In Cubed, Nikil Saval traces the beginning of the professional class of workers back to the clerks memorialized in Melville’s Bartleby, the Scrivener. Among the things he identifies: the earliest clerks saw themselves apart from other kinds of workers, and fashioned a distinctly different ideology about the nature of their work—one that is depressingly familiar:

Manual workers earned hourly or piece-rate wages, while nonmanual workers earned annual salaries. What this meant for white-collar workers, in an American economy beset by intense fluctuation in prices and frequent financial panics, was a measure of stability that manual workers never enjoyed. A small shift in power had begun to take place. If people who “worked with their hands” still assumed their possession of the world of things, clerical workers, those “working with their heads,” were now at the heart of capitalism’s growing world of administration and direction—close to power, if not exactly in control of it.

And so unlike “solidarity,” the key word of the European industrial labor movement that had made its way to England and America, the ethic that had begin to take hold among clerks was that of “self-improvement.” Clerical workers were uprooted from the close-knit world of families and farms, where knowledge was passed down from father to son. Other clerks were merely competition; they had no one to rely on but themselves. “The man who does not at least propose to himself to be better this year than he was last must be very good or bad indeed,” wrote the merchant’s clerk Edward Tailer in his diary entry on New Year’s Day 1850. There is, he continued, “no such thing as a stationary point in human endeavors; he who is not worse today than yesterday is better; and he who is not better is worse.”

Saval, Cubed, page 21

There is a certain kind of awareness that comes from realizing the tradition of using New Year’s to consider how you will improve your (working) self goes back at least one hundred and seventy one years. But what’s really telling about Saval’s analysis here is the way in which self-improvement is seen as antithetical to solidarity: focusing your energies on yourself, and beating out your peers, takes up the kind of space that collective action might have. And still often does, alas.

Among the threads that Saval traces in Cubed is that of the evolution from Taylorist “scientific management” (with its emphasis on mass-production and unskilled labor) to post-Fordist methods (with a stated preference for flexibility and innovation). But Saval notices that this supposed evolution rests on a shaky interpretation of the past:

Knowledge work itself came from a historic shift, one that [Peter] Drucker, like so many, traced to Frederick Taylor. But his version of history was marked by a curious and useful elision. In Drucker’s account, Taylor came upon a working world characterized by rote, nearly mindless, activity. It wasn’t so much planned as willed: the workers simply worked harder rather than “smarter.” Until Taylor, that is: “Taylor, for the first time in history, looked at work itself as deserving the attention of an educated man.” Drucker’s subsequent description of the insensate labor of unskilled men in factories draws almost entirely from Taylor’s portrait of them—and accordingly condescends to their abilities to plan and organize work. In actual fact, it wasn’t so. Before Taylor, work was already organized by teams of factory workers, who in large part had control over how they worked. The knowledge they applied to work was largely “tacit” in nature, agreed upon among the workers themselves and developed through a silent or coded language, rather than “explicit” (to borrow a famous definition from the sociologist Michael Polanyi). What Taylor sought in particular—indeed, what constititued his signal obsession—was to extract this tacit knowledge from the workers and install it in another set of people, the “industrial engineers.” Drucker called them the “prototype of all modern ‘knowledge workers’”—a plausible assumption but one that excised the tremendous amount of knowledge that already existed in the work process. (Taylor lamented that after being taught “the one best way,” workers had a stubborn tendency to return to their own ways of working.) It was a useful fiction, and a common one, that helped to uphold a new class of technicians and professionals as the masters of an ever more progressive society, dependent on the application of knowledge to work.

Saval, Cubed, page 196

I take a few things away from this: first, that the idea that manual laborers are unskilled is, of course, a lie—they have been deskilled, having had the use of their skills stripped from them and given to their bosses, supposedly the only ones who could be trusted with that information. But just as importantly, the notion that the modern-day workplace has done away with deskilling is itself a useful fiction. Hierarchical management structures across industries routinely serve to hoard planning and process decisions at the top. Even the language of “knowledge workers” elides who has power to declare which knowledge is right and actionable. Drucker’s vision of the modern workplace was glossier than Taylor’s, but it papered over the power structures rather than upending them.

Writing about the burst of new communications systems in the late nineteenth and early twentieth centuries, Nikil Saval notes that:

The invention of telecommunications allowed offices to be separated from factories and warehouses and in turn expanded the range of jobs available to office workers. Consider a mail-order company and its warehouse. Bosses and errand boys no longer had to conduct transactions in person; aside from the occasional messenger, it was now possible to send and receive information from your warehouse, or factory, or printing house, using Morse code or by picking up the phone. Within the office itself, the use of pneumatic tubes made it possible to run material between levels of a company, while Dictaphones, which could carry a smoke- and bourbon-ragged voice from a snug top-floor executive suite down to a harried typist marooned in the steno pool, made transcription rational and smoothly impersonal.

Paradoxically, this new capacity to communicate speedily and efficiently resulted in more and more work for workers to push through—more physical products, but also more paperwork (invoices, receipts, contracts, memos, profit-and-loss statements), which meant more typewriters, which meant more typists, which in turn meant more messages and therefore more messengers.

Things that were heralded as “laborsaving” devices gave rise to a whole new industry, and to more labor. As the great theorist of technology Marshal McLuhan put it in Understanding Media, “It was the telephone, paradoxically, that sped the commercial adoption of the typewriter. The phrase ‘Send me a memo on that,’ repeated into millions of phones daily, helped to create the huge expansion of the typist function. In Sinclair Lewis’ office novel The Job (1917), his protagonist, Uma Golden, suffers a kind of sublime dread when she confronts the expanse of supposedly laborsaving machines: “machines for opening letters and sealing them, automatic typewriters, dictation phonographs, pneumatic chutes.” But she’s surprised that “the girls worked just as hard and long and hopelessly after their introduction as before; and she suspected that there was something wrong with a social system in which time-saving devices didn’t save time for anybody but the owners.”

Saval, Cubed, page 40

(Emphasis mine.) Sound familiar? That same siren song of laborsaving technology echoed throughout the twentieth century, with vacuums and dishwashers and personal computers and 5G networks and on and on and on. Yet the average number of hours worked (both waged work and housework) hasn’t budged. The present-day fiction of so-called AI is the same story, second verse. The people urging AI on us won’t save labor; they will try to degrade that labor, to further alienate us from it, to demand that we work as long and hard and hopelessly as ever. Unless we refuse.

In 1898, Frederick Taylor was hired as a consultant by the Bethlehem Iron Company with the stated mission of improving the efficiency of the workers. It was there that Taylorism morphed from the wheedling ideas of an eccentric into canonical corporate practice. As Nikil Saval notes in Cubed, Taylor’s recipe for efficiency rested on a singular, and dehumanizing, foundation:

The key, [Taylor] would discover, was to take knowledge away from the workers and install it in a separate class of people.

Saval, Cubed, page 47

That being the managers, of course. Taylor’s model of workplace productivity depended entirely on deskilling, on the invention of unskilled labor—which, heretofore, had not existed. More than half a century later, long after Taylor died while gripping a watch, Peter Drucker would pick up the baton he left behind and intone about the arrival of “knowledge workers.” But his definition of this new class of workers rested entirely on Taylor’s stories of their supposedly unknowledgeable peers:

Drucker’s subsequent description of the insensate labor of unskilled men in factories draws almost entirely from Taylor’s portrait of them—and accordingly condescends to their abilities to plan and organize work. In actual fact, it wasn’t so. Before Taylor, work was already organized by teams of factory workers, who in large part had control over how they worked. The knowledge they applied to work was largely “tacit” in nature, agreed upon among workers themselves rather than “explicit” (to borrow a famous definition from the sociologist Michael Polanyi). What Taylor sought in particular—indeed, what constituted his signal obsession—was to extract this tacit knowledge and install it in another set of people, the “industrial engineers.” Drucker called them “the prototype of all modern ‘knowledge workers’”—a plausible assumption but one that excised the tremendous amount of knowledge that already existed in the work process.

Saval, Cubed, page 197

In other words, Drucker’s now-infamous formulation of knowledge workers only makes sense if you accept the premise that other workers do not themselves truck with knowledge. But that premise was the product of theft—an outcome of Taylor’s extraction rather than a natural or immutable fact of the work. Or, as Saval writes:

In this respect, it’s probably better to think of knowledge work as the name of a desire, or a hope, rather than an actual feature of the workplace.

Saval, Cubed, page 199

Perhaps it’s even better to acknowledge that there never were any knowledge workers. There have only ever been workers.

In the opening of Reclaiming Work, André Gorz writes provocatively:

We must dare to prepare ourselves for the Exodus from “work-based society”: it no longer exists and will not return. We must want this society, which is in its death-throes, to die, so that another may arise from its ruins. We must learn to make out the contours of that other society beneath the resistances, dysfunctions, and impasses which make up the present. “Work” must lose its centrality in the minds, thoughts, and imaginations of everyone. We must learn to see it differently: no longer as something we have—or do not have—but as what we do.

Gorz, Reclaiming Work, page 1

Gorz is writing long before AI hypemen started to promise the elimination of millions of jobs, but automation has always been the aim of capitalism, since the first steam-powered looms displaced experienced weavers and installed children in their place. Now, that automation is coming for “knowledge” work—for work that was hailed as somehow exempt from the degradation and alienation of all other work, but whose hailing was only ever a thin and shabby cover.

[The work that is ending] is, unambiguously, the specific “work” peculiar to industrial capitalism: the work we are referring to when we say “she doesn’t work” of a woman who devotes her time to bringing up her own children, but “she works” of one who gives even some small part of her time to bringing up other people’s children in a playgroup or a nursery school.

Gorz, Reclaiming Work, page 2

That is, the work that is being abolished is work that is institutionalized and exploited, that makes a profit for someone other than the worker. Work where there are rules and metrics, performance-improvement plans, stack-rankings, hastily arranged all-hands where executives perform the meanest apology after the latest round of layoffs while suggesting those workers had it coming (and you likely do, too). This isn’t “work” in the sense of making change or making a contribution to the world, but work as the price you pay for the privilege of keeping a roof over your head—absent any promise that the roof will always be there.

[I]t is precisely in the sense of self-realization, in the sense of “poiesis,” of the creation of work as oeuvre, that work is disappearing fastest into the virtualized realities of the intangible economy. If we wish to rescue and sustain this “real work,” it is urgent that we recognize that real work is no longer what we do when “at work”; the work, in the sense of poiesis, which one does is no longer (or is increasingly rarely) done “at work”; it no longer corresponds to the “work” which, in the social sense of the term, one “has.”

Gorz, Reclaiming Work, page 2

Poiesis is the emergence of something that wasn’t there before; it’s the work of creation. But so much knowledge work—even before AI started to be shoved down our throats—wasn’t creative but necromantic: raising up the de-fleshed bones of past successes and reconfiguring the joints into recognizable but beastly products with glossy marketing wrapped around them like an emperor’s cloak. The assertion that slop-makers will replace most jobs is a pathetic attempt at resurrection: only work that has been demeaned into the grave could be supplanted by such boring and obsequious ghosts. The real work has always been elsewhere. Perhaps it’s time we follow where it takes us and leave the dead to their tombs.