Surf's Up in Slop City

How should authors navigate a world with disappearing books coverage and a rising flood of AI slop books?

In case you missed it, I have a new existential and mythological flash fiction piece up in Strange Pilgrims called “Consider the Eagle.” And my not-yet-a-year-old novel Metallic Realms is still available in a bookstore near you!

There were two depressing bits of book world discourse this past week that I can’t help but link together.

Vanishing Book Coverage

The first was that the Washington Post is gutting its book coverage despite being owned by one of the world’s richest men who could fund the paper for many lifetimes without even denting his billions. The only slight silver lining one might see while squinting is that the Post is also gutting sports coverage, photojournalism, international reporting, and a dozen other things. So, this wasn’t about books coverage performing poorly—indeed WP book section editor Jacob Brogan said the traffic was “actually quite robust”—and seemingly just a matter of Bezos looking to cut tons of jobs.

This is the coldest of comforts. Adam Morgan noted the loss has brought the number of full-time salaried book critics down from seven to five. Five in a country of nearly 350 million. Most authors I’ve spoken to about book coverage have agreed it’s harder than ever to get coverage. A novel that even five years ago would get several prominent newspaper reviews may today get none. Recently, I was googling an interesting-sounding novel (from a well-known big 5 imprint) and could only find one trade review.¹ I mentioned this to an editor I know, who responded, “Oh yeah. It’s a fight for even trade reviews these days.”

Despite (or more likely because of) the decline in book sections, the remaining book reviews have a decreasing effect on sales. Another editor,² Sean deLone, recently published a great article on the institutional decay across culture. DeLone points out that it isn’t only snooty book critics whose influence is waning, even things like celebrity book clubs have declining influence on sales. Overall, he says, “Few institutional players have the power they once did to draw attention to a subject and drive demand for it. To “make” a book, a tv show, or band with coverage like a review or interview. The same disconnect between politics and community is happening for culture as the big organizations covering art are increasingly failing to deliver and, not helping the matter, are rapidly disappearing or defanged.”

DeLone offers up examples of what is taking the place of these institutions in the culture—parasocial relationships, hidden networks, podcasts, and Substack—that all seem correct to me. And there are reasons to believe that some of those have real advantages on the flawed and vanishing institutions they’re replacing.

But. It seems notable that the institutions are withering at the same time that readers arguably need more help navigating a publishing landscape that is being flooded with low-quality books.

The Slop Tsunami Cometh

That leads me to the second news item, Alexandra Alter’s article in the New York Times today about Romance authors who “write” their novels in under an hour by prompting AI. This is not about authors who might use GenAI in artistic or innovative ways. Such uses exist. This is about authors who, I think it is fair to say, have a business plan of flooding the Amazon Kindle store with a firehose of slop:

With the help of A.I., Ms. Hart can publish books at an astonishing rate. Last year, she produced more than 200 romance novels in a range of subgenres, from dark mafia romances to sweet teen stories, and self-published them on Amazon. None were huge blockbusters, but collectively, they sold around 50,000 copies, earning Ms. Hart six figures.

While we spoke over Zoom, an A.I. program she was running ingested her prompts and outline and produced a full novel, about a rancher who falls for a city girl running away from her past. It took about 45 minutes.

Thinking of writing this way is depressing to anyone who cares about books. And I do find it depressing. Although before that I would urge a little caution. First, let’s not grant this activity the title of “authorship.” No one who is pumping out a novel in 45 minutes, or publishing them at a rate of hundreds per year, is likely reading the AI outputs in full much less giving them the care and attention we expect of even a hack author. While this grift is book-shaped, it has far more in common with crypto pump and dumps, NFT rug pulls, and dropship t-shirt scams than anything related to literature.

Also…

The Story Smells a Little Off

I believe the Kindle Unlimited store is already flooded with slop. Much of the internet is too. I’m also certain some grifters are releasing slop books at the rate of hundreds a year. But is Hart actually “[selling] around 50,000 copies, earning Mrs. Hart six figures”? And is this actually a replicable business model that authors should pay attention to? I’m skeptical. Being skeptical of tech hype has been proven right a lot more times than wrong in recent years. Remember how the media spent the last few weeks freaking out about how “AI agents” on Moltbook were inventing languages and religions, and a few days away from becoming Skynet? Prominent AI figures were calling it “the most incredible sci-fi takeoff.” Turns out it was, actually, bullshit. A combination of the usual LLM pastiche and human-authored forgeries tricking gullible pundits.

As for Coral Hart, I don’t have access to her bank statements. Maybe she’s earned six figures just on AI books. But a quick look at the Amazon and Goodreads pages of her books show they have almost no ratings—and the few reviews that exist tend to be 1-star “this is AI slop!” comments posted after the NYT article. Can anyone sell 50,000 copies, even spread across 200 titles, and receive no ratings or reviews? The books seem to be part of the Kindle Unlimited service, so it is possible Hart is counting a “sale” as anyone who downloads the book for free. I’ve also seen it speculated that since Hart is South African the “six figures” could refer to South African rands, which are worth a fraction of US dollars. (100k rands = about $6,200.)

[Editing to add: a source in publishing emailed me to say they’d looked up Hart’s sales with an industry tracker and “Coral Hart authored books total 7, and her sales across all of them total 37 units. Inclusive of KU subscription reads. Fishy indeed!” The source noted that Hart has other pennames, but they couldn’t imagine they could add up to 50k.]

Hart’s books also seem to be sold in the Kindle Unlimited store, which does not involve selling copies per se. KU is essentially a Netflix for ebooks program in which users pay a monthly fee for unlimited ebooks and authors are paid a percentage of revenue on a per-page-read basis. But this is how the slop scam works. You flood KU with tons of titles tailored to different subgenre niches and search terms hoping each snags a few accidental clicks. Even if every single reader stops after a few pages, no one ever completes a single novel, and no one directly buys even one copy, you can still earn money.

So, maybe Hart is somehow working this clickbait KU scam without getting any real readers and thus no reviews. Even if true, that game is an obvious race to the bottom in which competing grifters will spew out ever-more titles in the hopes of stray clicks. Perhaps this is why Hart herself is already pivoting to a new scam: Through her author-coaching business, Plot Prose, she’s taught more than 1,600 people how to produce a novel with artificial intelligence, she said. She’s rolling out her proprietary A.I. writing program, which can generate a book based on an outline in less than an hour, and costs between $80 and $250 a month.

That is where the real money tends to come from. Other people’s dreams.

The Slop Tsunami Could Still Drown Us Though

Even if Hart’s income isn’t fudged, I think the danger of LLMs for literature remains their quantity, not their quality. I have never worried much about LLMs replacing authors in terms of quality. Perhaps I’m naive, but I think humans will always prefer art made by other people imbued with real experiences, ideas, and emotions. (If a tech company ever manages to produce a sentient Data-from-Star-Trek-type AI, then I’ll be interested in the art such an entity makes. That seems a long ways off.)

But even pure slop that no one enjoys can cause a lot of problems when it gushes forth in an unending stream. A few years ago, several prominent science-fiction magazines were inundated with AI slop stories. (Rumor was that these were traced back to TikTok hustle accounts suggesting you could make easy money selling SFF stories. Lol. Sigh.) It doesn’t matter that the stories were bad. The sheer volume of slop mucks up the whole process. Even quick rejects add up when you’re talking about hundreds, thousands, or tens of thousands of submissions.

The Flood of ChatGPT Crap About to Clog Up Publishing

How does the Kindle Unlimited store survive countless grifters flooding the store with slop books under various pen names and accounts? Unless Amazon figures out some sturdy safeguards, the entire KU will be drowned. Even if you don’t care about the KU, which candidly I do not, it is easy to imagine how a dramatic increase of quantity—in magazine slush piles, agent query inboxes, Substack posts, etc.—can quickly overwhelm what remains of our rusting pipelines.

Obviously, the AI companies will be happy to sell the cure to their own disease. AI software to sift through slush for magazines, queries for agents, and submissions for editors. It’s hard to see how this improves things for anyone involved, especially for writers. If you thought the publishing world was too focused on connections, networking, and nepotism before, then just imagine what it will be like if those are the easiest way to know someone querying you is at least a real human.

Remember that people reading submissions and queries are already overwhelmed³ and are forced to skim and skip just to keep up. What happens if this volume is multiplied ten times? Or one hundred times? Or a thousand? Even if you plan to skip editors and agents and self-publish, what happens when the number of books exponentially increases without any notable increase in the number of readers?

This has always been my question when AI evangelists pitch the idea of using GenAI to “produce 100x as much writing without extra effort!” What good does that do me, readers, or the world? Even before ChatGPT was released, I thought this was the central question of GenAI and literature. Here’s what I wrote in May of 2022, six months before the release of ChatGPT:

There are already far more novels published than anyone can read. Far more manuscripts filling agent inboxes than will ever be published. Even a novelist as famously productive as Stephen King only averages a book a year or so. Could an AI program help King produce 100 novels a year? Maybe. But it’s unlikely people would read that many, and even less likely that the publishing ecosystem could handle them. How many book reviews can be published? How much shelf space exists in bookstores? How many books can a publicist work on at a time? There are already far too many books written than we can handle. What the world needs is better books. More original books. More visionary books. But not simply more books.

How Do We Stay Dry the Sea of Slop?

I assume few people reading Counter Craft are interested in pumping out as many book-shaped ChatGPT outputs as possible to spam the KU store. If you want to just make money and don’t care about your writing, then slop world awaits. Get pumping. But if you want to do something else, then the question is how we stay above the waves of slop? And as readers seeking real books, how do we discover the books we’ll love when the ocean of slop stretches as far as the eye can see?

Why You Should Still Build Your Raft of Art in the Sea of Slop

I have to believe that most readers still want to read human writing. And I have to imagine art, real art, still matters. Perhaps not to all people or even to most people in our age of declining literacy and robot slop, but to enough people. One thing we have to do is not give a shit about crap. As I put it last year in a little cri de coeur on this topic:

It has nothing to do with you! Nothing. It is completely irrelevant to what you are trying to do. It doesn’t matter. Don’t waste your time on it.

I don’t merely mean that it is irrelevant in the sense that stressing about artist-hating AI bros or pop culture nonsense won’t help you improve your sentences or revise your novel-in-progress—although that’s certainly true. The more time we spend stressing and scrolling, the less time we have to engage in the art we love or work on our own artistic practice. But I also mean that those people are doing something completely different than you are trying to do. Yes, a hashtag-tailored/written-by-committee bestselling novel and Moby-Dick are technically both “books.” But only in the same way that the ExxonMobil corporate logo and Guernica are both “works of visual art” or that microwaving a frozen burrito and preparing a meal from scratch are “cooking dinner.”

I think this is an important mindset to cultivate for your own artistic practice and mental health. I also think it is literally true. If there really are readers who enjoy intentionally formulaic ChatGPT novels spewed out in 45 minutes, they are not the readers who were going to buy your experimental novel or personal project.

However, few of us merely want to write books. We also want people to read them. So, we cannot ignore the crumbling institutions or the waves of slop. We have to find some way to surf through the ruins.

Building Our Own Infrastructure

The best (though perhaps hardest) thing we could is to build our own infrastructure. This is part of what I wrote about in my last article about having a “punk rock” or “do it ourselves” ethos. If the old institutions are crumbling, that is also the perfect time to build new ones. I’m not saying this is easy. But, what choice do we have? Alternative literary infrastructure does exist. There are great indie presses, unique reading series, independent podcasts, and so on. The more the better though. Much of the mainstream and alternative infrastructures were eroded by the social media age. But, social media is itself dying now and is extremely susceptible to the slop flood. Now is the time to build bigger and better rafts.

In practical terms, authors will probably need to focus even more efforts on their own publicity and try to create things like podcast and newsletter tours for their books. I don’t think this is good. I just think it is. Smart editors and publicists will also tap into these networks. (I also think it would be smart for big presses to put more effort into their own platforms and communities. I expect readers will be looking for imprints to trust in the slop future.)

Here is a silver lining to all this: the parts of culture that are the most at risk of being drowned in slop are the parts that were already slop. It is the most formulaic work, unoriginal artists, and staid institutions that are at risk. Perhaps it is time for the nimble mammals to thrive as the old dinosaurs die.

Embrace the Human in Your Art

Here, I risk being a bit corny. But fuck it. I want to believe the way we counter automation is by embracing the human. We have to make the kind of art that computer programs can’t make. This could take many forms, not all of which I find thrilling. Still, I think those of us who don’t want to pump out hundreds of books will want to emphasize the human in authorship, content, and style.

Human authorship - This outcome doesn’t excite me but I expect that as the world is drowned in robot text that the importance of artist as public figure will only increase. I’m not sure the future will be friendly to anonymous Ferrante or private Pynchon type authors. I think we will see more parasocial relationships via TikTok videos, blogs tracking an author’s publication journey, Instagram posts of handwritten drafts, podcasts, newsletters, etc. Stuff that makes readers believe you are a real person who worked long and hard on your art.

I know that some AI believers are going to say that none of the above will prove human authorship. That in the future LLMs will be undetectable and the TikToks and podcasts faked. I expect that is true. I expect we’ll see some hoaxes and scandals too. I don’t think it changes the fact though. Readers will look for those markers of human authorship even if they aren’t infallible. And not just readers. Agents and editors may be even more concerned with this. Unknowingly publishing an AI book risks both backlash and potential legal headaches (e.g., around copyright law). Last year, I was chatting with a literary magazine editor who had put their magazine on hiatus for a few years. They mentioned how exciting it was to come back to the slush pile that had been closed since before ChatGPT: “I feel like I am sitting on this treasure trove of stories that are all definitely written by an actual human.”

Human content - Here I don’t mean content in terms of the events in the story, but rather the sense of autobiographical connection. This is the least exciting outcome for me. I quite enjoy many autofiction books, but despite how much of our culture treads all fiction as autobiographical. How some insist authors should be reduced to their identity categories and backgrounds and only write characters who mirror their experiences. Or how authors feel they have to reveal their private traumas and hardships to “prove” they are worthy of writing about a topic. There is a lot of bad and boring stuff that comes out of focusing on author biography. Still, I expect the popularity of those connections will only increase. That doesn’t necessarily mean autofiction. It may mean explaining how the orcs in your fantasy novel and gizmos in your SF story are inspired by things in your life. Or how you took this plot idea from your parents’ history and that one from your child’s experiences. Blah. But, I expect it.

Human style - I’ll end on a hopeful note. The most exciting response to slop world might be renewed focus on individualized human voices and style. The more you can create a unique style⁴ to distinguish your work, the better. If the world is full of unintentional robot text, maybe we will seek out writing that is highly and individually intentional. This could happen. It has happened before. AI evangelists love to compare LLMs to cameras and painting. But cameras didn’t destroy painting. The invention of cameras was followed by a flourishing of painting styles—impressionism, expressionism, surrealism, abstract expressionism, etc.—that no photography machine could capture. And when photography became an artform, it did so through photographers adding their own unique elements and artistic intention to the machine reproduction.

These are just some ideas and predictions. I don’t think every author will need to do them all. But no matter what, the slop wave is here and we’re going to need to grab a board or be washed away…

---

My new novel Metallic Realms is out in stores! Reviews have called the book “brilliant” (Esquire), “riveting” (Publishers Weekly), “hilariously clever” (Elle), “a total blast” (Chicago Tribune), “unrelentingly smart and inventive” (Locus), and “just plain wonderful” (Booklist). My previous books are the science fiction noir novel The Body Scout and the genre-bending story collection Upright Beasts. If you enjoy this newsletter, perhaps you’ll enjoy one or more of those books too.

1. “Trade reviews” are pre-publication reviews in industry magazines like Kirkus and Publishers Weekly that ostensibly help bookstore buyers and the like decide what books to stock.

2. Full disclosure: my editor for Metallic Realms and the forthcoming Haunted Hills.

3. Devon Halliday had a great, funny, and (from my experience) very true article on how slush pile readers judge things.

4. Yes, this might even involve using GenAI programs in some way, though not in the way that spews out 200 books a year.