The narrative that internet culture was male-only is empirically false. Women were 40% of The WELL in 1985, 90% of AO3, and 30% of 4chan. Fandom invented tagging, content warnings, and collaborative worldbuilding before any platform. The erasure is methodological, not historical.

The Women Were Always There

In 2019, Archive of Our Own – a fanfiction platform that is ninety percent female, built and maintained entirely by volunteers, hosting over twelve million creative works across more than sixty thousand fandoms – won the Hugo Award for Best Related Work. The Hugo is science fiction’s oldest and most prestigious award. AO3 beat out professional publications backed by major media companies. The platform had been built from scratch by a community of fans, most of them women, many of them without prior coding experience, in direct response to a corporate platform purging their work. It is one of the largest collaborative creative projects in the history of the internet. It does not appear in any of the canonical histories of internet culture.

Not in a footnote. Not in an appendix. Not in a “women in tech” sidebar. Absent. Sixteen million works of fiction, a pioneering tagging taxonomy, a Hugo Award, and the largest volunteer-run literary archive on earth – none of it qualifies as internet culture, apparently. Meanwhile, the Time Magazine poll hack – in which 4chan users rigged an online survey to spell out “MARBLECAKE ALSO THE GAME” – gets its own section in every account of internet history written since 2009. The poll hack was funny. Archive of Our Own is a civilization. One of these things gets documented. Guess which one.

This is not a guilt trip about representation. The narrative that internet culture was a boys’ club that women eventually entered, usually to be harassed, is not an oversimplification. It is wrong. Empirically, demonstrably, documentably wrong. The evidence sits in plain sight – in platforms and communities and archives that anyone could have checked, if anyone writing the histories had bothered.

They did not look. Here is what they missed.

I. The Wrong Story

The story goes like this: The internet was invented by men. Populated by men. The culture that developed on it – BBSes, Usenet, IRC, forums, imageboards – was created by men, for men, in a social environment ranging from oblivious to actively hostile toward anyone who was not a young, white, technically literate male. Women arrived later, tentatively, and were greeted with harassment. Minorities arrived even later and fared worse. The story of the internet is the story of a boys’ club that the rest of the world eventually crashed, and the resulting culture war is what we are still living through.

Simple story. Also false.

Not partially false. Not “true but incomplete.” False the way “Columbus discovered America” is false – there were already people there, they had been there for thousands of years, and the only thing “discovered” was a European’s awareness of something that had existed without him. Chapter 2’s taxonomy of trolling noted, in passing, that gender dynamics in online communities were more complicated than the received narrative allowed. That was the survey-level observation. This is the correction. The internet was not a male-only space that women entered. It was a mixed space from the beginning, and the narrative of male exclusivity was constructed after the fact by people who only interviewed the men.

Here is what the record actually shows:

# canonical_internet_history.py
# What the histories describe vs. who was actually there

canonical_narrative = {
    "BBS_era": {
        "described_as": "male hackers and phone phreaks",
        "actually_included": [
            "Stacy Horn (founded Echo, 1990, 40% female)",
            "Sister Mary Elizabeth Clark (AEGIS AIDS BBS, 1985)",
            "Women sysops running boards across the country",
            "TARDIS BBS: female moderators locked the male founder out of Ladies Only sections"
        ]
    },
    "early_web": {
        "described_as": "nerds building websites about Star Trek",
        "actually_included": [
            "Elizabeth 'Jake' Feinler (created .com, .edu, .gov, .org, .net)",
            "Radia Perlman (invented Spanning Tree Protocol)",
            "Gail Ann Williams (invented professional community management, 1991)",
            "Jude Milhon (coined 'cypherpunk', co-launched first public BBS, 1973)"
        ]
    },
    "platform_era": {
        "described_as": "4chan, Reddit, Something Awful",
        "actually_included": [
            "LiveJournal (majority female, 1999-2010s)",
            "FanFiction.Net (80% female, 12M+ users)",
            "Archive of Our Own (90% female, 16.5M+ works, Hugo Award winner)",
            "Tumblr (25% LGBTQ, 'queerest place on the internet')"
        ]
    },
    "hacker_culture": {
        "described_as": "Kevin Mitnick, cDc, L0pht, DEF CON bros",
        "actually_included": [
            "Susan Thunder (pioneered social engineering, testified before U.S. Senate, 1983)",
            "Jude Milhon (cypherpunk founder, Mondo 2000 editor)",
            "Joanna Rutkowska (created Blue Pill, built Qubes OS, rewrote hardware trust)",
            "Kristin Paget (cold boot attacks, live RFID cloning, Apple platform security)",
            "Maddie Stone (Project Zero, reverse engineers zero-days being used in the wild)",
            "Parisa Tabriz (runs Chrome security, manages Project Zero)",
            "Sarah Zatko (NSA mathematician, Cyber-ITL co-founder)"
        ]
    }
}

# The gap between these two columns is not a matter of emphasis.
# It is a historiographical failure.
for era, data in canonical_narrative.items():
    erasure_rate = len(data["actually_included"]) / (
        len(data["actually_included"]) + 1  # +1 for the described_as narrative
    )
    print(f"{era}: {erasure_rate:.0%} of the story is missing from the standard account")

The pattern is older than the internet. Hypatia ran the library at Alexandria. Dion Fortune built the Society of the Inner Light. Doreen Valiente wrote the liturgy that Gerald Gardner took credit for. The Hidden Fire traced this erasure through every esoteric tradition it examined and arrived at the same conclusion every time: the women were not absent from the work. They were absent from the credit. The internet is not an exception. It is a continuation.

The question is not “when did women arrive in internet culture?” Women did not arrive. They were already there. The question is: when did the people writing the histories decide not to see them?

II. Women in the Early Internet

The WELL – the Whole Earth ‘Lectronic Link, founded in 1985 by Stewart Brand and Larry Brilliant – is the community that every internet history book cites as the ur-community, the proof of concept that online interaction could produce real culture. Howard Rheingold wrote a book about it. The New York Times profiled it. It is, in the canonical telling, a place where bearded Grateful Dead fans and Bay Area technologists invented the concept of virtual community.

Women were forty percent of The WELL’s user base.

Forty percent. At a time when women were ten percent of the online population. That number does not appear in most accounts of The WELL. It did not appear in Chapter 3’s account of the BBS era either, because that chapter followed the standard sourcing, and the standard sourcing does not mention it. What does appear is Gail Ann Williams, except she usually does not appear either. Williams came to The WELL in 1990 as part of a CompuMentor volunteer project and was hired as community manager in December 1991. Over the next twelve years, she managed and trained over two hundred volunteer conference hosts. In November 1993, she organized the first “Hosts on Hosting” meeting to professionalize online moderation skills – one of the earliest known efforts to turn community management from ad hoc volunteer work into a discipline with methods, standards, and institutional knowledge. In 1995, she changed policies to allow user-created forums. She served as Director until the 2012 sale to community members.

Gail Ann Williams invented the job of professional online community manager. The entire industry – thousands of people at every major technology company on earth – traces its lineage to work done by women like Williams in the early 1990s. Almost nobody in the field knows her name.

Three thousand miles east, Stacy Horn was doing something similar and more deliberate. In 1990, Horn – a graduate student in interactive telecommunications at NYU – founded Echo: East Coast Hang Out. A BBS, launched before the World Wide Web existed. The first virtual community to take hold in New York City. Before “Silicon Alley.” Before Friendster. Before Facebook.

Horn understood something the technology industry would take twenty years to articulate and another ten to act on: a community with only men in it is boring. “I thought it would be a cooler place if there were women there,” she said in a 1996 interview – simultaneously the least pretentious and most accurate mission statement in the history of social platforms. She offered women twelve months of free membership. Created private spaces where women could talk and report harassment. Spoke to women’s groups about the internet. Taught Unix courses out of her apartment so technical illiteracy would not be a barrier. Required real names instead of usernames – a policy decision that Facebook would implement fifteen years later and claim as an innovation.

Echo hit forty percent female membership when women were ten to fifteen percent of the internet population. Horn did not achieve this by accident or by creating a “women’s platform.” She did the work: free access, safe spaces, education, real identities, content about literature and film and culture and sex rather than just technology. She built the template for every successful social platform that followed her. None of them cite her.

Elizabeth “Jake” Feinler, meanwhile, was running the actual infrastructure. From 1972 to 1989, Feinler directed the Network Information Systems Center at Stanford Research Institute, operating the Network Information Center for the ARPANET. Her group developed the top-level domain-naming scheme that structures the internet to this day: .com, .edu, .gov, .mil, .org, .net. First Internet “yellow pages” and “white pages” servers. First WHOIS server. Every time anyone types a .com address into a browser, they are using infrastructure Feinler built. She was inducted into the Internet Hall of Fame in 2012, and of her many accomplishments, the one she said she took the most pride in was being able to hire women – at a time when other industries would not give them a chance.

Radia Perlman invented the Spanning Tree Protocol in 1985. STP prevents network loops in Ethernet – the thing that makes it possible for large computer networks to function without crashing into themselves. So fundamental that Perlman has been called “the Mother of the Internet,” a title she dislikes, not because it is inaccurate but because she considers it reductive. She is right. She is not the mother of the internet. She is one of its architects, and the maternal metaphor is itself gendered diminishment. Nobody calls Vint Cerf “the Father of the Internet” as though all he did was provide genetic material.

Chapter 4 told the story of Usenet as the first forum wars – Eternal September, kill files, the collapse of self-governing communities under the weight of AOL’s open floodgates. What Chapter 4 did not tell was who else was in those communities. The women on Usenet were harder to see, not because they were absent but because they learned to hide. Male aliases. Gender-neutral handles. Locked communities, secret mailing lists, hidden newsgroups. SAWnet – the South Asian Women’s Network – was created in direct reaction to the marginalization and harassment of women on newsgroups like soc.culture.india. The feminist newsgroups alt.feminism and soc.feminism were among the most active on the network. Susan Clerc’s 1996 work “Estrogen Brigades and ‘Big Tits’ Threads: Media Fandom Online and Off” – published in Wired Women: Gender and New Realities in Cyberspace – documented women’s participation in online fandom while it was happening. The scholarship is rarely cited in mainstream internet histories, because those histories do not consider women’s participation to be part of the main story.

On IRC, a survey of 423 chat room operators found twenty-five percent were women. Significantly older than the male operators. They developed coping strategies. They survived. They moderated communities that male users relied on without ever knowing who was keeping the lights on.

By 2000, the internet’s user population was evenly split between men and women. The narrative that women “arrived” in the social media era is off by a decade. On The WELL, Echo, and LiveJournal, women had been the plurality or majority for years before that.

III. Fandom’s Invisible Empire

Here is a number that should restructure every conversation about internet culture: as of January 2026, Archive of Our Own hosts over 16,555,000 works of fiction across more than 76,780 fandoms. Ninety percent women. Sixty-nine percent LGBTQ. Maintained by volunteers. Funded by donations. No venture capital. No advertising. A Hugo Award. Built because a corporation destroyed the previous platform and the women who used it said “never again” and meant it.

The backstory is Strikethrough. Chapter 8 told the story of LiveJournal’s death as a platform. Here is what died with it.

On May 29, 2007, LiveJournal – then owned by Six Apart – suspended approximately five hundred accounts without notice. A conservative advocacy group called Warriors for Innocence had compiled a list of “problematic” journals and sent it to LiveJournal’s trust and safety team, who mass-deleted without investigation. The suspended accounts included pornish_pixies (a popular Harry Potter fanfiction community), literary journals, RPG blogs, and – in a twist that would be darkly funny if it were not also cruel – survivor support groups for sexual abuse victims, flagged because they discussed abuse. The only notification: a strikethrough through their usernames. Hence: Strikethrough.

Within hours, fandom_counts had ten thousand members. Within days, thirty thousand. The response was not a hashtag campaign or an angry blog post. It was infrastructure. Women who had been writing fanfiction and building communities for years redirected those organizational skills toward platform migration and, ultimately, platform creation. Immediate exodus: GreatestJournal and InsaneJournal. Long-term response: more ambitious.

In May 2007, the fanfiction author astolat – Naomi Novik, who had already published the bestselling Temeraire fantasy series under her legal name – suggested that fans needed “an archive of their own.” Not a corporate platform that could be pressured by advocacy groups. Not a commercial service that could change its terms of service on a whim. Fan-owned, fan-operated, no advertisements, no investors, answerable to nobody but the community. Within months, the Organization for Transformative Works was incorporated. By 2008, AO3 was in development. By 2009, open beta.

What the OTW built was not just an archive. It was an ecosystem:

Archive of Our Own: 16.5 million works and counting
Fanlore: A fan-run wiki documenting the history of fandom itself
Transformative Works and Cultures: A peer-reviewed academic journal
Legal advocacy: Active defense of fan creators under fair use doctrine

But the technical achievement that deserves the most attention is AO3’s tagging system. Information scientists call it a “curated folksonomy”: users create and apply any tag they want, and a team of approximately 160 volunteer “tag wranglers” links synonymous tags, maintains navigability, and keeps the system searchable at scale. How do you let millions of users categorize content without the taxonomy becoming unusable? Commercial platforms spend millions failing to solve this. AO3 solved it with volunteers.

INNOVATION_TIMELINE: What Fandom Built First
================================================

Content warnings / Trigger warnings:
  Fandom:    1990s (Usenet-era mailing lists)
  Industry:  2010s (Tumblr, then Twitter, then everyone)

Collaborative tagging taxonomies:
  Fandom:    2008 (AO3's curated folksonomy)
  Industry:  2014-2020 (various commercial platforms, still worse)

Community content moderation at scale:
  Fandom:    2000s (LiveJournal community moderators)
  Industry:  2010s (Reddit mods, Facebook groups)

User-generated ratings systems (G through Explicit):
  Fandom:    1990s (FanFiction.Net rating tiers)
  Industry:  2010s (app store ratings, Netflix content advisories)

"Don't Like, Don't Read" as community norm:
  Fandom:    1990s
  Industry:  Still trying to figure this one out

Platform migration as collective action:
  Fandom:    2007 (Strikethrough exodus)
  Industry:  2023 (Twitter/X exodus to Bluesky)

Fan-funded nonprofit infrastructure:
  Fandom:    2007 (OTW / AO3)
  Industry:  2023 (Wikipedia still begging)

Volunteer-run open-source platform w/ 75% female developers:
  Fandom:    2009 (Dreamwidth)
  Industry:  Still at approximately 1.5% female (OSCON 2009 data)

Dreamwidth deserves its own paragraph, because the number is too important to bury. Dreamwidth – a journal platform founded in 2008 by ex-LiveJournal staff Denise Paolucci and Mark Smith, built to serve the community LiveJournal had betrayed – had a development team that was seventy-five percent female. The open-source software community as a whole was approximately 1.5 percent female at the time, per OSCON 2009. Dreamwidth did not have a diversity initiative. Did not have a women-in-tech pipeline program. It had a community of women who needed a platform and built one.

The fanfiction communities were enormous long before AO3. FanFiction.Net, founded in 1998, had over 118,000 registered users by 2002 – one-third under eighteen, eighty percent female. By 2026, over twelve million registered. A 2015 survey confirmed the demographics had held steady: seventy-nine percent women, seventy percent between thirteen and twenty-nine. Not niche numbers. The numbers of a major cultural institution that has operated continuously for over twenty-five years.

The creative lineage goes back further. Joan Winston co-organized the first Star Trek fan convention in January 1972 with Elyse Rosenstein. Not just the first convention focused on a single franchise – the biggest science fiction convention held to that date, 3,200 attendees. Modern fandom convention culture, now a multi-billion-dollar global industry, traces directly to a convention organized by women in 1972. Jacqueline Lichtenberg, Sondra Marshak, and Winston published Star Trek Lives! in 1975, documenting a creative community already thriving before most of the men who would later “discover” internet culture were born.

Slash fiction – romantic and erotic fiction pairing male characters, originated by women in the 1960s Star Trek fandom with Kirk/Spock as the prototype – is not a curiosity or a footnote. It is a literary tradition with its own conventions, critique, quality standards, and cultural significance, a space where women and LGBTQ individuals explored identity, sexuality, and power dynamics that mainstream culture would not touch. By the 2013 AO3 census, slash fiction authors were ninety percent women and sixty-nine percent LGBTQ. The body of work is orders of magnitude larger than the creative output of most male-dominated internet spaces. Almost completely absent from mainstream internet culture histories, because the people writing those histories did not consider fanfiction to be “internet culture.” It was. It was more of it than most of what they did cover.

IV. Women in Hacker Culture

The person who coined the word “cypherpunk” was a woman.

That sentence should be enough to restructure the conversation, but it never is, so here is the rest. Jude Milhon – who went by St. Jude, because she had a sense of humor about lost causes – was a self-taught programmer, civil rights advocate, founding member of the cypherpunks, co-launcher of the Community Memory Project in 1973 (the first public computerized bulletin board system, predating personal computers), and co-founder and editor of Mondo 2000, the magazine that defined 1990s cyberpunk aesthetics. Before any of that, she marched from Selma to Montgomery. Her motto was “Girls need modems,” and when she said it, the statement was a provocation, not a platitude. Milhon did not need a pipeline program. She needed a modem. She got one. She used it to help build the philosophical and technical foundation of digital privacy.

The Community Memory Project, which Milhon co-launched in Berkeley, was a public-access computer terminal in a record store. Walk in, post a message, read a message. A BBS before the acronym existed, a social network before the concept had a name, built by a team that included a woman who would go on to coin the term for an entire movement in cryptography. The cypherpunks – the loosely organized group of cryptographers, programmers, and privacy advocates whose work led to PGP, Tor, Bitcoin, and the broader architecture of encrypted communication – got their name from Jude Milhon. She died in 2003. The movement she named is now a trillion-dollar infrastructure layer. Her Wikipedia article is shorter than most entries for minor cryptocurrency tokens.

Susan Thunder – born Susan Headley – was a phone phreaker and computer hacker in the late 1970s and early 1980s, one of the only women among the early phone phreaks and one of the earliest known female hackers. Her specialty was what she called “psychological subversion” – now known as social engineering, the art of manipulating people into revealing information or granting access. Social engineering remains the single most effective attack vector in cybersecurity. Every major breach that makes the news – every one – involves a social engineering component. Susan Thunder was doing it in 1980, before it had a name.

She testified before the U.S. Senate on hacker capabilities in 1983. Presented at DEF CON in 1995. Claire L. Evans described her as “a mythological figure in hacking history” and “the great lost female hacker of the 1980s.” She is “lost” not because she disappeared but because the historians did not look for her.

The list continues, and it is long enough to constitute its own argument:

Joanna Rutkowska presented Blue Pill at Black Hat 2006 – a hardware virtualization rootkit that could invisibly compromise a running operating system by migrating it into a virtual machine underneath itself. The name was a Matrix reference. The implications were not theoretical. She demonstrated that the entire trust model of modern computing – the assumption that the operating system controls the hardware – was wrong. The hardware could lie to the OS, and the OS would never know. She founded Invisible Things Lab, published foundational research on firmware and hypervisor security, and then did something almost nobody in security does: she built the fix. Qubes OS, released in 2012, is a desktop operating system built on the assumption that every component will be compromised. It isolates everything – networking, USB, applications – into separate virtual machines so that compromising one does not compromise the rest. Edward Snowden endorsed it. The NSA reportedly uses it. The woman who proved your computer could not be trusted built the only operating system designed to work anyway.

Kristin Paget was part of the Princeton research team that demonstrated the cold boot attack in 2008 – the discovery that encryption keys persist in RAM after power-off, meaning a laptop’s full-disk encryption could be defeated by freezing the memory chips and reading them in another machine. The entire full-disk encryption industry had to redesign its assumptions. At Shmoocon 2012, she demonstrated live RFID credit card cloning with a $50 device she built herself – swiping contactless credit card numbers out of the air from audience members’ wallets in real time. The contactless payment industry scrambled. She went on to Apple’s platform security team. She transitioned publicly while already one of the most recognized hardware hackers in the community, and the community’s reaction was largely “cool, anyway, about that firmware exploit.”

Maddie Stone joined Google’s Project Zero in 2019 after years at FLARE, FireEye’s elite reverse engineering team, where she tore apart malware for a living. At Project Zero she tracks zero-day vulnerabilities being actively exploited in the wild – not theoretical bugs, not CTF puzzles, but the actual weapons being used against real targets right now. Her annual “0-day In the Wild” reports are the industry’s definitive accounting of who is getting hacked and how. She reverse-engineers Android exploits, iOS chains, browser zero-days. She has presented at Black Hat, DEF CON, USENIX, and every other venue that matters. She is one of the best reverse engineers working today, in a field where “best” is measured by whether you can take apart a weapon you have never seen before and explain exactly how it kills.

Parisa Tabriz joined Google security as a college intern, went full-time in 2007, took over Chrome security in 2013, now manages Project Zero – the elite unit that finds and fixes zero-day vulnerabilities across all software, not just Google’s. She drove HTTPS adoption from under fifty percent of Chrome traffic in 2015 to over ninety percent by 2019. Her business card says “Security Princess” instead of “information security engineer” because it “sounded less boring.” Behind the title: one of the most powerful positions in internet security.

Sarah Zatko, former NSA mathematician, married Mudge (Peiter Zatko, one of the most famous hackers in history) in 2007. The profiles of Mudge mention her as his wife. Her algorithms power a novel software security scoring system she and Mudge co-founded at the request of the White House. The algorithms are hers. The system is hers. The profiles call her “Mudge’s wife.”

V. Women Who Troll

Thirty percent of 4chan’s users are women. Not an estimate from an external survey. Not an academic’s guess. 4chan’s own advertising demographic data. Hundreds of millions of monthly page views; thirty percent means millions of women. They are invisible for the same reason everyone on 4chan is invisible: anonymity is the default. But invisibility has been read as absence, and absence has been treated as proof the platform is exclusively male. One in three users is a woman. Nobody writing about 4chan mentions this because nobody writing about 4chan checked.

One in three. Let that number do its work for a moment.

The rule that made those women invisible – “there are no girls on the internet” – had a genuine philosophical kernel, worth acknowledging before we complicate it. The idea was meritocratic: in anonymous spaces, ideas are judged on their content, not their author’s identity. No gender, no race, no credentials, no appeals to authority. Just the argument. As a philosophical position, this was not nothing. In its best formulation, a radical egalitarianism that most physical institutions have never achieved. But the January 2026 DOJ release of the Epstein files adds a dimension that even the generous reading has to accommodate. If shadow actors with documented interest in platform manipulation had access to the architects of online infrastructure – Boris Nikolic to Jeffrey Epstein: “the potential for manipulation is huge” – then the anonymity that functioned as genuine philosophy was also, simultaneously, the structural precondition for invisible influence operations. You cannot astroturf a space where identities are verified. You can astroturf the hell out of a space where identity is philosophically prohibited. “There are no girls on the internet” enforced both a meritocratic ideal AND the conditions under which manufactured consensus is easiest to conduct undetected. The meritocracy of anonymity was real. It was also, potentially, cultivated infrastructure. This is documented evidence, not conspiracy theory – the files are public, the communications are on the record, and the question of whether anonymous spaces were only organic philosophy or also useful architecture for manipulation is one that the evidence now forces us to ask. (Cross-reference Chapter 6 for the broader pattern of how organic online norms can serve interests their participants never intended.)

Academic research on 4chan and gender has a specific, revealing bias: it overwhelmingly studies how women are discussed on 4chan rather than women’s direct participation. A paper from Southampton Solent University – “‘Be Me, 24yo Femanon’: Greentext Stories as Feminine Space on 4chan” – is one of the rare exceptions, examining women’s actual participation through the “femanon” identity. The default mode of academic inquiry treats 4chan as a space that acts upon women rather than a space in which women act. The research equivalent of documenting a city by interviewing only the tourists and ignoring the residents.

The broader data is even more instructive. Demos analyzed ten thousand tweets containing misogynistic language and found half were posted by women. Evita March and colleagues at Federation University Australia found women and men equally likely to troll other users on Tinder (2017, Computers in Human Behavior). A 2014 analysis of “slut” and “whore” on Twitter, covered by Rolling Stone, found women used those terms as often as men – in some periods, more often. The data does not show that women are trolling’s victims. It shows women are trolling’s full participants – creators, targets, and perpetrators in roughly the same proportions as men.

The dominant media narrative positions women exclusively as victims of internet culture: harassed, threatened, driven offline. The harassment is real, documented, serious. But the narrative is wrong about everything else. It denies women agency. It treats half the population as a passive surface on which internet culture acts, rather than as active participants who shape, create, and yes, troll. “Women are only victims” is not a feminist position. It is a paternalist one, dressed up in concern.

Women do not hack. Women do not code. Women do not shitpost. Women do not troll. In each case, the claim is false, the evidence is available, and the people making the claim did not look.

VI. People of Color Online

The narrative of internet culture as exclusively white is as wrong as the narrative of it as exclusively male, and it fails for the same reasons. Wrong places. Wrong people. Limited sample mistaken for the whole population.

AfroNet was an indie BBS network that sprang up in the late 1980s – not the late 1990s, not the 2000s, the late 1980s – a FidoNet-based system serving as a hub for message exchange among “likeminded Africans and African Americans in the US and Canada.” West Coast and East Coast sysops connected their communities. Localized their systems to suit their neighborhoods. Many ran their boards late at night after working day jobs. Not hobbyists on the margins of BBS culture. They were BBS culture – doing exactly what every other sysop was doing, building community infrastructure out of modems and telephone lines – in a parallel network that the white sysops mostly did not know existed and the historians mostly did not document.

The key figures have been excavated by Charlton McIlwain in Black Software: The Internet and Racial Justice, From the Afronet to Black Lives Matter (2019), one of the few books that treats Black digital culture as a continuous tradition rather than a reaction to white platforms. Arthur McGee created the first open-source list of Black online communities. William Murrell was one of the first Black engineers at IBM. Kamal Al-Monsour founded AfroLink software in 1988. E. David Ellington and Malcolm CasSelle created NetNoir, a portal for Afrocentric content that launched on Juneteenth 1995 on AOL, which invested half a million dollars.

“Once AOL added NetNoir, it became essentially the first cultural commodity bringing folks to the Web… Black content and Black cultural production was really at the core of what became the internet as we know it today.” Not an activist’s claim. A finding from McIlwain’s historical research. Black cultural production was not peripheral to the commercial internet’s growth. It was central to it.

WorldStarHipHop, founded in 2005 by Lee “Q” O’Denat from Hollis, Queens, was averaging 1.2 million unique visitors per day by 2011. Before Vine. Before Instagram. Before TikTok. It “afforded agency to black youth not allowed anywhere else on the internet” and fundamentally changed how hip-hop culture interfaced with technology – paving the way for Instagram Live, TikTok, and YouTube’s creator economy. “Distinguished and cohered by blackness”: a Black space, created by Black people, for Black cultural expression.

Black Twitter, which scholar Andre Brock dates to a 2008 piece by blogger Anil Dash, is not a subculture within a white-default internet. It is a cultural force that frequently drove broader internet culture – trending hashtags, memes that shaped internet humor, social justice movements (#BlackLivesMatter, #OscarsSoWhite), political organizing, critical pedagogy. The standard narrative treats Black online communities as derivative. They were often the source. Jack Dorsey called Black Twitter “such a powerful force,” which was accurate and also a remarkable admission from the CEO of a platform that consistently failed to protect its Black users from harassment.

The digital divide is real. Black adults are less likely than white adults to own traditional computers (sixty-nine versus eighty percent) or have broadband (seventy-one versus eighty percent). One in three Black, Latino, and American Indian/Alaska Native households lack high-speed home internet. But the digital divide is about infrastructure access, not cultural participation. Conflating the two – treating an access gap as an absence gap – has erased decades of Black digital culture. Black communities did not wait for the digital divide to close. They participated with whatever access they had, built their own spaces, created their own platforms, and produced culture that the broader internet consumed without crediting.

VII. LGBTQ+ Spaces: The Internet as Lifeline

In 1983 – a decade before the World Wide Web, two decades before social media, in the middle of the AIDS crisis, during an era when homosexuality was still classified as a mental illness by significant portions of the medical establishment – a Boston programmer named Steve Dyer created net.motss on Usenet. “Members of the same sex.” Deliberately coded so the newsgroup’s purpose was not immediately obvious to hostile outsiders. By 1984, a thriving community. For some, the first place they felt comfortable coming out. The first in-person meetup, motss.con, was held in San Francisco in 1988. The newsgroup still exists.

Earliest documented LGB-specific online community. Forty years ago. For queer people in 1983 – people who could lose their jobs, their housing, their families, and in some jurisdictions their freedom for being open about their sexuality – an anonymous text-based discussion group was not a convenience. It was a lifeline. The internet did not create queer community. It gave queer people a way to find each other that did not require walking into a physical space that might not be safe.

By 1999, an estimated one-third of AOL’s chat rooms were LGBTQ – a number so large that AOL was widely known as “GayOL” within the queer community. Not a marginal population using a niche service. A third of the chat activity on the largest internet service provider in the United States. If you were not queer, you probably did not know it was happening.

LiveJournal’s majority-female platform overlapped heavily with queer communities. Tumblr, launched in 2007, went further: users 193 percent more likely to be LGBTQ compared to other platforms. One in four identified as LGBTQ. “The queerest place on the internet” – not marketing, demographics. The reblog feature made identity construction participatory rather than performative: users built identities through accumulation, sharing, and remixing rather than static profile pages. An Australian survey of LGBTQ youth found they used Tumblr significantly more than their peers, reporting that it broadened their understanding of sexuality and gender.

When Tumblr banned all NSFW content in December 2018 – removed from the Apple App Store over child safety concerns – the ban disproportionately hit LGBTQ users and sex workers. Chapter 8’s roll call of dead platforms documented the corporate patterns; same pattern, different mask. Avery Dame-Griff’s 2019 research in Feminist Media Studies described Tumblr as “a trans technology” – a platform whose specific design choices had enabled trans community formation in ways other platforms did not. Traffic dropped to a third. A corporate decision made to satisfy Apple’s App Store policies destroyed a community infrastructure that had taken a decade to build. Strikethrough again – a corporate platform making a unilateral decision that devastates the community most dependent on it – except this time millions of LGBTQ users instead of thousands of fanfiction writers.

The linguistic influence of queer communities on internet culture is so pervasive that most people using the vocabulary do not know where it came from. “Yas queen” – 1980s New York City ball culture, where drag balls were a haven for young queer people of color, mainstreamed via a 2013 viral video of a fan’s reaction to Lady Gaga. “Shade” – ball culture, immortalized in the 1990 documentary Paris Is Burning. “Tea,” “slay,” “werk” – all ball culture, all created by Black and Latinx queer communities decades before the internet existed. The internet was the distribution mechanism. The culture was created by the most marginalized communities in America.

The dominant tone of the internet – ironic, exaggerated, self-aware, performative – has deep roots in what linguists call “verbal camp,” a citational style associated with queer discourse using echoic reference, irony, and exaggeration to create bonds between participants. When the entire internet communicates in a register of arch self-awareness, it is using a style queer communities developed as a survival mechanism. The culture flowed upward, from the most marginalized to the most mainstream, and the mainstream does not know it is speaking a borrowed language.

VIII. The Erasure Mechanism

The question that matters is not “were women and minorities present in internet culture?” – this chapter has answered that with names, dates, platforms, and statistics. The question is: why did the narrative erase them? The answer is not a conspiracy. It is something more banal and harder to fix: a methodological failure.

The journalists, the academics, the documentarians followed a reliable process. Identify the most visible participants. Interview them. Write about them. The problem is that visibility on the internet is not random. It correlates with specific behaviors: posting publicly, posting frequently, posting provocatively, posting in spaces journalists already know about. These behaviors correlate with demographics: the visible, loud, provocative participants whose names and handles could be traced, profiled, and quoted were disproportionately young, white, and male. Not because those spaces were exclusively young, white, and male – 4chan is thirty percent female, after all – but because visibility is not representativeness, and nobody checked the advertising demographics.

# the_erasure_algorithm.py
# How internet history gets written (simplified)

def write_internet_history():
    """Standard methodology for documenting internet culture.
    Bug: selection bias at every step. Feature: produces a clean narrative."""

    # Step 1: Choose which spaces to study
    spaces_studied = ["4chan", "Reddit", "Something Awful", "hacker_cons"]
    spaces_ignored = ["LiveJournal", "FanFiction.Net", "AO3", "Tumblr",
                      "AfroNet", "NetNoir", "Dreamwidth", "soc.motss"]

    # Step 2: Within chosen spaces, find the loudest participants
    participants = get_most_visible_users(spaces_studied)
    # Note: visibility != representativeness
    # Note: anonymity makes women invisible, not absent
    # Note: nobody checks the advertising demographics

    # Step 3: Interview them
    interviews = [talk_to(p) for p in participants if p.is_reachable]
    # Note: anonymous users are not reachable
    # Note: women using male handles are not identifiable
    # Note: Black BBSes are on a different network

    # Step 4: Write the narrative
    narrative = synthesize(interviews)
    # Output: "Internet culture was built by young white men"
    # Confidence level: high (all the interviews confirm it)
    # Accuracy: low (the interviews are a biased sample)

    # Step 5: Future researchers cite this narrative
    # Step 6: The narrative becomes the "accepted history"
    # Step 7: Anyone correcting it is "pushing an agenda"

    return narrative  # A history of who was loudest, not who was there

The bias compounds at each step. Academics study 4chan and Reddit – thousands of papers – and under-study LiveJournal and AO3. When fandom is studied, it gets filed under “media studies” or “fan studies,” not “internet culture,” as though women writing fanfiction on the internet are doing something categorically different from men posting on forums on the internet. The academic literature on “internet culture” is actually a literature on male-dominated internet culture, with female-dominated spaces segregated into a different discipline. A future researcher surveying the field would conclude – correctly, based on the available evidence – that internet culture was predominantly male. The conclusion is wrong. The evidence is biased. The bias was introduced at the beginning, when someone decided that 4chan was internet culture and LiveJournal was not. Sixteen million works of fiction lost to a filing error.

danah boyd documented a parallel version of this process in her 2011 paper “White Flight in Networked Publics?” Based on 103 formal interviews and 2,000 hours of observation across seventeen states, boyd found that white, affluent teens migrated from MySpace to Facebook while Black and Hispanic teens remained on MySpace. A fourteen-year-old white girl described MySpace as “ghetto.” Parents pushed their children toward Facebook based on class and race coding. The internet does not create a post-racial, post-gender utopia. It reproduces existing hierarchies. Including the hierarchy of whose stories get told.

The erasure is not malicious. That is what makes it so durable. Nobody sat down and said “let’s write women out of internet history.” The methods used to document internet history – interview the visible, study the loud, write about the spaces you already know – systematically excluded participants who were less visible, less loud, and in spaces the historians had not heard of. A feature of the methodology, not a choice of the methodologist. You cannot fix a bias you do not recognize as a bias, and the people who wrote the histories genuinely believed they were documenting the whole story. They were documenting the part that was easy to see. They treated it as the whole.

IX. The Corrected History

Put the erased participants back into the story and the history of internet culture does not change in emphasis. It changes in structure. Not the same story with more characters. A different story.

The original story is linear: male technologists build the infrastructure, male nerds build the culture, women and minorities arrive later and are either assimilated or harassed. The corrected story is ecological: multiple communities with different demographics, different norms, different creative outputs, developing simultaneously on the same infrastructure, often without knowing about each other. The men on 4chan did not know about the women on LiveJournal. The white hackers at DEF CON did not know about the Black sysops on AfroNet. The straight kids on IRC did not know about the queer adults on soc.motss. Each community thought it was the internet. None of them were. All of them were.

The corrected history looks like this: In 1973, a woman co-launches the first public computer network. In 1983, a queer community creates one of the first online social spaces. In the late 1980s, Black sysops build a parallel BBS network. In 1985, a woman invents the protocol that makes large networks possible, and another woman runs the system that will become .com. In 1990, a woman founds what may be the first social network and achieves gender parity a decade before the general internet does. In the late 1990s, women build the largest collaborative fiction archive on the internet. In the 2000s, they build a nonprofit to protect it, a tagging system that outperforms commercial alternatives, and an open-source platform with female developer representation fifty times higher than the industry average. In 2007, the queerest platform on the internet launches. In 2019, the fanfiction archive wins a Hugo Award.

At no point in this timeline is internet culture a boys’ club. At every point, there are women, people of color, and LGBTQ individuals building, creating, moderating, coding, and participating. The boys’ club narrative is not a story about the internet. It is a story about who got interviewed.

Get the history wrong, get the present wrong. The Fires of History built a taxonomy of trolls stretching from Socrates to Anonymous and got the types right – but even that taxonomy inherited the source bias, because the canonical examples were overwhelmingly male. Socrates, Voltaire, Swift, Andy Kaufman, moot. The women who trolled – Jude Milhon coining “cypherpunk” as a provocation, the Guerrilla Girls trolling the art world with gorilla masks, the fandom communities inventing content warnings and tagging systems that the commercial internet spent a decade failing to replicate – were there all along. The taxonomy was correct about the structure and wrong about the population. If internet culture was a boys’ club that women disrupted, the solution is to make the boys’ club more welcoming – diversity initiatives, codes of conduct, pipeline programs. If internet culture was a diverse ecosystem that got described as a boys’ club, the solution is different: stop describing it wrong. Document what is actually there. Stop filing women’s internet communities under “fandom” and men’s internet communities under “internet culture.” Stop studying 4chan and ignoring AO3. Stop treating visibility as a proxy for presence. Every community documented in this book – from the BBS sysops of Chapter 3 through Something Awful’s goons in Chapter 5 to the subreddit builders of Chapter 7 – had women in it. The correction is not additive. It is structural.

The women were always there. So were the Black sysops, the queer Usenet posters, the Latina drag queens whose vocabulary the entire internet now uses without attribution. The history of internet culture does not look like a boys’ club that women and minorities eventually entered. It looks like a diverse ecosystem described as a boys’ club by people who only talked to the boys.

The correction is not generosity. It is accounting. The books do not balance. They never have.

Prefer RSS? Subscribe here.