4chan Is Not a Monolith

The Fires of History Newsletter, No. 5

It is the evening of May 15, 2007, and somewhere on 4chan’s /v/ board, a user has just committed an act of cultural vandalism so perfectly calibrated that it will still be producing serotonin in strangers’ brains two decades later.

The setup is simple. Someone posts a thread: “GTA IV trailer just leaked.” The link looks plausible. You click it. And instead of Grand Theft Auto IV footage, you get Rick Astley – coiffed, blazered, performing a choreography that looks like a man trying to hail a cab while also having a religious experience – singing “Never Gonna Give You Up” at you with the full sincerity of 1987.

You have been rickrolled.

The user who did this was Shawn Cotter, who later told Vice he chose the song because he found a list of hits from 1987 – the year he was born – and the video struck him as inherently funny. There was no grand theory. A guy thought a music video was ridiculous, linked it where people expected something else, and the resulting bait-and-switch became one of the most recognized memes in the history of the internet. The music video has been viewed over a billion times. Rick Astley has leaned into it with the good humor of a man who understands he has been handed the most improbable second career in pop music history.

The whole thing started because a 4chan user thought it would be funny.

Now here is the part that every mainstream account of 4chan gets wrong: the rickroll did not come from /b/. It came from /v/ – Video Games. A completely different board with a completely different culture, different regulars, and different norms. But every article about the rickroll says “4chan” as if the site were a single entity with a single culture, the way you might say “Harvard” or “the Pentagon.” It is not. It has never been. And the refusal to understand this produces a distortion so fundamental that it corrupts nearly everything written about internet culture since 2008.

Seventy Boards, One Name

4chan, as of this writing, has over seventy boards. They cover anime, cooking, literature, music, fitness, fashion, papercraft and origami, science and math, transportation, weapons, wallpapers, travel, do-it-yourself projects, and dozens of other topics. Each board has its own culture, its own in-jokes, its own norms, and its own regulars.

/a/ (Anime & Manga) was the original board – the reason 4chan was created. Christopher Poole, a fifteen-year-old going by “moot,” launched it in October 2003 because he wanted a place to discuss anime that worked like the Japanese imageboards he admired. That is the full origin story. A teenager who liked anime.

/mu/ (Music) runs sophisticated recommendation threads and genre-taxonomy debates that would not be out of place in a musicology seminar. /lit/ (Literature) maintains reading lists and conducts close readings of literary fiction. /ck/ (Food & Cooking) shares recipes. /diy/ (Do-It-Yourself) helps people fix things. /po/ (Papercraft & Origami) is one of the slowest, most peaceful boards on the site, frequented by people who fold paper into intricate shapes and share templates. If your image of 4chan is shaped by CNN, the existence of /po/ will not compute.

And then there is /b/ – Random – the board with no rules except the global rules against illegal content. /b/ is the board that produces the content that makes the news. /b/ is the board that journalists write about. /b/ is the board that, in the public imagination, is 4chan.

/b/ is one board out of more than seventy.

The Monolith Problem

Academics Amanda Colley and Samuel Moore identified this in a 2020 study: the “4chan monolith problem.” The media, politicians, and most researchers treat 4chan as a single community with a single culture, attributing the behavior of /b/ or /pol/ (Politically Incorrect) to the entire site. It is the equivalent of writing about “Reddit” as if every subreddit were the same, or writing about “New York” as if Manhattan and Staten Island were interchangeable.

The monolith treatment is not just lazy. It is structurally misleading in a way that cascades through every subsequent analysis.

When a news organization reports that “4chan radicalized a mass shooter,” what they mean is that a specific board (/pol/) with a specific culture (openly white nationalist since at least 2016) was a site of radicalization. What the audience hears is that the website – the whole website, all seventy-plus boards, every user – is a radicalization engine. The person on /a/ discussing manga, the person on /ck/ sharing a recipe for pad thai, the person on /po/ folding a paper crane – all of them are, in the monolith framework, members of a radicalization cult.

This does not merely insult the majority of 4chan’s users. It makes the actual problem harder to understand and address. If the entire site is the problem, the solution is to shut down the entire site. If a specific board with a specific culture is the problem, the solution is vastly more targeted and more likely to work. The monolith framing is not just wrong. It is counterproductive. It ensures that every proposed response will be overbroad, underprecise, and ultimately ineffective.

The Architecture

To understand why 4chan is not a monolith, you have to understand the architecture, because the architecture is the argument.

Anonymous by default. No registration. No account. No username. The name field says “Anonymous” unless you type something else, and typing something else is actively discouraged. Over ninety percent of posts are made as “Anonymous” – a figure from Bernstein et al.’s 2011 MIT study analyzing over five million posts. This is not pseudonymity, where you build a persistent identity over time. This is anonymity. Every post starts from zero.

Threads expire. 4chan has a fixed number of thread slots per board. New replies bump a thread to the top. No replies, it sinks. When it sinks past the last page, it is deleted. Forever. The Bernstein study found that the median thread on /b/ lasted less than five minutes. Most content on 4chan exists for minutes. Some for hours. Almost none for days.

No algorithmic curation. No upvotes. No downvotes. No likes. No recommendation engine. The only way to express approval is to reply. The only way to suppress content is to ignore it. The feed is chronological, modified only by bumps. No algorithm decides what you see.

These three architectural features – anonymity, ephemerality, and the absence of algorithmic sorting – produce a radically different incentive structure than any platform-era social media. On Twitter, you build a following. On Reddit, you accumulate karma. On Facebook, you maintain a social graph. On 4chan, none of that exists. You cannot rest on reputation. You cannot leverage past success. Every post is evaluated in isolation, by people who have no idea who you are.

This produces both the best and worst of internet culture, and the mechanism is the same mechanism. The anonymity that frees people to be creative without social penalty also frees them to be monstrous without social consequence. Every attempt to explain 4chan that does not hold both of those facts simultaneously is an attempt that has already failed.

What the Architecture Produces

The meme lifecycle is a direct consequence of the architecture. Content is posted anonymously. If it resonates, it gets reposted, iterated, refined. If it does not resonate, it vanishes in minutes. There is no incentive to claim credit because there is no mechanism for claiming credit. The result is a Darwinian content engine: the funniest, sharpest, most resonant ideas survive not because their creators promoted them but because other anonymous users found them worth repeating.

This is how the rickroll happened. This is how lolcats happened. This is how rage comics, Pepe the Frog, wojak, and a thousand other meme formats originated or were refined on 4chan before migrating to the broader internet. The architecture – not the community, not the ideology, not any individual user – is what makes 4chan a meme engine. It is a machine for testing content at high speed with zero attribution.

It is also, for the same architectural reasons, a machine for producing content that no other platform would tolerate. /b/’s history includes every category of disturbing content you can imagine and several you cannot. The anonymity and ephemerality that make creative experimentation frictionless also make cruelty frictionless. This is not a paradox. This is how the mechanism works.

Why the Distinction Matters

Here is why I am insisting on this point five essays into a newsletter about internet culture: if you do not understand that 4chan is not a monolith, you cannot understand anything that happened on the internet between 2003 and now.

You cannot understand meme culture without understanding the architecture that produces memes. You cannot understand Anonymous without understanding the specific boards and threads from which it emerged. You cannot understand the radicalization pipeline without understanding that /pol/ is a specific board with a specific culture that is not representative of the other seventy boards. You cannot understand the free speech debate without understanding that the same architectural feature – anonymity – produces both the content you want to defend and the content you want to condemn.

The monolith framing lets everyone off the hook. If 4chan is just “bad,” you do not have to think about why the same architecture produces both the rickroll and gore threads. If 4chan is just “the dark web” (it is not, and it is not even close, but CNN has implied it), you do not have to explain why millions of ordinary people use it every day to discuss anime and cooking and papercraft. If 4chan is just “/b/,” you do not have to reckon with the fact that most of its boards are quieter, more civil, and more substantive than most of Twitter.

The easy story is that 4chan is a bad place. The true story is that 4chan is a machine, and like most machines, what it produces depends on what you feed it and how you use it. That is a harder story to tell. It does not fit in a headline. But it is the story that actually explains what happened.

This essay draws from Lurk More, coming fall 2026.

Next week: Alan Turing’s “Imitation Game” is not a test for machine intelligence. It is a formalized trolling problem. And recent AI research has proven it.

Source URLs