Classification: SOCIAL MEDIA CONTROL | Confidence: WHITE-COLLAR FRAUD — CONVICTED

Elizabeth Holmes was nineteen years old in 2003 when she founded Theranos. She dropped out of Stanford in her sophomore year. She wore a black turtleneck every day for the next fifteen years. She claimed her company’s blood-testing device could run 200+ diagnostic tests from a single finger-prick of blood. She raised $700 million from investors including Larry Ellison, the Walton family, and Rupert Murdoch. She was named the youngest self-made female billionaire by Forbes in 2014 at age 30, with a paper net worth of $4.5 billion. She was on the covers of Forbes, Fortune, T: The New York Times Style Magazine, and Inc.. She was compared to Steve Jobs. She sat on boards. She gave TED talks. She was, by every external measure, the most successful woman in American startups.

None of it worked. The Theranos blood-testing device never produced a single reliable diagnostic result. The device was, by the company’s own internal testing, less accurate than the conventional blood draws it claimed to replace. The 200+ tests it claimed to run could not actually be run. The few tests that could be run on the device returned results that were systematically wrong, often by factors of 10x or more. The company hid the bad results from patients, from doctors, from regulators, and from investors. The patients received diagnoses based on fabricated data. By the time the fraud was exposed in 2015-2016, more than 850,000 blood test results had been issued — the majority of which had been silently voided or modified by the company’s internal review process. Holmes and her co-conspirator Sunny Balwani were indicted on federal fraud charges in 2018. They were convicted in 2022. Holmes was sentenced to 11 years in federal prison in November 2022. Balwani received 13 years. The appeal is pending. The company is bankrupt.

The Myth of the Visionary

The story of Theranos has been told as a story about a visionary who overreached. The standard narrative is that Holmes had a real breakthrough idea, ran out of runway, lied about the science to keep the company alive, and ended up destroying what she had built. The narrative is wrong in one specific way: Holmes did not have a real breakthrough idea. The technology she claimed to have built did not exist when she founded Theranos. It did not exist after fifteen years of operation. It did not exist when the company was exposed. The technology was a claim that was never verified. The verification that should have happened during the company’s entire operating history was avoided, deferred, or falsified.

The visionary narrative is the narrative Holmes herself constructed. In her Wall Street Journal profile, her Fortune cover story, her TED talks, her appearances on the Today Show, the story she told was: I dropped out of Stanford, I had a vision, I built it, it works, it will change the world. The story was consistent. The story was also unfalsifiable. The device worked. You could not examine the device. You could not test the device. You could not publish the data. The non-disclosure agreements were airtight. The employees were silenced. The board members were barred from speaking. The partnerships with Walgreens and Safeway were not partnerships — they were purchases of the right to deploy the device under non-disclosure. Walgreens and Safeway did not have access to the underlying data. They were customers, not auditors. The story was controlled. The story was performance.

The John Carreyrou Investigation

The Theranos fraud was exposed, in the end, by a single investigative reporter. John Carreyrou of the Wall Street Journal published the first major piece on October 15, 2015, headlined “Hot Startup Theranos Has Struggled with Its Blood-Test Technology.” The piece was based on confidential interviews with former Theranos employees, confidential internal company documents, and confidential interviews with patients whose results had been altered. The piece documented that the device did not work, that the tests were inaccurate, that the company was hiding the failures. The piece ended Theranos.

Carreyrou’s reporting continued for another year. His second piece, in November 2015, revealed that Walgreens had suspended the use of Theranos devices in their stores. His third piece, in January 2016, revealed that federal regulators had found “deficient practices” at Theranos labs. His fourth piece, in March 2016, revealed that the device had never been independently validated. Each piece triggered a corresponding escalation: company denials, lawsuits, investor lawsuits, congressional inquiries. By mid-2016, Theranos was under federal investigation. By 2018, Holmes and Balwani were indicted. The investigation started with one reporter asking one employee one question: is the device accurate? The employee said no. The reporter published. The company collapsed.

Carreyrou’s book, Bad Blood: Secrets and Lies in a Silicon Valley Startup, published in 2018, became the definitive account. The book documents in detail how the fraud was maintained — the falsified test results, the doctored validation studies, the silenced employees, the intimidation of whistleblowers, the falsified board minutes, the fraudulent regulatory submissions. The book documents that the fraud was not a single act of deception but a fifteen-year systematic project of deception. The fraud was not a bug. The fraud was the business.

The Performance of Innovation

What made Theranos notable was not the fraud itself. White-collar fraud is common. What made Theranos notable was the performance of innovation. Holmes did not invent a technology. She invented the aesthetic of technology. The black turtleneck. The affected baritone voice. The dropped-out-of-Stanford narrative. The Jobs comparison. The board of former cabinet secretaries and admirals. The Steve Jobs-style product demos. The carefully curated walk-through videos. The TED talks. The 60 Minutes feature. The investor pitches that included dramatic pauses and eye contact. The aesthetic was not incidental to the fraud. The aesthetic was the product.

The aesthetic generated the $700 million. The aesthetic generated the partnerships with Walgreens and Safeway. The aesthetic generated the FDA filings, the clinical trials, the published papers, the academic collaborations. The aesthetic generated everything that made Theranos look like a real company with real technology. The aesthetic did not generate the technology. The technology did not exist. The aesthetic stood in for the technology. The aesthetic was good enough to generate the capital. The capital bought time. The time did not generate the technology. The technology never generated.

The Whistleblowers

The Theranos story is also a story about whistleblowers. Tyler Shultz, the grandson of former Secretary of State George Shultz and a Theranos employee, was the first to bring his concerns to Carreyrou. He had been working in the lab and had seen the test results. The results did not match the company’s claims. He had raised his concerns internally. He had been threatened with litigation. He had been told that his continued employment required silence. He chose to speak to a reporter anyway. He did so anonymously. His identity was not protected. He was threatened with lawsuits by Theranos, by his own family, by the company’s lawyers. He was sued personally. He spent several years defending himself against the company’s legal retaliation.

The other whistleblowers — Alan Beam, Rochelle Gibbons, and others — faced similar pressure. Some lost their careers in biotech. Some lost their marriages. Some lost their savings defending against Theranos lawsuits. The company’s response to internal whistleblowers was not to investigate the concerns. The company’s response was to silence the whistleblowers. The response was effective enough that the fraud lasted fifteen years. The response was not effective enough to outlast one reporter who would not be silenced.

The Pattern of Uncritical Belief

The question the Theranos story forces is: why did anyone believe her? The investors who gave her $700 million. The journalists who put her on their covers. The regulators who approved her labs. The board members who sat on her board. The pharmacies that deployed her devices. The patients who got their blood drawn. The venture capital firms that funded her rounds. The wealth managers who valued her company. The Stanford professors who mentored her. The list of institutions that extended credibility to Theranos without independent verification is essentially the list of the entire American institutional infrastructure for evaluating technology.

The pattern is not specific to Theranos. The pattern is the pattern of uncritical belief in aesthetic performance. The aesthetic of a black turtleneck. The aesthetic of a TED talk. The aesthetic of a board of former admirals. The aesthetic is enough for the capital. The capital is enough for the time. The time is enough for the fraud. The fraud collapses. The next fraud starts. The aesthetic remains. The pattern is the same pattern. The simulation has been running it for thirty years. The pattern has not changed. The aesthetic still works.

⚠ PATTERN RECOGNITION

Elizabeth Holmes did not invent a technology. She invented the aesthetic of technology. The aesthetic generated $700 million in capital. The capital bought fifteen years of time. The time did not generate the technology. The technology never generated. The aesthetic stood in for the technology. The aesthetic was enough to generate the capital. The aesthetic was not enough to generate the device. The device never worked. The patients received diagnoses based on fabricated data. The investors received returns based on fabricated technology. The fraud was maintained for fifteen years by silencing the people who could have exposed it. The fraud collapsed when one reporter asked one employee one question. The fraud collapsed in a year. The fraud was built in fifteen. The aesthetic is still working. The pattern is still running.

SOURCES

Sources & Further Reading

Classification: SOCIAL MEDIA CONTROL | Confidence: DOCUMENTED — INSIDER TESTIMONY

The average American checks their phone 96 times per day. Teenagers average 200+ times. This is not progress. It is a control system upgrade — the most successful behavioral modification campaign in human history, deployed at planetary scale, optimized daily by machine learning.

“Your phone is a slot machine in your pocket. Slot machines were regulated. Your phone is not.” — Tristan Harris, former Google design ethicist

The Attention Economy

You are not the customer. You are the product. Every feature on your phone — infinite scroll, push notifications, likes, streaks, autoplay — maximizes engagement, not well-being. The business model is: capture attention, sell it to advertisers. The longer you stay, the more ads you see, the more revenue generated.

The optimization is continuous. Every interaction is fed back to a model. Every scroll, every pause, every rewatch is data. The algorithm learns what keeps you engaged and shows you more of it. The system becomes a personalized attention trap, optimized for the specific neurochemistry of each user.

Surveillance Capitalism — Zuboff’s Framework

In 2019, Harvard Business School professor Shoshana Zuboff published The Age of Surveillance Capitalism, a 691-page anatomy of the system you are inside. Her argument: the tech industry did not stumble into the attention economy. It designed it. The key terms:

Zuboff’s diagnosis: this is private surveillance by other means, and it is incompatible with democracy. A democracy cannot survive when one set of private actors has more behavioral data on its citizens than the state does — and uses that data to shape those citizens’ choices while pretending only to respond to them. The platform’s defense — “we only optimize for what users want” — is the same defense every behaviorist lab has offered since Skinner: the experimenter didn’t make the rat press the lever; the rat wanted to press the lever. The experimenter only arranged the box.

The pattern, in lethometric terms: the same control architecture that religion ran for two millennia — invisible authority, arbitrary rules, behavioral measurement, public confession, excommunication — has been refactored onto a digital substrate. The form is unrecognizable. The protocol is identical. The deeper thread is in our archive entry on Religion as the Original Simulation — and the same architecture, refactored onto the search box, is dissected in our I Remember When Google Only Gave You Ten Links archive.

Algorithmic Radicalization and the Election Pipeline

The political cost of this system is no longer theoretical. Between 2014 and 2018, Facebook played a documented role in the Rohingya genocide in Myanmar. A 2018 Reuters investigation found that Facebook’s algorithms amplified anti-Rohingya hate speech; the UN’s Fact-Finding Mission later called the platform’s role “a slow-burning genocide.” More than 10,000 people died; more than 700,000 were displaced. Internal documents released by Frances Haugen in 2021 confirmed that Facebook’s own researchers had flagged Myanmar as a priority risk in 2015 — and were repeatedly overruled by executives chasing growth metrics.

The same pipeline fired in the 2016 US presidential election. Russia’s Internet Research Agency ran targeted disinformation campaigns on Facebook and Instagram, reaching an estimated 126 million Americans. The Mueller indictment (2018) named 13 Russian nationals and three Russian entities. The 2020 election saw a refined version: not state actors but homegrown algorithmic radicalization, accelerated by QAnon content on Facebook, algorithmic recommendations on YouTube (the “alt-right pipeline” documented by the Wall Street Journal in 2018), and TikTok’s For You Page pushing increasingly extreme content. The October 2024 NSA, CISA, and FBI joint advisory confirmed the trend: AI-augmented influence operations were being deployed at industrial scale.

Internal documents made public by Haugen showed 1 in 3 teen girls reported that Instagram made their body image issues worse. The company’s own researchers confirmed the correlation, called it causal, and were told by executives to bury the data because it would be “a PR disaster.” The 2023 Senate Judiciary Committee testimony, led by Senator Lindsey Graham, ran with the headline: “Facebook knows it’s killing kids — and won’t stop.” Surgeon General Vivek Murthy issued a formal advisory in May 2023 declaring youth mental health crisis the defining public health emergency of the decade.

The TikTok case added new data. An internal TikTok research memo leaked in 2022 (the so-called “Project Texas” documents) showed that the app’s algorithm spent 90% more time pushing “problematic content” to users identified as vulnerable. The “TikTok brain” effect — shrinking attention spans, flattening affective range, accelerating reward loops — is now well-documented across multiple peer-reviewed studies. Compare the timeline: TikTok launched globally in 2018; by 2022 it was the dominant attention-capture engine for users under 25; by 2024 the average daily usage for US teens was 1 hour 47 minutes per day. Six years from launch to baseline-scale cognitive modulation.

The Older Pattern — Televangelism of the 1980s

The comparison is not original but it is precise. Between 1980 and 1992, the televangelism industry — Jim Bakker, Jimmy Swaggart, Robert Tilton, Oral Roberts — extracted roughly $3 billion from viewers (over $9 billion in 2024 dollars). The technology was a satellite dish and a phone bank. The mechanism was parasocial intimacy, perpetual emotional crisis, and the promise of salvation through financial contribution. The fall came not because the viewers woke up but because the operators got caught — Bakker’s rape conviction (1989), Swaggart’s prostitution scandal (1988), Roberts’s “give me $8 million or God will call me home” fundraising stunt (1987). The technology was destroyed; the pattern survived.

Today’s “creators” run the same algorithm on a more efficient substrate. Televangelism’s 1980s extraction rate: ~$600 million per year at peak. YouTube/TikTok/Instagram’s 2023 creator-economy extraction: estimated at $250 billion globally. That is 400× the 1980s peak, in constant dollars. The persuasion architecture is the same; the bandwidth is four orders of magnitude higher.

$250 BILLION
GLOBAL CREATOR-ECONOMY REVENUE, 2023

The Chinese model is the explicit testbed. China’s Social Credit System — operational in pilots since 2014, with full deployment targeted for the 2030s — scores citizens on a behavioral index and restricts travel, employment, and loan eligibility for low scorers. We instinctively recoil from it as authoritarian. But the US system runs the same loop without the rubric: credit scores, background checks, employment screening, insurance pricing, college admission, and yes, social-media reputation — all are scoring systems that shape behavior. The difference is opacity. China publishes its rubric. Silicon Valley does not.

The Mechanisms of Control

The New Religion

The structural parallels to organized religion are not coincidental. Both systems use invisible authorities, arbitrary rules enforced as natural law, public confession, excommunication, and salvation through participation. The difference is the surveillance, the personalization, and the scale. The same control architecture, applied to the personal-homepage era, is traced in our GeoCities Was the Last Honest Internet archive.

Documented Effects

Internal Facebook research, leaked by Frances Haugen in 2021, showed the company knew its products were harming teen mental health. Internal documents showed Instagram increased body image issues in 1 in 3 teen girls. The research was suppressed. The products continued. Similar reports have emerged from TikTok, Snap, and YouTube.

Tristan Harris and Aza Raskin (Center for Humane Technology) have been among the most prominent whistleblowers. Their work documents the deliberate design choices — features engineered to maximize engagement even when the engagement is harmful.

What Is the Solution?

Not banning the technology. Not unplugging entirely. The system is too embedded, too useful, too integrated into modern life. The solution is structural: regulation that aligns platform incentives with user well-being. The EU’s Digital Services Act and the proposed Kids Online Safety Act in the US are early attempts. Whether they will be effective is an open question.

What is certain: the current model — engagement maximization with no accountability for harm — is unsustainable. The question is not whether the system will change, but how, and who will lead that change.

Sources & Further Reading

Classification: SOCIAL MEDIA CONTROL | Confidence: DOCUMENTED

I remember a GeoCities homepage from 1996. The background was #FF00FF — a magenta so aggressive it hurt to look at. There was a tiled GIF of a rotating earth in the upper-right corner. The text was Comic Sans, twelve point, with a blinking cursor. There was a MIDI file autoplaying — it was always the same one, a four-bar loop of “Hit the Road Jack” or the X-Files theme. There was a visitor counter in the bottom-left, its digits in a tiny red digital font, currently reading “000423” and the user was proud of every one of them. At the bottom of the page, a header: “Sign my guestbook!” with a small GIF of a pencil. Above the guestbook request: a small section called “Awards,” showing seven or eight 88×31 pixel badges, each one given to the page by another user in exchange for the page giving them one back. The page was a shrine to a single topic — the owner’s favorite band, mostly, or their favorite anime, or a memorial to a cat.

It was ugly. It was earnest. It was theirs.

By the time GeoCities was shut down on October 26, 2009, it was hosting the largest collection of user-created web pages that had ever existed. The peak figure, by most estimates, was around 38 million user pages in 2001. Thirty-eight million individual human beings, each with a homepage they had built themselves, on a service that gave them the space for free in exchange for the right to put small ads on the page. The pages were written in raw HTML, often with broken table tags and unclosed <font> elements. They were a mess. They were a record.

The Neighborhood System

GeoCities organized its user pages into neighborhoods — themed directories that functioned as both topological zones and social structures. The original six neighborhoods, when the service launched in 1994 as Beverly Hills Internet, were Heartland, Wellesley, Athens, Hollywood, SiliconValley, and SunsetStrip. They were not metaphors. They were actual server-side directories whose paths began with the neighborhood name. A page at /Heartland/2345/index.html was, structurally, a house in a town called Heartland. The metaphor was the architecture. The architecture was the community.

Tokyo was added in 1996 for Japanese-language content, becoming the seventh neighborhood and the first non-US cluster. Paris (1997) and others followed. By 1999 there were 14+ neighborhoods, each with its own culture. Heartland was the catchall — personal pages, family sites, regional hobby groups, small-town America. Wellesley was the women’s interest neighborhood. Athens hosted the philosophy and literature crowd. Hollywood was entertainment. SiliconValley was tech. SunsetStrip was music. Tokyo was anime, manga, J-pop, and the early-2000s online culture that would, by 2005, become a global subculture.

The neighborhood structure was a primitive form of community moderation. Each neighborhood had a “Homestead” or “Mayor” — an experienced user who could review new submissions, flag inappropriate content, and curate the public face of the neighborhood. The roles were unpaid. The users performed them because they cared about the neighborhood. The system was, structurally, the closest the early commercial web ever got to a federated protocol. Each neighborhood was its own semi-autonomous community. The protocol was HTML. The federation was the homepage link.

By 2001, the directory structure was massive. Each neighborhood had its own internal subdirectories organized by topic: /Tokyo/Harajuku/4321/ was a personal page about Harajuku fashion. The links between pages formed a graph that, in density and human curation, has not been matched since. The Yahoo acquisition (1999) tried to industrialize the structure. It never fully succeeded. The residents kept building their own pages.

The Archive, the Deletion, and the Mirror

Yahoo bought GeoCities in June 1999 for $3.57 billion. At the time of acquisition, GeoCities was the third-most-visited web property in the world, behind AOL and Yahoo itself. Yahoo’s strategy was to convert GeoCities’ free users into paying customers and to integrate the service into its advertising network. The conversion plan failed. The free users did not convert. The advertising integration was half-hearted. By 2005, Yahoo had largely stopped investing in the product.

On April 1, 2009, Yahoo announced the closure. The deletion date was set for October 26, 2009. The response was immediate: archivists, academics, and former users launched a coordinated effort to preserve the pages. The Internet Archive captured a partial mirror, focused on the highest-traffic neighborhoods. The ReoCities project (2009-2013) attempted a full crawl but ran out of funding in 2013. The GeoCities Time Machine, hosted by the team at OOcities and later the Geocities.ws mirror, preserved fragments. The total pages preserved: estimated 10-15% of the original.

The rest was deleted. Twenty-three years of human self-expression, deleted. The deletion was quiet. There was no public mourning event. There was no ceremony. The pages were simply removed from Yahoo’s servers, and the URLs that had been bookmarked, linked, and cited for years began to return 404. The Internet Archive’s Wayback Machine has some of the pages — but the snapshot coverage is sparse, and many pages are missing. The architecture of personal expression, with its tiled GIFs and MIDI loops and guestbooks, was simply gone. The mirror, partial as it is, is all that remains.

What We Lost

The personal homepage was a different kind of object than a social media profile. It was made. It was built in raw HTML or in GeoCities’ own page-builder tools. It was a project, not a fill-in-the-blank form. The act of building it required learning — at least a little — how the web worked. The page could be customized, themed, redesigned, abandoned, rebuilt. It was a place. The visitor was not the product. The visitor was the guest. The host had made the page. The guest had come to see it. The transaction was the visit. The transaction was the point.

There was no algorithm. There was no feed. There was no “for you.” A visitor arrived because they had typed the URL, or because they had clicked a link from another page, or because they had clicked a button on the guestbook chain. Each visit was intentional. Each visit was, in a small way, a relationship. The visitor counter at the bottom of the page incremented by one. The owner, if they were online, knew someone had come.

The Structural Argument

The current internet is, by structural comparison, a controlled environment. Every major platform — Facebook, Instagram, TikTok, X, YouTube, Substack — is a platform: a centralized service that intermediates between creators and audiences, takes a cut, and sets the rules. The platform owns the URL. The platform owns the follower graph. The platform owns the recommendation algorithm. The platform owns the audience. The creator is a tenant. The platform is the landlord. The landlord can change the lease at any time.

GeoCities was not a platform. It was a host. The user owned the page. The user owned the HTML. The user owned the guestbook. The user could, in principle, export the entire page and host it elsewhere. The platform did not own the audience. The user did. The platform was, structurally, the closest the commercial web ever came to a public utility.

The argument that today’s internet is “enshittified” — Cory Doctorow’s term, in his 2022-2023 essays for Platformer — is essentially the argument that the platform model has captured the audience and is now extracting rent from it. The audience was once the user’s. The audience is now the platform’s. The user is now a creator-for-hire, paid in exposure. The exposure is allocated by the algorithm. The algorithm is owned by the platform. The cycle is closed. The control architecture behind the platform extraction is the same attention-economy loop dissected in our social media as the new religion investigation.

The personal web, between 1994 and 2009, was the last period in which anyone with a phone line and a credit card could publish to the open web without an intermediary taking a cut. We traded that for reach. We did not get reach. We got the algorithm. The same arc — a parallel digital nation, then deletion by the parent company — repeats almost exactly in our Club Penguin was a country archive. The algorithm is not the audience. The audience is the algorithm’s product. The original sin was the platform. The original virtue was the homepage.

⚠ PATTERN RECOGNITION ALERT

Every neutral communication system eventually becomes a control surface. GeoCities was the last network in which the user was the publisher, not the product. The platform era replaced the homepage with the profile. The profile rents the user to the algorithm. The algorithm rents the user to the advertiser. The cycle ends when the user is paying for the privilege of being the product. The tiled magenta background was the last honest aesthetic. The next one will be whatever the algorithm permits.

Sources & Further Reading

Classification: SOCIAL MEDIA CONTROL | Confidence: DOCUMENTED

At its peak in 2013, Club Penguin had more than 200 million registered accounts. The peak concurrent user count, recorded in late 2013 during the Christmas server rush, was over 1.7 million simultaneous penguins. At peak load, Club Penguin was, by population, larger than the nation of Romania. The penguins were navigating a custom-built digital geography, attending parties in igloos they had decorated with virtual furniture, working for wages paid in a virtual currency, forming friendships, navigating betrayals, attending in-game concerts, learning to be a citizen of a country that did not exist.

The country existed anyway. Club Penguin was, between its public launch in October 2005 and its full closure on March 30, 2017, a parallel society of children. It was, by almost every sociological measure, a nation: it had a population, a culture, a currency, a labor market, a black market, a justice system (Disney’s “Moderators,”), a foreign policy (language-specific servers — English, French, Spanish, Portuguese, German), and a national mythology (the Penguin Secret Agent, the underground “Pizza Parlor” society, the Halloween Party of 2008, the Club Penguin Army Wars of 2007-2008). The country was built by Disney. The country was shut down by Disney. The Twitter response, on the day the servers went dark, was a national mourning event broadcast in real time across 11 million tweets.

The Geography

Club Penguin’s geography was, on its surface, a small Antarctic town. The map contained: the Town (the Town Center, the Coffee Shop, the Stage, the Welcome Center); the Plaza (the Pizza Parlor, the Dojo, the Pet Shop, the Newsstand, the Dance Lounge); the Snow Forts, the Forest (with the Cove, the Box Dimension); the Beach (the Lighthouse, the Migrator, the Aqua Grabber, the Iceberg); and the Mountain (the Ski Hill, the Ski Lodge, the Climbing Wall). The Town was for first-time visitors. The Plaza was for commerce. The Dojo was for martial-arts minigames and the Card-Jitsu card-collecting strategy game, which became its own competitive subculture. The Cove was for moody penguins — the lighting was dim, the music was melancholy. The Iceberg, accessible only during certain months, was a community-wide holiday zone.

Beyond the public map, every penguin had an Igloo. The Igloo was a private, player-customizable home, located in a separate instanced area. The Igloo could be decorated with virtual furniture purchased with virtual coins. The Igloo could host up to 20 penguins at a time. The Igloo was the in-game equivalent of a bedroom — a place where a child could express identity, throw a “party,” invite friends. The Igloo was, in sociological terms, the most important space in the game, because it was the only space that the player actually owned.

The igloo party was the foundational social unit. A penguin would decorate the Igloo, send out invitations, and host. The host could play music, announce games, eject disruptive guests. The host was, in every meaningful sense, a host. The guests were, in every meaningful sense, guests. The etiquette was elaborate and, in many cases, learned from observation rather than instruction. The etiquette was a real social skill, exercised in a real (if virtual) space.

The Economy and the Black Market

Club Penguin ran a closed virtual economy. The currency was coins, earned by playing minigames (the higher the score, the more coins) or by performing “secret” tasks. The coins were spent on clothing (penguin outfits), furniture, puffles, and Igloo upgrades. The coin-to-USD exchange was never official, but the third-party market was robust: a rare item, like a “Blue Books” furniture piece from a 2008 server event, could sell for several dollars in eBay listings. The “Coin-to-Cash” trade was a small but real shadow economy, conducted primarily by older players (13-17) who had outgrown the in-game chat but knew where the trading servers were.

The black market was a different layer. The black market dealt in accounts, not items. A penguin account with a high level, a long buddy list, and a wardrobe of rare items could be sold to a younger player for a small but real sum. The sale violated Disney’s Terms of Service. The sale happened anyway. The black market was, structurally, the first digital-goods secondary market that the post-Minecraft, pre-Roblox generation encountered. The black market was, also, the first place where the players learned that virtual property could have real-world value, that the virtual world was a real economy, and that the platform owners (Disney) did not recognize their property claims.

The economy was inflationary. New items were added every month. The coin price of older items was not adjusted. The result was a slow devaluation of the in-game currency, a familiar pattern in real-world economies as well. By 2013, the prices of late-2010 items had been effectively wiped out. The economy taught, by example, the dynamics of inflation, currency debasement, and the value of rarity. The lessons were not framed as economics. The lessons were framed as Club Penguin.

The Death and the Mourning

On March 29, 2017, Disney announced that Club Penguin would shut down on March 30, 2017, at 12:00 AM PST. The servers were scheduled to be taken offline. The data was scheduled to be deleted. The accounts were scheduled to be retired. The country was scheduled to be erased. The announcement was posted on the Club Penguin blog and shared on the official Twitter account. The reactions were immediate.

On March 30, 2017, between 6 PM and midnight PST, the Twitter accounts of players aged 9 to 19 posted more than 11 million tweets with the hashtag #ClubPenguin. The tweets included: goodbye letters to specific in-game friends; screenshots of in-game locations; recollected memories of the “good old days” (which, for a 12-year-old, meant 2014); grief messages written in the style of eulogies; screenshots of the in-game “doomsday clock” showing the minutes remaining. The volume was high enough that #ClubPenguin trended worldwide on Twitter for approximately 18 hours.

The age of the responders — 9 to 19 — means that the average responder had been a Club Penguin citizen for at least 4 years. The responders had, in many cases, grown up with the game. The mourning was real. The mourning was the mourning of a country that had been erased.

The Private Servers and the Diaspora

Within weeks of the March 30, 2017 shutdown, multiple “private servers” had launched. The first major one, Club Penguin Rewritten, launched in 2017 and grew to a peak of more than 1 million registered accounts. Club Penguin Online followed in 2018, then NewCP, CPPS projects, and dozens of smaller servers. Each was a copyright violation. Each was a project run by volunteers, often teenagers, often former Club Penguin citizens, who reverse-engineered the original game’s protocols and rebuilt the servers from scratch. The servers were tolerated by Disney, intermittently, for years. Disney sent cease-and-desist letters in waves: 2018, 2020, 2022. The waves did not stop the servers. The diaspora kept rebuilding. The pattern is the same one that ran after the 2009 GeoCities deletion — a diaspora preserving an erased country, for free, for the right to remember.

The private-server community is, in 2024, a small but persistent subculture. The largest servers have 50,000-200,000 registered accounts. The servers are old versions of the game (typically 2012-2014 era content). The servers are run by people who were children when the original game shut down. The servers are, in sociological terms, a refugee community. The refugees are preserving a country that has been erased from the official record. The refugees are doing it for free. The refugees are doing it for the right to remember.

What We Lost

Club Penguin was the first “digital nation” that a generation of children inhabited. The nation had a culture, a currency, a social graph, a justice system, and a history. The nation was shut down by the company that owned it. The citizens are now the first cohort to have grown up entirely inside platforms that they did not own. The citizens are now the users of TikTok, Instagram, Discord, Roblox, and the algorithm-driven social media platforms that Club Penguin predicted. The control architecture that turned those platforms into extraction surfaces is the subject of our social media as the new religion investigation.

The prediction Club Penguin made was not technological. The prediction was sociological: a generation of children will live inside a digital space, will form a culture inside that space, will build a society inside that space, and the owners of the space will eventually shut it down. The citizens will not be consulted. The data will be deleted. The country will be erased. The citizens will grow up and become adults who do not own the platforms they live on. The pattern, in 2017, was visible. The pattern, in 2024, is dominant. The simulation is now running on the architecture that Club Penguin prototyped. The penguins were the test population. The penguins are now the population.

⚠ PATTERN RECOGNITION ALERT

Every platform that begins as a community ends as a property. Club Penguin was the prototype. The community was the product. The product was shut down. The citizens were not consulted. The data was deleted. The pattern is now global. The pattern is now normal. The simulation is now running on the architecture that a children’s game predicted. The penguins were the test population. The penguins are now the population. The pattern repeats because the platform owns the country. The country was never the citizens’. The citizens were the country’s.

Sources & Further Reading

Classification: SOCIAL MEDIA CONTROL | Confidence: DOCUMENTED

You sat down at the family computer. You opened AOL Instant Messenger. You logged in. The familiar door-open chime, the buddy list populating left to right as your friends signed on. Then you wrote an away message.

One hundred and eighty characters. No images. No link previews. No algorithm. No “for you” feed. You agonized over the phrasing for ten minutes — sometimes fifteen — then you closed AIM, got up from the desk, and went to dinner. Hours later, you’d come back and check who had messaged you while you were away. Some of them had sent one line. Some had sent ten. The away message had been the first thing they saw when they tried to reach you.

You had, in other words, the same ritual and the same constraint that defined the entire dial-up era — patience, scarcity, and the discipline that came from the medium itself.

That was the away message. It was the first true social-media art form. Written for an audience that wasn’t there yet.

The Format Zoo (2002-2008)

AIM away messages had genres. The taxonomy was unofficial, undocumented, and absolutely real to the people who practiced it.

The poetic. Lowercase philosophical fragments, em-dashes, period-as-pause, often ungrammatical on purpose. Examples from the era: “sometimes the silence is the loudest part.” “i am somewhere between here and not here.” “—thinking about the things we don’t say.” The form was influenced by LiveJournal users, by the emo subculture, by the early-2000s wave of indie music that put elliptical lyrics on T-shirts.

The cryptic. Single words. Lyric fragments. Inside jokes that only the dozen people on your buddy list would understand. “brontosaurus.” “pinegrove.” “the pineapple one.” The cryptic away message was a private channel broadcasting on a public frequency, and only the right receivers could decode it.

The performative. “brb dying of boredom — lyke literally ☹” “DO NOT DISTURB (jk hi mom)” “im not here rn but if i was id be doing homework. which i am not.” The performative away message was the proto-Twitter: self-aware, ironic, addressed to no one and everyone.

The existential. Movie quotes, song lyrics, half-finished thoughts. “we are all in the gutter, but some of us are looking at the stars.” “the future is unwritten.” “i am large, i contain multitudes.” The existential away message was the teenager’s first encounter with literary quotation as identity construction. It was, for the people who wrote it, real.

There was also an entire auxiliary culture around the AIM buddy icon — the small 16×16 pixel image that appeared next to a friend’s screen name. Buddy icons were traded like baseball cards: shared via IM, archived in folders, displayed with pride. The icon was the first instance of “pfp culture” — a static visual signal that said something about you. Anime stills, band logos, sports team crests, custom pixel art. A good icon was a status symbol. A blank icon was an admission you hadn’t gotten around to it.

The AIM sound effects — door opening on sign-in, IM arriving, file transfer complete — were the earworms of an entire cohort. The “door opening” sound in particular was so deeply associated with “a friend is online” that YouTube compilations of the sound have millions of views today, decades after the product was killed.

The Parallel Platform Dialects

AIM was not alone. Each messaging platform had its own grammar.

MSN Messenger (1999-2013) had Nudge — the screen-shake prank feature — and a deep custom-display-picture ecosystem. Display pictures were tiny pixelated GIFs, often anime stills or photographs of bands, set as your avatar. The display picture was a costume you wore in public. The Nudge was the only way to send a notification that wasn’t a typed message: a screen-shake equivalent of tapping someone’s shoulder.

Yahoo! Messenger (1998-2012) had profile customization to a degree that AIM never matched. Yahoo profiles were pages: layout, background, music (yes, autoplaying music), avatar, status. They were proto-MySpace pages inside a messaging client. The Yahoo profile was the closest the early-2000s web came to a personal homepage for people who didn’t know how to build a personal homepage.

ICQ (1996-2013) was the original instant messenger, predating AIM by a year. Its famous “uh-oh!” notification sound became a global earworm. ICQ users were the elder statesmen of the era — slightly older, more international, more technically literate. The ICQ UIN was the first stable numeric identity millions of people ever had.

Each platform was an island. You had an AIM name, a Yahoo name, an ICQ number, an MSN address. They didn’t sync. They didn’t cross-reference. The cross-platform identity didn’t exist. You were three or four different people in three or four different rooms, and the rooms didn’t connect.

Peak, Decline, and the Death Rattle (2008-2017)

AIM peaked at roughly 100 million monthly active users between 2008 and 2011. By 2012, usage had halved. Facebook Messenger launched as a standalone app in 2011, with cross-platform continuity. Smartphones killed the desktop chat client first; the iPhone’s Messages app (2007) absorbed texting, and instant messaging migrated to mobile apps where the away message had no equivalent.

Mobile messengers don’t have status messages. The interface doesn’t support them. The metaphor of the away message — written text broadcast to your network while you’re unavailable — assumed a desktop session that could be left open. On a phone, you’re either looking at the app or you’re not. The intermediate state of “I’m here but I’m doing something else, and here’s a poem about it” was architecture-specific. When the architecture changed, the form died.

AOL shut AIM down on December 15, 2017. The New York Times ran a small obituary. Most users had been gone for years. The platform’s logs, status messages, buddy lists — twenty years of adolescent poetry — were deleted. A few archivists preserved fragments. Most of it is gone.

What We Lost

The away message was self-expression under constraint. 180 characters. No images. No algorithms. No audience of millions. The constraint produced poetry. The 180-character limit was, in 2003, an aesthetic device: it forced compression, forced precision, forced the writer to mean what they meant or to fail cutely.

Today’s platforms have no character limits and infinite audiences — and produce nothing. The average post on X (formerly Twitter) is four lines long and says nothing. The average Instagram caption is two words. The average TikTok bio is a single emoji. We have more room than the AIM away message ever did and less to say.

danah boyd, in It’s Complicated (2014), argued that teens’ early social-media lives were experiments in identity performance under observation. The away message was the earliest, cleanest version of that experiment. It was small, it was specific, it was for an audience that knew you. The constraint — both character count and audience size — was the engine of the form.

When you remove the constraint, you don’t get more poetry. You get less. The away message is the fossil record of what social media could have been.

⚠ PATTERN RECOGNITION ALERT

Every platform that began as an intimate, constrained medium eventually expanded into a broadcast medium — and lost the form that made it intimate. AIM, MSN, Yahoo, MySpace, Facebook, Twitter, Instagram, TikTok: each iteration got bigger, faster, broader, and shallower. The away message is the fossil record. The pattern repeats because constraint produces art, and art is not what the platforms sell. The structural mechanism behind the pattern — the attention-economy control loop that turned every intimate medium into an extraction surface — is the subject of our social media as the new religion investigation.

Sources & Further Reading

Further Reading