David Graeber's theory of technology

He was not so much a technological determinist as he was a social determinist.

[I wrote most of this most 6 years ago when I read David Graeber’s Bullshit Jobs as part of a book group. A recent conversation made me go back to it, finish it, and press “publish.” I hope you like it.]

The late David Graeber’s Bullshit Jobs is a book that revels in contradictions. It’s an unabashed polemic written for a popular audience in Graeber’s characteristically vivid style, but Graeber insists throughout that it is based on his “research.¹” Even if one thinks of it as a piece of anthropological research that is based on people’s experiences of their jobs, its findings are undercut by the fact that Graeber relies almost entirely on the testimonies of a self-selected group of people who follow him on Twitter; moreover, he takes the testimonies at face value, making no attempt whatsoever to even interpret them or read them against each other. Finally, Graeber’s thesis sits completely on top of a peculiar theory of technological change but that theory is left unspecified and implicit.

It is Graeber’s theory of technology I will talk about most in this post. I will argue that Graeber is not your garden-variety technological determinist; if anything, he is a strong social determinist. He is, at once, too instrumental about how technologies are invented (we can make whatever technologies we want; all we need is the will to do it), and too sweeping in his estimation of their effects (the right technologies can change everything). This mischaracterization of how technology works thus leads Graeber into sounding very much like your modern Silicon Valley entrepreneur, eager to disrupt sclerotic industries, despite his very different (radical, anarchist) politics.² It also, I argue, undercuts the whole reason he wrote Bullshit Jobs in the first place: rather than create an alternative to today’s world where he argues that “we have become a civilization based on work […] as an end and a meaning in itself,” he ends up reinforcing that very idea. 

Graeber’s socially determinist theory of technology

The most explicit statement of Graeber’s theory of technology is not in Bullshit Jobs (though as we will see later, it is crucial to its argument) but in a long, rollicking, essay he wrote for The Baffler titled “Of Flying Cars and the Declining Rate of Profit.” In this essay, Graeber opens with a simple question: why, he asks, are there no flying cars today even though these objects were imagined in a lot of American popular culture from the 1950s onwards? Rather than flying cars, why does today’s technological frontier consist of Facebook instead? 

Graeber’s answer is interesting. To begin with, he says there are two possibilities: either people in the 1950s didn’t know what they were talking about; or else, they did, but something came in the way. 

Why did the projected explosion of technological growth everyone was expecting — the moon bases, the robot factories — fail to happen? There are two possibilities. Either our expectations about the pace of technological change were unrealistic (in which case, we need to know why so many intelligent people believed they were not) or our expectations were not unrealistic (in which case, we need to know what happened to derail so many credible ideas and prospects).

Graeber thinks the second option is more probable. The expectations were realistic but something happened. But what exactly? According to Graeber, the powers-that-be decided that the technologies being developed were all-too-radical, and therefore diverted resources away from them towards less revolutionary technologies that were more likely to preserve the status quo. 

But if there was a conscious, or semi-conscious, move away from investment in research that might lead to better rockets and robots, and toward research that would lead to such things as laser printers and CAT scans, it had begun well before Toffler’s Future Shock (1970) and Gilder’s Wealth and Poverty (1981). What their success shows is that the issues they raised — that existing patterns of technological development would lead to social upheaval, and that we needed to guide technological development in directions that did not challenge existing structures of authority — echoed in the corridors of power. Statesmen and captains of industry had been thinking about such questions for some time.

What did these “statesmen and captains of industry” do? Graeber argues that they punted. Rather than going full-steam ahead with developing even better machines to replace workers, the titans of industry shipped manufacturing jobs to the global South instead. And the statesmen started to fund not real technologies but instead what Graeber calls “technologies of simulation” i.e. “mainly either medical technologies or information technologies.” This became even more true after the United States won the space race; once the threat of the Soviet Union’s technology was nullified:

the shift to research and development on information technologies and medicine was […] part of an all-out effort to follow the technological humbling of the Soviet Union with total victory in the global class war — seen simultaneously as the imposition of absolute U.S. military dominance overseas, and, at home, the utter rout of social movements.

In other words, the “statesmen and captains of industry,” together, decided to accomplish their twin goals of “US military dominance overseas” (the statesmen probably wanted this more but presumably the captains would swoop in later to make these regions their markets for consumption) and “the utter rout of social movements” (the captains probably wanted this more, especially to stamp out the workers movement). These twin goals were to be accomplished by shifting the objective of research and development from cars and machines to medicine and information technologies.

Graeber contends that these choices were political rather than economic — or as he puts it, the ruling class “systematically prioritized political imperatives over economic ones.” What Graeber is saying here, I think, is that the ruling classes were subverting the natural logic of capitalist production. Rather than using money to earn more money (the standard Marxian logic of the capitalist mode of production), they were choosing to develop technologies of simulation (focusing on waging class war, not profits) rather than invest in more efficient technologies of production (which could have alleviated class war and increased their profits). 

To summarize: 

At this point all the pieces would seem to be falling neatly into place. By the sixties, conservative political forces were growing skittish about the socially disruptive effects of technological progress, and employers were beginning to worry about the economic impact of mechanization. The fading Soviet threat allowed for a reallocation of resources in directions seen as less challenging to social and economic arrangements, or indeed directions that could support a campaign of reversing the gains of progressive social movements and achieving a decisive victory in what U.S. elites saw as a global class war. The change of priorities was introduced as a withdrawal of big-government projects and a return to the market, but in fact the change shifted government-directed research away from programs like NASA or alternative energy sources and toward military, information, and medical technologies.

Without explicitly articulating it as such, what Graeber is proposing here is nothing less than a theory of technological change. Such a theory would have to answer the following questions: Why are some technologies invented and not others? Why are some commercialized and not others? Why do technologies look the way they do? Why are they adopted the way they are? Why does the same technology get adopted differently? Why do some succeed (whatever the definition of “succeed”) and others don’t?

Graeber’s answers to these questions are straightforward. The ones with the most power are “statesmen and captains of industry.” These members of the ruling classes are omniscient in that they can predict both the feasibility as well as the social effects of any technology. Inventors have considerably less agency, especially so if they are dependent on the ruling classes for funding, employment, and workplaces (hence Graeber’s claim that the pioneers of the 18th century Industrial Revolution were more innovative because they worked in “small family firms” than the pioneers of the 19th century Second Industrial Revolution who worked for large bureaucratic corporations). 

Consumers and end-users are even further down the chain, and pretty much at the mercy of inventors and the ruling classes. Graeber dismisses as trivial the work required to commercialize a technology, of getting it tightly intertwined into the daily routines of consumers. Sure, there are “clever new ways of combining existing technologies (as in the space race) and new ways of putting existing technologies to consumer use” but none of this is true innovation. Moreover, it is a straightforward matter and consumers don’t really play a role in this process. In Graeber’s theory, what matters is making flying cars. That a world where flying cars are routine would need an infrastructure to support them (refueling stations, driving licenses, repair shops, traffic rules, driving schools, insurance frameworks for dealing with accidents, parking lots, etc.) and that the making of such an infrastructure is not entirely in the control of the ruling classes and the inventors, he pretty much dismisses as irrelevant. Irrelevant in the sense that he believes that if they want to, the ruling classes can bring about this infrastructure without too much effort.  

How should one classify this theory of technological change? I would argue that Graeber is less a technological determinist and more a social determinist. He is a technological determinist in some ways because he believes that technologies have straightforward implications: flying cars and housecleaning robots would liberate the working classes while the internet and medical technologies would not. But he is even more so a social determinist because he thinks the question of technological invention and adoption is entirely up to the ruling classes; they snap their fingers and redirect resources and get the technologies they want. If only the ruling classes had let it happen, we would have flying cars; but because they didn’t (their hatred of progressive social movements exceeded their appetite for profits), we have the internet.

Is Graeber’s theory of technology correct? Suffice it to say that this is not how most historians of technology conceptualize technological change. Within the social studies of technology, the answers to these questions are often intricate and hard to generalize. Scholars often begin from the question of patronage: whose vision of the future gets funded? They then analyze the contexts of invention: how do designers conceptualize the consumer? What specific choices do they make in order to get their technology to “work” (in a pragmatic sense)? Since no technology operates in its own universe, how do they strive to integrate it into the world in which it will be used? Finally, what do end-consumers themselves do with these technologies? The answers to each of these questions can be quite different for different technologies. No one group seems to actually know what’s going on. Patronage, profits, standards, all of these matter, but there seems to be no master-plan.

Managerialism, reborn, as feudalism

Of course, even in Graeber’s story, the ruling classes didn’t quite get what they wanted. He says that even after diverting resources to technological projects that were less politically salient for the working classes, the hoped-for technological breakthroughs didn’t materialize. There are no super-guns, we still don’t have missiles that can kill specific people at long distances, and while the internet is remarkable, he sniffs that “super-fast and globally accessible combination of library, post office, and mail-order catalogue,” no more. And the computer, for all that it has wrought, can’t actually think like a human. 

So what happened? In one word: managerialism. Somewhere along the way, says Graeber, the pursuit of technologies of simulation led to an iron cage of bureaucracy. We talk rather than do. We sell proposals for research rather than do the research. By making the research process into a bureaucratic maze and forcing the researchers to compete against each other, we have pretty much made sure that no breakthroughs will occur except the ones that benefit the ruling classes. “The tyranny of managerialism,” according to Graeber, “pretty much answers the question of why we don’t have teleportation devices or antigravity shoes.”

Common sense suggests that if you want to maximize scientific creativity, you find some bright people, give them the resources they need to pursue whatever idea comes into their heads, and then leave them alone. Most will turn up nothing, but one or two may well discover something. But if you want to minimize the possibility of unexpected breakthroughs, tell those same people they will receive no resources at all unless they spend the bulk of their time competing against each other to convince you they know in advance what they are going to discover.

It is the “tyranny of managerialism” that ties Graeber’s Baffler essay on technological stagnation to the argument he makes in Bullshit Jobs. Unlike the Baffler essay, where managerialism seems to be more of an unintended byproduct (though the text is ambiguous on this point), Bullshit Jobs argues that the ruling classes invented a certain kind of managerialism specifically to keep workers busy in unproductive tasks. Graeber calls this phenomenon “managerial feudalism.” He argues that just as feudal lords needed to create new jobs to secure some legitimacy from their serfs, so also, our ruling capitalist classes have created bullshit jobs to keep some workers occupied even as they increasingly siphon off profits to themselves. (Graeber says that bullshit jobs, which are “white-collar and salaried,” are different from “shit jobs,” which are “blue collar and by the hour.”)

“It’s as if someone were out there making up pointless jobs just for the sake of keeping all of us working,” Graeber says at the beginning of the book and by the time we get to the concept of “managerial feudalism” in Chapter 5, it’s clear that he means this literally. The ruling classes, in Graeber’s tale, are omnipotent; they can decide which technologies get invented; they can also create new jobs out of whole cloth.

Graeber’s evidence for his thesis — managerial feudalism as a cause of the rise of bullshit jobs — is based on a rather tendentious interpretation of the job statistics in the post-war era (see Chapter 5: Why are bullshit jobs proliferating?). He argues that while statistics show a rise in service sector jobs, the real service sector jobs have stayed constant, about 20% of the total jobs; the other non-real service sector jobs are all “information work” and almost by definition, mostly bullshit jobs created to keep workers occupied. 

Others have written in depth about why this interpretation of the job statistics is wrong and I will simply redirect you to those criticisms.³ In this piece though, I want to stick to what I began with: Graeber’s theory of technology and how it undercuts his whole point of writing Bullshit Jobs. 

Graeber’s theory of technology undercuts the ostensible point of Bullshit Jobs

His theory of technology leads Graeber to his very peculiar treatment of what he ends up calling “duct tape” jobs. “Duct tapers,” he says, “are employees whose jobs exist only because of a glitch or fault in the organization.” Graeber confidently asserts that:

Many duct-taper jobs are the result of a glitch in the system that no one has bothered to correct—tasks that could easily be automated, for instance, but haven’t been either because no one has gotten around to it, or because the manager wants to maintain as many subordinates as possible, or because of some structural confusion, or because of some combination of the three. [my emphasis]

Here is that theory of technology again: easily automated jobs that are not automated because the ruling classes don’t want to automate them.

Based on the testimony of a programmer called Pablo, Graeber argues that

Pablo’s main point is that with the growing reliance on free software (freeware), paid employment is increasingly reduced to duct taping. Coders are often happy to perform the interesting and rewarding work on core technologies for free at night but, since that means they have less and less incentive to think about how such creations will ultimately be made compatible, that means the same coders are reduced during the day to the tedious (but paid) work of making them fit together.

So making new software is good; getting different software components to work together, making them “compatible” is just duct-taping. It does not seem to occur to Graeber that duct-taping different software components might be hard and challenging work; that in a world in which software is always running, it is hard to completely replace it with something new. But Graeber’s theory of technology steps in: if managers wanted to, they could, but they don’t, hence the unnecessary duct-tape jobs.

All of which ultimately makes Graeber sound exactly like a classic Silicon Valley entrepreneur who wants to “disrupt” established industries because they are too inefficient and wasteful, and who thinks that a new technology will easily eliminate all that inefficiency and waste.

For example, when the journalist Kevin Carey took a tour of Silicon Valley education startups in 2012, he found entrepreneurs who adopted a transactional framework of education in which education was ultimately about "communication between people" (teachers and students) and the "exchange of information." Since these were "increasingly happening over the Internet" for most other domains—Amazon for shopping, Facebook for communication, etc.—they argued that it was only logical that the same thing would happen to schools and universities.

Carey finds one guy who “wants to use publicly available data like academic rank and grade inflation to standardize the comparative value of different college degrees, then allow people to add information about what they’ve learned outside of college to their baseline degree ‘score.’” He finds a woman who wants to “upset the $8 billion textbook industry with cheaper, better, electronic textbooks delivered through tablet computers.” Another guy wants to create the “first new elite American university in over a century” but online.

All of these entrepreneurs have a very similar theory of technology that Graeber does. Technology can simply be plugged into institutions and they will change; jobs can be easily remade or automated if the right technology is plugged in. Does it work that way? Maybe Graeber and the disruptive entrepreneurs are right but the social studies of organizations and technology suggests otherwise; organizations are more than just conduits of information; technology is not something that can just be plugged into organizational routines. I suspect this argument would not have impressed Graeber; he would have argued that organizations don’t change because the ruling classes don’t want them to.

Why did Graeber write Bullshit Jobs? He tells us why at the beginning:

I would like this book to be an arrow aimed at the heart of our civilization. There is something very wrong with what we have made ourselves. We have become a civilization based on work—not even “productive work” but work as an end and meaning in itself. We have come to believe that men and women who do not work harder than they wish at jobs they do not particularly enjoy are bad people unworthy of love, care, or assistance from their communities. It is as if we have collectively acquiesced to our own enslavement. The main political reaction to our awareness that half the time we are engaged in utterly meaningless or even counterproductive activities—usually under the orders of a person we dislike—is to rankle with resentment over the fact there might be others out there who are not in the same trap. As a result, hatred, resentment, and suspicion have become the glue that holds society together. This is a disastrous state of affairs. I wish it to end. […] If this book can in any way contribute to that end, it will have been worth writing. [my emphasis]

Graeber thinks that we have become a civilization based on work and when people hate their jobs because they are boring, we think they are bad people. The point of Bullshit Jobs is to argue that when people hate their boring jobs, it is because those jobs are boring, that they have been made boring by the ruling classes to keep workers occupied and stop them from revolting. But if we could free ourselves of the ruling classes, then all the boring jobs will be done by machines and all the human beings in the world can do interesting jobs that they will love.

It is unclear to me how Graeber thinks this is building a world where work is NOT an end in itself, where lives are NOT defined around work. In Graeber’s ideal world, work—or to be precise, productive and interesting work—is very much at the center of a life worth living. Graeber argues that this is fundamental to human nature beginning all the way when we are toddlers who discover with delight that we are capable of doing something; losing oneself in something productive, “from running a race to solving a complicated problem” is central to the human experience.

Of course, what undergirds this vision of the anarchist utopia is Graeber’s socially determinist theory of technology; all we need to do to create this world is simply build the technology that gets the boring stuff done so we can all have interesting jobs to do interesting things that delight us. This is also perhaps why Graeber found a soulmate in the right-wing venture capitalist Peter Thiel. What unites them both is that they think that the lack of technological progress in the last 75 years can be traced back to political leadership; both differ, of course, in exactly how and why the leadership failed.

But what if Graeber’s theory of technology is wrong and a safe, prosperous, innovative world actually turns out to be full of “boring” jobs that can’t be automated away easily? Take, for instance, the amazing safety record that American airlines have achieved in the US: no fatal crash since 2009. How did they achieve it? As reporter Andy Pasztor writes:

The astonishing safety record in the U.S. stems most of all from a sustained commitment to what was at first a controversial idea. Together, government and industry experts extracted safety lessons by analyzing huge volumes of flight data and combing through tens of thousands of detailed reports filed annually by pilots and, eventually, mechanics and air-traffic controllers. Responses led to voluntary industry improvements, rather than mandatory government regulations.

In other words, while improvements in engines and aircraft systems played a role, that role was minor. The real improvement in the safety record came from good old-fashioned managerial drive: asking pilots and mechanics and controllers to file “detailed reports.” You can see this system generating hundreds of what Graeber would call bullshit jobs (or at the very least bullshit tasks). But it has a clear dividend: safety.

1. So we get lines like this: “So not only has the hypothesis been confirmed by public reaction, it has now been overwhelmingly confirmed by statistical research” (page xxiv). Later, he describes his methodology and mentions that he “duly color coded” the testimonies he collected. It isn’t clear whether he’s serious or merely saying this to mock academics who take these things seriously. This distinction matters because later on Graeber argues that academics have become timid clerks who only make safe arguments and hide behind issues like methodology—rather than the swashbuckling warriors and innovators he would like them to be.

2. On the other hand, if you accept Greg Ferenstein’s account of the politics of Silicon Valley (and I do), then Graeber does share something with Silicon Valley entrepreneurs. Ferenstein has argued that Silicon Valley wants a “civil society completely oriented toward innovation. They don’t see conflicts between citizens, the government, big corporations, and other countries—just one big mass of people coming up with mutually beneficial solutions as fast as possible.” There is some resemblance between this and Graeber’s anarchist, but endlessly creative, ideal society.

3. For a completely different interpretation (and thereby a rebuttal) of the employment figures that Graeber uses, see these two pieces: Paul Thompson and Frederick Pitts on how Graeber’s distinction between productive versus unproductive jobs is too pat; and Jason Smith on Graeber’s misunderstanding of what the rise of the service sector is actually about.