In the corner of the internet we built—our little ThinkMyTime bunker with its hand-rolled reflections and late-night arguments—we used to mistake the silence between posts for breathing room, used to believe that if we just sharpened the sentences and kept the nerve, the world would keep meeting us halfway; but the numbers came back like a pathology report and they weren’t subtle: the audience didn’t drift, it evaporated, as if the door we’d been speaking through had been quietly replaced with a wall, as if the same questions we were sweating over had started being answered elsewhere by something that doesn’t sleep, doesn’t doubt, doesn’t age, and can manufacture “thoughtful” on demand—so now the blog reads less like a lighthouse and more like a distress flare in a sky full of satellites, still burning, still technically visible, but no longer the brightest thing many navigate by.
We used to pretend that writing was a kind of private weather. You sat down, you stared into the dark aquarium of your own skull, you stirred the sand, and whatever rose from the bottom—an argument, a confession, a poem with a pulse—was yours. It was slow, expensive, and in a quiet way heroic. Not because it was always good, but because it cost you something you couldn’t invoice: attention, courage, the willingness to be embarrassed in public under your own name. And then the machines arrived, polite as hotel staff and tireless as grief, and they didn’t just learn the grammar. They learned the posture. They learned the cadence of sincerity, the rhythm of “thoughtful,” the warm bathwater tone of a person who has read enough to sound informed and not enough to bleed. Overnight, the market filled with texts that looked like writing the way a mannequin looks like a body: correct silhouette, no heat.
People say AI is killing writing. That’s not accurate. It’s doing something colder. It’s making most writing worthless by making it easy. There’s a difference, and it matters. Killing is dramatic; it implies a victim and a crime scene. This is more like an entire profession waking up to find its product has been cloned and dumped on the street by the tonne, shrink-wrapped, discounted, “good enough” stamped on the box in six languages. The problem isn’t that machines can write. The problem is that readers—tired, distracted, permanently half-elsewhere—were already training themselves to accept writing as a utility. Give me the gist. Give me the summary. Give me the takeaway. They weren’t asking for a human soul; they were asking for a frictionless pill. Now the pharmacy is open 24/7.
Generic thoughtful prose is abundant and cheap because it always wanted to be. It was built out of reusable parts: an opening hook that nods to the zeitgeist, a paragraph that defines terms, a paragraph that acknowledges nuance, a paragraph that gestures at history, a paragraph that turns to the reader and says we should all be kinder, wiser, less polarised, more mindful. Sprinkle one metaphor, salt with one personal anecdote, end with a line that sounds like a closing door. That template existed long before the models; the models just industrialised it. They took the polite middle, the safe complexity, the mass-produced “balanced” that never risks being wrong in a way that costs you friends. They learned the rhetorical shrug: on the one hand, on the other hand, shades of grey, the world is complicated. They learned how to sound like someone who has looked into the abyss and found a content strategy.
And here’s the part nobody wants to say out loud because it makes you look vain: a lot of what we called “writing” was really performance of literacy. Being the person who could assemble coherent sentences was a kind of social advantage. In office corridors and comment sections, it signalled competence. It told strangers you were safe to listen to. But competence is not art, and literacy is not truth. When an algorithm can produce competent sentences at the speed of a firehose, competence stops being scarce. The badge loses its shine. Suddenly, “I can write clearly” is like “I can open a PDF.” Useful, not a personality. If your identity was built around being one of the few people in the room who could take messy feelings and turn them into paragraphs, you’re going to feel a little like a priest watching the congregation discover streaming spirituality.
The future of writing isn’t a future where nobody writes. It’s a future where writing, as a standalone value, collapses. The words themselves become almost weightless. You will see oceans of text, more than at any point in human history, and almost none of it will matter. It will be pumped into the world the way exhaust is pumped into the air: continuous, invisible, taken for granted, and slowly suffocating everything that once depended on clarity. The web becomes a foam of near-identical essays with different skin tones. Your search query returns not voices, but statistically plausible approximations of voices. Your feed becomes a mirror maze of “insights” that all point to the same empty centre.
Philosophical writing gets hit first because it was already living on abstraction, and abstraction is a playground for pattern machines. Philosophy on the internet—most of it—was not a laboratory; it was a lounge. A place to sit with your opinions and stroke them until they purred. A place to rehearse morality in public. A place to talk about meaning without committing to action, to talk about ethics without naming the body that pays the bill. When machines can generate ten thousand variations of the same “What is a good life?” essay before your coffee cools, the question doesn’t become less important. It becomes harder to sell. The supply overwhelms the demand. Not for meaning, but for paragraphs about meaning.
You can’t write as you were doing because the world that rewarded that kind of writing is gone. Not “changing.” Gone. The reader who used to click your blog post because it promised a thoughtful fifteen-minute walk through a concept is now being offered the same walk in thirty seconds, with bullet points, in their preferred tone, with citations, and with the option to ask follow-up questions without admitting they didn’t understand the original. They’re not betraying you; they’re behaving like humans under pressure. Convenience beats virtue almost every time. The people who still want the long walk will take it in a form that fits their lives: an audio essay while they cook, a newsletter in the morning, a curated book club that gives them belonging. A random blog tab in a browser full of open debts doesn’t compete.
Poetry is the interesting casualty because it’s where the lie is most seductive. People will tell you poetry is safe because it’s personal. Because it’s weird. Because it’s ineffable. And yet the machine can do a competent imitation of the ineffable, too. It can produce a poem that looks like longing, sounds like grief, and never once had to lose anything. That’s the uncanny horror: a text that carries the costume of emotion without the injury. It’s a karaoke of the human condition, perfectly on pitch, sung by something that has never had a throat. The reader will still feel something, because humans are generous like that. We project. We sympathise with triangles in cartoons. We cry at fictional deaths. So yes, AI poetry will move people. Not because it is true, but because the reader is.
The older idea of writing—someone in a room trying to be honest—relied on a scarce channel. You had to publish. You had to distribute. You had to earn attention. The gatekeepers were annoying, but they created a kind of friction. Friction is what separates signal from noise. Now the gates are wide open and the friction is gone, replaced by a different kind: the friction of being found. The friction of being believed. The friction of being worth the time. You’re not competing against other writers anymore. You’re competing against a machine that can produce an endless stream of plausible text tuned to the reader’s mood like a smart thermostat.
And because the reader can now have a custom essay written for them, the idea of reading someone else’s general-purpose essay becomes less rational. Why read your two thousand words on stoicism when a model can summarise stoicism, compare it to Buddhism, apply it to my breakup, give me ten exercises, and do it in the voice of a patient friend? The answer, if you’re honest, is that there was always an irrational component to reading: the desire to be in contact with another mind. To feel a presence behind the words. To be changed by a stranger who risked something. But that is not what most internet writing trained people to want. Most internet writing trained people to want results. It trained them to skim for utility. And utility is exactly what the machines provide, with a grin.
So the hard truth is this: a lot of writers are obsolete, and not because they are untalented, but because they were operating inside a category that turned out to be automation-friendly. “Thoughtful” was never a moat. “Balanced” was never a moat. “Well-written” was never a moat. Those are table stakes now, as basic as spelling. If your work can be described as “a smart person reflecting on a topic,” then congratulations: you’ve described a genre that can be generated in bulk, priced at near zero, and served hot. The market does not care that your reflection came from lived experience if you don’t bring the receipts inside the text. The market does not care that you struggled to find the right sentence if the output is indistinguishable from something that didn’t struggle at all.
And don’t hide behind purity. The real insult isn’t that the machine writes. It’s that you, too, are tempted to use it because it’s good enough. That temptation is the proof. You want the shortcut because you’re tired, because you have a life, because the internet is loud, because you have to publish to exist, because the metrics punish silence. The machine doesn’t need to “kill” writing; it just needs to offer a cheaper substitute and let economics do what economics always does. If you choose the substitute, you teach your audience to expect the substitute. Then you wonder why the real thing doesn’t sell.
The future is not illiterate. It is hyperliterate in the way a surveillance state is “safe.” Text everywhere, meaning nowhere. Everyone will be able to generate a novel in a weekend and almost nobody will be able to get a stranger to care about a paragraph. The value migrates away from the words and toward the things that words can’t fake cheaply: original observation, original data, skin in the game, a record of being right and being wrong over time, a voice that is not just a style but a set of commitments. The kind of writing that survives will look less like “content” and more like evidence. Less like vibes, more like a trail. Less like a blog post, more like an artefact.
If that sounds bleak, good. Bleak is honest. The romance of the solitary writer was always half myth anyway, and myths get shredded in periods of technological acceleration. What you can do now is decide what kind of writer you are. If you are writing to be admired for competence, you will lose, because competence is now cheap. If you are writing to sell comfort, you will lose, because comfort is now personalised. If you are writing to prove you are thoughtful, you will lose, because thoughtfulness has been mass-produced. But if you are writing to say something that cannot be said without you—without your specific history, your specific risks, your specific willingness to name what you saw and show how you know—then you’re not competing with a machine. You’re competing with your own cowardice.
And that’s the real reason we can’t write as we were doing. Not because the machines stole the keyboard. Because they stole the alibi. The old excuses are gone. “It takes time.” “It’s hard.” “I need inspiration.” Meanwhile the feed fills with instant, polished, plausible text that makes your drafts look like they’re wearing yesterday’s clothes. In that pressure, you either become a curator of machine mush or you become rarer: a person who writes slower, sharper, and more accountable than the tools. The future of writing is not a battlefield between humans and AI. It’s a sorting mechanism. It separates those who used writing to display intelligence from those who use it to expose reality. One group will be replaced. The other will be read by fewer people, more intensely, and they will matter more.
Most will not choose the harder path. They’ll call it adaptation and ship another warm, balanced, nicely structured essay with no blood in it. The internet will applaud with polite silence. The machines will hum. And somewhere, the remaining readers who still know what it feels like to be touched by a real mind will keep looking for the rare texts that don’t just sound human, but cost human.
This essay was produced using ChatGPT 5.2 Thinking for editing and critique.
