Digital Culture

The link your friend sent at eleven at night. The voice memo from the cousin. The screenshot of an article shared in a thread of seven people who all met in graduate school. That is where most of culture moves now. The stage everyone keeps writing about is mostly empty. The actual cultural transmission happens off-stage, behind a small green icon, and the audience knows each other’s children.

This page is a working archive of digital culture writing. Most of the essays gathered here circle the same observation. The internet stopped being a place a while ago. It is a layer now, sometimes the floor and sometimes the weather. The shift was small enough to miss in any single year and total enough to redraw the cultural map across a decade.

Group chats won the social internet. Not Twitter, not Facebook, not the open square that the early-2000s blog era thought it was building. The thing that actually colonised civic and cultural conversation was the closed thread. WhatsApp, iMessage, Signal, Discord, and the dozen smaller variants people keep open simultaneously. By the early 2020s, more cultural transmission was happening through dark social than through any open platform. The link you read because a friend sent it is the link a public platform never measured.

The shape of that retreat had a precedent that few people connected at the time. The eighteenth-century coffee house and the early twentieth-century salon were the original public spheres in the sense that political theorists later used the term. They were rooms, neither private nor official, where private people could come together as a public to discuss matters of common concern. Those rooms were specific historical accidents. They needed a new middle class, a printing infrastructure, and a working postal service. They did not last as their original form, and they did not need to. They became newspapers, periodicals, public lectures, and eventually radio and television.

That public sphere had already been hollowed out by the mid-twentieth century, swallowed by mass media that turned readers into audiences and citizens into consumers. The early commercial internet briefly looked like it might produce a new one. The blog era, the comments section, the open forum, all of it suggested a structural shift back toward semi-public conversation. For about ten years it almost worked.

Then the open platforms got worse. Not subtly worse. Visibly, structurally, deliberately worse. The platforms optimised for engagement, then for outrage, then for the ad-buying behaviour of the most disinhibited users. Posting in public stopped being conversation and started being product. People with anything careful to say noticed early. They moved their actual thinking into the chat.

The chat is more honest because the audience is smaller. You are talking to ten people, six of whom you have eaten dinner with. You can be tentative. You can be wrong. You can bring up a half-formed thought without it being scraped into someone’s screenshot collection for later weaponisation. The cost of public misspeech became so high that public speech selected for either bland safety or aggressive performance. Everything else moved off-stage.

The collateral is harder to name because it is structural. Public discourse, properly conceived, is a slow filtration system. Things rise into common awareness because enough people, in semi-public, have already made the case and have already been challenged on it. The group chat short-circuits the filtration. A take that is interesting in a chat of twelve people becomes the take that one of those twelve people quietly publishes a week later, fully formed, with the chat’s intermediate disagreement invisible. The op-ed has been pre-laundered. We are reading conclusions whose discussion has been privatised.

The public used to be where ideas went to be tested. Now it is where ideas go to be announced. The testing happens behind invitation. If you are not in the chat, you are not in the test. The fact that ideas appear in the public already polished is itself a sign that the public is not where the work is done.

The chat is, in many concrete ways, an improvement over what came before. Tentative thinking needs a low-cost venue. People are kinder when they know each other. The internet’s open spaces were never going to be a comfortable workshop for half-formed ideas. They could not survive the contact. The chat is the workshop. That is real progress. It is also why so much of what the public used to do has gone missing.

The public sphere was not just a venue. It was a kind of practice. It taught people how to disagree at scale. It taught the muscles of civic patience, the work of listening to a stranger long enough to find the strongest version of their position even when the position annoyed you. The group chat does not need those muscles. The chat selects in advance for people who already share the relevant priors. Disagreement, when it happens, runs on a track of pre-aligned trust. That is not a flaw of the chat. That is what makes it work, and it is also why the chat is not a public.

What replaced the public is not a single chat but a swarm of them. Most people sit in twenty or thirty active threads. Each thread is its own micro-public, with its own norms, its own running jokes, its own consensus. Cultural transmission is now a kind of weather system between these threads. Ideas migrate by being copy-pasted from one chat to another, gathering edits along the way, until they arrive in the open square already exhausted. The exhaustion is the tell.

Algorithms still shape what arrives in the chats, of course. The link that gets passed along almost always came from somewhere algorithmic first. The platform stack has not vanished. It has just been demoted to a feedstock layer. Open feeds produce candidate items. The chats decide which of those items count as cultural events. The platform metrics measure the wrong half of the journey.

This matters for digital culture writing in a particular way. Most of what gets called digital culture analysis is a description of platform behaviour. Posts, likes, follows, viral cycles, creator economies, algorithmic flattening, recommendation systems. All of that is real. None of it is the part of the culture that moves. The actual moving happens one layer below the analysis, where the metrics never reached.

Even the question of taste runs through this filter. Music discovery, once routed through radio and the record store, then through Spotify and YouTube, now runs through a friend who texts you a song. The recommendation that lands is the recommendation that came from someone who knows what kind of week you are having. The algorithmic recommendation is more like the weather forecast. Ambient, often correct, almost never decisive. The decisive recommendation comes through a private channel.

I keep a small list of pieces that turned out to matter. Almost all of them came to me because someone I knew sent them, not because an algorithm surfaced them. The pattern is not unique to me. It is the pattern of how taste actually moves through a population in 2026. The platforms know this. They are quietly trying to build private-channel mimics, encouraging shares to small groups, surfacing the link your friend already saved. They are too late. The infrastructure of attention has already migrated.

What is left for the rest of us is the small deliberate act of bringing chat-quality thought back into something more public, on purpose, knowing it will not survive the contact intact. Writing a careful essay. Hosting a small dinner. Running a comments section that someone actually moderates. Maintaining an old-fashioned mailing list. None of this restores the public sphere. It just keeps the muscle memory alive long enough for whatever comes next to find someone who still remembers how it used to feel.

The public did not die. The public moved. The writing collected on this page tries to track where it went, what travels in its place, and what the new infrastructure quietly costs us in the process.