If you let AI do your writing...

Johanna’s job as a producer of literary novels had finally been automated.

It was such a shock! For years, it’d looked like this wouldn’t happen—Johanna had believed human beings would always have some demand for high-brow fiction written without computer assistance. She had actually read zillions of thinkpieces to this effect: “AIs can never make art!” the articles had claimed.

Unfortunately, in the year 2060, they were proven wrong, and Johanna lost her career.

You might be confused. Why did it take so many decades for this to happen? After all, there’d been this scandal in 2026, where a well-known journal named Granta had printed a piece of AI fiction that’d been awarded a prize. And then after that another thirty whole years passed before literary fiction finally gave up the ghost? That seems absurd.

Well, to understand how the literary world responded to AI, I’ll give you a round-up of the state of play in 2059 (the year before Johanna lost her career).

At this point in time, there was something called the ‘e-novel’ that was essentially AI-generated. These were available mostly for free online, and some huge number of people were reading these books.

Then there was the mainstream novel. Increasingly, these were written by people with an e-novel background, but these e-novelists were forced by the publishing industry to go through a horrible rigamarole, where they ‘re-learned’ how to write their works (that had already been downloaded a zillion times online) in a more human style.

The idea was that these novelists had learned something about narrative from this e-novelist community, but to reach a broader audience, they also really needed to put the skin back onto the bones and write in a more human style.

But even after this re-education process had supposedly resulted in the composition of 100 percent human novels, these former e-novelists were still not particularly respected by the literary community, which viewed them as a suspect and tainted by their early use of AI.

Some of these mainstream e-novelists were so hungry for literary credibility, and they beat on the doors for years, asking you to accept that their mashup of Stephen King and Colleen Hoover was somehow special and unique. That their early AI use had just been a ‘learning tool’ and now they knew better and had somehow transcended their AI origins.

Johanna was doubtful. Yes, she had been entertained by a few mainstream e-novels, but she personally was not going to launder the reputation of some millionaire who published mass-market drek that, even if it was technically human, was still no better than it needed to be.

Because if these mainstream novels started getting prestige, then they’d be in competition with Johanna’s own category. She wrote literary novels—certified one hundred percent human.

How, you might ask, was this certification accomplished? AI detection software was pretty good, but not infallible. False positives existed, and if your work was detected you could always say the software had (in your case) erroneously flagged your work as AI-generated.

But the quality of the software didn’t matter. Because people in the literary world didn’t trust to AI-detection. Instead you just had to know, you had to perceive. That was exactly what was at stake after all, right? Whether you had the ability to appreciate whatever was distinctly human in this writing.

However, in practice, it was a lot easier to ‘appreciate the human’ if you knew beforehand that the writer was definitely a human. And this was really the product that the literary world provided to the reading public. Literary publishers gave you writing that, you were assured, was one hundred percent human, never tainted by an LLM.

The way it worked was that literary judgement was outsourced to small journals.

The literary world was still organized around dozens of these small journals, and all of these journals claimed to be anti-AI.

What these journals provided was the assurance that this writing was human. That was their product. The assurance. They provided that assurance by mostly being highly-networked, tied to a certain time and place (New York City, usually) and by sourcing contributors who had elite educational credentialing (i.e. a degree from Harvard) that, even in 2060, connoted at least the pretense that you knew how to read and that you valued literature for its own sake.

Because what was true of the journals was now true of education as a whole, right? A university could only succeed if it graduated students who could meet organizations’ incoherent AI-use guidelines. Many organizations (most of them, in 2060) believed that there were certain parts of your job you should always do with AI (and you’d be fired if you didn’t do them with AI) and that there were other parts of your job that you should never do with AI, because it was somehow morally wrong to let AI touch this particular part of your job.

This made no rational sense, but basically if you went to a better college, then you were equipped to somehow manage these contradictory demands, because…that’s what colleges taught. And, moreover, that’s what being a smart striver was in 2060. It was figuring out ways to use AI without losing the pretense that it was really a human being in charge.

In practice, writers who’d gone to Harvard were extremely good at exuding the impression that their work was fully human and partook of the same kind of inspired creative frenzy that’d produced all the best literary production before 2022.

Anyway, these journals in 2060, were filled with writing from people who genuinely believed that they were producing the best that was doable. In private, maybe they sometimes felt inadequate about how their writing didn’t hold a candle to the best writing of the past, but they told themselves that lots of people throughout history had felt inadequate.

At this point, in 2060, the literary world was knee deep in the Neo-Lernerian movement, a return to the idea of unmediated experience. This movement was centered around one particular journal, called The Harvest, whose fiction section was full of writers trying to convey the rhythms of real life, using their carefully off-the-cuff anecdotes and shaggy storytelling.

Johanna found it very charming, honestly, because it was a return to the period when she herself (she was seventy-five) had come of age. And, not incidentally, this meant some of her work was coming back into vogue and was due to be reissued by a reprint press [in my mind it’s NYRB, but maybe that’s too self-serving. Hmm, whatever, let’s make it NYRB in this fantasy]. Yes, Johanna’s debut literary novel, a neo-Lernerian opus set in 2010s San Francisco, was being reissued as a distinctive period-piece by the New York Review of Books’s Classics imprint.

It was such a sweet moment for Johanna, even though a part of her protested that her novel wasn’t particularly influenced by Ben Lerner. She’d been much more influenced by Sally Rooney, but now for some reason that influence was very out of fashion (probably due to the author’s stance on various developments in the Middle East) so suddenly you had to pretend like Rooney wasn’t the defining influence on a certain kind of early-21st century novel. A huge portion of literary novels in that era were marketed on the idea that they would appeal to a Sally Rooney-loving reader, who wanted Rooneyesque doings, but with slightly non-Rooney characters.

Some writers had successfully channeled that Rooney magic, and others, like Johanna herself, hadn’t really managed it, because they’d never understood it. Personally Johanna had never gotten why so many women felt drawn to the passive doll-like protagonists that filled the novels of Sally Rooney and her most successful imitators.

In any case, Johanna’s book was mostly defined by its failure to be Sally Rooney. But nobody would read it (or pay any attention to Johanna) if that truth were to be known, so she was happy to pass it off as an early Neo-Lernerian work.

Then, unfortunately, all the accumulated resentment in this system caused it to collapse.

Basically, there was some group of young people who’d really aspired (even after all this time!) to make it as young literary novelists. But they’d told themselves that it didn’t pay the bills, which meant you needed a day job, so they’d gotten really boring jobs in tech (everything was tech now, but some jobs were more tech-like than others, and the more technical jobs often paid more, especially because people were so paranoid that it was precisely these more-technical jobs that would eventually get replaced by AI—like, try convincing some kid in 2026 that they really ought to invest in getting good at computers, and then you’ll see what the tech-talent pipeline problems are like in 2060).

Anyway, these young people lived in New York City, and they considered themselves cultured. However, they did have a very earnest quality that was distinctly off-putting to the New York arts people. These techie wannabe writers just hadn’t been acculturated into what remained of the artistic establishment. They didn’t speak the right words. All their references came from books and from forum posts, not from discussion with other cultured people. They didn’t quite know what was ‘in’. Like, they didn’t truly understand, not really, that the literary world was just about social power and not quality. They were still naively reading the top Neo-Lernerian journal, The Harvest, and sending in their short story submissions, not knowing that they basically had no chance of breaking in, because they didn’t have an MFA from NYU.

One of these tech-culture people had spent a lot of time in e-novel communities when she was a teenager. Then, as an adult, she had learned that e-novels were inferior to literary fiction, but she wasn’t sure that she totally believed it.

Anyway, at some point this woman, Indira Sen, was reading The Harvest, and it inspired her to post online:

“Is it just me, or do many of the Neo-Lernerian pastiches that I read in The Harvest not seem to be particularly good? That journal just feels like an e-novel community that isn’t working particularly well. Like, obviously there is some demand on the part of the readers, but the authors aren’t really producing a very good imitation. Everything they produce would get downvoted immediately on fourth-way” (this was the name of an e-novel platform).

She developed her thesis at length (“The Harvest as malfunctioning e-novel community”).

This provoked a lot of backlash from the circle around The Harvest. The editors and primary contributors were too high-minded to notice the Indira Sens of the world, but there were some hangers-on who wanted to curry favor with The Harvest, and one of them, Alice Mah, wrote an essay entitled “The e-novelist as malfunctioning human being”).

In its most-quoted passage, Mah wrote:

“Sen’s own essay offers an illustrative take on the dangers of letting AI do your thinking for you. The metaphors of AI have now become baked into human beings, so they launder AI through their very thought processes, and through their conceptions of art itself. Sen sees The Harvest as something like an e-novel community—as a temple to human consumption. She believes that there is some pre-existing need that The Harvest is attempting to fulfill. She doesn’t understand that the aim of the journal is to create new needs and to open up spaces that even Lerner himself had not yet explored.”

Sen felt this was very unfair, and she thought for a long time about her response. Many of her tech-culture friends took a crack at trying to defend Sen, and she felt like their responses were a bit clumsy. She particularly hated one viral post that was in the style of a faux-socialist takedown of The Harvest.

She hadn’t wanted to take down the magazine. She had just wanted to understand why it wasn’t fulfilling its own aim. The purpose of the magazine was what it did. Right now what it did (at least when it came to the magazine’s fiction offerings) was produce poor imitations of Ben Lerner.

Her question was, “Couldn’t these imitations be better?”

So in her epic, fifteen thousand word response, she first had to explain to Alice Mah how an e-novel community worked. That it was grouped around certain influences, and that it often operated off of shared community parameters that were often explicitly set forth in its charter. And then she had to explain what the participants in an e-novel community got from this activity.

The e-novel had developed, after all, as a reaction against traditional fanfiction. Many creators had wondered why it wasn’t possible, in our fanfictions, to imitate the style of the original as well.

One obvious way to closely imitate the style would be too use AI to mimic the patterns of the original, but with new romantic pairings and new story elements.

However, it was hard to get attention on traditional fanfic sites for work that was too obviously written with AI, so fourth-way had been started as a site for people to share their experiments in producing fanfictions that used LLMs to hew more closely to the style of the original. And within this community, there’d arisen many thoughts about what was good and what was bad. In this community, human attention was at an absolute premium. One good commenter could sway entire fandoms. Creators wrote to attract the attention of that one good reader.

The people who were willing to sift through all this content prided themselves on their sagacity. And they created a set of standards, meta-standards, for people attempting to write good imitations. These meta-standards were the baseline—if adopted, they prevented a work from evincing that uncanny corporate AI voice (i.e. “it’s not X, it’s Y—and that’s so valid”).

The whole point of the e-novel was that it provided an organized way of talking about style. It was a community for people who cared about style and thought that the style added something to the original, and that it was meaningless to do a fanfiction if you didn’t also copy the style. They really prided themselves on having a lot more taste than traditional fanfiction fans and writers.

Most e-novelist communities were organized around popular commercial fiction properties, but, when she was a teenager, Indira Sen had wandered into a very unique community that was devoted to the work of a literary novelist who used to be extremely popular, Miranda July.

There was apparently a group of women so inspired by Miranda July’s All Fours that they had created e-novels of themselves going on similar zany adventures. And in this community there had been a lot of debate about whether what they liked about All Fours was the zaniness or the very unfiltered quality. Like…if you were equally unfiltered, but about a less-zany life, would it still be compelling?

This community was very dedicated, and they had produced some great stuff that Sen had found comforting when she was a girl. Like, as a teenager, Sen had developed this fear (somehow) that she’d fall in love and get divorced and then be dried-up and unloved forever. She hadn’t even fallen in love once yet, and she was terrified of divorce. But something about this e-novel community had really resonated with her. And that had been her earliest community on the internet, where she’d first launched her own effort (an All Fours pastiches about a nerdy Indian girl in a suburban American high school, of course) that’d been praised perhaps disproportionately to its merits.

The All Fours community was very unique in that it had the e-novelist vocabulary, but it was narrowly-focused on a work that was so different from almost every other e-novelist community. This meant there wasn’t much cross-leakage from other communities.

In the All Fours community, people were very quick to call out pastiches that were too normie, too influenced by other sources. And as a result, it’d demonstrated what it would look like to have e-novels that drew from a self-consciously literary source (i.e. that drew upon the markers of 20th and 21st-century prestige fiction).

That community was the place where she had started to think, seriously, about whether it was possible to break the perpetual logjam between ‘human’ and ‘AI-assisted’ writing and create work that was recognizably excellent, but which came from this community where open AI use was the norm.

She recounted this history, and then she closed her effort-post by writing:

I am asking you if the false dichotomies between human and AI—norms that have been abandoned in so many of the other industries—might not be hampering this industry as well. That if there is an authentic human need for a certain kind of self-consciously high-brow writing, then perhaps that need can be satisfied using the same technologies that satisfy other human needs,

By the time Indira Sen hit ‘post’, she already understood the next steps. She had to use the ethos of the e-novelist world to produce a magazine that was better than The Harvest.

The problem, she realized, was that e-novelists were much too cautious. They were so used to being guided by ratings and by shares and kudos and comments, that that they couldn’t take the leap and produce something that was startlingly different from anything an e-novelist had done before. Because what would be the point? Who would read that?

It was actually a very difficult task to recruit other people for her magazine. She originally had a few co-editors, but none of them totally understood what she meant by the ‘e-novelist’ ethos. The difficulty surrounded that term, ‘community parameters’. She was the only person who understood that The Harvest had something unwritten that was very akin to an e-community’s parameters. She had read The Harvest so deeply that she had fully imbibed their conception of what it meant to be good, and she agreed that this style of journal, with its sneering takedowns, its earnest, searching essays, its strong political principles, and its avowed commitment to a highbrow (yet oddly accessible and entertaining) art was indeed good! She just felt it was possible to do it better than The Harvest did.

Her journal was called Hang The Bastards! The first section, Hangings, was composed of a series of first-person plural responses to contemporary news events. In each entry, some corporate business dealing was made transparent, and the blurb ended by saying, Hang The Bastards!

Then there were longer essays, which were all about how artificial intelligence was depraved and horrible and its users should be shot and also late-capitalism was somehow to blame. And, finally, a series of six pastiches of Lerner, which she was exquisitely proud of, since they each transported Lerneresque writing to different demographics and different parts of the country. She really expected these to be praised.

When the first issue of Hang The Bastards came out, it attracted zero notice from the mainstream media world or even from the other small journals.

But it did get a response from one guy, Neil Stein, who immediately understood what she was getting at. He wrote a long, exorbitant post praising the genius of this magazine, and then he said:

“Okay, but the essays are no good. Like, I think the problem is that because most e-novelist communities don’t do nonfiction, you don’t really understand the norms for this kind of essay. And, moreover, you can’t just do sneering socialist takedowns—they also need to be honest somehow. Like, this journal is pro-AI, so it shouldn’t pretend it’s not, just because it’s imitating a type of journal that is almost always very anti-AI.”

This really got her thinking, and with her next issue, she was prepared with some red meat. She took aim in this issue at a forthcoming NYRB re-issue of a book by Johanna Kapoor. In Sen’s essay, “The classic as mediocre e-novel”, she eviscerated Johanna’s book, demonstrating exactly how Johanna had tried and failed to emulate Sally Rooney, and, moreover, demonstrated the reasons she had failed. And she went on to identify a much superior novelist, now forgotten, named Raven Leilani, who had succeeded in her time because she met the unspoken community parameters for this kind of novel.

In subsequent issues, Sen really tried to develop these theories. She posited that every classic, throughout history, has created a penumbra. It’s created an audience of people who were hungry for more works like that classic. And that this then creates a secondary community of impassioned creators. If this secondary community is good at meeting the needs of the fans, then the classic itself endures. It becomes a classic, because it is at the center of a community. And then some people who are influenced by this community go on to create for other communities or write in other milieus. That classics have an influence when they’re internalized as influences by young people.

The ideas were somewhat inchoate, but they provoked a lot of thoughtful discussion online. Lots of people took seriously the challenge to produce an ‘e-novelist’ literature. This resulted in a burgeoning slush pile. After all, her magazine (although tiny) was probably the only journal where a truly unknown fiction-writer could get attention from sophisticated readers. Her journal didn’t bar AI-authored work, which meant there were a lot of submissions, but she had the skills, honed from years of participating in e-novelist communities, to sift through these submissions (with AI help). The biggest thing here was just the community parameters. If people didn’t care to make sure their submissions fit the parameters, then they were out. She was only interested in fiction that was within the Lernerian tradition. And very, very few e-novelists really understood the Lernerian tradition well enough to produce a passable imitation: most attempts immediately seemed rote, like they’d been produced with minimal and thoughtless prompting, so they were an easy pass.

It’s true that sometimes people pushed her on the narrowness of her boundaries. Many people said these criteria were sexist, and that it was wild that she didn’t allow Miranda July pastiches. She allowed this, because she felt she had a good sense of when one of these pastiches was good and when it was bad, but she continued to disallow things that were just too far outside her tastes—no Teju Cole or Garth Greenwell, because she just didn’t understand those guys as well.

And she got stuff that she thought was really excellent. She realized now that the Miranda July influence was good. It was such a trick to detail some insane sexual fantasy, but do it in a very off-beat way so it was believable that it might’ve actually happened. The deal with these stories was that they couldn’t seem too nakedly titillating or they’d seem unrealistic, but they also couldn’t be too embarrassing or deflating, because then that was no fun—too many neo-Lernerian writers acted like they had discovered some big truth when they announced that actually sex is boring (in fact, this is something everyone knows—one reason for declining birth rates).

The major trick was that you needed to make the encounter simultaneously hot and so embarrassing that you wouldn’t tell anybody about it unless it was true.

One great story was Hannah Wen’s “The Duel”, about a college student named Hannah who developed a crush on the sixteen-year-old girl that she was tutoring in math, and how they had this very long, drawn-out, mildly-erotic relationship centered around this teenager’s obsession with the French mathematician Evariste Galois, and how this girl, this younger girl, was so obsessed with Galois, so drawn to him erotically (even as she failed to be a very good math student) that Hannah kept trying to get into Galois herself, just to be there with this teenage girl in her obsession. It was cringe-inducing, but also delicious.

She couldn’t reliably factor a quadratic, but her room was full of math books, from which she painstakingly copied out equations in fountain pen into her notebook, explaining them back to me using words learned by rote from a chat window. She burbled at me for hours with shining eyes, and I couldn’t help it, over time, I kept saying, “Wow! That’s great! That’s so fascinating!” Because, although I knew it was an act, the act seemed somehow so sincere, so wholesome, so disconnected from material gain, that it actually, in some weird way, seemed better than really learning math.

This story went viral as an example of ridiculous AI writing. It was roundly mocked and picked-apart and discussed, and then Hannah really leaned into the bit, by, essentially, pretending it was true. She had such a schtick, this Hannah Wen, where she claimed everything she wrote was just a story, but it all just seemed so true, because it was of a piece with her own moody, awkward personality—over the course of a year she built up a huge following and became a literary celebrity!

She got so big that a reporter for The Atlantic asked her, point-blank, “Are you writing with AI?”

“I don’t know,” said Wen.

“What do you mean?”

“I type at the computer. Words come out. I feel like the words are true, even though they’re not. People like them. I don’t know…”

The reporter followed her to her house and watched her write for twelve hours. He was transported by how free and uninhibited this little Asian girl seemed at the keyboard. It bore so little resemblance to what writing looked like for any of the colleagues of this man, The Atlantic reporter, and yet he couldn’t help thinking that her freedom and her easiness, her swagger, maybe bore some relationship to the great writers of the past.

Then of course there was a huge backlash, where a prominent former e-novelist (turned mainstream novelist) wrote:

“You’re just describing an e-novelist! This is just what it looks like! Oh my god, I am losing my mind. I cannot believe The Atlantic is just columbusing what I and so many other novelists were explicitly told we must never admit to doing if we wanted to keep our careers.”

That was all great publicity for Indira Sen. Her magazine got pretty popular, and she began getting invited to trendy publishing industry events.

Over time, this journal grew somewhat stale. Indira moved on to become an editor at Knopf, and her successors at Hang The Bastards! held too tightly to the neo-Lernerian parameters and didn’t expand outwards to find new influences. They didn’t really grasp that she’d only made the magazine neo-Lernerian because that was a way of getting the attention of the mainstream world, creating something that they would recognize as potentially being good, and that eventually the fashions change, and it’s good for magazines to change too.

But in any case Hang The Bastards! was succeeded by other journals, many of which took HTB as their explicit inspiration. Some of these journals succeeded, and some didn’t. The old journals that were explicitly anti-AI lingered for some years, but it took so much work to create a fully-human journal that it wasn’t really worth the effort, if it wasn’t backstopped by prestige.

As for Johanna, that canny old fraud managed to reinvent herself yet again.

Johanna started a livestream, where she’d take suggestions from the audience and write a story in real time, over the course of several hours. She became a genuine phenomenon, and although her book of these stories (entitled Real Time) sold only a few thousand copies, she accumulated tens of thousands of devoted fans and never failed to sell out her public events.

The situation after Hang The Bastards! was roughly the same as before. It was very hard to convince other people that your writing was good. It was very hard to convince anyone to read anything, particularly if it looked very different from anything that’d come before. If there arose a journal that any group of people, no matter how small, seemed to genuinely be interested in, then that was worthy of comment.

The only difference was…nowadays, some of these journals contained writing that, its authors claimed, might be written with the help of AI.

Great, that wasn’t disqualifying anymore, but…it was still hard to get people interested. You needed to make it similar enough to something people already liked, but you had to put your own spin on it. Then once you had your initial audience, you kept innovating, feeding off the energy from the audience, seeing where you could push them, and where they would push you. Hopefully that energy grew at a level that was commensurate with your efforts and career aims. But, inevitably, that energy would start to dwindle, and then you’d have to decide whether this pursuit was really worthwhile anymore.

Because…audiences always flagged. Everything flagged. Everything died out. And when things died, they left behind only traces that, hopefully, inspired other people in the decades to come.

Before it sputtered out, Hang The Bastards! shocked its readers by breaking its only taboo. A young critic, Georgia Hsu, wrote about getting really inspired by Johanna’s livestream, and about experimenting with a practice that Georgia called autopublication. Essentially, you started with a prompt—some piece of writing you wanted to produce—and then, without any AI assistance, you wrote a response that attempted to fulfill the prompt (as if you’re the LLM-responding to yourself). Then you erased the prompt, and you simply published your own attempt, without ever consulting an LLM.

Many HtB readers felt like this was going too far, because it implied that what they were reading, Georgia’s essay, was essentially unaided, was her original human composition.

But many other people found nothing remarkable in this idea.

In her response, written later that year for a much larger magazine, Indira Sen wrote:

“Speaking purely as a matter of process, I think that that autopublication has much to recommend to it. Especially when it comes to nonfiction, I find that once I’ve gone through the process of figuring out what I want to say, then I worry the risk of an AI distorting my meaning is much higher than the benefits I’d get from its input. Furthermore, autopublication ensures that you truly know your ideas and stand behind them. You will have to answer, after all, for anything that you write. We have seen, too often, authors that are inarticulate onstage and don’t seem able to carry forward their ideas into the real world.

At the same time, I would hesitate to acquire any author who tries to turn ‘autopublication’ into their brand. In my view, human minds have to compete on a level playing field with each other. I have no desire to return to the world I was born into, where some writing is privileged because the author claims that it is was written without aid, using their own mind alone. That layer of privilege had, by the time I founded my magazine, become nothing more than the exercise of social power. Because of their pedigree, some people’s writing was presumptively held to be pure, while other peoples’ was presumptively held to be forever-stained by large language models. And the effort to maintain the illusion of pure human authorship had resulted in the inability to talk seriously about what authorship genuinely entails, and to what degree we can honestly claim that a given author is an auteur, driven by their own vision, rather than a craftsman, who is trying to give a rarefied experience to a small group of people with highly-specialized tastes.

Many of her fans were horrified by this statement on Sen’s behalf. They felt like she had sold out. Like she was claiming that actually it’d all been about the principle, and now she was saying, “Really my own work was all human, all the way through.”

Sen thought differently. She felt it was good for people to understand that raw human output had some value, particularly as part of an author’s broader project and overall development. But in truth this statement was also a trial balloon. It seemed, after everything, that there was still a lot of interest in the idea of ‘human authorship’. She had killed it once, for fun. But now maybe it would be fun to resurrect the idea as well, just to see what happened. After all, that’s why we wrote, wasn’t it? For fun.

---

Source

Acknowledgements

I religiously read John Pistelli’s Weekly Readings, where he gives his take on the week’s discourse. It’s written in a very strange style, which I initially found hard to decipher, but after three years I’ve become adept at understanding Pistellinese.

Anyway, after I wrote this story, I realized it was clearly inspired by this paragraph from a post by Pistelli (about the future of written word in the age of the LLM).

I renew my longstanding prediction of how this will all actually work out, accusations aside: lowbrow work will be almost totally cranked out by machine, middlebrow work will promote the artisanal as a value while being largely machine-tooled behind the scenes, and the highbrow and counterculture will divide themselves between a wholly artisanal and a wholly machinic avant-garde.

My story is basically just a five-thousand-word workup of this one paragraph.