In this short book from 2009, W. Brian Arthur, an economist and complexity theorist, gives a compelling and “organic” description of technology. He looks at the purpose of any created artifact from shoes to airplanes and tries to find what they have in common, where the whole is going, and why we should care. His conclusions are profound, although easy to understand, and they have important implications for the future.
His core thesis is that technology has three characteristics: 1) technology is combinatorial, 2) every technology has sub-parts, and 3) technology is a purposed application of a principle of nature. Let’s look at those three a little closer.
The first characteristic is the most important for understanding the future; it is another way of stating Kurzweil’s principle of accelerating returns, or Kevin Kelly’s idea that technology feeds on itself, that every artifact co-evolves in an environment made up of all other artifacts.
Combinatorial is a fancy word for what every kid learns with tinker toys: there are a lot of ways to put pieces together. In mathematical terms, if you have n elements, you can put them together in 2n different ways. With 10 elements there are a thousand combinations, with 20 there are a million, and with 30 there are a billion, and with 40 there are a trillion.
Now, change your perspective, and think of an element as the types of thing that go into a single working technology. This is Arthur’s second characteristic, that every technology has sub-parts. Take a bicycle for example. Wheel, chain, pedal, spoke, screw, derailleur, crankset, brakes, handlebar, fork, saddle, chainring, grips, shift lever, hub, tube, pump, light. That’s for a simple artifact. Add in all the others – autos, houses, electrical systems, baseballs, and the way that the elements of all these can be put together is practically infinite.
That’s at the level of what we see. Go down a layer to molecules and chemistry, and we have infinity times infinity in the number of possible combinations.
Possibilities are limited only by our imagination. What limits imagination? That’s a different question – but one thing that stimulates it is the combinatorial and fertilizing interaction of people in coffee houses that Brian Cowan describes in The Social Life of Coffee. As a matter of fact, I’m in the Tupelo coffeehouse right now, with 20,000 people swirling around me in the annual Run to Feed the Hungry in Sacramento.
As the number of possible combinations increase, the time that it takes to put them together in a new way decreases. With more to play with, the imagination takes off. So events in the sciences and technologies are speeding up by their very nature, quite apart from anything having to do with communication or “speed” of a particular technology.
And yet, the digitizing of information and the rapid-fire way that we can send an idea, book, video, or 3D shape across the Net means that the pace of change has taken off in the last twenty years. At a more profound level, any process can be converted to a digital form and combined with any other; or physical processes in different places can be linked together through sensors and networks. Digital information and code are the new tinker toys, and they can be put together to play with DNA, the design of a space station, the design of a brain – or anything else. With the advent of worldwide networks, there are essentially no limits to what can happen in the future, nor any way to predict how fast it will happen. Technology as a whole, what Kelly calls the “technium,” has become like an individual immune system, capable of creating trillions of unique combinations, a kind of second nervous system for the world.
This all begins to sound like biological evolution: combination, recombination, new features. Technical evolution is different from biological, however, because it doesn’t have the same limitations. Classical evolutionary theory limits change and variation in organisms to a single species. And for the big ones like ourselves, changes happen on a time frame of a million years. That’s one reason that we’re blind to the future; we look back a thousand or five thousand years, and humanity looks pretty much the same.
Technology can and does gleefully share its “genes” across species. You can add a GPS to a bicycle, or a light to your book. This is the equivalent of grafting wings onto a person, or giving yourself gills and echolocation. Since we can try out possibilities in virtual space, using computer simulation, a billion generations can be tested and tried in a matter of hours or days.
The development of a particular industry or technology can be predicted with some accuracy, simply because it is already “locked-in” to a set of parts, manufacturing processes, distribution, and uses. But predicting the development of technology as a whole is very different. Apart from predicting the trajectory of a single technology, it is almost impossible to see what combinations of ideas and things will emerge in the future – all the more so, I would add, because in the next decade a general-purpose 3D constructor (on analogy with David Deutsch’s “quantum constructor”) with connectivity to worldwide networks will be increasingly common. At that point the friction for moving a technology to everyone on earth becomes negligible.
Go back in time a thousand years, to 1012 AD. What new technologies, what new combination of scientific ideas were likely to happen in one human generation? Or even several generations? Not much. There weren’t many tinker toys to scramble together; not many ways for the technologies of the time to combine. Fast-forward to 2012. With infinite combinations, the window of predictability has shrunk to 5-10 years.
Arthur defines a domain as a group of technologies which are related to what one would call a field in academics: electricity or hydraulics or aviation. Thus there are the single technologies of integrated chips, screens, software, information science, but they are all part of the larger domain of computer technology.
This recursiveness, this fractal “self-similarity” at different levels of scale, has lots of implications. For one, every technology (biological or non-biological) is itself composed of sub-classes, and at the same time is a member of a larger entity. In the last chapter Arthur notes that for the first three centuries, machines were seen as mechanical, in comparison with organic life. This was obvious at a glance – nobody would mistake a bicycle for a hedgehog – and a result of mechanistic philosophy. Descartes and others, reinforced by the Newtonian success, saw the world ultimately as a pure, mechanical, clockwork creation.
But as technology shrinks to the nano scale, and becomes more imbued with information (senses, motors, intelligence), the qualities of life and mind are increasingly evident in technology itself – at every recursive level, not just the individual thing. This is Kelly’s point as well: taken as a whole, the entire technium seems to have a goal – that of increasing complexity, interconnectedness, sentience, diversity.
Because of sensors and actuators (the simplest description of robotics or machine intelligence), we are entering the era of “collections of things-in-conversation-with-things that sense their environment and react appropriately.” Add nano: “In fact, not only will these systems in the future be self-configuring, self-optimizing and cognitive, they will be self-assembling, self-healing, and self-protecting.” In other words, they will meet the most common definitions of living beings, and perhaps sentient beings as well.
Arthur’s third characteristic of every technology is that it “is a means to fulfill a purpose: a device, or method, or process…. A device always processes some thing; it works on that thing from beginning to end to complete the needed task.” This can be a sequence of physical steps (such as gears and switches) or steps in a process. In that sense, “A technology embodies a sequence of operations; we can then call this its ‘software.’ And these operations require physical equipment to execute them; we can call this the technology’s ‘hardware.’”
This software/hardware view, encompassing process and embodiment, allows us to see similarities among, for example, a bridge, a word processor, and a contract. Every “thing” in nature, at any scale, embodies processes and the means of carrying them out. If information theory is correct that a universal Turing machine can model any process, then we have now entered an era when relationships can be discovered or created between any two things in the universe – say, a thought and a galaxy, or a cat and a government.
Because every technology embodies some principle of nature, some way of applying understanding to the world, and these principles can be explained as scientific theories, physical technology embodies ideas. “It is because ideas that originate in science become digested over time into the bodies of technology themselves… These two cumulate together, and as this happens science organically becomes part of technology.” The panpsychism of science.
This means that to understand technology as a whole is to understand the science of the time – or vice versa, to misunderstand or be ignorant of how things work is to fail to understand the science, and one’s time. Since science, technology and culture are rapidly co-evolving, this also means to be ignorant of the very direction of events. The whole evolves together, and exhibit the qualities of “open-endedness, indeterminacy, and the emergence of perpetual novelty.”
Arthur speaks of how economic theory largely arose in relation to a specific era, the domain that involved large industrial processes (the industrial revolution). The idea of the invisible (Darwinian) hand of Adam Smith and the supply/demand theory of John Maynard Keynes arose in that environment, using mental models shaped by large industry. The problem, as we have seen in the recent market crash, is that “its form and structure – change as its technologies change.” So that “our interpretations of the economy must change constantly over time.”
Yes. And I would say the same is true of every area of life: as the systems around us evolve, we need to adapt our thinking, our worldview, or risk falling further and further out of touch with reality.
We may also fail to appreciate technology or science in esthetic terms. Even though design in engineering “is a form of composition, of expression” as much as music or art, the compositions in technology “are largely hidden. They tend to lie on the inside… hidden within some casing, or within lines of algorithm code, or within some industrial process, and not at all visible to the uninitiated.”
This fact that the complexity is hidden means, I think, that people will continue to be surprised as developments accelerate into the future. Few understand that the smart phone that they play games on is really a powerful universal computer that can in principle emulate any process whatsoever – including the processes of thought.
Technology can be seen as a pattern. “Each era.. is a pattern, a more or less self-consistent set of structures in business and industry and society set in place by the dominant domains of the day.” The dominant domains of our time are finance, forms of government, business processes. The invisible pattern, for most, is scientific discovery. In the largest sense, the pattern of an era is emblematic of the entire whole of culture, technology, and thought – the combination of zeitgeist, mind-plex, and scientific-technological complex that makes up the inner and outer world of the people of the time.
In an interview with American Scientist, Arthur calls technologies the “orchestration of phenomena,” and that as such they are an incredible wonder. Yet “it’s like having magic carpets at our disposal, and we have no idea how to fly.” In my opinion we are close to an inflection point in this respect – not necessarily a singularity in the sense of physical transformation of humanity or the physical world, but a tipping point in the functions of specific things that render them astoundingly magical. Artificial intelligence is the most likely, but nanotechnology, quantum dots, virtual/augmented reality, biotechnology, even simple robotics or natural language processing may be our next magic carpet..
And then what? Does humanity co-evolve with our creations, reaching ever more complex heights? Arthur, along with David Deutsch, Ben Goertzel, Ray Kurzweil, George Church, and others, believes so. The genie cannot go back into the bottle; it will continue granting wishes for ever.