In his recent essay in NOEMA, “Why is Everyone so Sad in Tech?”, Aaron Horwath paints a picture of a wave of people taking up pottery, knitting, and old-timey analog hobbies, while confessing that they suspect the jobs they’ve built careers around might not mean much. Horwath borrows a term, “workism,” from Derek Thompson’s 2019 essay in The Atlantic, describing a generation of high earners seeking the meaning and community from work that earlier generations found in religion or vocation. This sounded eerily familiar to my own experiences as a person who makes websites for a living.
Workism (a disease I may have acquired and could still struggle with) felt like the genesis of an earlier concept, one foundational to modern capitalism, “Taylorism”.
A time to every purpose
In 1881, at twenty five years old, Frederick Taylor conducted a time study at the Midvale Steel Company plant. He stood with a stopwatch and measured how long each motion of a worker’s task took, then broke the work down into its smallest repeatable parts. This became the basis for Taylor’s management theories, and Taylorism was embraced as a “scientific management system” for optimizing worker efficiency and productivity. A worker specialized, focused on a single motion, and the “unnecessary” movement was designed out of the job. Agile, waterfall, sprint planning anyone?
At the same moment, an equal and opposite reaction was taking shape. William Morris and the Arts and Crafts movement looked at the same mass production Taylor was optimizing and saw something being lost. Morris argued that the disappearance of artisan craft and quality was dehumanizing, that it severed the connection between a maker and the object they made, a connection he believed was essential to human fulfillment. He also argued that workers deserved pride in their craft and fair wages for it, and that beautiful, well made objects shouldn’t be reserved for people who could afford to commission them.
Master of my craft
Alongside mass production, growing energy demands, and a run of impressive engineering feats — suspension bridges, the internal combustion engine, the Eiffel Tower — came the professionalization of science and engineering fields. Scientists became public figures who had opinions that mattered with the credentials and associations to prove it. Modern genius was celebrated and mythologized, a cult of personality.
Industrialization and specialization created real opportunity, but they also created a new class: the person recognized as belonging to a specialized group with the authority to say what counted as correct or valuable work. That class, coincidentally, was also a white and male one. Professionalizing expertise wasn’t only about codifying knowledge, it institutionalized whose knowledge counted.
AI potentially tweaks that inheritance. Where the old expert class gatekept authority behind credentials, an AI tool hands anyone the tone of authority on demand — the confident structure, the ability to spellcheck and make grammatically correct sentences, immediate fluency in any language or topic. That sounds like democratization but we still need to be aware that the underlying judgments and information presets still come from somewhere fallible. They are built into a model trained by a small group of people, with a cultural bias for particular conclusions and meaning.
And access to knowledge is still specialized, “knowledge workers” are untouchable acolytes to the rarified world of data. Or, we thought we were, until AI threatened to take our jobs.
Here comes the sun (or not)
The epic story of industrialization is Man v. Nature (yeah, we won, good job everyone), it’s also a story about what aspects of the social fabric were subsumed by new industries and factory work: our relationship to land, family proximity, church, a legible community, a slower and shared sense of time. Telegraphs and trains (and, of course, later, phones, faxes, and the World Wide Web) collapsed the sense of time and space which had once structured how we related to each other and our intake of information.
Factory shifts collapsed something else, our relationship to the sun. Workers on night and early shifts could no longer rise with daylight and sleep at dark, so a trade sprang up to solve the problem: the “knocker-upper,”1 who walked the streets before dawn tapping on windows with a long pole to rouse people in time for a shift.
Modern life had begun to speed up our experience of time.
Whistle while you work
With this new sense of time as a resource to optimize rather than a thing to live in, being busy became a kind of evidence, proof that you were a contributing member of society, that your life had value (this belief is deeply enmeshed in our culture as evidenced in persistent rhetoric that dehumanizes people who rely on government services, or who are unhoused).
Devices and screens have taken that compulsion and 10X’d it. They’re built to be always on and always available. This intentional design feeds a specific, gamified fear of missing out: the sense that we can send one more email, refine one more process, finish one more task, and then we’ll be caught up. That feeling is an illusion, because work online doesn’t ever end. It’s self propagating. There’s no bottom or finish line.
The company store is the algorithm
If the work itself doesn’t offer meaning, what’s left is what the work can buy. A factory worker earned enough, sometimes, to shop at the company store. What we can afford with the money we’ve traded our time for starts to tell a story about our place in the world. Do we rent or own? Bike or drive? Shop at Whole Foods or Aldi? Consumption becomes a substitute for meaning and self-worth, and it also becomes the public-facing mask we identify with.
Craft against the machine
When a process can be automated, whether by an assembly line or an AI tool, the value of the experienced human hand behind that work will lose out to speed and efficiency, whether or not that speed and efficiency is genuine or only perceived.
It still takes real time to train a tool to produce something that reads as thoughtful. But eventually, the rough edges are worn away. Imperfections smoothed over. Choices that push in a different or difficult direction remain untraversed in favor of what becomes a more predictable outcome. The result is homogenous, which is the goal of a program that maximizes efficiency. A product built to be reproduced at scale was never meant to be unique or memorable.
There’s a bitter irony here, one Morris ran into himself. He wanted beautiful, well made objects available to everyone, not just the wealthy. But hand craft doesn’t scale, and his own furniture and textiles became luxury goods that only wealthy people could afford. The value of art is a weird and illogical anomaly and the distinction between what is craft and what is fine art, whose work matters, and whose is forgotten is an undecidable problem. So, back to AI –
The drawback of a friction-less experience and surface-level fluency is that the process is no longer the product. The ease of “looks good to me” is desirable because it keeps the machine running comfortably along it’s familiar track and we are rewarded with the warm glow of productivity.
Love / Hate
People have opinions about AI. Some are all in, experimenting with new tools and models, genuinely enjoying the inherent possibilities. As users, we don’t generally see the cost behind the results that are magicked up for us (the water, the labor and exploitation built into training these systems). We just get the impressive sounding answer to our request, usually in seconds. Others reject the whole premise of AI or just don’t feel they have any use for it. But, whatever our feeling around it, it’s changing us as well as the literal landscape.
In Charlotte Brontë’s 1849 novel Shirley, a struggling mill owner in Yorkshire introduces labor saving machinery to his textile mill, which leads to widespread unemployment and violent unrest — a fictionalized version of the real Luddite uprisings of 1811 and 1812. Writing a generation after the event, Brontë was looking back at what the machines had cost the community while acknowledging that progress was necessary for the industry to survive and compete.
The mill workers had experience and skills that stopped being of value the moment a machine could do approximately the same labor, faster and more cheaply.
In 2017, Jeff Bezos made comments widely read as support for a universal basic income, in the middle of a broader wave of tech leaders — Musk, Zuckerberg, and others — raising similar concerns about automation and job loss. Scott Galloway, writing in The Four: The Hidden DNA of Amazon, Apple, Facebook, and Google, read that endorsement as something closer to an admission: the person building the robot-run warehouses had already seen where the work was going, and in his vision of the future, there won’t be enough jobs left to sustain the current workforce. But to keep growing and expanding, as successful business requires, they (aka. we) need to be given credits in order to keep feeding the machine (It’s people!).
Faultily faultless, icily regular, splendidly null, Dead perfection, no more.
— Lord Alfred Tennyson (perhaps anticipating AI)
As a freelance designer and developer, AI has expanded what I can competently pursue on my own, when I know what I want and I know what to ask for. The value of my experience and knowledge of underlying processes and behavior inform how I work with my tool (say “Hello” Claude!).
What I keep noticing, though, watching these tools work: they trend toward a smoother, middling result. They can’t make the leap of logic that’s a little nonsensical, even absurd, that opens up a new direction entirely. That part of our conversation with AI remains for our own human discernment.
I trained as a painter, and what I learned there wasn’t really about producing finished art objects. It was about building technical fluency with materials and developing a voice so that a decision could become instinctive. The eye and hand are trained to recognize the right unplanned moment when it shows up, because it can’t be scheduled or repeated on command. Cartier-Bresson called it the decisive moment: his years in Cubism gave him the technical grounding to catch spontaneous, incompatible moments as they intertwined into something coherent, a skill his camera couldn’t do for him no matter how good the lens was. Art, science, and technology have always cross pollinated this way, as inspired thought and experimentation feeding each other. That’s the genius, the human spark, and it can’t be automated.
An AI tool is built to give you the most likely next word, the average of everything it’s seen before. Average isn’t the same as good, but it’s fast, and fast is what we’ve been trained, by a century+ of Taylorism, by devices that never turn off, by the low hum of always having one more thing to do, to mistake for “good”.
“What is adequate journalism if not great journalism?” – Bill McNeal
I could be wrong, I could be right
The concept of a “living room” evolved from the communal hearth where a family unit or tribe would convene to share their stories and connect with their past, present and future. Stick a 75″ LED screen in there and it’s now a “TV room” – the warm glow of connection shed for the blue pulse of being together, alone.
Tools don’t just do the work you give them. They generate more work in their own image. A dishwasher doesn’t save you from dishes and a washing machine doesn’t finish the laundry — they change how you perceive the cost/benefit of that work. The object intuits the demand. Give a tool the ability to generate, draft, and revise, at no apparent cost, and you won’t do less work. You’ll do more of it, faster, at a lower bar, right up until you can’t tell the difference between something you made and something that was made adequately, on your behalf, while you were busy multi-tasking.
The cost of convenience is the loss of time we shouldn’t have to account for — daydreaming while we hand wash a few dishes, taking a wrong turn that leads to an unexpected discovery, making the mistake that becomes invention.
Why is “perfection” (aka. control) a goal? Isn’t mess and disorganization more closely aligned with how we actually operate? We are not machines (or gods), we will never perfect our output because absolute control and predictability are unnatural, static, and false.
Perfection is like a supermarket banana, it’s a seedless clone that will never evolve into anything new. Doing it wrong is the experience we need.
Song titles mentioned:
- Harder, Better, Faster. Stronger – Daft Punk
- Turn! Turn! Turn – The Byrds
- Master of My Craft – Parquet Courts
- Here Comes the Sun – The Beatles, performed by Richie Havens
- Whistle While You Work – Snow White and the Seven Dwarfs
- Sixteen Tons – Tennessee Ernie Ford
- Love and Hate – Michael Kiwanuka
- Rise – Public Image Ltd.
- Together, Alone – Crowded House
- I first read about this in the book Four Thousand Weeks, by Oliver Burkeman. ↩︎
