ai

The Spirit Is Anti-AI, But the Flesh Uses Claude

By in ai, conspicuous crusading, foolosophy, technology

When Paul Kingsnorth launched his Writers Against AI campaign, Alan Jacobs’s response was glib: “I would rather call it Writers for Writing, but sure, sign me up.” The pledge itself was not vague. It asked writers for three things: to promise that they would not use AI in their work as writers, would not support writers who did, and would support only entirely human-made work. A bright line, in other words — until Jacobs noticed that his own use might place a toe on the wrong side of it.

His is not some trivial edge case. Running dictation through Claude to punctuate and clean it is part of the writing process. Uploading six years of idea journals to Claude for thematic analysis, triage, and recovery of unwritten projects is part of the writing process. In plain English, he is using AI as a low-level editorial and developmental assistant. You can easily argue that’s legitimate. I certainly think so. I think many sane people would say it is. But if so, then the campaign he endorsed was foolishly fundamentalist from the start.

He writes, after acknowledging that he does not meet the pledge “by the letter,” “I’m okay with that.” Of course he is. The problem is not that he changed his mind; we should always encourage that. The problem is that he helped advertise a standard of suspicion first, and only afterwards granted himself the interpretive charity it denied to others. Moral severity for thee, exemptions for me: that is how these things always go.

Actually, let me revise that. I don’t think he changed his mind, because I don’t think he thought much about it in the first place. His earlier endorsement was a knee-jerk impulse; once he finally considered it carefully, he realized it was more complicated. He did not first weigh the structure of the campaign, its incentives, its likely abuses, its presumption of suspicion, and then revise his view. He seems to have begun with the gratifying emotional impulse — Down with AI — and only afterwards noticed the practical difficulties. It is a small but exquisite irony that the author of How to Think apparently endorsed a public anti-AI pledge before thinking through whether he himself could satisfy it.

Personally, I couldn’t care less if, or how, writers use AI. I just think that if you’re going to lend your voice to a public shaming/cancellation ritual, you owe it to the likely victims to spend a few minutes thinking about exactly what it is you’re supporting. The problems with this sort of campaign were obvious from the start, even to an idiot like me. A public pledge with no cost to entry is tailor-made for empty grandstanding and virtue-signaling. That is exactly the kind of thing a seasoned academic, especially one old enough to have watched a dozen waves of performative moralism sweep through institutions, ought to recognize on sight. The status incentives were obvious. The temptation to furtive noncompliance was obvious. The drift toward bot-hunting and guilt-by-aesthetic-association was obvious. None of this required prophetic powers. It just required a functioning sense of how public righteousness works.

You do expect this sort of thoughtless moral enthusiasm from excitable young people who still imagine that public declarations are the same thing as judgment. From a sixty-something Christian humanist, you expect some tragic sense — some awareness that campaigns of purity do not stay pure, that standards weaponized against bad people are soon applied promiscuously, and that people are often least reliable precisely when they are enjoying the sensation of being obviously right.

Postmodern Prometheus

By in ai, foolosophy, technology, unintended consequences

Julie Beck:

This is a major transformation, a sudden and dramatic shift in which millions of people are seeking companionship from machines that they formerly could have gotten only from other humans. Yet in some ways, AI companionship is a logical destination for the current direction of human friendship. Social chatbots provide the semblance of a kind of friendship that many people already want, or at least have gotten accustomed to: one that’s on demand, low effort, and completely personalized.

As societies grow wealthier, safer, and more mobile, human beings become freer from inherited limits. We are no longer bound in the same way to the religion of our fathers, the customs of our town, the loyalties of our neighborhood, the companionship of whoever happens to live within walking distance, or even, now, to other people as such. That freedom is real. No honest person wants to romanticize the old world too much. Pre-modern life was often narrow, coercive, parochial, and hard. Most of us would not truly choose to return to a village where our horizons were fixed by birth and where options could be counted on one hand. (Maybe Paul Kingsnorth would.)

The old world imposed solidarity by necessity. The new world requires solidarity by principle. It’s one thing to help your neighbor because he is your neighbor and there is no alternative social universe to defect to. It’s another to help him when you have a thousand other options, all more pleasant, more convenient, and more tailored to your preferences. It’s one thing to remain loyal to a faith because family, custom, and community make defection costly. It’s another to remain loyal when belief has become a buffet, and one may freely take the consolations while discarding the commandments, sacrifices, and obligations that offend one’s convenience.

Choice exposes our weakness. Left free to arrange life around comfort, convenience, self-protection, and personalized reward, we often do exactly that. We evade difficult friends. We thin out our obligations. We “set boundaries” in ways that suspiciously resemble mere selfishness. We select from religion what soothes or flatters us and leave behind what commands us. And now, when the friction of actual friendship becomes too tiring, too slow, or too unpredictable, we can turn to a machine that asks nothing, needs nothing, and exists entirely on our terms.

AI companionship offers the dream of friendship drained of mutuality. It is the logical destination of a culture that increasingly asks of every bond, every institution, every loyalty: what is this doing for me? And yet nobody can live well for long in a world made entirely of that question.

This is the tragic dilemma of modern freedom. We have become so successful at escaping necessity that we now have to recreate, by conviction, the disciplines that necessity once imposed for free.

I Was Knocking On Your Ear’s Door But You Were Always Out

By in ai, moralizing, technology

Matt Milliner:

Recently I had a three-hour conversation with a woman on a plane who was struggling with a family issue, and we actually got somewhere. She thanked me for the time, confessing she had almost gone to AI the night before. I was so glad to have taken the place of the machine, even though it cost me some alone time. Now that I know AI’s counselling allure experientially, I aim to resist it all the more. It is an act of human defiance to be there for each other anyway, to continue to fill the spaces the machines would all too willingly fill in our absence.

That’s great. I mean it sincerely. Yet it must be said: his example is a lottery win. It depends on a three-hour flight, the right seatmate, the right mood, the right person willing and able to talk, and a man who has the temperament and perspective to play volunteer counselor on command. Many people don’t get that lucky. They’re not sitting next to Professor Pastoral on a plane. They’re sitting alone in an apartment after a shift at a dead-end job, with a phone, a headache, and nobody who wants to deal with “thinky stuff.” “Be there for each other anyway” is beautiful as an aspiration, but as a prescription it assumes a social world that lots of people do not inhabit.

I’m reminded of the parable of the drowning man: The flood rises. A man refuses a rescue boat — “God will provide” — refuses a second boat — “God will provide” — refuses the helicopter — “God will provide” — and then drowns. In heaven he complains, and God says, I sent you two boats and a helicopter. Milliner admits that he tried to “pray that prayer” of thanks to God for sending such a helpful tool to him, but somehow couldn’t do it. The lofty advice here is: refuse the imperfect help and wait for the ideal help. Don’t use the chatbot; go find real community; do it the “right” way. Okay. When your actual life contains no such community, or contains it only in meager doses, what then? Intellectuals, whether right or left, religious or secular, all tend to gravitate toward the neat, tidy, elegant theory of life. As a miso-intellectual, I’m more interested in the real-world casualties that get left out, the savaged body parts littering the floor around the bed of Procrustes.

This is just a problem of scarcity. Attention is scarce. Thoughtful feedback is scarce. Adult life is triage. If you have the option of real human counsel, by all means, take it. If you don’t, and the choices are to grind alone inside your skull, pester busy people who won’t respond, or use a limited tool that is available and “good enough,” then yes, many people are going to take the boat that actually shows up.

Fit Audience Though Few

By in ai, technology, the geist of the zeit, writing

A journalist who isn’t running around like Chicken Little clucking over AI and the end of literacy? Well, if that don’t beat all.

Alan Jacobs noted last year, with regard to the doomsaying over declining readership, that a lot of our panic is covert nostalgia for an unusually favorable moment we mistook for normal. For most of literary history, the definition of “literacy” was merely the ability to sign your name. Virgil, Augustine, Dante: none of them expected wide readership, because serious literacy, as we define it now, was never common. So when people moan about a “post-literate age,” Jacobs says the phrase is basically nonsense: literacy is more widespread globally now than at any point in human history. What’s shrinking, at least in the American context, is not “literacy” as such, but a particular postwar distribution of it.

That postwar distribution is what I think of as the mid-century middlebrow: a brief, flattering climate where reading felt more democratic than it had ever been. The G.I. Bill and mass higher education expanded the literate public; print culture was cheap and widespread; television hadn’t yet monopolized every idle minute. Will and Ariel Durant spent forty years producing an eleven-volume history of civilization written for a non-academic audience. Earl Nightingale, a radio speaker and author, produced short common-sense exhortations on motivation and personal development with epigraphs from Wordsworth, Goethe, and Samuel Johnson, on the clear assumption that the average reader would know and care who they were. The Book-of-the-Month Club featured authors like Evelyn Waugh, George Orwell, Winston Churchill, E.B. White, and Graham Greene. In that window, it was possible to entertain the pleasing illusion that “everybody” was at least casually acquainted with the great books, or with their respectable, middlebrow cousins. Then history resumed.

What resumed, exactly? Competition. First on the demand side: TV, then the internet, then social media, then streaming; each a better-engineered rival for attention than the book ever had to face at scale. And now we’re getting the new pressure on the supply side: AI can manufacture endless “good-enough” text — serviceable summaries, plausible op-eds, competent product copy, adequate homework, infinite sludge, all at negligible marginal cost. Journalists, authors, academics, librarians, teachers, grad-school adjacent types, and hardcore readers — the very people reacting like Chicken Little — form a dense, self-referential ecology online. Unusually verbal, unusually literate, and unusually invested in the status of the written word, they quote each other, react to each other, and amplify each other’s anxieties. It can feel like “the public square,” but it’s mostly just a professional subculture dominating the stage, assigning themselves all the starring roles. The numerous readers of Stephen King, devotional literature, or celebrity tell-alls are “excluded” from the conversation, in the idiotic parlance of our time. Look, I personally know enough intelligent women who mostly read “romantasy” by the truckload, enough so that I’ve started reconsidering whether women should have ever been allowed to learn to read in the first place if this is all they’re going to do with it.

We can debate that another time. Either way, the point is, many if not most people simply don’t give a damn about literature or the self-important conceit of the writer-as-seer, communing with the soul of the world and the spirit of the age. Reading for pleasure and cultivation may well go the way of horsemanship, which was an essential skill for middle and upper classes for millennia; now reduced to a niche interest. In the world that exists, you can either lament, moan, and propose protectionist policies to keep a certain class of writer in business, or accept that “excellent prose” is an acquired taste, maybe even a minority taste, and write something worth reading for whoever still has the appetite. Not everyone will. But the existence of a minority has always been enough to keep real literature alive. At the very least, face our fate with dignity and stop whining.

The Bees Plunder the Flowers, But Afterward They Make of Them Honey

By in ai, technology, writing

Jesus, what a bunch of knock-kneed, lily-livered chickenhearts. This fast-growing genre of AI-lament has the whiny tone of Hudson in Aliens: “Game over, man, game over!” Calm your asses down. Writers have survived printing presses, typewriters, radio, TV, the internet, Twitter, and the death of attention itself. I’ve never thought of myself as an optimist, but I seem to have a little more faith in the resilience of human creativity than most. If “AI style” becomes too prevalent the way “2010’s newsletter-style” did, or even the “single-sentence/paragraph clickbait” style that Muni (unwittingly?) reproduces here, those allergic to it will find a new way to express themselves.

Yes, yes, I know — the romantic “writer-as-shaman” conceit. The artisanal mystique. The sacred cave of the soul where the author spelunks for phrases and ideas only he could have formed, unaided by anything but suffering and candlelight. I mean, sure, make it a luxury signal if you want. Put NO BOTS on the dust jacket the way rock bands used to announce NO SYNTHESIZERS in the liner notes. But synthesizers didn’t go away; they won. And AI will too, because this isn’t (X) a moral debate — it’s (Y) an incentives problem. People tend to find their way to the path of least resistance. The tool saves time, smooths friction, and does a serviceable, affordable job at multiple tasks: amanuensis, research assistant, brainstorming-partner, proofreader, editor. People will use it. If you turn that into a sin, you won’t (X) reduce usage; you’ll just (Y) make everyone lie about it. Until, inevitably, some loud anti-clanker crusader gets caught with his pants down in a hotel room with Claude, Gemini, and a bag of white powder and does the full Jimmy Swaggart tear-bath: I have sinned; I have fallen; please forgive me.

You’re not a shaman. You’re a scavenger. So am I. We’re all raccoons, opossums, and vultures with pretensions to artistry. We’ve always worked by conversation, overheard remarks, half-remembered books, accidental plagiarism, and imitation-bad-enough-to-look-original. We steal rhythm from a sentence we loved ten years ago. We borrow a metaphor from a friend and forget where it came from. We build mosaics out of whatever the world coughs up. AI is not the end of writing; it’s just another pile of material in the landscape — another junkyard where you can find a useful hinge, a bright shard of glass, a piece of wire you can turn into something that actually holds its shape. The only real question is whether you can still make it yours.

When the Punishment Begins

By in ai, moralizing, technology, the geist of the zeit, the madness of crowds

By the time the church bell quits tolling, they’re already gathering at the gaol. The rumor has traveled faster than any rider: dark doings among the scribes. Something unclean has happened in the scriptorium. A man of the guild! one of them, trained in the sacred arts of ink and ledger! — has been seen consorting with the enemy. The crowd presses closer, anxious and thrilled, crossing themselves as if the very idea might infect the air. They mutter about corrupted ink and devilish machines. And there! Pangram, Botfinder General, straight-backed, solemn, with his shiny new device that turns speculation into a number. He announces his judgment: under interrogation, the scribe’s work has been shown to bear the Devil’s watermark. That’s all they require. Relief loosens their faces. Now they can go home feeling righteous, already anticipating tomorrow’s sport: the culprit in the pillory, the mob’s clean-conscience holiday, rotten vegetables thrown in the name of virtue.

This is what it looks like in a despoiled information ecosystem when people start hunting for scapegoats. Social media already fractured the old reality-brokering order; now AI arrives and further upsets the balance of power. Video can be faked, audio can be faked, and worst of all for the scribes, text can now be convincingly mimicked too. Nobody trusts anything or anyone. It’s a Lord of the Flies media environment, and everyone’s ready to kill over who gets to wield the conch of authoritative speech. The old agenda-setters feel their entire way of life slipping away, and set about like modern Luddites in a futile effort to smash the new narrative-weaving machines. In that atmosphere, even reasonable people start craving a single line between real and artificial, human and machine.

I’m only superficially familiar with René Girard, but I couldn’t help thinking of him here. Fear wants an object. Anxiety wants a face. Suspicion wants judgment and a sentence. One one level, this episode is about the particulars of journalistic ethics. On a deeper level, this is about the anxieties many people feel as their place in the world feels threatened. We’re anxious over what it even means to be human in such a cyborg world, especially when technology can convincingly mimic even the most quintessential human activities, like art and craft. Ironically, we outsource our uncertain judgment to…more machine analysis, more software. Pangram, and other similar bot informers, will serve as the authority demarcating the real from the artificial. Most people won’t know how Pangram works, or how reliably it works, if at all, but it doesn’t matter. We just need to believe it does what we need it to, and then we feel relief. The world makes sense again, and now we can focus on punishing the scapegoats: the species-traitors who collaborated.

Immoral Voltron

By in ai, moralizing, technology

In “Mechanical Man,” Lou Barlow sang, “When I said I understood, I only knew where to stand.” He seemed to anticipate the conformist idiocy of online discourse. Not “what is this thing?” but “what vibe is this thing?” Tocqueville noted this two hundred years ago: when you remove custom and aristocracy as sources of authority, people don’t start thinking independently; they just start nervously measuring themselves against their peers. We’ve stopped forming views and started presenting ourselves in stances. Understanding the object is old-fashioned. The point is to locate it on the moral map as fast as possible, because the map is the social world and nobody wants to be caught in the wrong zip code.

Once the sorting happens, the story writes itself. A complex technology becomes a one-stop portal for every available sin: data centers in poor areas, poisoned water, higher utility bills, racism, “classicism,”  fascism, the death of art, and the reduction of your mind to sludge. It’s the inverse of Berlin’s monism: not “all goods must fit together in a harmonious whole,” but “all evils must fit together in a malevolent whole.” Snap on racism. Snap on classism. Snap on pollution. Snap on fascism. Congratulations; you’ve formed IMMORAL VOLTRON. Now you don’t have to argue about particulars; you just point at the monster and scream.

The trick, of course, is that these mad-libs incantations don’t enlighten; they discipline. Once AI is declared contaminated, any hesitation becomes complicity. To question one link in the chain is to “reveal yourself” as the kind of person who would defend the whole chain. Berlin’s emphasis on pluralism, tragic conflict, and incommensurable goods dies here, because pluralism requires the adult ability to say: some harms are real, some benefits are real, and people can disagree on how to balance them without being cast as villains. But stance culture can’t afford that. It needs a single verdict, a single villain, a single shaming label. Tradeoffs are messy; exorcisms are tidy.

There’s Terrible Trouble Behind It!

By in ai, fresh hell, technology, the geist of the zeit

Amelia Miller:

For adults, we need warnings that clearly convey the serious risks. The lessons that took tobacco regulators decades to learn should apply to artificial intimacy governance from the start. Small print disclaimers about the effects of smoking have been rightfully criticized as woefully deficient, but large graphics on cigarette packs of black lungs and dying patients hurt sales. The harms caused by A.I. companions can be equally visceral. The groundbreaking guardrails that Gov. Gavin Newsom of California signed into law last year, which require chatbots to nudge minors to take breaks during long sessions, are a step in the right direction, but a polite suggestion after three hours of A.I. conversation is not enough. Why not play video testimonials from people whose human relationships withered after years of nonstop chat with bots?

When I was a kid in ye olden days, reformers were forever trying to save us from ourselves. Smoking will kill you. Drunk driving will kill you and someone else. Drugs will melt your brain. We had assemblies with “scared straight” guest speakers, posters of tar-blackened lungs on the walls of health class, PSA’s, and (God help us) McGruff the Crime Dog records with bouncy little earworms about every substance known to man. The theory was that if we were properly educated, if the warning labels were vivid enough, if the videos were grisly enough, we’d make good choices forever after. And yet people kept smoking. People kept driving drunk. People kept getting high. Not because they lacked information, but because they were stressed, bored, lonely, impulsive, self-destructive, curious, or just plain human.

Now the new panic is emotional intimacy with AI, and the new solution is…warning labels and more regulation. “This is not a substitute for genuine relationships.” “ChatGPT can make mistakes.” And here some newly-minted reformer wants to make teenagers watch video testimonials from recovering chatbot junkies, like driver’s ed films of crumpled cars. But McGruff is the perfect emblem of the whole genre: you’re trying to solve a messy relational problem with a cheesy piece of mediated messaging. You even risk the pink-elephant effect — by foregrounding the forbidden thing (“crack cocaine!” “romantic bots!”) you plant it in the mind with an oddly vivid presence, while the adults congratulate themselves for having “done something.” It’s outsourcing the hard work of parenting and community to a cartoon mascot and a compliance checklist.

None of this is to say that nothing can or should ever be improved. But there is no design standard, warning label, or stern authority that can make life safe and painless. People will always try to jerry-rig comfort out of whatever’s available when their actual lives are starving for it. If you don’t want Johnny turning into a dope fiend, you can’t fix that with a McGruff album. If you don’t want Johnny “marrying” a chatbot, you can’t fix that with a disclaimer. The solution, such as it is, was never primarily informational. It was always relational. And wherever we go, there we are.

I do look forward to today’s cartoon characters producing an album warning kids to not fuck an android sexbot.

Relligiosi

By in ai, conspicuous crusading, moralizing, technology, the geist of the zeit, the madness of crowds, writing

Relligiosi

That’s what I call those who take it upon themselves to convince others and make them similar to themselves, make them think in the same way.
I abhor this kind of mental usurping, the great task of political parties and confessions.
What matters for you, doesn’t matter for me. Because it matters for you, it doesn’t matter for me.
To understand oneself, it’s essential (in my eyes) to think differently.

— Paul Valéry, Cahiers / Notebooks 1

Via Paul Kingsnorth. And what does this pledge entail, you may ask?

Do you support the campaign? Go rape yourself, that’s my response. Even answering that question concedes too much to the commissar asking it. It’s none of your goddamned business how I feel about AI use. I don’t owe you a pledge, a badge, a logo in my sidebar, or an affidavit about my writing process. If I read something and it feels synthetic, I’m free to close the tab and avoid that writer in the future. That’s how grown-ups handle taste, without forming a militia and a blacklist.

What I can’t stand, as always, is the missionary spirit animating this entire self-appointed class of jumped-up hall monitors: the righteous salesmanship, the demand for public alignment, the insinuation that silence is complicity. This “download a free logo and join us” routine is a ritual with no barrier to entry for people who like the feeling of conscience without the burden of thought. Once the ritual exists, the next stage is guaranteed: bot-sniffing as sport, comment-section inquisitions, and purity audits, as though the only point of reading is to detect the Devil’s watermark.

This is the online HOA, or the Concerned Citizens Council, coming to the door with clipboards and smiles, asking you to sign. And underneath the smile is the implied threat: Do you want us to make it known that you refused? Because that’s how these things work. There is no neutral ground in a holy war. The whole point of the ritual is to sort the righteous from the wicked, and to make “I decline” sound like a confession. Even if you happen to agree with this particular campaign, the same petty functionaries will be back knocking on your door next week, with clipboards and smiles, expecting your participation in the next loyalty test. Whenever you finally decline, the smiles will vanish, and you’ll see them for what they really are.

The most a campaign like this can do is create another platform for virtue-signaling (as if we needed more) and another excuse for resentful mediocrities to claw at higher-status targets under a banner of public morality. Watch how quickly “protecting the craft” turns into guilty-until-proven-innocent accusations. It invites the exact type Valéry despised: the mental usurper, the person who can’t just choose their own life, but must conscript yours into their crusade. No thanks. Read what you want. Write what you want, how you want. And leave me out of your religion.

Take This Heavy Burden

By in ai, foolosophy, technology

Blair Effron:

There is little doubt A.I. will be transformative. And yet, for all the disruption it promises, I am struck by how much will remain unchanged. The most consequential decisions in business have never been about processing information faster or detecting patterns more efficiently. The most salient concerns are questions such as what kind of enterprise a firm should aspire to be, what culture it should embrace, what risks it should tolerate and how its leaders can plan when the path forward is unclear. These are questions of judgment, and judgment cannot be automated — at least not any time soon.

True. But it can be abdicated instantly, and there’s nothing people like more than handing off responsibility for judgment. They’ll love a machine that makes decisions feel inevitable. AI doesn’t replace judgment so much as provide a frictionless way to dodge it: the model recommends… and suddenly nobody is choosing, they’re “following the data.” If things go well, leadership “leveraged cutting-edge AI.” If things go badly, “the model was wrong.” Judgment won’t be automated. It will be subcontracted, with plausible deniability baked in.

Yippee Cahiers!

By in ai, books

For much of the past year, I’ve been making an effort to seek out more aphoristic, fragmentary, short-form writing. I want material that’s more suggestive than declarative, stuff that allows me to play off of it in my own writing. One recommendation that came up was Paul Valéry’s five-volume notebooks, Cahiers. In his book In a Cardboard Belt!, Joseph Epstein suggested that the first two volumes were “more literary and, in a distant sense of the word, personal” than the remaining three volumes, which were more scientific and philosophical, so I set my sights on those, and will let one of my aristocratic patrons secure the other three at some point.

Unfortunately, each volume is typically priced at eye-watering sums. Patrick Kurp confirmed that they are marvels and taunted me with the fact that he lives in civilization and thus can borrow them from the university library whenever he wants, but I, out in the sticks, couldn’t even secure an inter-library loan for them. I resigned myself to setting price alerts in hopes of pouncing upon a used copy accidentally listed for $10.00 instead of $100.00, but my hopes weren’t high.

Don’t listen to the anti-AI fanatics. ChatGPT found me copies from the University of Toronto Press for 60% off, which, with the conversion rate, amounts to them practically giving the books away for a song. It appears that they don’t ship directly to the States, but thanks to my contacts north of the border, I can work around that. I will not be denied!

We Made Ourselves a Pillar, We Just Used It As a Crutch

By in ai, aphorisms, foolosophy, the geist of the zeit

James Geary:

This is the awesome power of aphorisms, and it is what sets them apart from the trite sound bites of social-media influencers and the platitudes of self-help gurus: Aphorisms do not make you feel good about yourself. That Burrill saying doesn’t offer an easy fix or a neat solution. In fact, it doesn’t offer a fix or a solution at all. It is a great example of aphoristic intelligence at work. It makes us reconsider what we’re doing, or not doing, with our lives.

I’ve enjoyed Geary’s books about language in general and aphorisms in particular (here and here for those interested), and I certainly don’t begrudge the man for promoting a reissued edition of one of them, but this is The Atlantic, whose audience proudly sits right on top of the bell curve, so of course the topic has to be pre-chewed and spooned out in the consistency of Gerber’s baby food. Hence, “Aphoristic Intelligence Beats Artificial Intelligence.” You like to read literature, and maybe aphorisms? You feel anxious about AI? You’re a subscriber to the premier magazine for the stupidest smart people in the professional-managerial class? Well! Come into my parlor, said the spider to the fly…

I love aphorisms, and I would agree that the best of them have a compressed depth that lends itself to different interpretations as opposed to pat answers, but this kind of framing is itself simplistic and, well, platitudinous. Not all aphorisms are about confronting us with our shameful failings. Gary Saul Morson, for instance, uses “Yogi Berra-isms” as an example of what he calls “wise folly,” sayings that sound dumb on the surface but contain surprising depths. And personally, when I read La Rochefoucauld skewering courtly-salon society, I come away thinking less about my own moral culpability and more about how maybe the French Revolution had some good points. Either way, the point is, this quasi-macho attitude — “I read difficult, confrontational literature; I shine the light of moral interrogation into the dark, wretched corners of my soul” — can itself be a form of self-flattery, and self-flattery is what we should be on guard against, in whatever guise it presents itself. The Devil can cite The Atlantic for his purpose too.

Solzhenitsyn famously said that the line dividing good from evil runs right through the heart of every human being, and that this is why schemes which promise to neatly divide vice from virtue and quarantine the former are doomed to failure. Everything in life can present a choice between honesty and self-flattery. You can read aphorisms for shallow reasons and retain nothing; you can use AI intelligently and creatively. In fact, most if not all of the moral sorting mechanisms we come up with are complicated by messy reality like that. There’s no sanctuary from shallow thinking, only the necessity of vigilance. The minute you start congratulating yourself on being immune to shallowness is usually when it has you by the throat.

The Long, Foul Rant That Is Our Moment In History

By in ai, moralizing, socmed, technology, the statusphere

My thesis, again, is a simple one. The information technologies of the twenty-first century have enabled the public, composed of amateurs, people from nowhere, to break the power of the political hierarchies of the industrial age. The result hasn’t been a completed revolution in the manner of 1789 or 1917, or utter collapse as in 1991, but more like the prolonged period of instability that preceded the settlement of Westphalia in 1648. Neither side can wipe out the other. A resolution, when it comes, may well defy the terms of the struggle. None is remotely visible as I write these lines.

— Martin Gurri, The Revolt of the Public and the Crisis of Authority in the New Millennium

The internet didn’t invent deception, of course, but it did make mediation the default condition of social life. Most of us have probably heard the old joke from the New Yorker: on the internet, nobody knows you’re a dog. Once you live through screens, you’re always inferring a reality you cannot directly inspect. Even when nobody is lying, the medium itself encourages curation, performance, selective lighting, selective framing. A fragile trust is inherent from the start.

Social media monetized that uncertainty. It turned identity into a product and status into a scoreboard. It incentivized us to treat the self as a brand and attention as a currency — quantified, gamable, and therefore permanently vulnerable to counterfeiting. Filters, lighting, editing, branding, metrics; none of it is “fake” in the simple sense. It’s real, but optimized. And optimization is the prelude to suspicion, because it blurs the line between honest self-presentation and calculated manipulation.

As Gurri explains in his excellent book, the basic claim that institutions can certify reality — this is true; this is false; these are the rules; this is what happened — has been weakened. That doesn’t mean institutions are always wrong, or that the public is always right. It means the power to end arguments by proclamation has decayed. The public can keep disputes permanently open. Competing narratives can be built cheaply, distributed instantly, and defended by swarms of “people from nowhere.” Arguments metastasize, every claim begins to feel like propaganda, every correction begins to feel like power protecting itself, every disagreement starts to carry the tone of betrayal. You don’t merely think I’m mistaken; you think I’m captured, paid, brainwashed, malignant, or stupid. And once that adversarial posture becomes political weather, it spreads. “Chaos rampant, an age of distrust. Confrontation, impulsive habitat,” the metal band Slayer sang, describing hell, social media, or both.

For a while, we still had reliable tells. Spam looked like spam. Machine writing sounded like a machine. Bad Photoshop had a visible sheen. The reassuring thing about fakery was that it was legible.

Generative AI is an accelerant because it collapses the tells. It produces competent imitation at scale — text, images, voices, short videos — often good enough to pass the first glance, sometimes good enough to pass the second. So the aesthetic question (“is it good?”) gets replaced by a forensic question: “Who made this, how, and do they deserve credit?” When outputs get cheap, the status war moves to authorship and provenance. Everyone becomes a customs officer stamping passports of authenticity.

It’s common to say “online isn’t real life,” and for a long time that was comforting because in-person reality could still function as the ultimate lie detector. Filters might beautify a photo, but presence would reassert itself up close. You can curate your profile, but you can’t curate your face at close range under ordinary lighting on an ordinary day. Digital editing can’t firm up a weak handshake.

Except now the body itself is becoming a kind of interface, not so much because bodies are “fake” than because the meaning of a body is less legible. PEDs and GLP-1s create an actual result, not a mere illusion. The suspicious question isn’t “is that physique real?” It’s: “What does it signify?”

For decades, we treated physical transformations — lean bodies, muscular bodies, big lifts, dramatic before-and-after stories — as costly credentials. We read them as moral evidence: discipline, self-control, sacrifice, will. You could argue about how fair that moral reading was, but it was socially legible. It made sorting easy. Now, a person can be leaner because appetite has been pharmacologically altered. A person can be bigger and stronger because recovery and hypertrophy are pharmacologically assisted. Now the old inference (“therefore virtue”) is unreliable because the method is opaque. And humans hate unreliable inference. We hate it the way we hate a loaded die: not only because it might be unfair, but because it makes the game hard to interpret.

Then comes the new complication: AI doesn’t arrive as just another tool, like a better wrench. People talk to it. It answers. It imitates personality. It can console, flatter, argue, roleplay. So even if it has no intentions in the strict sense, it reliably triggers the human perception of agency. A new force enters the already-chaotic field — something that can amplify messages, imitate people, generate culture, automate persuasion — and the public can’t agree what kind of thing it is. Friend, enemy, mercenary, or merely a new industrial capability. The disagreement itself becomes part of the atmosphere. Europe got a “German Question” in 1871. We’re getting an “AI Question” now.

Our atmosphere of suspicion is less paranoia than a rational adaptation to broken shortcuts. When cheap trust signals stop working, people do what they always do: they demand receipts (“prove it,” “disclose it,” “show your drafts,” “are you natural?”), they invent new tells (“AI slop,” “natty or not,” “Ozempic face”), and they fight to rewrite the moral code. It encourages moral entrepreneurship. Not everyone is seeking truth; everyone is seeking advantage. Authenticity becomes a weapon because authenticity is a status currency in a world where the old currency is being inflated.

It’s sort of like musical chairs. Technological change, as it often does, has set the music playing again, and people are circling around warily, hoping to be well-positioned in the scrum for a high-status seat once it stops. Eventually some settlement will arrive. Possibly, as Gurri says, one that defies the terms of the struggle. But for now, the defining feeling is that the old shortcuts for telling what’s real are failing, and we haven’t yet agreed on new ones.

The Soft Hum of Suspicion

By in ai, technology, the geist of the zeit, the statusphere, writing

Sam Kriss:

If you’re anything like me, you did not enjoy reading that paragraph. Everything about it puts me on alert: Something is wrong here; this text is not what it says it is. It’s one of them. Entirely ordinary words, like “tapestry,” which has been innocently describing a kind of vertical carpet for more than 500 years, make me suddenly tense. I’m driven to the point of fury by any sentence following the pattern “It’s not X, it’s Y,” even though this totally normal construction appears in such generally well-received bodies of literature as the Bible and Shakespeare. But whatever these little quirks of language used to mean, that’s not what they mean any more. All of these are now telltale signs that what you’re reading was churned out by an A.I.

A specter is haunting the severely-educated — the specter of AI usage in writing. All the powers of old media have entered into a holy alliance to exorcise this specter; the NYT Magazine and the Atlantic, the Iowa Writer’s Workshop and Turnitin, university integrity offices and Reddit bot-sniffers. And so here we have Sam Kriss, yet another spoiled-rich-kid communist who wants to burn down the system that gave him his rarefied aesthetic education (and who nonetheless still harbors bourgeois dreams of inheriting his parents’ “massive house“) swirling a vintage line from Virginia Woolf under his nostrils as he informs you plebs that he was already bored with ChatGPT in 2022, before you even knew it existed.

Seriously, though, the people most likely to see AI text as contamination are the people whose social identity is bound up with text: the literary humanities set, the professional writers, the custodians of taste. For everyone else, it’s just a convenience. An appliance. Maybe that’s part of the outrage: automation was supposed to take the warehouse and the assembly line, not saunter into the poetry workshop and start doing an impression of “voice.” The machines are stealing the wrong jobs.

Most people have never treated everyday writing as sacred in the first place. We flatter ourselves that Christmas wishes “come from the heart,” but come on: normal people use templates because they have lives. Mass-produced holiday cards work for their intended purpose, and no one regards them as a civilizational threat. If anything, the sincere and authentic are the ones who refuse the template and send a detailed, self-important “year in review” that makes everyone vaguely regret having an inbox. Before chatbots, the socially-striving could always reach for Bartlett’s book of quotations; AI just makes the rhetorical prosthetics cheaper and faster. And of course, anything that steals fire from Olympus and gives it to the masses is, by definition, sacrilege.

We are changed by what we change. Good ol’ Auden. It’s the best description I know of what’s happening to public language right now. People imitate their peers constantly. So of course we’re going to imitate the newest voice that surrounds us, even when it isn’t a person. If the environment becomes saturated with AI phrasing, we will begin to speak it the way you catch an accent. From a strictly evolutionary angle, that’s neither good nor bad. It’s adaptation. It’s the social animal looking for a snugger fit.

But there’s always a counter-movement. A stock phrase can be useful right up until the moment it starts to sound insipid from overuse. Then we drop it and reach for something else, not because we’re heroes of authenticity but because we’re bored. “AI voice” will eventually become a class marker the way corporate-speak already is: safe, competent, emotionally pre-chewed. And as that dialect hardens into “middlebrow, typical, mediocre,” anyone who wants to sound alive will have to push harder for distinction — more particularity, more risk, more texture. So it goes: the mass adopts a style because it’s easy, and then a minority flees it because it’s everywhere.

These pieces about AI writing all share the same undertone: suspicion. A kind of low-grade paranoia that turns reading into auditing. You can’t just take in a paragraph anymore; you have to scrutinize it for tells, like a customs agent checking passports. “Tapestry,” “delve,” the syrupy uplift, the eager sincerity. ChatGPT, like Kilroy, was here. Maybe that’s justified, maybe not. But it’s corrosive either way. Even when you’re right to be wary, living in that posture constantly makes you meaner, narrower, less amused.

And it isn’t just AI. It’s a broader collapse of trust in cheap signals. I don’t have any grand analysis or conclusion; I’m just observing what appears to be the prevalent mood. A sincere paragraph used to be a mild indicator of effort. Now it might be a one-click artifact. So the audience becomes forensic, and the writer becomes defensive, and both sides grow touchy. It reminds me of the obsession with PEDs in sports: post a physique photo or a lift total and you’ll be swarmed by amateur detectives hunting for signs of steroid use. Enhancement spreads (or the suspicion of it spreads), and suddenly every impressive result is presumed guilty until proven otherwise. The honest have to perform extra authenticity; the dishonest learn to fake the performance; and the spectators get to feel clever for “catching” someone.

We are changed by what we change. We built machines that can produce endless fluent prose, and now we’re building a world in which fluent prose is no longer evidence of a mind behind it, only evidence that a machine exists. Reading starts to feel like border control.

Word Tords

By in ai, language, the geist of the zeit

I might have guessed. Most trendy words gradually grow in popularity; this excrescence became omnipresent almost instantly. Part of that is because of its association with AI, but it quickly became a much more general term of derision. I mostly see it used on Twitter, which I browse because, as Elon Musk said when he bought it, it functions as something of a town square in the digital age. I must admit, though, that the sheer maddening repetition of “slop” in our digital town square has given me persistent fantasies of climbing a nearby bell tower to…jump. (What? What did you think I was going to say? Sickos.)

I think what I hate isn’t slang, or argot, or jargon as such. Those can be vivid, funny, even clarifying. What I hate is autopilot language, what I’ve started thinking of as meme-phrases: prefab little utterances engineered for online life, easy to repeat, easy to recognize, and rewarded mostly for being the correct noise to make at the correct moment. “Gaslighting.” “Psyop.” “Read the room.” “Not a good look.” “Quiet part out loud.” “Take a seat.” They’re verbal reaction GIFs: not necessarily stupid, but too often substituted for an actual sentence.

Clichés, whatever their faults, at least express a thought. In a way, they’re victims of their own success: they’re true enough and useful enough that they get repeated into linguistic satiety. Unlike many, I’ve never had any problem with “It is what it is.” It’s a perfectly serviceable way to express resignation or acceptance.

The meme-phrases that irritate me are different. They don’t spread because they’re true; they spread because they’re socially efficient. They’re low on meaning and high on utility as a signal. “Read the room” isn’t an insight, it’s a rebuke — an attempt to shame someone into conformity with the group’s mood. “Gaslighting” gets used indiscriminately to mean everything from ordinary lying to actual, sustained manipulation. “Slop” started as a particular insult for subpar AI churn, and within minutes it became a universal solvent: anything I dislike, anything I’m tired of, anything I want to dismiss without describing.

What drives me insane isn’t that language evolves. It’s the mindless repetition, the obvious parroting. You can watch the phrase spread through a feed the way a yawn spreads through a room. People aren’t choosing words; they’re selecting a noise that marks them as current and aligned. The phrase could be replaced by a grunt, squeak, or hoot and do the same job of indicating displeasure or disapproval.

And yes, I admit I might simply be a curmudgeonly asshole: I’m getting irritated at a sign of what a social species we are. It’s like getting annoyed at the sound of a flock of ducks quacking. Why would I expect each duck to sound unique or interesting? Yet — irrationally, stubbornly — that’s exactly what I expect from people. I want to hear thoughts, not noise. Ah, well, it is what it is.

«‹ 3 4 5 6