A Feature Not a Bug: Managerialism as Operating System

REVIEW ESSAY
Fatal Abstraction: Why the Managerial Class Loses Control of Software
by Daryll Campbell
W. W. Norton, 2025, 320 pages

With Boeing’s launch of the 737 MAX, the company promised a high-tech update to the world’s most popular plane. The model contained a software feature new to commercial aviation: the Maneuvering Characteristics Augmentation System (MCAS), which promised to automatically resolve emergencies caused by too much lift without any help from a human pilot. It seemed a sound enough program, that is, until October 2018, when a faulty sensor on Lion Air Flight 610 mistakenly activated MCAS and initiated a dive, overriding the pilot and causing the plane to crash, killing all 189 passengers and crew. Only five months later, the same software malfunction on Ethiopian Airlines Flight 302 led to the deaths of an additional 157 people.

What happened? In one sense, what happened was a terrible design process. Given the software’s complexity, which would cost pilots precious time to master, Boeing’s designers had opted not to tell pilots or airlines about the program, wiping all references to it from training manuals. MCAS activation was sophisticated and unlikely enough, they reasoned, that there was no need for pilots to know about it, much less to spend time getting trained in its operation. The pilots had been intentionally made ignorant of the program, so they couldn’t understand the problem.

But there is a deeper sense of mystery to this question. How is it possible for a technological system to come into existence—one that is as huge and complicated as the world’s largest aerospace company—in which agents utterly lack the information they need to perform their roles, no one can understand why things happen as they do, and it is unknown where blame should rest when disaster strikes? (Only one employee, a technical pilot, faced criminal charges, and he was acquitted. Boeing recently reached a deal with the Justice Department to avoid criminal prosecution.¹)

The question is hardly unique to Boeing. Everywhere one turns today, one confronts some maddeningly opaque system—unreachable customer service call centers, government services marked by inscrutable forms and procedures, smartphones’ ever-increasing intrusion between us and whatever we are actually trying to do—the operations of which defy comprehension by any mere mortal, even those supposedly in control.

This feeling of dread borne of organizational entanglement is hardly new. Kafka’s study of impenetrable bureaucracies, The Trial, turned one hundred this year. Two decades earlier, Max Weber was already describing the “iron cage” built by modernity’s rationality and calculation. But this old frustration has been given a new color by computer technology and its ever-more total digitalization of our lives, which we are condemned to waste away in a hunch while we struggle with buggy but indispensable apps. We might even yearn for the physicality of an iron cage, now obsolesced by a crueler virtual one. Josef K. was prosecuted for an unknown crime, but at least he got to speak to someone in person; we get a chatbot. There is, we suspect, some vast design surrounding us, the unintelligibility of which is matched only by the incompetence of its authors.

Darryl Campbell’s Fatal Abstraction: Why the Managerial Class Loses Control of Software seeks to unravel this mystery and explain the unexplainable digital webs around us. Campbell is both a veteran of the tech world (having worked at Amazon, Uber, Tinder, and other firms) and a chronicler of it (as a former reporter at The Verge). Armed with this experience and perspective, he tells the story of the rise of “managerial software” and the takeover of the tech industry by a class of narrow-minded, bungling elite managers. Fatal Abstraction gets at something essential by placing the unholy marriage of managerialism and software at the center of modern tech’s dysfunctions, but he misattributes their cause. For the unity between software and the managers, between the Computer Revolution and the Managerial Revolution, is no arbitrary, reversible thing but the reflection of an enduring tension between the growth of information and the need for control. Software has not been co-opted by managerialism; it is co-constitutive with it.

The World the Managers Made

Campbell takes as his starting point the Managerial Revolution observed by James Burnham in the early twentieth century. The subject has received much attention in recent years but bears a brief summary here.

For Burnham, as for Campbell, the increased output and regularity of production and distribution in the modern economy introduced a degree of complexity and need for coordination that small, privately owned firms could no longer accommodate. In their place arose industrial giants whose centralization of communications, decision-making, and capital, and whose “separation of ownership and control”² between nominal property owners and the actual controllers of corporate assets, signified a decisive shift from one economic system to another. Capitalism’s reliance on the market mechanism and the unity of ownership and control over the means of production had given way to a new order, one based on deference to the “visible hand”³ of the vertically integrated corporation and the centralized power of the advanced bureaucratic state.

The expanded scale of these new enterprises required a kind of mastery beyond that of the details unique to each industry. It demanded a knowledge of the more conceptual, intangible matters of corporate operations, organizational structure, and financial accounting—matters that applied to any large enterprise, regardless of what it did. To fill the need for this type of expertise, a new character in modern society had to emerge, representing, in Campbell’s words, “a new command-and-control layer at the very top of the company.”⁴

The members of this class specialized in abstraction: rather than focusing on the mundane details of the particular products or services they sold, they floated above in the realm of efficiency and organizational strategy. As Burnham put it, “through changes in the technique of production, the functions of management become more distinctive, more complex, more specialized, and more crucial to the whole process of production, thus serving to set off those who perform these functions as a separate group or class in society.”⁵ The manager might not know how his factory’s widgets got made, but he could read a balance sheet and present quarterly earnings before a board.

Campbell somewhat oversimplifies this story, collapsing into one concept Burnham’s managerialism and the later ideology of neoliberalism. Claims that “professional managers cannot prioritize profits over anything, because in their view there is nothing else worth considering,”⁶ and that managerialism “defines good and bad only in terms of impacts to cash flows and profits,”⁷ relate more to the Friedman-esque vision of business in the 1970s and 1980s than to the problems of organizational bigness that the earlier managers arose to solve. One can read the preeminent business historian of the Managerial Revolution, Alfred D. Chandler Jr., or Burnham himself, without encountering this all-encompassing celebration of money. Indeed, Burnham thought that as capitalism faded away, so would the profit motive. He also understood himself to be describing a transformation that had already been underway for decades. Campbell muddles the history by fusing this long‑term economic transformation with a more recent ideological development.

Nevertheless, his definition of managerialism offers an effective enough starting point from which to critique the tech industry: “the notion that every part of a business can be abstracted into its financial components, and therefore optimized according to scientific principles.”⁸ Managerialism’s power and danger lie in how it forces every domain of human action into the procrustean bed of a generic, contentless administration. “The uniform flattening allows professional ‘managers’—who lack special expertise in what an individual business actually does, but who have significant expertise in the operation of business in general—to parachute into any operation and quickly understand how to balance the competing demands of customers, suppliers, employees, and investors.”⁹ A manager can survey a business proposition with just enough understanding to know how to make a quick buck, cutting personnel waste here, synergizing best practices there, and evaluating key performance indicators before jetting off to the next inefficient venture.

Now, the world does not lack criticisms of the modern cult of efficiency, or of businesses’ prioritization of short-term gains, or of the infuriating insertion of software into, and predictable fragilization of, everything from cars to refrigerators. But Campbell’s argument goes beyond complaints about getting, spending, and complicating. His more novel contribution is to tie managerialism, which dates back more than a century, to the rise of software in recent decades, arguing that their union is uniquely destructive and accounts for today’s inscrutable but inescapable systems.

From Bad to Worse: The Managerialism of Bits

Managerialism, in Campbell’s estimation, was always bad on its own, with its “invocation of non-values such as optimization and the reduction (but not elimination) of collateral damage, the concealment of difficult moral questions behind bland technical jargon, and above all, the dismissal of our common humanity in favor of the inexorable logic of efficiency.”¹⁰ But under what we might call the managerialism of atoms, at least the managers’ greedy grasp was confined to the physical world, establishing some basic guardrails on their actions.

Software demolished those guardrails, enabling an unbounded managerialism of bits. Because software collapses the physical boundaries that constrained the movement of capital, goods, and communication, and can be deployed anywhere on earth instantaneously, managerialism’s focus on efficiency can strike faster than ever. Products are released before they are proven safe, and updates are pushed on users before they are proven functional. Under managerialism, it doesn’t matter whether a defective app traps users in a dead end, or whether a malfunctioning software update proves deadly, as long as profits are going up.

Accompanying the collapse of time and space is software’s abstraction and frictionlessness. Software worsens the intrinsic managerial flaw of putting people in charge of processes they don’t understand: it enforces a knowledge gap between worker and manager—between the man who knows the product and the man who knows the organization making that product and the industry it operates in. The mistaken premise that “an understanding of a technology’s revenue model is the same as understanding the technology itself”¹¹ was well captured in a claim by the don of scientific management, Frederick Winslow Taylor: “the workman who is best suited to handling pig iron is unable to understand the real science of doing this class of work. He is so stupid that the word ‘percentage’ has no meaning to him.”¹²

This illusion of understanding was problem enough when managers were observing, in-person, workers laboring in factories to produce physical items. But software’s intangibility allows managers to delude themselves even more thoroughly into thinking that they understand their company’s products. As “the most frictionless conduit between the human mind and the external world ever devised,” software lacks the hurdles that force us to slow down, assess our actions, and avoid mistakes.¹³ Now, a manager need only push a few buttons to wreak havoc.

The result, Campbell contends, is our maze of computer programs that don’t work but which we nevertheless feel forced to use, culminating in disasters that the managers, in their ersatz control, don’t know enough about to prevent. Managerialism’s agency problem, which ignores long-term risks and declines in functionality in favor of short-term gains, is thus coupled with an epistemic problem—a failure to comprehend the very systems and risks that one is supposed to be overseeing. Software’s formal features induce managers to drop their guard and use it carelessly, and software’s content intensifies the resulting problems, producing today’s “dangerous union of corporate blindness and technological power.”¹⁴

Fatal Abstractions in Action

Most of Fatal Abstraction consists of case studies that aim to extract from various instantiations of managerial software an account of the thing itself. Besides the Boeing disasters, Campbell covers Uber’s self-driving cars, PowerPoint, social media, and artificial intelligence—with varying degrees of success.

The Boeing story is the most compelling for Campbell’s thesis, given the clear role of both management and flawed software, and their ruinous consequences. In Campbell’s telling, the MAX crashes were the culmination of a decades-long decline precipitated by Boeing’s obsession with efficiency. The MAX “was not intended to be the highest-quality or highest-tech product in its category, only the cheapest.”¹⁵ Managers’ obsession with lowering costs and hitting absurd deadlines, despite their weak grasp of the engineering challenges facing products, necessitated cutting corners and forgoing due diligence, especially around safety. In any organization that comes to rely on cheap kludges and that follows the “one unbreakable law of project management: something can be done fast, cheap, or well, but you can only pick two,”¹⁶ disasters such as the MAX crashes are all but guaranteed.

Other case studies follow a similar pattern: a managerially run tech company wants to make money over everything else; managers foist a new technology on their business, or their customers, which they themselves barely understand, and which barely works; and finally, calamity ensues. Uber, for instance, saw software as the key to getting self-driving cars on the road as fast as possible: as the head of its autonomous vehicle program explained, “We don’t need redundant brakes & steering, or a fancy new car, we need better software . . . . To get to that better software faster we should deploy the first 1,000 cars asap.”¹⁷ The result was the death of a pedestrian just months after Uber halved the number of humans required to monitor autonomous cars’ sensors.¹⁸ As with the Boeing crash, culpability here is obscure, perhaps by design: the single safety operator involved in the crash was charged with negligent homicide, but Uber itself faced no criminal charges.¹⁹

Some case studies seem to have been selected less because they support Campbell’s thesis of the perfidy of managerial software than because they permit him to recount his own experiences in the tech industry.²⁰ His criticism of PowerPoint as “a weapon of executive deceit,”²¹ for example, allows him to recount his fifteen years at companies that used PowerPoint presentations to mislead investors, customers, and employees. But Campbell offers no reason to view the software as intrinsically complicit in the fraud. If the use of PowerPoint by Theranos’s Elizabeth Holmes, FTX’s Sam Bankman-Fried, and McKinsey consultants teaches anything, it is that a sucker is born every minute—not anything about the particular nature of managers’ usurpa­tion of software. Indeed, this reviewer had never been tempted to defend PowerPoint until seeing it defamed here as an accomplice in Colin Powell’s erroneous presentation to the United Nations Security Council about Saddam Hussein’s weapons of mass destruction.

For all the promise of Campbell’s concept of managerial software, much of Fatal Abstraction offers what amount to rather typical examples of tech companies’ malfeasance, or at least absurdity. Campbell tells a hilarious story, for example, about “Uberversity” struggle sessions to convince employees that no criticisms of the company were well founded. But such ridiculous behavior can be laughed at easily enough without any theory of managerial software. Campbell implicitly concedes as much when he traces Uber’s crashes not only to the company’s “scale [and] its financial needs,” but “above all [to] the cult-like allegiance that it demanded from its employees.”²²

Nevertheless, although Campbell’s criticisms of particular tech companies are not always compelling, he is right to identify a profound connection between managerialism and software. He also has his finger right on the pulse of today’s pervasive sense that no one, least of all the peddlers of the latest mandatory upgrade, knows what’s going on. The tech industry has become, to borrow from economist Dan Davies, one big “accountability sink,” with no one to clearly and justly punish.²³ What, then, is to be done?

Hacking Towards Utopia

If the problem is managerial software, there would seem to be two potential solutions: get rid of the managers, or get rid of the software. Campbell hardly considers the second option, opting instead for overthrowing the managerial class and, rather than destroying their works, rebooting them with a new ideology. We need “software that is not shaped by managerial imperatives”²⁴—a tech industry in which safety, collateral harm to society, and other qualitative values are incorporated into product development and deployment, rather than being dismissed as nonexistent due to their absence from the balance sheet. Such a governing philosophy would offer “an approach to software that values modesty alongside ambition, safety alongside speed, and our shared human qualities alongside purely financial qualities.”²⁵

It would also entail overhauling the industry’s entire ethos. Campbell sees the nonstop grindset and pitiless hours characteristic of the tech world as unavoidably contributing to the industry’s loss of control. Accidents are inevitable if everyone is too focused on hitting quarterly targets and pushing out the next update to stop and check whether the software actually works. For Campbell, such precaution is strictly incompatible with the absolute elevation of efficiency: “managerial companies have only one strategic impulse . . . get big, do it fast, and make as much money as you can, no matter the consequences.”²⁶ The needed reforms, therefore, cannot happen as long as managerialism remains in place.

Of course, the managers aren’t going away without a fight. But Campbell is skeptical that more aggressive regulatory or antitrust policy, or reforms to corporate governance, are possible at this point; the biggest tech companies are just too powerful for government regulators to rein in. Instead, if we are to “rescue software from the managerial class altogether,”²⁷ the tech workers of the world will need to unite in essentially hacking their own products, contriving to add “some additional safety logic”²⁸ beyond what the managers mandate. After all, managers don’t understand the products they are overseeing, so they’d never catch the workers who jam their own software to save lives, even if they also reduce sales, thereby ensuring that software reaches the user only once it has proven trustworthy. Campbell’s proposal is a sort of reverse Bridge on the River Kwai: here, the managers are the ones ordering the construction of a faulty, dangerous bridge as quickly as possible, and it’s up to the POW coders to undermine the mission and secretly make the bridge safe.

There’s something intuitively appealing about this argument: if a technology is being used in a bad way, why not just start using it in a good, non-managerial way? After all, software “has no wants or desires of its own. It can only do what it’s told by the managerial class, which acts through the tech workers that it oversees.”²⁹ But this would be to fall for the fallacy, in Marshall McLuhan’s words, of the “technological idiot,” who sees technologies as neutral tools that can be equally used for good or for ill, with no necessary connection between the nature of the technology and its social consequences.³⁰

Elsewhere, however, Campbell points to just such a connection, intimating a truth that undercuts his hope for a digital world without managers, namely that the two elements of managerial software form a package deal, united by functions and incentives stronger than a Big Tech cash grab. There is a provocative suggestion implicit in Campbell’s argument: that there is an elective affinity, to use Weber’s own term, between managerialism and software, based on their shared abstract nature, promise of universal applicability, accommodation of scale, distaste for particularity, and demand for rationalization. The linkage that Campbell perceives between the two is no fluke of avarice but rather an indication that they serve a similar purpose and offer mutually reinforcing modes of engaging with the modern world.

Managerialism Is Eating the World

Campbell’s insistence that software has been hijacked by managerialism, and that we can have one without the other, supposes that the managers’ embrace of software has been a purely contingent affair. Managers have been managing for a long time, one might think, and so when the information economy took off and public attention and financial returns shifted from the tangible to the intangible economy, they simply kept doing what they were doing, only now online. Even when he writes that “high technology and managerial control were inextricably linked,” he claims that such a link exists only “in the eyes of the corporate world,” not in reality.³¹ Campbell thus approaches, only to ignore, a profound congruence between managerialism and software, missing a deeper reason why the former has seized upon the latter: software is the perfect expression of the managerial worldview and the ideal instrument for its mission, making the Computer Revolution but the latest chapter in the older, underlying Managerial Revolution.

What Campbell misses (perhaps because he merges the administrative challenges that managerialism arose to address with neoliberalism’s obsession with profits) is software’s tremendous power to increase what James C. Scott called the “legibility” of the world—the degree to which some discrete object can be “identified, observed, recorded, counted, aggregated, and monitored,” as the result of an intentional effort to organize and make manipulable a mess of complexity.³² The computer, with its promise of translating everything into the uniform, comprehensible format of ones and zeros, promised a legibility hitherto unattain­able, a legibility naturally alluring to the manager, even aside from any financial promise. Taylor, with his then seemingly sophisticated, but now terribly crude, stopwatch method to boost worker productivity, would have bowed before the awesome power of the Excel spreadsheet, a program that would have been more apposite than the PowerPoint slide for a critique of tech-powered “quant-maxxing.” The computer has proven to be the most powerful mechanism thus far created for imposing the very kind of order from chaos that has always been the managers’ objective.

The relationship between software and managerialism, between the Computer Revolution and the Managerial Revolution, is thus not one rooted in mere revenue-chasing—but rather in the manager’s recognition of a similar function and (if we may anthropomorphize a human artifact) a like mind. The manager succeeds by mastering the form, not the content, of an industry and by forcing every foreign detail into the same homogeneous template. What could be better than the world of bits’ promise of leveling all the complicated particularity of the world of atoms? This has always been the task of the managers: to, in Campbell’s words, “make sense of a world that has grown beyond our capacity to understand.”³³ Something similar has always been the task of the computer, going back, at least, to Claude Shannon’s foundational work reducing all communication to standardized, interchangeable bits.

In the same way, the proponents of what Scott called “high modernism” observed a world teeming with variation, accident, and incommensurability—unplanned cities, emergent linguistic quirks, and organically evolving naming practices—and sought to replace them through the forced establishment of a clean, fixed, rational design. “Seeing like a manager” is little different from, in Scott’s formulation, “seeing like a state.” Indeed, Campbell sounds like an updated Scott when he writes, “The whole point of abstracting the real world into the digital one is to impose order upon it: to find the patterns that are too difficult for humans to grasp, the corrections that are too subtle for us to make.”³⁴ Campbell’s hapless bosses and inept tech overlords exhibit the same assumptions as Scott’s modernists: that reality can be decomposed into individual, manipulable parts; that processes can be optimized through measurement; and that subjective human judgment can be replaced with objective systems.

Lest this relationship between information technology and managerialism appear too “ideological” or “superstructural,” history also shows a profound material connection between the two. In fact, it would be a surprise if software were anything but seized upon as an instrument of managerial control, given the longer histories of IT and management—histories which, as historian James Beniger has shown, are practically coterminous.³⁵ As he argues, the unprecedented speed of, and corresponding spike in, information and communication during the mid-to-late nineteenth century precipitated a “crisis of control,” that is, “a period in which innovations in information-processing and communication technologies lagged behind those of energy and its application.”³⁶

Beginning with transportation in the 1840s, then spreading to distribution, production, and eventually marketing by the 1880s, the American economy found its capacity to process information overwhelmed by the Industrial Revolution’s volume and speed. With every part of the economy—matter, energy, and information—moving ever faster, and with few extant institutions, the market mechanism included, proving capable of meeting the demand for system stability and coordination, society’s wheels risked coming off.

This crisis was eventually solved by a “Control Revolution,” defined as “a complex of rapid changes in the technological and economic arrangements by which information is collected, stored, processed, and communicated and through which formal or programmed decisions can effect societal control.”³⁷

Such changes in information processing came in two forms. There were technological innovations: among others, the typewriter, stock-ticker, dictating machine, punch-card tabulator, mimeograph, and photostat, all of which helped to accommodate the growth in production, distribution, and consumption. And there were new types of social structures to capitalize on these technologies: modern bureaucracies, coordinated large-scale business enterprises, and a class of managers to oversee the increasing rationalization of society.

What this emphasis on a dialectic between information and control contributes to the usual telling of the Managerial Revolution is the fundamental role played by technological change, with its tendency to both cause and solve crises of informational coordination—to both create a need for, and make possible, a new layer of administrative control—in facilitating the rise of the managers.³⁸ For the entirety of the Managerial Revolution, there has been a mutually reinforcing relationship between new technologies of information processing and managerial control. The relationship long predates, and likely will outlive, Campbell’s bogeymen at Microsoft, Meta, OpenAI, and today’s other tech giants.

The rise of software and the intangible economy ushered in a new but continuous episode in this history. When Beniger published his history in 1986, the world’s computational capacity could accommodate a few billion gigabytes of information.³⁹ Today, that figure is measured in the hundreds of zettabytes.⁴⁰ (One zettabyte equals a trillion gigabytes.) The computer-driven information explosion is exploding the very language we use to describe it: in 2022, the new prefixes ronna (10²⁷) and quetta (10³⁰) were coined to accommodate the demands of IT and the “exponentially growing size of the global datasphere.”⁴¹
In the face of this informational onslaught—essentially, a new crisis of control—not just the continuation but the acceleration of the managerial project became necessary, resulting in another iteration of the Control Revolution that is as monumental and complex as the demands of the digital age.

As Beniger observes, “All else being equal, increases in power will always result in increases in speed, which in turn increases the need for control and hence for communication, information processing, programming, and decision,”⁴² the same abstract technical functions that managers have always performed and sought to perfect. For all of Campbell’s attention to software’s unrivaled speed, he fails to appreciate that with greater speed comes the need for greater control, and hence for greater management. Hoping for the one without the other, for software without managerialism, is a mug’s game, for the two constitute intertwined elements in a single Control Revolution.

What remains for some future Beniger to write is the history of this unfolding phase of the Control Revolution, showing how the computer created both a new crisis of control and the means by which an adapting managerial class could overcome it to secure its place in the new technological reality. In light of the coeval growth of information, technologies to process such information, and managers to administer such technologies since the Industrial Revolution, Campbell’s “managerial software” is something of a pleonasm, representing merely the acceleration of an old and established trend. Even the “information overload” that Alvin Toffler warned of in his influential 1970 book Future Shock is not really a problem particular to the computer age but rather a recurring challenge for all industrialized economies. Now, we have come full circle: managers embrace software because the control crisis offered an undeniable opportunity not just to enjoy strong financial returns but to reassert administrative dominance over the economy and society, in what is a faithful extension of high modernism’s enterprise. Campbell is correct to observe that “Managerialism is a doctrine of control.”⁴³ But he misses its essential corollary, that software is a technology of control.

Detailing this relationship, which Campbell occasionally suggests but never brings out into the open, would be more useful for understanding the political economy of tech, and the reasons why the proliferation of software and the digitalization of our lives have felt so inexorable, than moralistic criticisms of managers’ greed. It has often been said, following venture capitalist Marc Andreessen, that “software is eating the world.”⁴⁴ But in light of the ongoing Control Revolution, of which Fatal Abstraction’s case studies are but a few recent episodes, the truer statement would be that managerialism is eating the world. Software is but the latest instrument in the long-standing managerial project to establish systematicity over a world of mounting complexity, which paradoxically makes control seem both necessary and impossible.

Iron Cage: The App

Seen from the longer perspective of revolutions both managerial and technological, dreams of liberated software, unconstrained by the stern hand of the manager, are likely to end in disappointment. Managerialism is not just one ideology among many, or a bit of malware disrupting the information age, but the operating system that allows the entire “techno‑capital machine” to run.

Here, Campbell might have engaged more seriously with Burnham himself, who scorned such hopes for a society without some top layer of control. For Burnham, no elimination of the class position currently filled by the managers was possible, in the broadest sense that there would always be a cohort of elites running the show from the top of every society, including modern industrial society.

Certainly, not every elite has been as dully focused on the blinkered “neoliberal” priorities of recent decades’ tech CEOs. But any plot for a revolt by the tech proletariat, a takeover of the machine that yet leaves the machine intact, will fly in the face of the enduring connection between the growing complexity of the machine and the need for a master of that machine. As Burnham himself wrote, “When an organization grows to a certain size and when its aims have a certain scope and importance, the conduct of the organizational affairs becomes itself a considerable activity. . . . The organization is able to keep alive and to function only through its leaders.”⁴⁵ For all the computer’s reshaping of society, even it has not overturned the iron law of oligarchy: “Whatever social changes occur, whatever happens to economic relations . . . organization will remain, and through organization an oligarchical rule will be perpetuated.”⁴⁶

Still, for all the managers’ durability, it is worth remembering that for Scott, as for Campbell, the high modernist pretense of knowledge is ultimately just that. Managers have seen in software a technology they think they can understand; as at least the better case studies in Fatal Abstraction show, the managerial class loses control of software and does not actually comprehend the programs it claims to run. Campbell’s documentation of botched tech rollouts and overridden safety concerns is the mirror, however conceptually foggy, of Scott’s history of high modernism’s wreckage, from the Great Leap Forward in China to agricultural collectivization in the Soviet Union to forced settlement in government-planned villages in Tanzania. In each case, the modernist’s and the manager’s struggle to overlay a messy reality with a neat order worked—but only for a time, with the extent of the system’s eventual collapse proving proportional to the depth of the illusion of control.

We are witnessing the warning signs of another such collapse today, as the public pulls back the curtain and finds that the experts no longer understand the systems they run. Having outsourced the provision of many public goods to the tech sector and made basic participation in public life dependent on access to its products, and having also made expertise the criterion of legitimacy, we lack any apparent resources for shoring up the prevailing regime and restoring coherence. The institutions of public life may still possess nominal authority, but they have forfeited intelligibility: citizens and consumers encounter frail user interfaces for unknown backend systems. Each of Campbell’s disasters, by publicizing our rulers’ shoddy work, has done its part to dispel the image of competent administration and undermine any claim to legitimacy. And the disasters will continue until morale improves; in the months since Fatal Abstraction was released, a different model of the Boeing 787 has crashed, killing at least another 270 people.

Perhaps, then, the hubristic managerial project of ceaseless standardization, abstraction, and quantification will eventually come to an end, incapacitated, as are each of Scott’s and Campbell’s case studies, by its own internal contradictions. Burnham himself would be the first to say that no economic order or social structure is permanent, that, in fact, each is more fleeting than most realize.⁴⁷ But until such a terminal event should occur, the modern triad of revolutions—Managerial, Computer, and Control—and therefore today’s unfathomable yet unavoidable managerial software, are here to stay.

This article originally appeared in American Affairs Volume IX, Number 3 (Fall 2025): 69–84.

Notes