The Second Machine Age: Work, Progress and Prosperity in a Time of Brilliant Technologies by Erik Brynjolfsson and Andrew McAfee (W.W. Norton & Company 2014), 320 pages.
The subject of this book is the “second machine age,” in which “computers and other digital advances are doing for mental power — the ability to use our brains to understand and shape our environments — what the steam engine and its descendants did for muscle power.” The key technological building blocks for this second machine age are already in place, even if they aren’t yet completely mature.
Although these technologies are the culmination of several decades of gradual development, they have only recently reached the threshold of a fundamental phase transition in the way the economy operates. The first part of the book is a walk-through survey of the various technologies involved in the second machine age and the reasons we’re just now approaching a phase transition.
Second-machine-age technologies are exponential. But as Ray Kurzweil has pointed out, the early states of exponential growth are deceptively modest. He uses the example of the inventor who, offered a reward by an emperor, asked that one grain of rice be placed on the first square of a chessboard, two on the second, four on the third, and continuing to double the number through all sixty-four squares. While the numbers get large in the first 32 squares, they’re still the kinds of numbers we encounter in the real world. By the inventor’s arrangement, there would be about four billion grains of rice on the 32nd square—“about one large field’s worth.” It’s on the second half of the chessboard that things get weird, with the final amount of rice equaling the total agricultural output of many earths.
Thanks to Moore’s Law, machine-age technologies have been moving from square to square every couple of years since World War II. And we’re just now getting onto the second half of the chessboard, where things really take off and the elbow of the curve bends into a straight vertical line.
The new technologies are also combinatorial. Incremental improvement in the capabilities of existing technologies can bring them to the threshold of an exponential increase in their possible number of combinations. For example, Google’s Chauffeur project has produced results several orders of magnitude better than previous attempts at autopiloted cars a few years back, despite the fact that the same building-block technologies existed back then. Incremental improvements in those technologies, interacting synergistically with each other, enabled an exponential increase in performance. And continuing incremental improvement in building-block technologies increases the stock of off-the-shelf technologies available for developers to recombine at will to fit particular needs.
We’re rapidly approaching a range of cheap, modular, general-purpose, off-the-shelf building-block technologies sufficient to enable a near-infinite number of custom mash-ups for any conceivable purpose.
Reading Cory Doctorow’s novel Makers, a near-future story about open-source hardware hackers and micromanufacturing, Bruce Sterling commented that there was “hardly any engineering. Almost all of this is mash-up tinkering.” It’s just the recombination of modular components. But that’s just the point, Doctorow responded. “It’s not that every invention has been invented, but we sure have a lot of basic parts just hanging around, waiting to be configured.” The result is that “we now inhabit a world where knowing something is possible is practically the same as knowing how to do it.” (Doctorow, “Cheap Facts and the Plausible Premise,” Locus, July 5, 2009.)
Murray Bookchin’s essay “Toward a Liberatory Technology,” written in the 1970s, quoted something Vannevar Bush had said back in 1955:
Suppose, fifty years ago, that someone had proposed a device which would cause an automobile to follow a white line down the middle of the road, automatically and even if the driver fell asleep…. [His] idea would have been called preposterous…. But suppose someone called for such a device today, and was willing to pay for it, leaving aside the question of whether it would actually be of any genuine use whatever. Any number of concerns would stand ready to contract and build it. No real invention would be required. There are thousands of young men in the country to whom the design of such a device would be a pleasure. They would simply take off the shelf some photocells, thermionic tubes, servomechanisms, relays, and, if urged, they would build what they call a breadboard model, and it would work. The point is that the presence of a host of versatile, reliable, cheap gadgets, and the presence of men who understand all their cheap ways, has rendered the building of automatic devices almost straightforward and routine. It is no longer a question of whether they can be built, it is a question of whether they are worth building.
There was no little exaggeration in what Bush wrote almost 60 years ago. But today it is true almost beyond his (and Bookchin’s) wildest imaginings.
And with all the rest of this, throw in the fact that we’re rapidly approaching — if we haven’t already reached — artificial intelligence capable of running all these mash-up machineries with sufficient flexibility and discretion to replace human operators. The result will be the automation of human labor and the elimination of existing job categories to an unprecedented degree, very likely extending to the majority of work-hours and creating an enormous mass of unemployed or underemployed people.
That brings us to the next point. Despite the overall benefits of the new technologies, their benefits tend to be distributed unevenly, leading to what the authors call a “spread” in “wealth, income, mobility, and other important measures” between those at the top and bottom. And they warn that the spread in benefits from new technology —which they view as the natural outcome of increased productivity —will accelerate as the second machine age goes on, “unless we intervene.” It generally happens that “a relatively small group of people … earns most of the income from … new products or services.”
Of course the new information technologies, with almost zero-marginal cost, are also destroying GDP by radically deflating prices (just look at the 40-percent drop in revenue from music sales between 2004 and 2008 alone, or the 100-percent price drop for people who now choose Wikipedia over Britannica). But how are people to buy goods and services, even drastically cheapened ones, if they don’t have any income from jobs at all?
Part of the answer is that, under the existing model of corporate-state capitalism, the goods and services are still not cheap enough. They’re not allowed to be. The main reason so many benefits of new technology are monopolized by those at the top is that the state enforces, well, monopolies. Economic ruling classes are able to enclose the increased efficiencies from new technology as a source of rents mainly through artificial scarcities, artificial property rights, and entry barriers enforced by the state.
And perhaps the most important of these artificial property rights is so-called intellectual property, which the authors take for granted.
In reality, the natural course of affairs absent such state-enforced monopolies is not for a small group to monopolize most of the income, but for market competition to socialize all of the productivity gains in the form of lowered prices.
One of the most interesting things about the new technologies of this generation — the technologies of both ephemeral, small-scale physical production and more- powerful information processing — is that their basic logic undermines the scarcity logic by which ruling classes of the past have extracted rents from society. They render large concentrations of land and capital increasingly irrelevant, and put the potential for ownership and control of the means of production in the hands of ordinary people.
There’s one big difference between the effect of the steam engine in the first Industrial Revolution and the effect of cybernetic technology today. All the labor-displacing and impoverishing effects of steam power resulted from the facts that (1) steam engines and the kinds of machinery that ran off them were enormously expensive, and beyond the resources of individual workers or small groups of workers to acquire; (2) the resources for purchasing such machinery had been concentrated by a long historical process of enclosure and other state-enforced expropriations of land from the peasantry to a relatively small propertied class; and (3) the transaction costs of aggregating the small individual savings of ordinary working people into capital to finance cooperative production were enormous — even when they weren’t preempted or foreclosed by prohibitive state restrictions on freedom of association.
In contrast, new technology today is not only radically cheapening the means of both physical and mental production, but also eliminating the transaction costs for crowdsourced financing (or crowdfunding, defined by Oxford Dictionaries.com as “The practice of funding a project or venture by raising many small amounts of money from a large number of people, typically via the Internet”).
New technology is a source of “unemployment” for people who currently have “jobs,” rather than simply being tools for ordinary people to use to produce for their own subsistence and trade with others, if we assume the persistence of a framework in which production is carried out by “companies” that control access to the machinery. The authors cite a study which found that
companies used digital technologies to reorganize decision-making authority, incentives systems, information flows, hiring systems, and other aspects of their management and organizational processes. This coinvention of organization and technology not only significantly increased productivity but tended to require more-educated workers and reduce demand for less-skilled workers.
Note the assumption: the “company” owns the machines, decides how to organize production, and decides what workers it needs to hire to carry it out.
But the new production technology is making those assumptions obsolete. Let’s start instead from the assumption of a subsistence farmer. If he comes up with a new way of doing things that enables him to produce the same amount of corn with half the work, he doesn’t lament all the work that it will put him out of. That’s because he owns both the farm and the final product, and internalizes all the benefits.
All these predictions that big industrial companies will organize assembly lines with CNC routers, cutting tables, drill presses, and 3-D printers, with robots that transfer unfinished goods between them, ignore one thing: the open-source versions of most of those CNC tools can be built for under $1,000 each, and are entirely within the means of small neighborhood cooperative shops manufacturing for local barter-exchange networks in return for the output of other shops, of home microbakeries and microbreweries, surplus garden produce, babysitting and barbering, and the like. As John Robb wrote on his Twitter feed, “You can either compete with technology for a job, or use it to help you make a living outside of a job. Your choice.”
The main source of continued corporate control of the production process is all those artificial property rights such as patents, trademarks, and business licenses, that give corporations a monopoly on the conditions under which the new technologies can be used. But “intellectual property” is becoming increasingly unenforceable (take another look at that 40-percent drop in music revenue), and corporations that are going bankrupt from a collapse in aggregate demand and that will suffer from disintegration of their supply and distribution chains in the face of $15/gallon fuel (my projection) probably aren’t going to have the resources to notice or care about garage producers who fill in the gaps — let alone do anything about it.
And in the meantime the state barriers ordinary people face from the other direction — barriers that put an artificial floor under the cost of subsistence by criminalizing vernacular building technology (by means of building codes written by contractors) and criminalizing home-based enterprise (through zoning and occupational licensing) — will also become unenforceable as fiscally strapped local governments find themselves faced simultaneously with record homelessness and unemployment and record numbers of abandoned, foreclosed, and condemned vacant housing.
In other words, the solution is not a universal basic income or other expedient to reallocate the wealth that naturally accumulates in the hands of the few, or a Japanese-style school system to turn everyone into more-valuable corporate human resources. And it’s not massive new subsidies to “infrastructure” to make centralized, hierarchical firms serving large market areas artificially profitable (as railroad land grants and the Interstates did for mass-production dinosaur corporations). It’s to destroy all the state-enforced monopolies that shift wealth to the few in the first place, prevent the relocalization of production, and give corporations control over employment.
The upshot all these wonderful new technologies that Brynjolfsson and McAfee write about — and they are wonderful! — will reach their full potential not within the framework of the existing corporate power structure, but within the new economy that arises from the ruins of this one.
This article was originally published in the January 2015 edition of Future of Freedom.