[I’m afraid this is a long one. It may not fit into a single email, but I didn’t want to try to separate it into two parts, given how long it’s taken me to finish. If you want the full thing, you’ll need to open it in a browser.]
I’ve finally arrived at the final episode of this series, a mere 5 months (and 6 installments) after I began. More than anything else, I’m looking forward to being able to close the multiple tabs that I’ve been holding onto for most of that time. During that stretch, though, it’s been gratifying to see some of the blind faith surrounding artificial intelligence begin to shift in the opposite direction. And that’s made this post a little easier to sort through—it’s given me a chance to focus in on one thing in particular, which I’d describe as platform ensloppification, the way that the tech industry has seized upon Ai as a means of increasing the asymmetry that Cory Doctorow identified as enshittification (I’ve begun seeing the term used without attribution, a sure sign that it’s caught on enough for people to use in normal conversation).
But first, I want to zip quickly through one more stage, the move from social media to platform. Film Crit Hulk published a fantastic essay this week about Ai (“AI: The Apocalypse of Intent”), where he distinguishes between “learning tools” and “dependence tools.” (As you might infer, too much of Ai hype positions it as the former to disguise the fact that it’s the latter.) When it comes to social media, I think there’s an analogous distinction to be made, and I’d frame it a bit differently: social media as they emerged in the mid-2000s were above all communication tools/media. And I’d go so far as to include search engines in there. These sites were built to allow folks to connect with each other, to share links, texts, pictures, videos, and the sites were pitched to us as conduits, as invaluable tools for (YUCK) putting the world at our fingertips. Certain corporations captured market share, others didn’t. But it wasn’t enough to win, particularly when winning came with shareholder expectations of “number go up.” That’s when these tools began reshaping themselves as platforms, putting the lie to their idealistic public messaging and devoting themselves full time to the pursuit of profit. They made their services worse, eliminated competitors, captured regulatory statutes and agencies, and turned themselves into fiefdoms designed to extract as much value from their users as possible. These sites that seemed so promising once upon a time devolved into dependence tools.
And that’s where we’ve found ourselves in the past decade or so, watching these platforms slowly suck the life out of our discourse, our cultures, and our lives. But it gets worse.
Slopify
As social media began its invasion of our traditional media ecosystem, one of the open secrets1 of that process was the degree to which these platforms—like Twitter, Instagram, YouTube—were gameable. The rise of the creator/influencer economy was driven in part by this falsehood: you “won” at these platforms by accruing “followers,” and the more followers you had, the more likely you were to attract attention from sponsors. This is box office logic2: you could track the popularity of analog media based on how many people purchased, watched, listened, etc. But offline, there was some amount of minimal effort required in registering approval. (Not that those systems weren’t also gameable, but it cost resources to do the gaming.) Online, you can “follow” someone on social media with a single click, never engage with their content again, and still contribute to their celebrity status. And honestly, “you” belongs in quotes as well, because these platforms are rife with bots3, artificial accounts that are used to game platform dynamics.
Despite occasional claims to the contrary, platforms have little interest in separating the wheat from (Ai) chaff. In addition to generating influencers/celebrities, these platforms have a vested interest in portraying their sites as active cultural exchanges, even if much of that activity is artificial. This motive goes hand in hand with the fear of missing out (FOMO!) that these platforms cultivate4. And I think that for a long time, we’ve just treated this as one of the unwritten rules of social media (like “if you’re not cheating, you’re not trying”), part of the “famous for being famous” flywheel that social media incentivizes5. As I’ve talked about before, this “economy” produces occasional lottery ticket winners, but that “influence” is pennies compared to the billions of dollars that the platforms themselves rake in by encouraging the myth that this path is available to anyone.
Even that’s not enough, though: “number” must go up, and asymmetry (enshittification) marches on. Maybe the best example of this comes from a platform I don’t write about much (because I don’t use it), Spotify. I’m drawing on Liz Pelly’s book Mood Machine: The Rise of Spotify and the Costs of the Perfect Playlist, which I’m currently reading, but Harper’s published a substantial excerpt from the book and I recommend it.
If you thought “Crush” did a poor job of reading the room, then wait until you read Pelly’s account of the rise of “ghost artists” on Spotify. The platform, which began as an (oligopoly-backed) alternative model to the disruption that Napster wreaked upon the music industry, cloaked itself in bullshit about “leveling the playing field” and making it easier for artists to connect with audiences. Pelly documents the numerous pivots that Spotify made during its early existence, before arriving at the mid-2010s, where “the service was actively recasting itself as a neutral platform, a data-driven meritocracy that was rewriting the rules of the music business with its playlists and algorithms.”
And if by “rewriting the rules,” we understand them to mean “destroying anything outside of the major labels,” then that’s accurate, I suppose. Pelly documents how Spotify was driven by their data to shift their focus to playlists, which were increasingly populated not by the work of actual artists but by what the company called Perfect Fit Content (PFC). PFC is cheap, work-for-hire content generated by underemployed studio musicians for “production companies” (owned in large part by Spotify itself as well as its investors) that then “sell” it back to Spotify where it’s algorithmically boosted by the platform’s “Discovery Mode.” If you go onto the platform looking for a random playlist, there’s a good chance that you’re not supporting actual artists but the platform itself. As one musician told Pelly,
The money seemed pretty good at first, since each track took only a few hours. But as a couple of the tracks took off on Spotify, one garnering millions upon millions of streams, he started to see how unfair the deal was in the long term: the tracks were generating far more revenue for Spotify and the ghost label than he would ever see, because he owned no part of the master and none of the publishing rights. “I’m selling my intellectual property for essentially peanuts,” he said.
Little wonder, then, that Spotify’s CEO was so excited about the potential of Ai; the platform is undoubtedly already retooling to save itself the cost of those peanuts with a subscription to Suno. In the meantime, the platform is increasingly crowding out any possibility for a music industry beyond major labels and social media lottery tickets. Ted Gioia’s essay on how our culture has flipped from new to old is instructive here: over the course of a generation, we’ve gone from an 80-20 ratio in favor of new music, movies, shows, books, etc., to a complete flip of that, where our culture is dominated by retreads, reboots, covers, and leftovers. Gioia is perversely optimistic, because he believes that this model is unsustainable. Pelly, by comparison, raises the question of whether there’ll be anything left once we finally object to this extraction.
This treatment of music as nothing but background sounds—as interchangeable tracks of generic, vibe-tagged playlist fodder—is at the heart of how music has been devalued in the streaming era. It is in the financial interest of streaming services to discourage a critical audio culture among users, to continue eroding connections between artists and listeners, so as to more easily slip discounted stock music through the cracks, improving their profit margins in the process.
It’s tempting to think of Spotify as the latest technology applied to a music industry that once included radio, stereos, American Bandstand, the Walkman, CD players, and iPods. But it’s more realistic to understand Spotify as the brand name on the compactor featured in Apple’s commercial.
Friendslop
Even though most of us now understand Facebook’s Feed as a slowly boiling pot of sponsored content and data harvesting, there are still some vestiges of its original purpose, which was to provide a similarly “neutral platform” allowing us to connect with our families and friends. But Facebook has undergone its own 80-20 flip, in favor of targeted ads, Ai slop6, and the like. Millions of people still dance through this minefield of crap to connect with one another, as they must in order for Meta to profit from the advertising money that forms the vast majority of its earnings. Facebook has been “optimizing for engagement” for well over a decade now, making use of the information and strategies that they shared with Cambridge Analytica. You might well ask how much more abusive and enshittified the platform can become.
Well, the good folks at Meta have an answer for you. At this point, the only unoptimized content circulating on Facebook is the contributions that we ourselves provide. In an effort to close that particular loop, the next stage for Facebook appears to be creating ghost friends for us. After all, according to Zuck, the average American has fewer than three friends (?!), and after decades of hard work seeding our country’s loneliness epidemic, Facebook is here to provide a solution to the problem that it’s helped create. As Dave Karpf remarked at the time, “It’s like Mark Zuckerberg heard about Zombie Internet Theory and decided it was a feature, not a bug.”
You might recall Meta’s attempt to start deploying Ai characters in early January of this year—the response was so overwhelmingly negative that they quickly deleted them. Given Zuck’s remarks, though, I think it’s fair to assume that they’ve been redeployed without any fanfare. Most of the time when I visit FB, my Friend Suggestions are mostly people with whom I have significant social graph overlap, folks who have 30, 50, or even 100+ friends in common with me. But on a few occasions this year so far, I’ve gone in and seen “suggestions” to connect with accounts that I have no ties with whatsoever. Are all of these Ai “friends”? Probably not, but everything we’ve learned about Facebook over the years suggests that some of them probably are.
This vision of Facebook has a lot in common with Spotify’s trendline. Instead of a platform where you can share stories and pictures with the people you know outside of the site itself, it’s becoming more and more like a personalized Matrix, a full-scale simulation of social life that can be prodded and shaped to convert us into stacks of batteries that power Meta’s C-suite dreams of world domination. Or as Karpf puts it, this vision “sounds hollow, pathetic, sad. A world where we have four times as many AI friends as real friends is a world where we have just, collectively, given up.”
Enclosed
One of the most prominent use cases (besides cheating) for generative Ai right now appears to be chatbot companions7, a growth area that is shockingly unregulated when we consider how many of these companies are advertising themselves as solutions to loneliness, mental illness, etc. It’s hard for me not to see a connection here, one that’s rooted back in that idea of “dependence tools.” Ai characters are the micro, while platforms like Facebook, Instagram, Spotify, and the like are the macro, cities populated almost entirely by said “characters,” all optimized to maximize our engagement, keep us penned in, and cut us off from other humans or any roots we might have in reality.
Last fall, LM Sacasas wrote an essay where he answered the question, “What’s the most urgent task before us?” And his answer, appropriately enough, is where I want to end both this individual installment as well as this series of posts:
Resist the enclosure of the human psyche.
I’ve been sitting with that essay for months now. While it certainly made sense to me at the time, it was after the turn of the year when we began hearing more about FB’s phantom friends, Spotify’s razing of the music industry, et al. For me, it seemed increasingly prescient. Sacasas draws an analogy with the enclosure of the commons, the process by which natural resources (land, forests, rivers, et al.) were turned into extractable commodities by the few who were said to own them. Citing the work of Eula Biss, he explains that “The nature of ownership changed within the newly set hedges of an enclosed field, where the landowner now had the exclusive right to dictate how the land was used, and no one else belonged there.”
It’s no great leap to see “enclosure” as one of the dominant themes of our time. A while ago, I wrote about the conversion of our material goods into paid services:
I haven’t used a CD player to listen to music in decades, or a DVD player to watch movies in nearly as long. I don’t “own” those things, even though I’ve paid for them. They are instead hostage to the distribution channels through which I access them, an access that I pay for monthly, but ultimately have no control over. If I stop paying the fees, I’ll have nothing but memories to show for it. The same is true of much of the software that I use, and increasingly, the vehicles we drive, the appliances we live with, and the institutions we rely upon are all dependent on the uninterrupted “goodwill” of a small handful of corporate overlords. We’ve even all but lost the right to repair the things we do buy.
It costs me more (and sounds kind of dumb), but I still order print copies of the books I really value, even though I read them on my Kindle. At any moment, it’s entirely possible that the tens of thousands of dollars I’ve invested in that collection could vanish, depending on the whims of someone at Amazon. Amazon effectively owns those files, and my access to them depends upon my begrudging willingness to keep paying them for it.
Think about what that means for people who are already relying upon generative Ai for friendship, for therapy, for life advice. All of these companies, whether their products are currently free or not, are fostering dependence (if not addiction), and at some point or another, the rent will come due. It’s distributed unevenly, but some people are literally paying monthly fees for a simulation of social life. This is, quite literally, the enclosure of the human psyche playing out before us. And we’re eagerly participating in that process when we should be raising questions. As Nitasha Tiku remarks in the podcast interview I linked above, these are the same dark patterns and consequences that we experienced with social media.
The unceasing stream of notifications and pings, the persistent way even the built environment beyond the screen hails us—all of this is just the necessary operation of the engines of value extraction efficiently at work on the raw material that is the human psyche. When the enclosure of the psyche is complete, we lose the right to wander and roam and loaf about in thought, just as the enclosure of the commons restricted freedom of movement and disdained economically unproductive but life-affirming forms of leisure.
Sacasas writes that “One reading of AI is to see it precisely as a further development of the enclosure of the psyche,” but I think I’d go further than that. I think that we’re already seeing evidence that Ai isn’t simply the latest stage so much as it a phenomenon that will overwrite, fortify, and intensify the enclosures that we’re already struggling with. It’s not the temporary exploitation of an asymmetry that will even out eventually, but rather a suite of tools by which they can be made permanent. And that actually scares the crap out of me more than Ai-doomer fantasies about Skynet. My concern is that we’re “debating” the Torment Nexus from our position within the Torment Nexus.
And I think that’s all I have to say for the moment. I may write up a quick postscript that includes a couple of things I didn’t have room for, but we’ll see. I’m just happy to be able to turn my thoughts elsewhere now. And in the meantime, if you’re interested in reading the other installments in this series, I’ll be adding this linkline to each of them:
I don’t think anyone truly believes anymore that these platforms are “natural” ecosystems of public opinion (“digital town squares!”). At the same time, the scope of the manipulation remains hidden to most of us. Exhibit A is the astroturfed groundswell of “support” for Justin Baldoni, a man unknown to just about everyone a mere two years ago.
“Box office” is how we understand movies and theater, but it’s the same principle by which we track best-selling books, albums, tv shows, etc.
It’s a few years old now, but this account of Dan Woods’ work is instructive: he estimated that as many as 80% of accounts on Twitter were bots. As an experiment, he created a new account and spent just $1000 dollars on bot services—in a short time, he was able to gather 100,000 “followers.” See also Framed: A Villain's Perspective on Social Media, by Tim O’Hearn, who was part of this shadow industry.
I’ve yet to see an article about how FOMO has mutated lately into IYKYK (“if you know, you know”), but I’ve been seeing the latter a lot more, and it feels like a smirky version of the former.
Caroline Calloway provides a pretty strong soup-to-nuts example of this, for instance.
If you’d like a peek into how bad it’s gotten, Futurism caught wind of an SEO/Ai Slop purveyor bragging about how effectively he targets women over 50 on Facebook and Pinterest (the latter is especially useless these days, thanks to Ai).
Sherry Turkle recently described Ai companions as “the greatest assault on empathy” she’s ever seen, which is interesting to pair with Elon Musk’s shit take about how “the fundamental weakness of Western civilization is empathy.”
One can generally find an ok DVD player at the charity shops for twenty or thirty bucks... right next to a dump bin of old DVDs. Which to my experience, is a better interaction than the 'dump bin' of 'streaming' that'd pour out of the screen if I 'engaged' with it.
Really good!