Average Is Over: Powering America Beyond the Age of the Great Stagnation by Tyler Cowen (Dutton 2013), 304 pages.
In Average Is Over, George Mason University economist Tyler Cowen delivers good news and bad news with nearly equal enthusiasm. Basically, artificial “intelligence” (AI) is aggregating the “knowledge of the entire world” and intruding everywhere, ready to overturn our lives, laws, and customs. DARPA (Defense Advanced Research Projects Agency) sails electronic seas, and GPS follows us through retail stores. Given enough ever-better supercomputers, “we” (some of us) will see like a state and know as much as God. The NSA certainly believes that.
As AI makes goods cheaper, money flows to scarce factors: land, intellectual property, and quality labor. Machines will put humans out of work, but Cowen wobbles a bit on how many. Still, each military drone needs 160 human minders (be happy), and he who increases AI “value” wins. Marketing will provide the “hyper-meritocratic” elite with false friends (it’s a job), and others can serve them as gardeners, maids, et cetera. In this two- layered society, exclusionary corporate teams will dominate ever-finer divisions of labor and select employees for conscientiousness. Few jobs lost in the recent (or current) “recession” are coming back; any that do, will pay less. Unemployment, Cowen says, is significantly higher than is usually admitted. Wages have stagnated since 1973. Owing to AI, from 2009 labor productivity per hour has gone up. So has self-employment. By now Cowen has begun making lengthy digressions on computer chess.
Robot managers will “grade” professionals’ work and help doctors rid themselves of uncooperative patients. Robo-credit checks are here. AI contributes to war, surveillance, and drones (be happy): here, probability analysis decides someone’s death. (Robotic Just War Theory is surely next.) There is AI dating advice and AI bother without end. Humans do have a little time to adjust to the new order, in which the cost of these new machines will set wage rates. Cowen wisely avoids Singularity and even notices that the whole human body is needed for the “brain” (he may mean mind) to work properly. (He almost discovers Aristotle’s concept of hylemorphism, the idea that substance comprises form and matter.) In all the general AI enthusiasm, the saving (and limiting) notion of GIGO (“garbage in, garbage out”) is not found in this book.
Outsourcing and immigration, Cowen writes, haven’t caused American job and wage losses, even if “offsetting gains” in the corporate state are hard to find except for those going to capital. There are real regional losses. “Educated” machines are the cause. (Cowen’s peculiar plan of keeping the jobs here and putting foreigners in them seems incoherent.) In Europe Cowen foresees the “‘hollowing out’ of various regions” — a conclusion long ago reached by André Gorz in A Strategy for Labor (1967) and Farewell to the Working Class (1982). The socialist Gorz condemned it, while Cowen feels duty-bound to welcome it. Cyber Utopia thus entails turning southern Europe into southwest Florida writ large — supported by tourists, retirees, and others not working. Internationally, cyber education is a brain-drain accelerant — an obviously good result, if money made by American entities is our only test.
But not all is well: workers in two domestic sectors (health and government) still enjoy some security, given the lack of wholesome, international market “tests.” No one can tell what “value” they add, since it doesn’t benefit large corporations. And inefficiencies arise when too many people are secure.
Under AI, Cowen says education will be cheap, with competition driving “price down close to costs.” AI will even “grade” essay questions! Instructors will be untenured adjuncts, more or less. Robo-snoops will police students using scanners and sensors. Profits to AI are up. But that stronghold of job security that Cowen spies in universities must not stand. The old model was tolerable when teachers were cheap, but now professors must become “motivators,” “entertainers,” even “impresarios.” AI education will be “more like the Marines,” an inspiring model to be sure. Motivators will presumably help rich people find their Inner Midas.
Science marches on
Average science is over, too. Science is about prediction, control, and “understanding our world.” (For the last item, metaphysics might be much more useful.) Science is all about teamwork as scientists merge with machines, subordinating “the individual scientist” to corporations and the state. So-called empiricism prevails. Computers will do “their own research.” An advanced machine-generated “theory” of cosmology might well mean nothing to humans. But if it’s unintelligible, why isn’t it an untheory? Cowen worries (probably needlessly) that the public’s “long-term loyalty to scientific reasoning” may wane.
Under the new “hyper-meritocratic” social contract, the Stagnated shall have cheap education and heaps of entertainment. The U.S. fiscal crisis will force cuts in Social Security, Medicare, et cetera, but (apparently) never in military spending. The United States will not default on the national debt. Organized groups will shift costs onto others. The answer for the 85 percent of nonwinners is Texas, or Mexico, or imported Brazilian favelas. Texas has cheap housing (low rent), “job creation,” and little zoning. People can move to such places but are not forced to do so: note the economist’s trump card. (If you block all exits from the maze but one, the rats “voluntarily” go to that one.) It will all end well: an aging population won’t be revolutionary, and most “envy” is local. (But, as left-libertarian author Kevin Carson notes, justified anger at having the rules all rigged is not “envy.”)
For complacent economists, facts exist mainly for hauling offstage. Exercises in economic history, journalism, et cetera normally end, rather jarringly, right where everything turns out for the best. Possibly as insurance, Cowen plays the inevitability card (perhaps intending to demoralize opponents and skeptics?) and allows little escape from the radiant future.
Preliminary observations
A strange sort of economic honesty (wage stagnation, job loss) runs through the book, as witness (perhaps) to a kind of social myopia. Cowen establishes no “merit” in his technocrats. He mentions “value” fairly often, but can measure it only by (state) capitalists’ profit levels and high incomes for upper-middle class techies. By contrast, work done by me or thee is seldom “tested” rigorously enough. “Unproductive,” we shall have no pie. Machines got so “intelligent” as the book went on, that I thought for sure Cowen would declare Oskar Lange (with his 1960s computers) winner of the socialist-calculation debate, but the announcement never came. Instead, Cowen settles for the new regime of capital accumulation based on turning frozen ideas into capital. Since they generate big profits for worthy capitalists, these collectivist hives are unobjectionable.
Finally, and worse luck, Cowen endorses the American empire under other names. The United States has “logistical abilities, and a strong trade-protecting military,” and a North American century looms. The United States must be dominant (no reasons given). His unworried accounts of AI use in TSA scanners, drones — and the whole electronic Panopticon — reflect a strongly Hamiltonian conception of political economy. (Philip Mirowski terms Cowen a neo-liberal rather than a libertarian.)
Objections historical
Cowen wants to leave moral judgments on such political-economic matters to Emersonian individuals. We shall take our own look. From about 1900, American corporate managers and engineers, aided by competition-suppressing intervention, undertook, quite deliberately, to build complex operations into machinery to rid themselves of skilled workers. If World War I was a great success for those corporate cadres, World War II was a triumph, made permanent by the Cold War. Historical inevitability, heavily subsidized, could now set in, full-tilt.
Corporate chieftains largely shared the state’s ambitious managerial projects: knowledge was power and atom bombs were physics-on-parade. Cyborg (man-machine) sciences began in World War II gunnery studies, and from 1947, Air Force ideologues and the RAND Corporation were major players. With taxpayers paying for research and development, electronics and communications were among the most subsidized industries in U.S. history.
Here is Big Science and teamwork; here is the stable where AI was born, as a close cousin of the atomic bomb, game theory, and all the “rat” theories (such as Rational Actor). Of course the corporate participants could innovate and make money.
Objections philosophical
Bags of electrons organized by mathematical Platonists are quite impressive, really, but are not intelligent, whatever increasingly robotic human-Americans wish to believe. Already in 1835 Andrew Ure found in the wooden, iron, and brazen structures of the cotton mill a “perfection of automatic industry” amounting to “an intelligent agency.” (Emphasis supplied.) In a book full of computers apparently teaching, playing, composing, doing, seeing, et cetera, Cowen observes, “it isn’t the computer ‘thinking’… [it is] consulting memory” or engaging in “pure calculation” (not even that). He is quite right. Numerous scientists and laymen would prefer it otherwise, and Theodore Roszak (Where the Wasteland Ends) gave evidence of their existence way back in 1972.
Truth be told, programs, “code,” et cetera, amount to manmade Aristotelian forms. The once-useful word information has almost no lingering connection with meaning, knowledge, or wisdom. In America we don’t work along such unprofitable lines. Information now refers mainly to atomic “bits” fit for mathematical manipulation. (Compare that with individuals, atoms, signs, transactions, et cetera in other fields.)
A much older project is at stake. Sir Francis Bacon’s New Science aimed at subduing nature by abandoning the assumptions of Scholastic science. Alas, projects to control nature invariably lead to controlling people. Baconians were ready.
Computer mania may mark an era, but it also continues the ages of the Artificial Man (Leviathan), who has the bodily strength of many, and the Artificial Merchant-and-Craftsman (the corporation), who uses the savings of many and the bodies of many others. At long last, those two characters, working in tandem, are offering us Artificial Mind in their headlong, simultaneous pursuit of materialism and abstraction — materializing Kant’s project of living in our heads, while etherealizing anything actually material. Bacon wanted men to think “as if by machinery.” The madness is in the method. After two centuries of what Edmund Burke called the age of “sophisters, economists, and calculators,” we may need to rethink some things.
Perhaps Karl Marx was polemical and unscientific when he called capital goods dead labor; yet it might be useful to think of robotic capital (AI) as dead thought whenever folks go on about “machine intelligence.” In a computer program the erstwhile thoughts of a corporate team now lie entombed in various loci (software, hardware, et cetera). Just as dead labor confronted living labor in factories, so does dead intelligence confront the living, while its keepers badger us with updates.
With AI’s help, deskilling has advanced from the machine shop into the front office, and now besieges the holdouts: lawyers, doctors, academics, and possibly others. Of course under AI, as under older forms of corporatism, revenues attributed to machines go not to them but to their owners.
As hype and ideology, cyber-mania brings to mind America’s railroads — colossal, state-subsidized boondoggles, on which some sharp owners and promoters got rich, while spreading ideological mystification. Is it accidental that a recurring archetype in American literature is the confidence man?
Objections merely human
“Sorry, we just won’t retrain to work for you. Retrain your own damned self. And the machine you rode in on.” So some persons might say in a candid moment.
There are good reasons to reject the grave new world under consideration. The rent is too damned high; worse, the rents are too damned many. In Average Is Over the rents are too damned absent: Cowen notices only inefficiencies and pockets of offensive, undeserved security. “Cartel” comes up once, but only in reference to universities. It is odd that the word “proletarian” nowhere occurs in the book, given that so few people seem to possess their own means of production these days. At AI’s dizziest heights, it is perhaps enough that one is splendidly paid. Menaced with Cybertopia, we might well ask how we can avoid being railroaded into it: a good question, which I reckon AI can’t answer. Skilled human labor may be needed.
This article originally appeared in the June 2014 edition of Future of Freedom.