What Was the Internet ?
►https://www.bostonreview.net/articles/what-was-the-internet
L’état des choses dans l’internet vu par quelques personnes assez intelligentes et sincères.
6.8.2026 by Siva Vaidhyanathan, Avery Dame-Griff, Joanna Walsh, Chad Wellmon, Cory Doctorow, Matt Lord - As AI overtakes the web, five writers reflect on how far it’s fallen and what comes next.
Joanna Walsh’s latest book is Amateurs! How We Built Internet Culture and Why it Matters.
For as long as it’s been around, we haven’t been sure how to feel about it. The utopians had their favored metaphors: cyberspace; a global village; the electronic frontier, digital commons, or new public sphere. If American literature, as Emerson said, is “in the optative mood,” the same has been true—maybe truer—of so much thinking about the internet.
But it didn’t take long for disillusionment to set in. A few years after British computer scientist Tim Berners-Lee invented the World Wide Web at CERN in 1989, critics began to sound the alarm on the “Californian Ideology”: the insidious creep of corporate greed in the guise of bohemianism. Since then, the story of the web’s decline has paralleled that of Silicon Valley and the “new economy” of neoliberal globalization it stood for. Social media, once a vehicle of “participatory culture” and “networked protest,” has been enshittified. Big Tech has gone red (in the GOP, not communist, sense). Attention merchants lock us into doomscrolling while platform capitalism, the backbone of today’s national security state, extracts, surveils, and kills. Despite valiant efforts by activists and reformers, it’s been the dystopians’ world—or at least the realists’—for some time now: an age of techno-feudalism, digital colonialism, the new Jim Code.
With the rise of AI, however, some new mood seems discernible, at least around the edges: the elegiac. A lot of the web is being swept away. In May, Google overhauled its iconic search box for the first time since 2001, signaling its all-in on the “intelligence age.” As chatbots hallucinate and hide their sources, search results become harder to browse, and digital life becomes ever more atomized, online access to information looks more imperiled than ever. Wikipedia has managed to survive—a miraculous exception that proves the rule—but it’s staring down new threats. Boston Review has been online since 1995, but our search engine traffic is falling, like everyone else’s. If work is published on the net and no one can navigate to it, does it leave a trace?
In other words, is this the end of the web itself? And if so, is some kind of eulogy in order?
“Just as I’d hoped, [the web] has enabled a flourishing of human creativity and self-expression,” Berners-Lee writes in his recent memoir, This Is for Everyone. But “to ensure the web agents serve people—not corporate profits, not governments, not themselves, but people—it’s critical that we develop systems today that put the human first.” We asked five humans who’ve been paying attention what to make of this moment. Read on for their reflections: about how the web used to work and what it used to feel like, the strange and sorry thing it’s become, and the good that’s worth remembering when so much of what we love is dying, or already long gone.
—Matt Lord
“Infrastructural Imperialism”
by Siva Vaidhyanathan
“The Scale’s the Thing”
by Avery Dame-Griff
“Net Positives”
by Joanna Walsh
“That Obscure Object, Desire”
by Chad Wellmon
“Managed Decline”
by Cory Doctorow
Inside a Google data center in Dalles, Oregon, October 2012. Image: Connie Zhou/Google/ZUMAPRESS.com
Infrastructural Imperialism
Siva Vaidhyanathan
In October 2009, Google cofounder Sergey Brin took to the op-ed page of the New York Times to defend a project that had, by then, attracted lawsuits from the Authors Guild and the Association of American Publishers and more than a little suspicion from librarians and scholars of the web. The subject was Google’s bold, global plan to “digitize all books.”
No one had asked the company to launch Google Books; it seemed not to be doing it for any immediate boost in revenue. Instead, Brin claimed, they were looking out for the good of humanity. “The vast majority of books ever written are not accessible to anyone except the most tenacious researchers at premier academic libraries,” he wrote. He referred to a flood that had damaged the Stanford library stacks in 1998—the very libraries he and cofounder Larry Page had once accessed as graduate students—as proof of print’s fragility and the moral urgency of digitization.
All those centuries of books, in hundreds of languages: it was never about the good of humanity. It was about laying the groundwork for generative AI.
It was a lovely piece of writing in its way, and I don’t doubt that Brin believed every word of it. This was digital philanthropy offered up in the now-classic register of Silicon Valley benevolence: a hot, young company, incensed by the slowness and oldness of something, stepping in to do what creaky institutions couldn’t or wouldn’t do themselves, and asking for gratitude rather than payment. But Brin didn’t reveal what all that scanned, machine-readable, multilingual text was really for. Now we know: his company was compiling the largest collection of digestible text in human history in order to build and train a large language model. All those centuries of books, in hundreds of languages, drawn from dusty library stacks: it was never about the good of humanity. It was about laying the groundwork for generative AI—and that was the plan all along.
That technology is now ruining the World Wide Web, the collection of documents, images, and video that we so unwisely entrusted Google to manage for us. Whether one calls this a bait-and-switch or simply the foreseeable convergence of a data-maximalist research culture with an unprecedented act of data acquisition, the effect is the same: the project that Brin sold to the public as an act of preservation turned out to be infrastructure for a radical act of enclosure—creating the very system now replacing the search engine and the open web itself.
In those years I wrote a book, The Googlization of Everything, warning that Google had become something the world had never seen before: a private company entrusted, almost by accident and almost entirely without public deliberation, with the organization of the totality of human knowledge. I called this condition “Googlization,” and I broke it into three overlapping projects. First, Google was Googlizing us, shaping our habits of attention and self-disclosure through the surveillance of search. Second, it was Googlizing the world: mapping streets, scanning books, and laying fiber with the swagger of a nineteenth-century railroad baron. And third, it was Googlizing knowledge itself—inserting a ranking algorithm between every citizen and the record of human thought, and doing so with a design so elegant and an interface so spare that few people thought twice.
Revisiting that argument now, I realize something has changed that I didn’t fully anticipate: Google has stopped pointing. For its entire history, whatever its crimes and sins, Google Search performed one basic benevolent function: it sent our attention from Google to some source that most likely helped resolve a question or indulge an interest. You typed a query, and the company’s ranking algorithm—proprietary, opaque, contestable, but ultimately effective and useful—served up a list of doors onto the wider web. You clicked, and you left Google’s house to visit someone else’s. That mere act of departure, unglamorous as it sounds, was the thread connecting search to the older ideal of the public sphere. It was the residue of an Enlightenment premise: that citizens benefit from encountering a plurality of sources, weighing them, and forming judgments amid the friction of disagreement. Or, at least, Google wanted us to believe that. Many of us did.
Google has now severed that thread. Type a query into Google today and more often than not you are met with a pile of slop: a synthesized paragraph, stitched together by an LLM, that purports to answer your question so directly and bluntly that you have little reason to go anywhere else. Google calls these paragraphs “AI Overviews.” A more expansive version, “AI Mode,” dispenses with the pretense of links altogether and simply converses with you. By 2025, AI Mode had reportedly surpassed a billion monthly users, with Google boasting that query volume was doubling every few months. The infrastructural imperialism I described in 2011—the steady, unaccountable extension of one company’s dominion over the mechanisms by which we find things out—has entered its final and most totalizing phase. Having organized the web, Google now proposes to replace it.
It’s worth being precise about what has changed. The shift from ranking to rendering is not a matter of degree; it’s a matter of kind. A ranked list of links, however manipulated by SEO and however biased by commercial incentive, still preserved the plurality of the web as a formal feature of the experience. You at least saw ten different domains, ten different voices, ten different institutional pedigrees; the very act of scanning them—this one a newspaper, that one a blog, this one a government site, that one a forum thread—was a small, constant education in source criticism. The ranking was Google’s judgment, yes, but the sources remained legible as sources. An AI Overview suppresses all that. It takes the labor of dozens of writers, editors, researchers, and institutions, blends it into a single fluent voice with no discernible author, and presents it as if it had emerged from nowhere in particular. Links are demoted to a small cluster of citation chips, easy to ignore, appended as a footnote or afterthought.
It used to be that when you clicked, you left Google’s house to visit someone else’s. That mere act of departure was the residue of an Enlightenment premise.
In other words, AI Overview offers the seductive promise that the messy, effortful, contestable business of finding things out—comparing sources, noticing bias, tolerating ambiguity, occasionally being wrong and having to revise your view—can be replaced by a single, confident, algorithmically fluent paragraph. It substitutes the appearance of consensus for the reality of disagreement, laundering the gamut of human knowledge into a single, bland, authoritative-sounding voice with a penchant for hallucination. We have traded a system that was biased and commercially distorted for one that is biased, commercially distorted, and prone to simply making things up, all while sounding more confident than any of the sources it cannibalized to produce its answer.
Recent field research illustrates what’s at stake. A Pew study found that when an AI Overview appears, only about 8 percent of users go on to click a traditional search result, compared with 15 percent when no overview is present, and a mere 1 percent of users click through on the links embedded in the overview itself. An Ahrefs analysis published early this year found that AI Overviews correlate with a 58 percent reduction in click-through rates on top-ranking pages—nearly double the decline measured less than a year earlier, suggesting the effect is not a one-time adjustment but an accelerating trend. Seer Interactive, an industry analytics firm, clocked a 61 percent year-over-year decline in organic click-through for queries that trigger an overview. And in Google’s own AI Mode, the zero-click rate—the share of searches that end without the user visiting any external site at all—has been measured as high as 93 percent. Across the whole of Google Search, industry trackers now place the zero-click share at roughly 60 to 68 percent of all queries.
These are not marginal effects, and they reflect a bid for control that goes beyond Google Books. Unlike a library’s back catalog, the web depends on continuous revenue and continuous attention to survive. A book, once digitized, sits inertly on a server; a newspaper, a recipe site, a small nonprofit’s explainer page, or a local news outlet needs the next visitor, and the one after that, in order to keep publishing at all. Google’s summarization apparatus consumes the work of the open web in order to render the open web unnecessary to visit.
Of course, Google’s overviews and AI Mode are trained on, and generated in real time from, the work of publishers, journalists, encyclopedists, bloggers, forum contributors, recipe writers, and millions of other ordinary web authors—the same open web that Google’s original ranking algorithm was built to organize. Classic search occasionally (if unfairly and unevenly) rewarded source publishers with traffic: that reciprocal arrangement was the thing that made the commercial web possible for two decades. The generative layer breaks that circuit. Google now ingests the open web’s output to manufacture an answer that keeps the user inside Google’s own interface, denying the originating site the visit—and the ad impression, and the subscription prompt, and the newsletter signup—that used to be the whole point of writing something worth ranking.
Publishers have not taken this quietly; among other things, they’ve launched a bevy of lawsuits. In apparent recognition of the problem, Google has rolled out cosmetic fixes—a “Further Exploration” module tacked onto the bottom of an overview, offering a few curated links after the user’s question has already been answered above the fold. What should worry us here is not merely that individual publishers are losing revenue, painful as that is. It’s that the ecosystem of incentives which made the open web worth writing for is quietly being dismantled by the same company that once depended on that ecosystem’s vitality to make its own product useful.
From this vantage, Google’s summarization engine is parasitic, in the biological sense of the term: it requires a living host—the open web’s continuous production of new, verifiable, human-authored information—even as its design slowly starves that host of the resources needed to keep producing. A minority of publishers may survive by walling off their content entirely, blocking Google’s crawlers and rebuilding direct relationships with readers through subscriptions and newsletters; surveys already suggest roughly a third intend to try. But a web fragmented into paywalled enclaves is not a public sphere either. It is simply Googlization’s endgame arriving by a different road: a public information commons dissolved, whether by extraction or by retreat.
Personalization only compounds the trouble. I spent a good portion of The Googlization of Everything worrying about the way Google’s ranking algorithm quietly tailored results to a profile built from your search history, your location, your device, your prior clicks, so that two people typing the identical query might receive subtly different worlds. Eli Pariser would later popularize this concern as the “filter bubble,” and the phrase stuck because it named something real: a public sphere that only pretended to be shared, since each of us stood in a slightly different room of it, seeing a slightly different arrangement of the furniture.
Generative synthesis intensifies that condition in a way that personalized ranking never could. A ranked list, whatever its distortions and limitations, had discrete, checkable, external referents; you could, if you were curious, click past the first result and see what the second and third had to say. AI Overview offers no equivalent. Two users asking the same question may now receive two totally different yet confidently worded paragraphs, neither bearing any obvious sign of its own contingency, and neither offering the reader an easy way to notice that a different answer exists at all. There’s no way to study this phenomenon systematically, so no way to regulate it.
One of the arguments I made in 2011, and the one I think holds up best, is that Google’s dominance was not merely a story of technical prowess or market cunning: it was a story of public failure. In the 1990s and 2000s, governments, universities, and libraries—the institutions that historically bore responsibility for organizing and preserving public knowledge—declined to, or were prevented from, building the digital infrastructure plainly required by an age when democracy seemed ascendant and universally desirable.
We have traded a system that was biased and commercially distorted for one that is biased, commercially distorted, and prone to simply making things up.
Into that vacuum stepped a private company with venture capital, engineering talent, and an advertising business model elegant enough to make the whole apparatus appear free. We got a useful tool and soon became too dependent on it. It was a tool answerable to shareholders rather than citizens, optimized for engagement and margin rather than for the health of the public sphere. Since then the public failure has only compounded, and with the tacit blessing of law.
In August 2024, a federal court found, at long last, that Google had maintained an illegal monopoly in general search—a finding that vindicated much of what critics of Googlization had been arguing for a generation. But when Judge Amit Mehta issued his remedies in two rulings last year, he declined to order the most consequential structural fix available to him. He let Google keep Chrome, the browser that funnels a majority of the world’s search queries into Google’s own box, and he rejected a “choice screen” that might have interrupted the defaults nudging billions of people toward Google without a moment’s reflection. Instead he imposed modest adjustments such as a bar on exclusive distribution contracts and a requirement that Google share certain search-index and interaction data with competitors. These non-penalties arrived at the exact moment when the more urgent transformation in search—the move from links to synthesis—was already well underway, and the remedies say and do nothing about it. By the time the courts catch up again, the open web’s economic base may no longer be around to protect.
When I chose the subtitle And Why We Should Worry back in 2011, I meant something more specific than free-floating anxiety about technology. I meant that citizens of a democracy have a stake in how knowledge gets organized, verified, and circulated—and that this stake is too important to be left entirely to the incentives of a single advertising company, however well-intentioned its engineers.
This argument applies with even greater force to the present, because the stakes have moved from access to synthesis. Back then the worry was that Google decided what links you would see; today the worry is that Google decides what you read, in a Google-generated paragraph with no visible author and no stable citation trail—and does so at a scale, and with a fluency, that makes the underlying labor of the open web nearly invisible, even as that labor is being consumed to produce the answer. To be sure, the open web was never a commons—we shouldn’t get nostalgic for it. It was a mess of vitriol and scams peppered with some remarkable human expression; Google just hid the mess deep on the second page of links. Still, as nasty as it all was, it was better than what comes next.
Image courtesy of Avery Dame-Griff
The Scale’s the Thing
Avery Dame-Griff
When I grew up in the mid-1990s there were always two computers in our house: one for my father, a systems administrator, and one for me, always one of his hand-me-downs. One such specimen is captured in the photo above, taken in December 1995. The machine wasn’t exactly top-of-the-line, but for a seven-year-old it was a pretty nice setup. Thanks to my father’s job, we also had a home internet connection much earlier than many of my peers did. I remember us sitting side-by-side on weekend nights, him working through an issue with a work server while I was talking with strangers online in random chatrooms.
If you used enough early text chat, you’ll remember one of the most common questions from that era: “age/sex/location?” or, as it quickly became abbreviated, “a/s/l?” The query originated earnestly enough—a way for others to learn more about you and find points of connection—but at some point I had a realization: I could lie. No doubt, many did. I never lied outright, because I’ve always been a bit of a rule follower, but I did spend much of my youth online avoiding any reference to or mention of my gender. People just assumed I was male, and I did nothing to disabuse them of the notion. Looking back, I recognize this habit as part of my larger trans experience. Once I became friends with trans folks online, I began to understand the discomfort I felt with having to reveal my “real” gender: the online was, in fact, my “real” gender.
Our current internet landscape has a totally different scale—we are connected to more people than ever before, and yet we are profoundly isolated from them.
For years I’ve been collecting memories of the early internet. I’ve heard about lifelong friendships begun on mailing lists and forums, emails that kept couples connected before cheap long-distance calls, the passive aggressive navigation of teen friendships via AIM away messages, the forging of bonds over voice chat while gaming. The breadth of formats highlights just how much digital communication has evolved over the past forty years. We don’t hear much anymore about the Bulletin Board Systems (BBSes) that closed en masse in the mid-1990s when Americans embraced the World Wide Web. The once-omnipresent America Online—better known as AOL—survived the shift to WWW only by abandoning its walled garden model and becoming just another internet service provider. On the Web, a variety of digital spaces had their moment—message boards, Yahoo! Groups, LiveJournal—until first Facebook and then other social media companies slowly absorbed all communication online. Usenet, first launched in 1980, hung on longer than most, but by the 2010s the number of active communities had shrunk drastically.
I think it’s no coincidence that all these memories, my own included, are of organic interactions online, often at a small scale. Those of us who experienced the earlier days of the internet seem to be fondest of the small moments when we made a surprise connection or learned something about ourselves—usually in the small, often volunteer-managed digital communities that fostered strong interpersonal connections, both online and off. Folks who first met on local BBSes, mailing lists, or Usenet often connected at regular in-person get-togethers. The possibility that you might have to face someone at next month’s meet-up created a very different social dynamic: you might be less likely to flame them, among other things. Our current internet landscape has a totally different scale—we are connected to more people than ever before, and yet we are profoundly isolated from them.
The story of this transformation has been told many times, and shifting incentives are a big part of it. As Megan Sapnar Ankerson discusses in Dot-Com Design (2018), commerce drove popular adoption of the Web once most users could access visual browsers like Netscape. After the dot-com crash of the early 2000s, e-commerce became even more of a focus as companies scrambled to find new ways to recoup their losses. Looking at popular media coverage from the era, one sees a slow but unmistakable shift in depictions of the internet user from communicator to consumer. The Web was “the greatest social revolution since the automobile,” Popular Mechanics declared in 1994. And just as cars drove the growth of suburban shopping malls, the internet would stoke consumption—and advertising for consumption—in ever-greater amounts.
Key to all this was reducing friction: those moments online that require users to slow down, consider why they’re doing something, or take real effort to complete a task. Everything is faster, easier, and less mediated now, and we’re still grappling with the change. As technology scholar Jason Farman recently argued in Slate, “frictionless design has helped usher in technologies that manipulate behavior in ways that aren’t always visible to us” and “mask systems that we neither understand nor control.” Journalist Ryan Broderick has dubbed this moment the Clear Channel Internet, “where everything has been so thoroughly corporatized that nothing ends up in our feeds by accident anymore.” While Broderick’s focus is viral content, I would extend his framework to the modern internet platform per se. Tune in to any iHeartMedia station; no matter the genre, the listening experience is frictionless. It may not be exactly to your taste, but it’s consistent and predictable. It isn’t exactly Muzak, but it’s no Vintage Obscura Radio. Most important, it’s all advertiser friendly: last year iHeartMedia controlled 40 percent of radio ad revenue in the United States.
After two decades of innovation in digital attention-mongering, all the major social platforms today function on the same principle as mass media: keep the user engaged as long and as frequently as possible in order to sell ads. Along the way, frictionlessness has been extended from design to user-generated “content” itself. Platforms today want content that keeps users engaged (including through outrage) but isn’t so enraging or uncomfortable—doesn’t create too much friction—that it drives them away or alienates advertisers.
Of course, what counts as the wrong kind of affective friction, and for whom, has fluctuated over time—a function of changing social norms, commercial incentives, and political realities—but digital queer existence has always been precarious. LGBTQ communities were only tentatively embraced in the mid-1990s once platforms realized they were a possible source of revenue. When our appeal as a targeted ad market dries up, we become a kind of industrial byproduct that platforms can repurpose to fuel more user “engagement.” Last year Meta’s Chief Marketing Officer and VP of Analytics, Alex Schultz, went so far as to present this repurposing as an opportunity to shore up LGBTQ rights. Defending the company’s new, vastly pared-down moderation and hateful conduct policy—which among other things now permits users to post “Trans people aren’t real” and “They’re mentally ill”—Schultz wrote in an internal post, “I personally feel that actually the shock of friends and family seeing me receive abuse as a gay man helped them understand the hatred that exists and hardened support.” In this calculus, frictionlessness becomes one group’s prerogative over another: bigots enjoy the comfort of a seamless platform experience, while their targets are asked to experience a little friction and suffering in service to their cause.
Indeed, the only time social media platforms introduce friction these days is when they want to discourage users from producing content, and they increasingly rely on automated systems for content moderation for two reasons. First, the sheer scale of content produced is massive, so automation reduces the burden on content moderators. Second, decisions made by automated systems don’t introduce the same level of friction. Humans want to make the right decision, so sometimes we take our time to consider the situation. (This experience is perfectly captured by the game Moderator Mayhem.) Of course, users are typically given the option to appeal a platform’s moderation decision, but the process is labor-intensive and opaque, and you never get to talk to a human. Intentionally or not, the process thus deters users from making an appeal—it’s easier just to accept the decision and try making your posts more platform-friendly, if you can see how. But some users encounter so much friction it’s just easier to avoid talking about unapproved topics altogether. Last year one reproductive health group was reportedly advised by a Meta employee “to simply move away from the platform entirely and start a mailing list,” since “bans were likely to continue.”
Like contemporary platforms, the utopian comfort of Le Guin’s Omelas is only possible because someone else suffers.
Meta denied sending this message. But as I consider what our online future could be, I think of this warning and am reminded of the ending of Ursula K. Le Guin’s classic short story, “The Ones Who Walk Away from Omelas.” Like contemporary platforms, the utopian comfort of Omelas is only possible because someone else suffers—in this case a single child, living forever in darkness and misery. Everyone learns this at a certain point and must make a choice, to stay or leave forever for the unknown: “The place they go towards is a place even less imaginable to most of us than the city of happiness. I cannot describe it at all. It is possible that it does not exist. But they seem to know where they are going.”
But it’s not enough simply to walk away from today’s version of social media. In Afrofuturist N. K. Jemisin’s story “The Ones Who Stay and Fight,” there is no exploited child in the utopian city of Um-Helat. Instead, there are social workers whose sole job is to keep anti-utopian ideas like inequality and exploitation from entering society. While suffering is not fully eliminated in Um-Helat, its citizens prioritize the good of the community. Jemisin says she conceived of the story as a response to those who interpret “Omelas” as advocating totally abandoning current capitalist society—something impossible for those without generational wealth and with “nowhere to walk away to.” Um-Helat remains a utopia “as long as they keep at bay the idea that somebody’s got to suffer.”
I don’t have one all-encompassing prescription for my vision of a future better internet. What I know is that our current digital landscape wasn’t inevitable. A healthier social ecosystem online is possible. When our infrastructure is oriented toward the small scale rather than toward growth at all costs, we get those internet memories people have shared with me over the years—talking one-on-one, getting to really know and care about each other. There will always be friction when we interact and build community; the goal can’t be to eliminate it altogether. Rather, we should recognize how it can productively strengthen our investment in each other. Walking away is just the first step on the path to a different future we must collectively make together.
Detail from cachemash #185. Image: Hugh Manon via Flickr
Net Positives
Joanna Walsh
It was 2014, and I hadn’t yet published my first book. I opened my email app and wrote to the editor of 3ammagazine.com, an online literary journal:
You have a great mag but I can see you don’t have a dedicated fiction editor and, though you regularly publish fiction, the mag doesn’t have a consistent approach. How about you appoint me as fiction editor and I can fix all that?
Or something like that. I could search for the original email, but something tells me not to try. This kind of hubris only made sense on a web that no longer exists. Maybe I’ve changed; or maybe the internet has, or maybe it’s the position of independent online literary magazines.
I’d never met 3:AM’s founder, British writer Andrew Gallix, IRL, so I was surprised when the head of a journal whose tagline is “whatever it is, we’re against it” wrote back and said yes. The few bylines I had were mostly for unpaid work in free online mags. I was new, I had hope, and I worked hard. I wrote calls for submissions, read slush piles, selected work, wrote rejection and acceptance emails, edited and uploaded pieces, designed pages using HTML, sourced artwork (forming relationships with artists), and publicized the work via the mag’s social media accounts. I published two stories a month, which may not sound like much, but it took up a lot of my time and I worked for free.
On Twitter, I met friends, lovers, allies, enemies, funders, and employers. It was this atmosphere that made me believe, for the first time, that my writing might be published, might be read.
Never work for free, I hear from many fellow writers now. And they’re right. Sort of. I’d say, never work for free for anyone who’s being paid to produce any product they’re going to charge for. Never work for free to produce something that can be produced commercially. What I did at 3:AM was outside that. Because we couldn’t pay I made a commitment not to run anything I could imagine being published in a commercial venue. (Not so) sadly I was wrong: I published experimental writers, some of whom not only went on to be published in the mainstream, but who changed what the established world of literary fiction could be. I published early work by Eley Williams (who went on to be my 3:AM successor) and Isabel Waidner. I’m proud of what I did there.
Why did I do it? Like most artists, for mixed motives. I wanted to create a space of a particular kind, a space I didn’t see elsewhere. Like the writers I published, I also wanted to expand my chances of being published in the world of paid writing. What 3:AM offered, what I offered—to them, to myself—was a chance to dodge some of the gatekeepers of the mainstream literary world.
I “became” a writer on early Twitter, which I joined in 2008. I had been working as an illustrator and ran a drawing blog, which eventually became more writing than drawing and, though it managed to win a few awards, I had no education in writing, no contacts, and, with full-time care responsibilities and limited financial resources, little chance to attend readings, form connections, or get anywhere near any literary scene. But back then, online was a scene. Twitter was a largely uncommercial space, but it already had enough subscribers for me to connect with publishers, published writers, and writers who were just starting, who swapped reading and publishing recommendations, advice, and support—writers like me. It was crucial that I met the whole range because any emerging writer needs both peer support and opportunities. On Twitter, I met friends, lovers, allies, enemies, funders, and employers. It was this atmosphere that made me believe, for the first time, that my writing might be published, that my writing might be read.
2014 was a long time ago in internet years. What has changed?
With the enshittification of social media, particularly Twitter, where so many writers found a natural home, the algorithm now fails to deliver a predictable feedback loop. Tweets from accounts with tens of thousands of followers are downlisted unless the account-holder pays. But, with no advertising budget and no built-in reach, indie literary mags depend on social media. Tobias Ryan, an editor of minor literature[s], a “collaborative and not-for-profit” online journal, told me that “changes at Twitter in particular made maintaining a presence there more difficult and finding an audience almost impossible. They deactivated Tweet Deck, which logistically made things harder to manage, as well as blocking automatic tweets through WordPress, who host minor lits. That just made everything more difficult/time consuming.”
As platform owners consolidate and monetize, they also limit links that take readers offsite. “Algorithms changed to de-prioritize external links,” says Ryan, “with the aim of keeping users on site, so views fell through the floor.” What’s more, “when Musk took over people didn’t want to be on Twitter anymore (intensifying after the Trump election in 2024). First people fled to Mastodon, then Substack, then Bluesky—what had been a pretty coherent audience splintered. It became necessary to keep up multiple accounts, none of them garnering the views and interactions Twitter once had.”
Today there may be more of what the art theorist Claire Bishop calls “disordered attention,” due to a bewildering proliferation of platforms hosting huge volumes of both human and AI-generated “content,” but the open web is shrinking. Where search engines once provided a list of links to human-written articles, many readers are now content with an AI summary. But to blame readers, or even platforms, would be to offload responsibility for a deliberate and politically motivated war on the arts and humanities by governments and institutions across the English-speaking world.
The withdrawal of grants from independent publishers and magazines—particularly Trump’s devastating cuts to the National Endowment for the Arts, which supported so many independent venues for publication, online and offline—means that authors have to look elsewhere for income. As state support dries up, writers are back in the age of the patron—or, rather, Patreon. To paraphrase the well-worn quote: Times are bad, children no longer obey their parents, and everyone is writing a Substack. The original line is usually attributed to Cicero and sometimes even to Socrates. Search Google Books and it comes up in plenty of volumes (including many published pre-internet), though none by Cicero—an indication that the pre-AI publishing world was not so different from our own.
But if fake quotes existed before AI, editors now have to wade through ever-enlarging submissions piles that aren’t only slush but slop. This year Granta withdrew from its partnership with the Commonwealth Short Story Prize after unconfirmed speculation that this year’s winning story was AI-generated. The publishers of some indie magazines, already engaged in a labor of love, are losing trust in their submissions, and losing heart.
Even pre-AI the slush piles were getting bigger. There are more MFA grads now than there were in 2014, though U.S. and U.K. governments are doing their best to fight this, axing arts and humanities degrees even when they run at a profit. This slashing of liberal arts departments deprives the world not only of writers but critically engaged readers.
What does this mean for emerging writers now? Paywalls are going up again. Panicked by AI, underfunded traditional publishers are shutting the gates to new writers, and those gates are being kept. Literature is being re-professionalized, and too much of the little space that remains can only be occupied by writers with the connections or money to buy the time to stay in the game.
The opportunity to participate in the arts, not only as spectator but creator, was the promise of Web 2.0. It was a promise very imperfectly kept, but one we must remember.
A number of them will be excellent writers, so what does this matter? The narrower the social source of stories told in any medium, the more people are excluded from our political and social imaginary—the right to “main character energy.” Having our stories told for us, however expertly, by a narrower social cohort, is no compensation. We need a culture that allows us to see each other, and that also allows us to see ourselves.
The right to create our own culture doesn’t automatically fix politics. But the right to creative expression not only facilitates human rights: it is a right in itself. It is something humans do, and its limitation is a form of oppression. Why else do so many people who can make money elsewhere still dream of writing, painting, making films, and still use that money to pursue those dreams? In the current climate, writers, editors, and readers will have to look to ways of writing, publishing, and reading beyond the mainstream and beyond the monetized or monetizable.
The net is still a resource. Remove the net and you remove chances for writers whose work is considered too marginal, or who are too personally marginalized, to easily enter the mainstream. Ryan is concentrating on IRL events: “in the context of online entropy,” he says, “it’s good to be with people and share moments, but. . . . Doing things in-person is probably no better or worse than doing things digitally. Our guiding line has always been wanting to make stuff happen with the people whose work we like and respect. That’s what we would consider a net positive.”
It remains an internet positive. The opportunity to participate in the arts, not only as spectator but creator, was the promise of Web 2.0. It was a promise very imperfectly kept. But to fight for a future expansive arts culture, it is necessary to remember and understand what has been achieved.
Image: Gino via Flickr
That Obscure Object, Desire
Chad Wellmon
It’s true: I miss the old Google Search results page, bare but for its blue links and a few ads. For a fleeting historical moment, Search seemed to make real that old dream of a universal library, providing smooth passage to archives, books, and databases—the kinds of learning I’d always coveted. It was both a pathway through the open expanse of knowledge and a portal to surprise, a tool for figuring out what you didn’t know you didn’t know and where and from whom you could find it. So significant was this experience for so many that to search, for a time, simply was “to google.”
Looking back on a quarter century of Google Search and wondering what to make of its AI successor, I can appreciate, if not affirm, my nostalgia. In those blue links I saw footnotes, and in the ranked results I saw discernment; when I first encountered Google as a graduate student in the early 2000s, I recognized something of the self I was working to become and the institution I thought would get me there. Google cultivated this identification with scholarship and the university; its two founders, while in graduate school themselves, even contrasted “advertising oriented” search with their own efforts in the “academic realm.” But now that intellectual work and knowledge institutions are everywhere under threat, the company too, it seems, has moved on.
Liz Reid, Google’s head of Search since 2024, recently described the earlier results page as a “construct,” a response to an older Web built upon webpages, links, and documents. Google’s newer “intelligent” search responds to a very different Web, and a different division of labor and knowledge. The company’s goal is no longer to point us in the right direction—to orient us in the vast public conversation that is hypertext, to set us off on open-ended inquiry with others—but to close the loop of our own intellectual desires by guessing at what it is we want and giving it to us all on its own. Where Search was once a matter of pointing, it is now a matter of presentation.
We’re only just beginning to understand the consequences of this shift. What I’ve realized belatedly, myself, is that things were headed this way all along.
To appreciate what Google accomplished, you have to imagine a world not yet saturated by search. Before the explosive growth of the Web in the mid-1990s, document collections and databases were comparatively bounded, centralized, standardized, homogeneous, and organized—the province of experts and professionals, searchable through specialized syntax, typically with the help of technical mediation. Even early automated information retrieval systems, such as Cornell University’s System for the Mechanical Analysis and Retrieval of Text (SMART) of the 1960s, assumed relatively well-controlled collections. The early Web, by contrast, was distributed, fragile, heterogeneous, and, if at all, only locally organized. Its documents differed not only in language and content but also in format, style, size, and trustworthiness, and they could be queried not only by experts but by general users.
In those blue links I saw footnotes; I recognized something of the self I was working to become and the institution I thought would get me there.
These very features that made the Web exciting—openness, decentralization, proliferation, the lack of gatekeepers—also provoked fears of disorder, excess, and unreliability. Jon Kleinberg, a computer scientist at Cornell, expressed this ambivalence in 1997 when he called the Web an “intricate form of populist hypermedia.” He stressed the need for a filter to determine “authoritative” and “definitive” documents. But how could authority, quality, judgment, and value be conceptualized in the anarchy of the Web?
This problem was answered—in Kleinberg’s own HITS algorithm, and Sergey Brin and Larry Page’s PageRank—by adapting the quantification of academic citation developed by figures such as Eugene Garfield, Francis Narin, and Gabriel Pinski, who founded the field of bibliometrics in the 1960s and 1970s. Within this new search paradigm, a page was judged important if it was linked to by other important pages; authorities were identified by hubs, and hubs by the authorities to which they pointed. No claim was made about the truth or intrinsic value of particular pages. Instead the importance of a page was measured by its relation to other pages, mining older standards of legitimacy—expertise, scholarly review, cultural authority—from link structure.
In this way, PageRank quantified the social judgments assumed to be latent in the Web itself. This founding ideal fit a moment when garbage hadn’t yet overtaken the internet—when the quality of the average webpage was relatively high and trust in institutions hadn’t collapsed. It offered a distinct advantage over human-curated or keyword-based search engines—like AltaVista and Yahoo!—at the same time that it disciplined the “populist” Web, holding out the promise of an objective, neutral, and trustworthy technique for taming digital disorder.
From this vantage, Google’s recent embrace of “intelligent” search is less of a rupture than it might seem. Instead of a reversal of its original project, it’s an intensification of it. From the start, Google treated the Web less as a collective good, a public library or scholarly commons, than as a vast and unruly data source that could be transformed into property and profit. Though it welcomed comparisons to libraries and universities, Google was never either of these things, and never capable of sustaining the distinctive practices and values of scholarly communities. It was always a highly capitalized, competitive business in the market for data at scale.
And yet, it’s not surprising that Kleinberg, Brin, and Page would see in Web links a version of academic citation. Bibliometric citation analysis itself emerged as a response to another crisis of information overload: the postwar proliferation of scientific literature and the ascendance of Big Science. By tracking citational chains—who cited whom—and converting those relations into quantified metrics, citation analysis helped scientists decide what to read and help the federal government and universities decide what to fund. Google adapted this logic to the Web: like its bibliometric forebears, as historian Kevin T. Baker has explained, the company converted a marker of intellectual relation into a simplified measure of value.
I didn’t understand this at the time. What Google offered was a familiar form of authority, one that reminded me of older ways of searching. Already overwhelmed by spam and big banner ads, I wanted a Web that was as orderly and accessible as the libraries and journal collections I browsed. Just as the orderly arrangement of stacks didn’t rule out spontaneity, clicking through links served up by Google felt like following footnotes, tracing paths through an index that both organized and surprised. In reality, the system was increasingly governed by captured values. Garfield’s “solution to the problem of scientific information overload,” Baker writes, “became Google’s solution to the problem of internet information overload, and it was gamed in the same ways.” Since citation became synonymous with “credit” and “endorsement,” the metric became a goal in itself. The result, Baker observes, was to generate ever “tighter feedback loops.” The system Google built to organize the Web begot the very thing that exhausted it: SEO.
Like academic citation systems, PageRank was recursive: it grounded authority by referring to already recognized authority. That recursion could appear legitimate as long as the social institutions and systems behind it were trusted. Once that trust weakened, the same logic began to look less like social discernment than like capture. As the postwar liberal consensus about the value of higher education—that it is a good too good to be questioned—continues to erode, the appeals of universities to their own internal measures of academic prestige and disciplinary expertise (academic rankings, publication numbers, citation counts, prizes and honors) can seem less like manifestations of a professional judgment that demands public deference and more like just another example of elite self-interest.
But there was a key difference. Unlike universities and institutional science, Google recognized early on the threat that value capture posed to search. Its own skepticism about the PageRank ideal and the Web itself pushed it beyond the document-centric model. The ideal search engine, Brin suggested, would not merely process and organize the world’s information; it would “understand” us and give us what we want. The perfect search engine would not model the Web. It would model us. “We’re really making an AI,” Page reportedly told tech journalist Kevin Kelly in 2002.
Over the next two decades, as Google made good on that confession, the company steadily abandoned the PageRank ethos, turning its search engine into a complex, constantly changing system of algorithms, databases, and human interventions. Predictive text systems such as “Did you mean . . . ?” (2001) and autocomplete (2004) treated search not as fixed-token matching but as correction, prediction, and inference. Caffeine (2010) and Panda (2011), among the most significant updates to Search, conditioned relevance and value on a range of new variables: freshness, location, user behavior, query reformulations, human raters’ judgments, and other proxies. Then came the Knowledge Graph (2013), with more than 500 million objects and more than 3.5 billion relationships. Results pages began to display knowledge panels, summaries, key facts, and related entities. Searchers no longer had to visit and compare websites beyond Google; Google itself delivered a ready-made synthesis with its own proprietary semantic infrastructure. PageRank remained a signal in all this, but authority came increasingly from data that Google accumulated, organized, and evaluated.
“Let Google do the searching for you” is telling us the Web is too degraded, our institutions are too untrustworthy, and we ourselves are too exhausted to search. Maybe we are.
By 2017, in the wake of Trump’s first election, Google’s transformation of Search was converging with talk of disinformation, “fake news,” and a more ambient crisis of trust in a “post-truth” era. In response to always-evolving attempts to “game” the system, Google promised “structural” changes: revised Search Quality Rater Guidelines, more human evaluation, and adjustments to the signals feeding its ranking systems. Later that year, Google researchers published their influential Transformer paper (aptly titled “Attention Is All You Need”), introducing what would become the technical hinge between earlier iterations and LLM-based search. Traditional search, Google engineers summed up in a 2021 paper, imposed artificial restrictions by offering references to documents rather than directly answering information needs. It subjected users to an unnecessary and “significant cognitive burden.” Google should aim, they concluded, no longer to improve the index—to contribute to the health of the Web by making it navigable—but to transcend the indexical function of search itself. Today’s AI-driven “intelligent” search encloses the labor, judgment, and intelligence of the Web within a consumer interface that can generate answers directly—and capture the value of doing so.
The promise to “understand exactly what you want” and “give you the right thing” makes this transformation explicit. We sometimes know what we want, but often we don’t; indeed the most satisfying and fruitful inquires often begin in ignorance. Google presumes either that our desires are already clear and that it can satisfy them more efficiently, or that it knows our desires better than we do. In either case, AI search keeps us enclosed within the frictionless, decontextualized domain of this privately owned attention leviathan—all the better to capture and monetize the social knowledge and labor on which it depends.
The all-in pivot to “intelligent” search thus marks much more than a technical innovation. It belongs to a broader crisis of epistemic authority, institutional ruin, and social conflict across the whole of U.S. society—the universities, newspapers, libraries, community and civic organizations, and other social structures that knowledge creation and dissemination rely on. When one of the richest companies in the world implores, “Let Google do the searching for you,” it is telling us that the Web is too degraded, our institutions are too untrustworthy, and we ourselves are too exhausted to search. Maybe we are. But we should be clear about what this entails: the conversion of public crisis into private profit.
Those classic blue links reminded me of the unambitious, prosaic intentions of much earlier search technologies, from early modern indexers and commonplacers to twentieth-century documentalists, librarians, and information retrieval specialists. These figures understood search as the cultivation of particular practices within particular communities. Compared to the epochal pledges of Google, Open AI, and Anthropic, their ambitions may seem mundane, but they helped sustain shared cultures of knowledge production that remain worth sustaining today. Search, in this sense, has always meant an open-ended exploration that propels people beyond themselves and their own settled desires toward the possibility of intellectual surprise and ends that aren’t already fixed.
The history of search is full of warnings: the index is not the book, the footnotes aren’t the page, the map is not the territory. At the core of these admonitions is a fear that search might come to substitute for reading, writing, thinking itself. The PageRank ideal, despite its limitations and eventual failures, preserved something of these older practices. It rested on a reasonable premise: search technologies can aid but cannot replace the messy, context-bound, all-too-human creation of knowledge. With the collapse of that shared premise, the old fears have a new and very real target.
Image: Getty Images
Managed Decline
Cory Doctorow
If I want to make myself really angry about the enshittified state of the internet, I don’t fume over pages of slop-drenched, spam-drowned Google results; nor about the fake-review garlanded sub-Temu garbage on Amazon. I don’t even rage against Elon Musk’s child porn–choked, bot-infested Twitter.
No, when I want to get good and apoplectic about the internet, I think about policymakers. Specifically, I think about all the policymakers I’ve met and corresponded with and begged with and pleaded with and raged against in my twenty-five years with the Electronic Frontier Foundation.
These policymakers are the true authors of our misery. It was they who made decisions that had the foreseeable and foreseen outcome of ushering in the enshittocene, this era in which all of our technology is turning into shit. The thing is, we told them exactly what would happen if they made those policy blunders, and they made them anyway, and now they walk among us, disclaiming any responsibility for systematically rigging the game so that the worst people with the worst ideas make the most money.
It was policymakers who made decisions that had the foreseeable and foreseen outcome of ushering in the enshittocene—this era in which all of our technology is turning into shit.
It’s as though we used to put down rat poison, and we didn’t have a rat problem, and then these guys told us that we had it all wrong and we didn’t need the rat poison after all. Now, rats are eating our faces off, and these guys refused to accept that their pro-rat policies had anything to do with it. They say we’re living through the Time of the Rat, in which the great forces of history and the iron laws of economics have combined to create rats of unheard-of fecundity. When we point out the rats have bought and shuttered the rat poison factories, they get defiant: “So what? We’re not using rat poison anymore. That’s just economically rational!”
Take competition law. Forty years ago, Reagan’s economic advisors from the Chicago School declared that monopolies were efficient. In other words, if you arrive at a juncture in which everyone buys the same products from the same firm, you should conclude that we all just love that company’s products so much that we’ve flocked to them. If you accept this premise, well, then it would be perverse to use public resources to punish that company.
So we just. . . stopped. For decades, we allowed companies like Google and Facebook and Amazon to violate competition law. We let them buy their major competitors, and we let them swallow up any small company that might eventually become a competitor. We let them engage in “predatory pricing”—selling goods below cost in order to drive rivals out of business. We let them tie up retail channels with exclusivity deals, and we let them tie goods and services together. All of this is plainly illegal, but because we were told “monopolies are efficient,” we sat back and watched as these companies violated the law and captured their markets and turned the internet into five giant websites, filled with screenshots of the other four.
The thing is, a company that’s too big to fail is also too big to jail. Once these companies captured their markets, they also captured their regulators. When a sector has a hundred medium-sized companies in it, it is eminently governable. All those competitors erode each other’s margins by raising wages to attract the best workers and by lowering prices to capture the most customers. That leaves them with precious little surplus capital to piss away on regulatory adventures—and even when they turn their efforts to lobbying, they can’t possibly sing with one voice. You can’t get 100 companies to agree on anything, much less a regulatory agenda for their industry.
But when an industry cartelizes into a handful of companies, they find it trivial to capture their regulators. In 2024, Joe Biden’s Federal Trade Commission finalized a rule called “Click to Cancel” that banned the practice of making it very easy to sign up for a service but nearly impossible to cancel it (and the recurring bill it sends to you). In 2025, Donald Trump’s FTC let the rule die. Trump enjoyed the backing of the Wall Street Journal, which ran more than a hundred editorials condemning Biden’s generationally talented FTC chair, Lina Khan. While the Wall Street Journal isn’t the only firm that benefits from the death of Click to Cancel, it does have one of the internet’s most egregiously byzantine cancellation processes.
Think of how tech giants spy on us all the time and use the data to hurt us in terrible ways. The totally unregulated data-broker sector lets predators target marketing categories like “seniors with dementia” and sell credit card history to nursing agencies that use it to lower the hourly wages offered to the most indebted workers.
Why does the data-brokerage industry exist? Because it can. Congress hasn’t passed a new consumer privacy law since 1988, when Reagan signed a bill making it illegal for video store clerks to leak your VHS rentals. Every form of privacy invasion invented since Die Hard was in cinemas is legal. That’s what regulatory capture looks like.
If they weren’t beholden to the power of industry, policymakers could have passed a privacy law at any time in the past thirty-eight years. It’s not like there was a shortage of things to regulate. From deepfake porn to kids being targeted by social media algorithms to racial discrimination in online housing, lending, and job markets, the failure to pass a privacy law for two generations has inflicted enormous pain on hundreds of millions of Americans while making a small number of corporate executives very, very rich.
Regulatory capture isn’t just about preventing laws that get in the way of profits. A sector that captures the regulatory apparatus becomes fused with the state. In fact, tech giants don’t just exercise power over the law; they also enforce law, but only against smaller, newer tech companies that are small and scrappy enough to care whether you like their products. Tech’s favorite weapon here is a bizarre kind of intellectual property law called “anticircumvention.” Under this regime, it is illegal to modify a device that you own if the manufacturer has decreed that the device must not be tampered with. America got its anticircumvention law in 1998 as Section 1201 of the Digital Millennium Copyright Act (DMCA 1201), which established a new felony for “bypassing an access control” that carries penalties of a five-year prison sentence and a $500,000 fine for a first offense.
What this means is that if a digital device that you own is designed to steal your data and/or your money, you are not allowed to alter that digital device in order to protect yourself. Apple blocks Facebook from spying on iPhone owners, but it gathers exactly the same data that Facebook wants, using the phone it sold you, and uses that data to target ads to you. Thanks to anticircumvention law, it is a felony for me to sell you a software patch for your iPhone—the phone you paid $1,000 for, which belongs to you—to stop Apple from stealing your data.
America got its anticircumvention thanks to the efforts of Bruce Lehman, who was the Clinton administration’s intellectual property czar. When Lehman proposed this law, he was laughed out of the room. So, in Lehman’s own words, he did “an end-run around Congress” by going to the United Nations and getting anticircumvention law embedded in a treaty. Congress then passed DMCA 1201 in order to bring the United States into compliance with this treaty obligation.
Lehman was warned at the time that making it illegal for you to decide how you used your own property was a nightmare scenario—in particular, that it would encourage companies to lock their products to consumables (which is how we got $10,000/gallon printer ink); to dominate markets for complementary goods (which is how we got app stores that limit which software you can install and charge the providers of that software thirty cents on every dollar you spend in their apps and games); to lock out repair, to spy on us, and to downgrade their products after we buy them and then charge us subscription fees to use the features that came bundled with the product when we bought it. In short, Lehman was warned that anticircumvention was a moral hazard, an attractive nuisance, a license to enshittify with impunity . . . and he pulled out every stop to get it through Congress anyway.
These are the things I dwell upon in my dark hours. Not the moral deficiencies of the ketamine-addled Zuckermuskian mediocrities who’ve appointed themselves our permanent tech overlords. Awful as those men are, they are merely behaving in ways that the policy environment both permits and encourages. They are not the ultimate causes of enshittification: they are the effects of an enshittogenic environment. Force them out and they will be replaced by someone every bit as awful.
Nor do I dwell upon you and your deficits. I categorically reject the idea that the internet turned to shit because you chose the wrong services. Anyone who says, “If you’re not paying for the product, you’re the product” is helping Big Tech rip you off and ruin the world. Giant companies don’t abuse you because you don’t pay them enough money. They abuse you because they can get away with it, and if you pay them and they can get away with making you the product, then they’ll make you the product anyway. A company that is too big to fail and too big to jail will always be too big to care.
Awful as the ketamine-addled Zuckermuskian mediocrities are, they are not the ultimate causes of enshittification. They are the effects of an enshittogenic environment.
If we bear any responsibility as individuals, it is this: sometimes, we are foolish enough to buy into the excuses for enshittogenic policies and support them. DMCA 1201 was supported by many advocates for creative workers, who thought that banning circumvention would stop people from pirating media. It absolutely, categorically failed to do this (as was foretold at the time), but for people who had a problem, any solution—even a backwards, ineffective, harmful one—was something worth promoting.
And it keeps happening. In the decade since Trump signed SESTA/FOSTA, a bill nominally aimed at preventing sex trafficking by making online publishers liable if their platforms were used by pimps and other odious creatures, platforms simply banned anything remotely pertaining to consensual sex work, forcing sex workers out of the space where they screened clients and back into the street. Pimping, long in retreat, is now a growth industry again; sex trafficking is still alive and well. No one heeded the warnings about the obvious, foreseeable outcome of this badly conceived non-solution to an actual problem. Instead, we embraced the politician’s syllogism: “Something must be done. This is something. I am doing it.”
Today, many people who have good reason to hate Big Tech have thrown their lot in with a new enshittogenic policy: age verification. “Age verification” doesn’t exist. There is only identity verification: that is, strongly tying every packet sent or received on the internet to the real-world identity of living people. Children have lots of problems with the internet, but these problems will not be solved by requiring the most enshittified tech platforms in history to gather and retain even more data about everything they do—and everything everyone else does, too, just in case we turn out to be under age.
It’s not good enough. The internet is the system we use to wire together the world, the communications tool we will need to fight back fascism, survive the climate emergency, and end genocide. We can’t afford to make common cause with the policymakers who don’t care how enshittogenic their ideas are, so long as those ideas are politically expedient.
#médias #capitalisme #Sachzwang #monopoles #www #web_2 #AI #histoite #bibliothèques #GAFAM #politique #Charaktermaske
