in biology. We could start to see bacteria, germs and so on. With the coming of electron microscopy, if I got that right, we saw yet other things. With the coming of x-ray crystallography technology, using certain type of mathematics called ladder technology as well or method, these technologies brought the ability by the 1950s or so to look at the structures of organic molecules. In particular, a 1953 or so, the DNA. And so, new sciences…

Arthur: I don’t want to play down the amount of brilliance or insight of people like Crick and Watson, Galileo and Newton. But, science builds from observation and from thoughts and throw in genus, if you want. A deep insight. But science builds in no small way from its technologies. Now modern technologies, you mentioned quantum computers, you mentioned other technologies. GPS that responded to special and general relativity. Technologies for sure are guided by science, they’re designed using scientific principles. Certainly, last 100 to 150 years wasn’t very true before that. So, technologies, sophisticated ones are built out of understanding phenomena that science has uncovered. That science has uncovered and understood those phenomena using technologies or instruments. So there’s what I would call a lovely cycle, positive feedback cycle. Technologies give science new instruments and new methods, science gives technology understanding of a new phenomena and technologies benefit. It goes in a circle. We can’t say that one comes first, the other come second. They are mutually reinforcing and they both beautifully mutually supporting.

Jim: Yeah, I think that’s really important for people to understand, is that science is a creature of technology. I had this guy, Michael [Strobinson 01:02:44], who’s a philosopher of science back in January, really interesting guy. And he always talked about the theoretical cohort, which he gave the famous example of verifying general relativity with the expedition to the South Atlantic. And he talks about all the things that went wrong with that, and how dependent upon it was, to how the telescopes were mounted, and how the film worked or didn’t work, and then the mathematics of error correction of which there were some dubious games played and all this sort of stuff. How the science is never as clean as you think it is. It’s deeply and intimately involved with the technology necessary to make a reading, to gather the data, or to do the experiment.

Arthur: Yeah. You don’t want to look too closely into the sausage machine in some cases, yeah.

Jim: Yeah, even when that went up, and he goes, “Well, the guy got the right answer, but for the wrong reason, bro. It turned out his instincts were right though was science was a bit bogus, right? Okay, let’s move on to another. We’re getting short on time here, unfortunately. So it’s an conversation. Another really important part of your theory, is that evolution or change happens at multiple levels simultaneously. And one example, I think it’d be kind of fun to explore is the automobile. At one level, a car is not much different than when I was 10 years old and first became obsessed with them, 1964, right? They look sort of similar, they do the same thing, they’re not a hell of a lot faster than they were then. Better gas mileage, et cetera. But back in those days as a 10 or 11 year old, I could actually work on the car.

Jim: I remember changing the engine in a VW with my father, I remember replacing the generator regularly, adjusting the points and changing the spark plugs and all that sort of stuff. You try to do that today, boy, you’re in for a major disappointment, right? You open up the hood of a car today, you go, “What the hell?”. Essentially, every single thing in the car has changed since 1964. And I was thinking about that, I go if they would change for their own reasons, let’s think about electronic fuel injection. Used to be, the carburetor was the weak spot. One of the idiot skills I developed when I was 15, was how to rebuild a carburetor on one of my father’s 100 dollars piece of shit cars, right?

Jim: The carburetor was always the weak link, right. And it was like a three hour job to rebuild a carburetor on a 65 Chevy that had 120,000 miles on it. But it was a pain in the ass, and that was kind of an archaic skill and it was real expensive. If you had the store-bought mechanic doing it, we certainly couldn’t afford that. Then somewhere along the line, people realized that was a weak link and it was also not so good for fuel economy. And so we came up with an electronic fuel injection, which then became a subsystem, and then we got electric power steering came along, first electric assisted and then eventually, fully electronic steering. People bitch about that on lack of road field and what have you but frankly, it makes for more reliable, less error prone, less catastrophic errors, et cetera, antilock brakes, another good example. And yet the car sort of does the same thing. It takes you to the grocery store, takes you to work and back.

Jim: But then, this… I was thinking about this as an example. But all those precursors that happened for competitive reasons, call it Darwinian or most evolution. Somebody had the anti-lock brake, and people said, “Oh, shit! That’s a good idea. We better get that or we’re going to be left behind”, right? There’s electronic power steering, people especially ladies really liked that. You don’t need strong mechanical force anymore to turn the wheel and all these things happen. Guess what they set up. It’s set up the necessary prerequisites now for the self-driving car. I realized that. That if all those other things hadn’t happened for their own reasons, a self-driving car would not have been in within reach.

Arthur: The first instance of any technology that I’ve ever seen is it tends to be really crude. Imagine when cars came along, they looked a little more like steam engines in the 1870s, 1880s. And there’s a guy often the had steam boilers in those days. Guy had to shovel coal to heat up the steam engine, the driver. So that he heated up the steam engine, and if I remember the French rightly, that is ‘chauffe’ is to heat something up, so he was the chauffer and we still use that.

Jim: I love that and I know that. I love that.

Arthur: So things were primitive early on and overall, a car as you point out hasn’t changed that much and roughly 100 years or 120 years of its use. It still has four wheels, it still has a steering wheel which came along not right at the outset, it still has brakes and some form of accelerator and so on. What you’re talking about, and I totally agree with, is that a technology from the outside may not look that different, but all the parts are changing. There might be a better part, a stronger part or more reliable one that’s usable. So the old idea of carburetor is swapped out and a new, more reliable version is swapped in.

Arthur: Similarly, with all kinds of timing apparatus for sparks and pistons, those are improved and swapped out. So there’s an odd thing going on here. I’ll use a sociological example. If you just stare at a car over 100 years, over all that period you would see that on the outside it looks sleeker, it looks… No weave colors. It’s not… I’m old enough to remember when the early old cars were black and that running boards and so on. Only one color. So if you look at zebras or whatever, the giraffes, they look much the same and probably don’t change that much over eons. But with technologies, it may look the same, four wheels, some sort of cabin to sit in, kind of driving cockpit, all that looks quite similar, but everything, all the parts and pieces…

Arthur: [inaudible 01:09:00] but everything, all the parts and pieces are swapped, improved, much better materials and something that I call structural deepening happens. You might have only one way to break saying 1910, maybe rods and pistons that are somehow attached to the wheels. Now there may be some sort of sophisticated electronic brake system. So where you just had one simple type of sub-technology for the brakes, now it’s a system. And that system depends possibly on computers, there’s safety braking built in, et cetera, et cetera. So you get more parts, more sophisticated ones if the standard inside subsystems don’t operate that well. So the overall car looks fairly well the same nicer looking these days, I think. But the insides, the innards of the car, the basic anatomy has got stronger, better, more interactive, and far more complicated. The structure itself complicates or deepens because you’re trying to do more things in the same space.

Arthur: So it’s a story of development. It’s not unusual. And this is true just about all technologies. Jet engines, in the time of Frank Whittel, there was something like 70 or 80 parts to a jet engine. When I called up somebody at Boeing, I was told this was 2000 agers. “So how many parts would a modern jet engine have?”, said Pratt and Whitney, get back to me to this 22,000 parts. That’s because all the innards are getting more sophisticated and so on. And that’s again like many things in life. Our understanding becomes more nuanced, more sophisticated. And if things don’t work, we throw out old ideas and substitute new ones.

Jim: Somehow the market disciplines, all that, right? The fact that I could, as a teenager, could easily wrench on a four cylinder Volkswagen or a 300 cubic inch straight sex, 1962 Chevy. You know, that was considered to be of economic value at one point, right? So you didn’t have to go pay the overpriced mechanic. But now nobody cares about that because the cars are more reliable and all the other kind of good things that you mentioned. It came from that structural deepening. A lot of the roughness is now gone. And the unreliability.

Arthur: One of the things that happens is that if some sub-system is used often, might have say 54 parts to it. If it’s used again and again and again, over several models and years it becomes modularized, becomes its own thing and it’s separately manufactured and it may have a cover on it. It may not be accessible to amateur mechanics, and it may only be accessible if your trained by Mercedes or Audi or whoever it is. And you’re properly trained. You have the proper tools. So as the lesson here is that as inner parts are used again and again, in the same configuration, the tendency is that they become modules.

Jim: Yeah, of course the classic example is individual electronic components at one point, resistors, transistors, condensers, et cetera, capacitors. And then we had integrated circuits, right? And then we had full computers on a chip with input/output devices and everything else. At each level of encapsulation more and more and more, it goes into a black box, which is unfixable. I can remember fixing transistor radios back in the day. Literally you could replace that transistor if it went bad. Try to do that on a computer chip. You can’t, right? (laughs).

Arthur: I remember building a transistor radio. (laughs) No way. These days there’s absolutely no way I could do that.

Jim: And yet we get so much more value by this encapsulation. You can now buy a microprocessor with some IO stuff for about $3 for a small-scale embedded computing system with surprising power, right?

Arthur: Yeah. So modularity is a key property in these technologies. I want to mention, can I bring up a topic here so that we don’t miss it, if I may?

Jim: Sure, let’s go ahead. We’re getting close to time. So let’s have one or at most, two more topics.

Arthur: Okay. Well, I just have one topic on my list that I don’t want to miss. And that is that so far in this whole discussion, I’ve been talking about individual technologies. So the jet engine or the computer are the whole collection of technologies, the hundreds of thousands, or maybe even millions of different individual technologies. But there’s a layer in between I mentioned in passing and that is that technologies tend to arrive in clusters or groups. In the book I call these clusters domains. It’ll become clear why. And as I began to look at the history of technology, I realized that different phenomena tend to be understood or captured or used at different times. On the 1600’s, we began to understand optical phenomenon pretty well. Time of Newton, for example, telescopes, telescopes with mirrors, microscopes, et cetera. And so we had the optical technologies, maybe 100, 150 years later, we began to get chemical technologies.

Arthur: It took a hundred or a couple of hundred years for those to come out the kind of technologies that might make, gosh, sodium carbonate. If I recall, what’s the reaction that does that? The Solvay process. So industrial chemicals, then by the 1840s, we’re starting to understand electrical phenomenon. We start to get electrical motors and electrical equipment by the 1880s. What I want to emphasize is that there are clusters or bodies of technology that come along as we mine into and understand these clusters of phenomena. And so in the standard economic way of looking at things, we say, oh yeah, the economy improves when a new production technology comes along. So when the example I mentioned was when the Bessemer process came along, whenever that was 1875 or so, then steel-making gets cheaper and better and better quality. Then another process comes along a basic oxygen or whatever that is, and we do better. But I began to realize something more subtle and at a middle layer, is really happening.

Arthur: In the history of technology, we get these clusters or groups of technology arriving over a period of decades. So take for example, from about 1905 on through to the 1960s or so, we had electronics arriving and it was pretty primitive at first. And there was primitive radios and eventually in the 1920s and 30’s, primitive television and maybe recording devices and microphones and things like that. But what I want to emphasize is that this was no single invention. This was a cluster of inventions using different electronic phenomena to very specific purposes. And that’s not that marvelous or deep. What is interesting, I think is that the story of innovation and the economy isn’t just, we’ve got this technology. An improvement on that gives you a prime, better technology, a double primer, third technology to do exactly the same thing. Yes, there’s plenty of that goes on.

Arthur: But I think the big story in the economy is when these clusters or groups of technology come along and they sweep across the economy and they make a real difference. So let me take the whole railroad set of technologies. That comes along 1825, 1830. This is in Britain on through maybe to 1860s and to 1900 as it builds out in different countries. That consists of knowing how to make steel, to lay down rails. How to do railway locomotives, how to signal from one station to another. So it would include wire based telegraphy, telegraph systems in the 1840s, et cetera, et cetera. A large group of technologies. So that railroad cluster of technologies or demand, or body of technologies is available for what? For transporting goods or passengers. Before railways, if you wanted to get goods from say London to Bristol, it would take about two weeks.

Arthur: There’d be ox-drawn carriages lumbering over possibly hilly roads or rocky surfaces. It would get there, but it would take quite a long time. I remember reading The Economist magazine in the archives here at Stanford, and it was marveling. It said this was dated about 1839 and said a new economy is about to come along. The phrase they used was “a new economy” and with railways, and pretty soon within three, four years, we will be able to transport goods from London to Bristol in three and a half to four hours. Now that turned out to be exactly on target. Of course, lots of bumps along the way, no puns intended. But what I want to point out as such when these big bodies of technology come along, say modern synthetic biology or digital technologies or computation. When these huge bodies of technology come along, they sit there and businesses in the economy, don’t adopt them.

Arthur: You don’t say I’m adopting electronics. They encounter these bodies of technology and they pick bits of what they want and they combine their own functions with the new technology. So go back to about 1970 or so, businesses in that era, banks in particular knew how to do accounting. They knew how to keep books on accounts and you had to do transactions, transfers, all kinds of things. Computers were coming along. And so they began to realize that they could track and add and subtract and do arithmetic on their accounts. They could store the information and they could feed in new information. And so modern computer based banking was born as an encounter between commercial banks or consumer banks as well. And the new computational devices, IBM mainframes and things like that. So I think that the bottom line in this story is that an awful lot of what’s interesting about technology comes about when there’s a new body of technology available and you can see whole industries walking around it and wondering what they should pick up.

Arthur: So digital technologies, to fast forward to 2020 or so. We have blockchain. We have Bitcoin if that ever settles down a bit. New types of currency systems. We have sensors everywhere. We have big data as a result of all the sensors. We have ways to handle that modern computing. And we have marvelous telecommunication systems, largely using satellites and fiber optics. And so take any industry, let’s say health systems, and they’re standing there and say, oh, we used to do this. Doctors would fill out forms. Forms would go to insurance companies. Insurance companies would send more paperwork back and so on. Now, a lot of these things can be digitized and just have conversations between one mobile device or one device and another. And so suddenly we have this highly interactive platform based healthcare system. Maybe some images are taken here in Palo Alto at my local clinic.

Arthur: They’re transported by fiber optics and satellites to Mumbai. They’re read there, not just by human beings, but by algorithms. And the word comes back, we suspect this that and the other. And normally in these MRI images you sent, all of this is automated. So my basic point here is that the economy keeps morphing and changing as industries adopt these new possibilities each at their own rate and combines what they do doctoring maybe, or diagnosing with the possibilities and the new domain. So we’ve had this with canals. We had it with railways. We’ve had it with the electronics. The story, now I think 2020 is that the digital technology is that body of technologies has morphed and changed a great deal. And it’s just sitting there and different companies and different industries and different grades are finding ways to use it, whether it’s in China or whether it’s here in the US or here where I’m standing and Silicon Valley.

Jim: That was a great one. I would have picked, I think the same one from my list of topics. And I’m going to come back at you with one sub-component of the idea of technology domains. And this is from your section on how domains evolve. And this is a direct quote, “real advanced technologies on the edge, sophisticated technology issues, not from knowledge, but from something I will call deep craft.”

Arthur: (laughs)

Jim: This is really important, right? This is hugely important if you’re actually trying to be in the business of technology, right? So tell us about your idea of “deep craft”.

Arthur: So I’m trained as an engineer. I’m standing here literally in PARC used to be called Xerox PARC which has been traditionally one of the players in advanced technology and invented if I recall, trying to think, the LasrerWriter, the graphic user interface, the Ethernet, there’s a long list of things that got invented in this rather small building. And I began to notice I’ve been here quite a long time as a visiting researcher here. And I began to notice that engineers don’t work by saying, “I know the science, I can turn a crank. I can do the mathematics. So turn this out.” It was much more subtle and sophisticated than that. People would walk down the hallways and say to somebody colleague, maybe at the other end of the building, “I’m looking for a three micron etching device. I think you were using that two years ago. Where would I lay hands on one?”

Arthur: Or I’m thinking of doing whatever Santa Fe Institute, a genetic algorithm on this. I remember doing that and I was able to phone John Holland, who’d invented the genetic algorithm. Said, “John, are these the right parameters?” And John said, “no, no, no, no, no, no. We use that for mathematical purposes. Let me tell you how the recipe really works.” So I began to notice that real expertise in super advanced technology is not just science. It’s not just mathematics. It’s a matter of knowing what works, knowing what doesn’t work what’s been tried in the past. It’s a matter of knowing who knows about all this and maybe befriending them and getting advice from them. It’s very human and it resides in people’s brains. In the book. I call this “deep craft”.

Arthur: Other people pointed this out in the past. There’s nothing new about this, by the way. If you go to the town of Cremona in Italy, the Stradivari family and the Amati families in the 1700’s understood how to make violins. They understood which trees at which seasons should be harvested. They understood how that wood should be dried and cured and what varnishes were needed and what thicknesses were appropriate. None of that’s quite scientific. As I point out in the book, it’s a bit closer to cordon bleu cooking and Paris.

Arthur: You need to know the basics. You need to know the grammar. You need to know how it works, but above all you need experience and you need to know what doesn’t work and what to try if things go wrong. That’s inside people’s brains and that’s deep craft. And I began to realize that’s why Silicon Valley can do tech, whether it’s biotech or nanotech or driverless cars, autonomous technologies. And it’s not as easily done say, in the big island and Hawaii, not because people lack education or smarts, but they don’t have that build up of experience.

Jim: Yep, indeed. A hugely important point. In fact, our former colleague at Santa Fe Institute, Woody Powell has done a lot of work on that. Especially with respect to biotech. And call them communities of practice or deep craft. It’s a real thing and people dealing with it in the real world need to take into consideration. Well, we didn’t quite to everything in the book, but I think we hit most of the high points and we certainly gave people a good sense of what it’s about. If you’re a venture capitalist, and I know there’s some of those who are out there and listen to us and some entrepreneurs or some people just live in this strangely technological world of ours, and you liked what you hear here, go get Brian’s book, The Nature of Technology: What it Is and How it Evolves.

Arthur: Thank you so much, Jim. Great to talk to someone who knows a lot about technology and that’s you. Thank you very much.

Production services and audio editing by Jared Janes Consulting, Music by Tom Muller at