Ancient Mysteries - The Internet’s Most Disturbing Mysteries Get Even Darker | Vol. 15

Episode Date: September 25, 2026

How much stranger does the internet become the deeper you look?The internet has preserved some of the most unsettling, mysterious, and difficult-to-explain moments of the modern era. Some began as broadcasts seen by thousands of people. Others appeared as obscure photographs, cryptic recordings, strange websites, or videos whose origins became mysteries of their own.In this documentary, we explore another collection of disturbing stories from around the internet — separating documented events from speculation, online theories, and mysteries that remain unresolved.We revisit the infamous Max Headroom signal hijacking, when television broadcasts in Chicago were interrupted by a masked figure and the people responsible were never publicly identified. We also examine the terrifying 2018 Hawaii false missile alert, which sent an emergency warning to phones across the islands before officials confirmed that no missile was actually approaching.But the deeper we go, the stranger the stories become.From unusual photographs with disputed explanations to mysterious recordings that spread across online communities, we look at how fragments of information can become enormous internet mysteries — especially when their original context disappears.One of the strangest examples is 11B-X-1371, a cryptic video filled with coded imagery and hidden messages that triggered years of online investigation. Its creator eventually came forward, but questions surrounding the project helped turn it into one of the internet’s most memorable viral mysteries.Along the way, we examine what is actually known, what remains uncertain, and how easily fact, rumor, coincidence, and speculation can become intertwined once a mystery spreads online.And sometimes, discovering the explanation only leads to another question.Watch to the end as we explore another collection of strange broadcasts, unexplained images, mysterious videos, and disturbing stories preserved by the internet.💬 Which mystery in this episode did you find the hardest to explain?🔔 Subscribe for more documentaries exploring internet mysteries, unexplained events, strange history, lost media, and the darker corners of the modern world.

Hey there. That flat ringing on a line where nobody picks up is the sound of this whole episode. A TV broadcast hijacked by a guy in a rubber mask. An entire state told a missile was incoming by accident. A 911 call taken down perfectly and passed to nobody. A photo found on a parking lot that nobody could ever attach a name to. One pattern connects all of it. The gap between information exists and information arrived. No demons. No cursed files. The scariest villain tonight is a drop-down menu, and it costs 38 minutes of people saying goodbye to their families. Fair warning, some of these have no ending. Not a cliffhanger, there is just nothing there. Almost 40 years and the whole internet on the case, still nothing. Hit like if that drives you

crazy too and drop a comment with the city you're watching from. Let's get into it. November 22nd, 1987 was a Sunday, and in Chicago that meant one thing. The Bears had played, the Bears had won, and roughly 3 million people were sitting in front of a television, waiting for a man in a suit to show them the highlights. 9 o'clock news on WGNTV, Channel 9. Sports anchor Dan Rhone is walking the audience through the bears beating the Detroit Lions. Normal night, normal broadcast.

The kind of segment that gets produced, aired and forgotten roughly 300 times a year. And then the screen dies. Not static. Not a colour bar. not the friendly technical difficulties card that stations kept loaded for exactly this kind of moment. What appears instead is a person in a rubber mask standing in front of a sheet of corrugated metal that is spinning behind him, and the audio is nothing but a mechanical buzzing sound, like somebody recorded a wasp inside a filing cabinet. The mask is Max Headroom, which requires a small amount of context if you were not alive in 1987.

Max Headroom was a computer-generated television host, who was not actually computer-generated at all, because the technology to do that properly did not exist yet, so instead the actor Matt Fruhe sat in a chair for hours, getting fitted with latex prosthetics and a fibreglass suit so he could pretend to be a digital man with a glitching voice. He was everywhere that year. He had his own show on ABC. He sold Coca-Cola. He was the face of the future as imagined by people who had never seen a laptop. So that face is now on Channel 9, bobbing side to side saying nothing intelligible for about 25 seconds. Then it vanishes and the news comes back and Dan Rohn, to his eternal credit, delivers one of the great live television reactions in American history by telling the audience that if they are wondering what just happened, so is he.

Here is what happened, and it is genuinely more interesting than the mask. Television stations in 1987 did not send their signal from the studio directly into the air. The studio sat in one building and the transmitter sat somewhere else entirely, usually on top of the tallest structure available, because height equals reach. In Chicago, that meant the John Hancock Centre and the Sears Tower, two buildings that were essentially enormous antenna stands with offices attached. To get the program from the studio to that transmitter, stations used a microwave link, a tight little beam of signal shot across the city from one dish to another.

That link was the weak point, and everybody in the industry, knew it. A microwave receiving dish does not have opinions. It does not verify credentials. It does not ask who is sending. It simply listens on a frequency and forwards whatever is loudest into the transmitter, and the transmitter obediently blasts it across five states. So if you happen to own or borrow a microwave transmitter and you knew the exact frequency, and you could aim a dish at that receiver from somewhere with line of sight, you could simply be louder than the actual television station. The building's own signal loses. Yours wins. Congratulations, you're now WGN. WGN's engineers, being professionals who were physically present at their transmitter, figured this out within

seconds and switched their studio link to a different frequency, which knocked the intruder off the air. That is why the first hijacking lasted less than half a minute. It was ended by a guy who knew where the buttons were. Two hours later, the same person came back and picked a much softer target. At around 1115 that night, WTTW, Channel 11, Chicago's public broadcasting station, was airing an episode of Doctor Who, specifically a serial about a lighthouse where things go badly for everyone involved. UTTW's transmitter was on the Sears Tower and crucially it was unmanned at that hour. Public television did not have the budget to keep an engineer sitting next to a transmitter at 11 at night on a Sunday, because why would it? Nothing has ever gone wrong.

There was nobody up there to change frequencies, nobody to pull a plug, nobody to do anything at all. So when the signal got overpowered a second time, the intruder got the floor for a full minute and a half, uninterrupted, and used it to produce 90 of the stranger seconds ever broadcast on American television. The mask is back. The corrugated sheet is spinning behind him again, which was a deliberate joke, because the real Max headroom appeared in front of a rotating geometric background. and this was the version you could build in a garage for the price of some roof flashing. The audio this time is not just buzzing. There is a voice under there, run through so much processing that it sounds like someone gargling a modem, and people have spent nearly four decades arguing about individual syllables the way scholars argue about ancient manuscripts.

What you can pick out is a series of jokes, and this is the part I find genuinely fascinating because it tells you a lot about who did this. He mocks a WGN sportscaster by name. He hums the theme from Clutch Cargo, a cheap 1960s cartoon famous for using real human lips superimposed on still drawings, which is a deep cut even by 1987 standards. He holds up a can of Pepsi and makes a comment about it,

which is a punchline aimed squarely at the fact that Max Headroom's entire public identity was selling Coca-Cola. He puts on a rubber glove. He complains about being a nerd, or being called a nerd, or being surrounded by nerd. depending on which transcript you trust and how much you have destroyed your own ears listening to a degraded VHS rip at maximum volume and then the finale which I have to describe carefully for the sake of this video staying monetised the masked figure bends over a second person wearing what appears to be a maid outfit enters the frame and administers a spanking with a fly swatter there is some yelling the signal cuts out doctor who returns the lighthouse continues to go badly for everyone involved now take a step and look at what actually just happened, because the comedy of it obscures how serious it was.

Two federally licensed broadcast facilities were seized on the same night, in the third largest media market in the country, by a person who was never identified. This was not a prank call. This was not somebody defacing a website, which was not even a concept yet. Under federal law, this was unauthorized operation of a radio transmitter and willful interference with a licensed broadcaster, and the FCC made it extremely clear that it was treating it as such. The numbers being thrown around were fines up to $100,000 and up to a year in prison, which in 1987 money was the kind of penalty that ends a career, a marriage, and a lease all at once. The FCC's Chicago field office investigated. The FBI took an interest. WGN and

WTTW handed over everything they had, which was mostly the tapes, and the investigation produced in total nothing. No arrests, no charges, no named suspect. Not then, not in the 1990s, not in the age of high-resolution scans and audio forensics and an internet full of people whose entire personality is solving other people's cold cases. So let us talk about why, because the reasons are specific and they explain a lot about how mysteries survive. Reason one is that the equipment pointed at professionals, and professionals know how to not leave receipts. Everyone who has ever looked at this case seriously arrives at the same conclusion. Whoever did this understood broadcast engineering. They knew that the studio to

transmitter link existed, which most viewers did not. They knew the frequency, which was not printed in the newspaper, they knew which building to aim at. They knew that overpowering the beam was possible and roughly how much power it would take. And critically, they knew that WTTW's transmitter would be sitting up there alone at 11 at night, which is not information you get from curiosity. That is information you get from working in the industry. or from being extremely close to someone who does. The FCC essentially said this out loud at the time and it created an uncomfortable situation

because the pool of suspects was a professional community in a city where everyone had a friend at another station. Reason two is that the physical evidence was, from the beginning, almost entirely useless. The mask was a mass-produced Halloween item sold in stores across the country because Mack's headroom was a merchandising machine that year. You cannot trace a product that was on every seasonal rack in America. The corrugated metal was a building material. The fly swatter was a fly swatter.

The clothing was ordinary. The only unusual object in the entire production was the microwave transmitter itself. And a microwave transmitter is not a thing you leave in the trunk of your car after committing a federal crime. It is a thing you drive somewhere far away and get rid of. Or, if you borrowed it from work, a thing you quietly put back on the shelf where it lived, at which point it is not evidence at all. It is inventory. Reason three is, timing and this one is underappreciated. This happened in 1987, which is the technological equivalent of committing a crime in a fog bank. There is no digital trail because there is nothing digital to leave a trail in, no cell phone towers logging a location, no purchase history in a database,

no search queries, no group chat, no cloud backup of anything. If four people planned this, they planned it in a room, out loud, and then they went home. The entire case exists as a videotape in a rumor. Reason four, and this is the one that fascinates me most, is that the crime was designed to be untraceable in its content as well as its execution. Think about what is not in that 90 seconds. There is no manifesto. There is no political demand. There is no cause, no group name, no religious message, no threat, no ransom, no ideology. It is pure vandalism with a sense of humour, and that means investigators had nothing to work with motivationally. When somebody hijacks a broadcast to tell you about a cause, you have a suspect pool, people who believe in that cause.

When somebody hijacks a broadcast to hum a cartoon theme and get hit with a fly swatter, your suspect pool is everyone in Chicago with a sense of humour and a friend at a TV station, which is a demographic you cannot subpoena. And that is what makes this case different from basically every other broadcast intrusion in history, which is worth walking through, because the comparison is where the mystery gets its shape. The most famous American case before this one happened in April 1986 when a satellite dish salesman in Florida named John McDougal got fed up with cable companies scrambling their signals and pricing his customers out of the market. He worked nights at a satellite uplink facility, so one evening, alone on shift, he pointed

a dish at HBO and overrode the movie with a text message complaining about subscription prices, signed with the name Captain Midnight. It lasted about four minutes. He was caught within months. Why? Because the character generator that produced that specific font was rare, because the uplink facilities capable of doing it were a short list, and because he had been talking about it. He pleaded guilty, paid a $5,000 fine and got a year of probation, and he got off lightly because he was polite about the whole thing. The lesson is that he was found through his equipment and his mouth, which are the two things the Chicago intruder appears to have kept perfectly under control. Then there is the case in the United Kingdom from 1977, where an announcer's voice

on a regional news broadcast was replaced by someone claiming to represent a galactic authority delivering a warning about humanity's weapons. Six minutes of alien diplomacy. Never definitively solved either, but there is a critical difference. That intrusion carried a message with an identifiable subculture attached to it, which gave investigators a direction to look, even if they never got there. The Chicago case gave them a direction of absolutely nowhere. Every solved case follows that same shape. Someone hijacks a signal to say something, and the thing they said is the thread you pull.

Chicago said nothing. Chicago is a joke with no punchline attached to a person, and jokes do not have a paper trail. Which brings us to the internet era and to the part of this story that is, in its own way, as strange as the original broadcast. Starting in the mid-2000s, This case became a permanent fixture of online mystery culture.

The video was uploaded, re-uploaded, restored, analyzed frame by frame, and had its audio pulled apart by people with legitimate signal processing skills who were doing it for free at 2 in the morning because that is what the internet is for. Multiple independent attempts at cleaning the voice track have been made. The results are charitably contested. Everybody hears something slightly different, because the source is a tape of a tape of a broadcast that was already distorted by design. and at that point you're not doing forensics, you are doing audio astrology.

In 2010, a post appeared on Reddit from a user who claimed to have known the people responsible, describing two brothers from the Chicago area with the technical skills and the exact sense of humour on display, and giving a detailed account of how it supposedly came together. The post is extremely well written, internally consistent, and has circulated for over a decade as the leading theory. It has also never been confirmed by anyone with a badge or a document, and I want to be honest with you about the standard of evidence here, because this is the point where a lot of coverage of this case quietly slides from reporting into vibes. A compelling story told by an anonymous

account is not a solution. It is a lead. Journalists have followed it, including a fairly serious effort by a technology publication that tracked down and doorstep the people named, and the result was denial and dead air. And this creates the actual permanent problem with the Chicago case. The statute of limitations expired a very long time ago. Whoever did this cannot be prosecuted. They could walk into a new studio tomorrow, sit down, describe the equipment, produce the mask from a duffel bag and reconstruct the entire night in detail,

and the legal consequence would be zero. They would probably get a documentary deal out of it. In Toronto, every arrival is a statement, and nothing says it better than this. Cadillac Optic was the number one selling luxury EV in Canada for 2025. Find your rhythm across a seamless 33-inch display and an immersive 19-speaker AKG surround audio system. This city demands agility and optic delivers with precision

to make every drive extraordinary. Let's take the Cadillac. Find out more at Cadillac canada.ca. Luxury sales claim based on S&P Global Mobility Canadian New Vehicle Total Registrations for calendar year 2025 for the Cadillac definition of luxury. Visit Bellow. BetMGM casino and check out the newest exclusive.

The price is right fortune pick. BetMGM and GameSense remind you to play responsibly. 19 plus to wager. Ontario only. Please play responsibly. If you have questions or concerns about your gambling or someone close to you, please contact Connects Ontario at 1-866-531-2,600 to speak to an advisor, free of charge.

BetMGM operates pursuant to an operating agreement with Eye Gaming Ontario. In 1987, the intrusion was a nuisance story. It ran in the Chicago papers for a few days, the FCC made threatening noises, and then it went away, because there was no mechanism for a weird 90-second video to travel. You either saw it live or you did not, and if you did, you spent the following decade telling people about it and being politely disbelieved. Then the internet was invented and the tape got out, and suddenly a minute and a half of Chicago television from 1987 was in front of a global audience of tens of millions, permanently forever. available at any hour to anyone. The hijacker wanted an audience of Chicago viewers for 90 seconds. What he got, entirely by accident and entirely without consent, was the largest audience any broadcast intruder has ever had, running continuously for decades, complete with a fandom.

Somewhere out there is a man in his 60s or 70s who did a stupid thing in a rubber mask on a Sunday night, and whose stupid thing has now been watched more times than most feature films, and who cannot tell a single person at a dinner party about it. without ending his own anonymity. Naturally, being the internet, people also did the thing the internet always does, which is to try to make it darker than it is.

There are theories that the audio contains a hidden message. There are theories about who the second person in the frame was. There are periodic claims of deathbed confessions that never come with names, dates or documents. What the evidence actually supports is much simpler and much funnier. A small group of technically skilled people with a specific and slightly juvenile sense of humour pulled off a genuinely impressive act of engineering

in order to hum a cartoon song and get whacked with a fly swatter on public television and then, unlike literally every other person who has ever done anything like this, they shut up about it, forever. That is the whole trick. The technical part was the easy part, the discipline was the miracle, and that is what makes it the perfect opening for tonight, because it is a story about a signal arriving in exactly the wrong place

with nothing useful attached to it. A message got through, at scale, to millions of people, and carried no recoverable information whatsoever, which is a specific kind of failure. Now let me show you the opposite kind, where the message carried a great deal of information, all of it perfectly clear, all of it completely wrong, delivered to every human being in an American state simultaneously at 8 in the morning. Saturday, January 13, 2018. Hawaii. Ordinary weekend morning in a place that gets a lot of ordinary weekend mornings. At 807 in the morning, every mobile phone in the state went off at once with the specific shriek reserved for emergency alerts, the one you cannot silence and cannot ignore,

and the text on the screen said, in capital letters, that a ballistic missile threat was inbound to Hawaii, that recipients should seek immediate shelter and that this was not a drill. It also went out over television and radio, interrupting whatever was on with the emergency broadcast tone, which meant that within roughly 60 seconds, the entire population of the state, plus every tourist on every beach, had received the same instruction through every channel they owned. To understand why this landed the way it did, you need the context of the previous year. Through 2017, North Korea had conducted a series of missile tests, including launches that demonstrated ranges that made the Pacific extremely interesting to everyone living in it.

The rhetoric between Pyongyang and Washington had escalated to the point where, where the two heads of state were essentially trading insults in public like a rap battle with nuclear stakes. And Hawaii, which is roughly 4,500 miles from North Korea, sat in the middle of that ocean doing the math. So in December 2017, for the first time since the Cold War, the state reactivated its attack warning sirens. Actual air raid sirens tested monthly, statewide. Public officials went on television to explain what to do. The state calculated that if a missile were launched, detection and confirmation would eat up most of the flight time, leaving the public somewhere between 12 and 15 minutes to get inside and get low. The government published guidance, schools held briefings, hotels, updated

procedures, and every single resident of Hawaii spent the holidays quietly filing away the fact that if their phone ever buzzed with that particular message, they would have about a quarter of an hour. Which means that on the morning of January 13th, when the phones went off, nobody in Hawaii treated it as a glitch. They had been trained, formally and publicly, to treat it as real. That preparation, which was entirely reasonable and arguably overdue, is exactly what turned a software mistake into 38 minutes of statewide terror. What people actually did in those minutes is the part that stays with you. Highways became a problem immediately, because the official guidance was to get indoors, and a person doing 70 miles an hour on an elevated freeway is not

indoors and cannot become indoors. Drivers abandoned cars in traffic lanes and ran. A father in Honolulu filmed by a bystander lifted a manhole cover and lowered his daughter into a storm drain because a storm drain is underground and underground is what the guidance said and there is no other shelter option on a residential street in a state that never built fallout bunkers because it is a tropical archipelago and not a Cold War rust belt city. People went into bath-tubs and closets and stairwells, golfers on courses kept playing because a meaningful number of people simply refused to believe it, which is its own kind of correct. Hotel guests were funneled into basements and parking structures, the university sent students into interior corridors, and an enormous number of

people did the only thing left when there is no shelter within reach, which was to call and record messages for the people they loved, because if you have 12 minutes and nowhere to go, the last useful thing you own as a phone. There are recordings of those messages that have since been made public by the people who sent them, and I am not going to walk you through them in detail, because the value of this story is not in the anguish, it is in the mechanism. But I will say that the emotional content of that half hour was completely genuine. Whatever happened on the administrative side, the fear was real, the goodbyes were real, and a meaningful number of people in Hawaii spent 38 minutes fully believing they were in the final minutes of their lives. So here is the mechanism, and it is

so mundane that it is almost offensive. At the Hawaii Emergency Management Agency, at around 805, a shift change was underway. A supervisor decided to run an unannounced drill on the incoming day shift, which is in principle exactly what a competent emergency agency should do. Surprise drills are how you find out whether your people can actually perform under pressure rather than just perform on a form. The drill was executed by playing a recorded message over the phone line to the warning officers. That recording was bookended by the word exercise, repeated at the start and the end, which is the standard convention for making it unmistakable that a scenario is not real. In between those bookends,

however, the recording contained the language of an actual alert, because the point of a drill is to simulate an actual alert, and that language included the phrase that this was not a drill. You can already see the shape of the accident. One employee, by his own account, did not register the beginning of the recording, heard the middle of it, and concluded. included that an actual missile threat was in progress. Everyone else on shift understood it was an exercise. He did not. And this is where the interface enters the story as the true villain of the evening. The alert software presented a drop-down menu.

In that menu, the options were listed as short text labels, and the label for the drill version of the ballistic missile alert, and the label for the real, live, goes to every phone in the state version of the ballistic missile alert were nearly identical strings of abbreviations, sitting adjacent to one another in the same list, not colour-coded, not separated, not on different screens,

not behind different permissions. Two neighbours in a drop-down, differing by one word, one of which sends a test to a log file, and the other of which tells 1.4 million people to run. He selected the live one. The system, to its modest credit, then asked him to confirm,

but the confirmation was a generic prompt asking whether he was sure, which is the least useful safety mechanism ever designed. because a person who believes a missile is inbound is extremely sure. That confirmation dialogue was not protecting against a person who did not mean to click. It was standing in front of a person who absolutely meant to click and was wrong about the world. He confirmed. The alert went out, then came the part that turned a two-minute error into a 38-minute one.

Within about three minutes, the agency had confirmed with military command that there was no missile, no launch, no threat of any kind. So the state knew, almost immediately, that it had just terrified its entire population. The problem was that knowing and telling are two completely different systems, and only one of them had been built. The emergency alert infrastructure had a button for sending. It had no button for taking it back. There was no pre-written cancellation template, no approved retraction message sitting in the system ready to go, no procedure for what to do in the exact scenario that had just occurred, because nobody had planned for the possibility that the agency itself would be the emergency.

Cancelling required composing a new message, getting authorisation for it, and pushing it through the same federal alerting pipeline, and doing all of that while the building was on fire in a metaphorical sense, and the phones were ringing in a very literal one. Meanwhile, the fastest communication channels available were the informal ones. The agency put a correction on social media at around 8.20, which is 13 minutes in, and that only reach people who are on social media. rather than in a storm drain. A member of Congress from Hawaii tweeted that it was a false alarm

within about 12 minutes which helped anyone following her and nobody else, and the governor of the state who had been told within two minutes that the alert was false did not post a correction to his own social media accounts promptly and later explained why in a moment of accidental honesty that I think about constantly. He did not remember his login password. The head of state government sat on the single most urgent piece of information in the Pacific and could not deliver it because of a forgotten password, which is the most relatable failure in the history of public administration, and also, in that specific half hour, an actual public safety issue. The official correction, the one that went to every phone the same way the first one had, arrived at 845,

38 minutes after the alert. It stated that there was no missile threat or danger to the state and that it was a false alarm. The federal investigation that followed did not conclude that a bad employee ruined a good system. It concluded that the agency lacked reasonable safeguards, which is regulator language for the system, was arranged in such a way that this was going to happen eventually, and the only variable was who would be sitting in the chair. The findings included the ambiguous drill recording,

the absence of a two-person authorization requirement for the highest severity alerts, the interface design, and the total absence of any correction procedure. the employee who sent it was dismissed. It also emerged that this was not his first confusion between an exercise and a real event, which raises an obvious question about why he was still in that seat, and that question lands on management rather than on him. The head of the agency resigned, the employee's supervisor left as well, and the employee himself, who declined a formal federal interview but did speak publicly afterward, maintained that he had been completely convinced the threat was real, which is, if you accept it.

both an explanation and an indictment, because a warning system that depends entirely on one tired person correctly hearing the first three seconds of a recording is not a warning system. It is a coin flip with a government logo. The fixes were the obvious ones, and they arrived fast because nothing motivates a bureaucracy like international humiliation.

Two-person verification for missile alerts. Drill options clearly separated and labelled. Pre-written cancellation messages loaded and ready. A named officer with authority to issue a correction immediately without a committee. Cleaner interface design so that the two most consequential menu items in the state are not sitting next to each other looking like twins. And as if to prove the problem was never one guy in Hawaii, Japan's public broadcaster pushed a false missile alert of its own three days later, delivering the same shock to a country that is considerably closer

to the launch site. Two independent nations, one week, same category of failure. It was never about the individual. It was about the fact that we built systems capable of of speaking to millions of people instantly, and then attached them to interfaces designed with roughly the same care as a hotel booking form. There is one detail I saved because it summarizes the whole thing better than any analysis. Months before the alert, the agency had hosted press photographers, and a photograph published in the aftermath revealed a workstation with a password written on a sticky note, in view, in the frame, in a facility responsible for warning a state about incoming nuclear weapons. It was not the cause of anything that happened that morning.

It just tells you the ambient level of rigour, and it is the perfect companion image to a drop-down menu with two identical looking options in it. So put the two halves of this together, because they answer the same question from opposite ends. In Chicago, a broadcast reached millions of people flawlessly, and carried nothing anyone could use, and 38 years of investigation have extracted no identity from it. In Hawaii, a broadcast reached over a million people flawlessly and carried a message of maximum possible clarity that was entirely false, and the correction, which was known within three minutes, took 38 to arrive.

One of those was done on purpose by someone who has never been found. The other was done by accident by someone whose name we know, and whose worst moment was a menu, which brings us to the third failure in this block, and it is the smallest one. in the sense that it involved a single telephone call, one operator and nine minutes. The date is January 17, 2008, and the place is Northport, Florida, a town on the Gulf Coast that mostly exists in that pleasant Florida way,

where the streets are wide, the houses are new, and nothing happens. Denise Amber Lee is 21 years old. She has two small boys, a two-year-old and an infant, and she comes from a law enforcement family, which matters later in a way that is almost unbearable. because it means the people looking for her that night were not strangers to the process. They knew exactly how a search works. They knew exactly what should have happened.

That is the specific kind of knowledge you never want to have while your daughter is missing. That afternoon, a man named Michael King, 36, an unemployed plumber with a green Camaro, entered her home. When her husband came back, the door was open. The children were alone and unharmed, and Denise was gone. There is a version of this story that focuses on King, and I am not going to tell that version, because he is the least interesting person in it, and he has had enough airtime. He was arrested, convicted, and has spent the years since then in a Florida prison, which is where he will die.

That is all the space he gets. What matters is what happened on the phone lines over the next several hours, because Denise Amber Lee was not lost, she was not invisible, she was not in the trunk of an anonymous vehicle on an anonymous road with nobody watching. She was, for a stretch of that evening, one of the most reported people in South West Florida, and the information about where she was and which direction she was moving reached emergency services repeatedly from multiple independent sources in real time. The first source was Denise herself. At some point during the drive, with King briefly out of the vehicle, she got hold of his cell phone and called

911. Think for a second about the composure that takes. She is 21, she has been taken from her own house. She has a window of maybe a minute, and she uses it not to scream but to deliver data. She gives her name, she gives her address. She says she has been taken by a stranger. She's trying to hand the operator everything a search needs, and she's doing it while listening for the sound of a man coming back to the car. The call ends when he returns. It came from his phone, which meant that number was now in the system, and the technical process of pinging a cell phone to a rough tower location was already possible in 2008. Not the pinpoint blue dot you have on your screen right now, but good enough to say which stretch of which county to flood with patrol cars. The second

source was a relative of Kings. He turned up at a family member's property that evening asking to borrow a shovel, a flashlight and gasoline, which is possibly the single most alarming shopping list in the English language. The relative was uneasy enough about it that a call was made to 911 to report the behaviour, along with the detail that a woman's voice had been been heard from the car. So now emergency services had name, a vehicle, a location, a witness account, and a stated intent that no reasonable person could misread. Deputies were looking for that Camaro. Northport Police were looking for that Camero. The information was moving. The machine was, in the technical sense, working, and then came the third source, and this is the one

this chapter is actually about. Her name was Jane Cowleski. She had no connection to anyone involved. She was simply driving that evening on a busy road, stuck in the ordinary evening traffic of a Florida town, when the car next to her caught her attention. There was a hand slamming against the inside of the rear window. There was screaming coming from inside the vehicle, loud enough to be heard through two sets of closed windows in moving traffic, which should give you some idea of the volume. She saw a child's hand, or what she believed was a child's hand, and she did the thing that every public safety campaign in the world begs you to do.

She called 911 and then she stayed. She stayed on that line for roughly nine minutes. Nine minutes is a very long time to hold a phone to your ear while driving. She followed the vehicle. She described it. She described the direction of travel. She gave the road, the turns, the lane, the movement. She narrated, in real time, the exact position of the exact car

that two separate police agencies in that region were at that moment searching for. She was, functionally, a live tracking device that had volunteered. for the job. The call was answered. The call was recorded. The information was written down. The operator was calm, professional and unfailingly polite. And that information never reached a single patrol unit. Not one. It was never broadcast over the radio. It was never entered into the system in a way that put it in front of the officers driving around that night with the description of a green Camaro taped to their dashboards. There was no relay to the neighbouring jurisdiction actively running the search. The report Jane Kowalski spent night.

nine minutes assembling in traffic at personal risk arrived at a desk and stopped there. The failure is a stack of small things, which is always how these things go, and I want to walk through them, because this is not a story about one lazy person. It never is. Start with geography, which in American emergency services is less a physical fact than an administrative religion. Southwest Florida in 2008 was a patchwork of adjacent agencies with adjacent jurisdictions, each with their own dispatch center, their own radio channel, their own software, their own procedures and their own institutional pride. A car being driven at 50 miles an hour crosses the boundary between two of those kingdoms in about four minutes. A piece of information crossing the same

boundary required a human being to pick up a separate phone and call another building. There was no automatic sharing, there was no shared map, there was no unified incident record. The vehicle could move faster than the paperwork about the vehicle, which is a sentence that should never be true and was true constantly. Add the volume problem. A dispatch centre on a weekday evening is not a quiet room with one glowing screen. It is a person wearing a headset handling multiple lines, multiple radio channels, multiple screens and a queue that does not stop for anything, including the fact that the call currently in your ear happens to be the most important call of the decade. Every call sounds urgent to the person making it. That is what a 911 line is. The operator's job is to sort genuine emergencies

from a river of noise and the sorting mechanism is training and the training at the time was, in many places across the country, whatever your particular agency felt like providing. Which brings us to the actual scandal underneath this case. In 2008, in most of the United States, there was no mandatory certification standard to be a 911 operator. Let that sit for a moment. But hair professionally in Florida, you needed a license involving hundreds of hours of instruction. To be the single point of contact between a screaming stranger and the entire emergency response capability of the state, you needed to be hired. Training varied wildly from one centre to the next. Some agencies ran excellent programs. Others handed you a headset, sat you next to a veteran

for a couple of weeks, and wished you luck. There was no national flaw, no universal protocol, no required continuing education, and in a lot of places no formal quality review of calls after the fact, which means nobody was checking whether the information going in was coming out the other side. So the recording of that nine-minute call became something that had never really existed before, a perfectly preserved minute-by-minute document of an emergency system failing while sounding completely normal. Nobody yells on that recording. Nobody panics. There is no dramatic breakdown, no moment where you can point and see. say, there, that is where it went wrong. It is a calm conversation between a frightened citizen

and a professional receiving her report in an appropriate tone of voice, and at the end of it, everyone hangs up satisfied that the correct procedure has been followed. Denise Amber Lee was found two days later in a shallow grave off a boulevard on the edge of town. She had been killed the same night. Her family did something with that recording that I still find remarkable, because the ordinary human response to that tape would be to never listen to it again. Instead, her husband and her parents built an organisation around it, named after her, dedicated to a single, unglomerous cause, making sure that the person who answers the phone when you dial three numbers is actually trained to do it. Not more police, not tougher sentencing, not another

law about the crime itself, training standards for dispatchers, which is possibly the least marketable public safety cause imaginable, and also, in terms of lives per dollar, one of the most effective. The Foundation started running conferences and courses. They took the recording into training rooms and played it for working dispatchers, which is a brutal thing to sit through, and also the only thing that actually changes behaviour, because a policy memo has never once made anyone care, and nine minutes of audio does it in one sitting. They pushed for legislation in Florida requiring public safety telecommunicators to be certified, trained and held to a standard, and it passed. The model spread. Agencies across the country adopted formal call review,

meaning someone now listens back to a sample of calls specifically to check whether the information collected actually made it to the field. The technical fixes moved as well. Modern dispatch increasingly runs on systems where a call takers note populate a shared record that neighbouring agencies can see, where location data flows from the phone rather than from the caller's ability to read street signs while pursuing a kidnapper and where a report involving a vehicle in motion generates an alert across jurisdictional lines

instead of dying in one building. None of this is exotic technology. It is the same category of infrastructure that lets you split a dinner bill with a stranger's phone in three seconds. It simply took a case like this one to make anyone spend money on it. There is a detail here that I keep circling back to.

Jane Kowalski did everything right, every single thing. She noticed which most people do not. She acted, which fewer do. She followed, which is genuinely brave. She reported clearly and stayed on the line for the length of an entire pop album while driving through evening traffic. If you designed a training video for the ideal witness, you would be describing her behaviour beat for beat.

And it made no difference at all because the last link in the chain, the invisible one, the one no public can ever mentions, did not close. That is the uncomfortable thing sitting under this whole story. Every campaign tells you that if you see something, you should say something. Nobody tells you what happens after you say it, and the answer is that it depends entirely on the internal procedures of an office you will never see, staffed by people you will never meet, using software you cannot evaluate. You're trusting a black box, and until this case, almost nobody

outside the industry had any idea how variable the contents of that box were. And so the worst piece of audio in this entire story is not the screaming. It is not the abduction. It is a composed professional voice confirming that yes, the report has been received. Everything has been noted, thank you for calling. Received and noted. The information existed in full, in the right building at the right time, in the hands of the right agency. And it never crossed the last 12 inches between a desk and a radio, which is a strange kind of horror, because it has no villain in it at that point, just a gap where a procedure should have been. Now let me take you to the opposite extreme of that same problem, where the information was correct, the warnings were loud,

the scientist were right, and the failure was that human beings simply refused to accept what was about to happen to them, up to and including a man who told the world he was not leaving his house, and then did not leave his house. Twenty-eight years earlier, on the morning of May 18, 1980, in Washington State, the largest landslide in recorded history was was about 10 seconds away from starting. Mount St. Helens had been putting on a show since March. A magnitude 4 earthquake in the middle of that month announced the reawakening of a volcano that had been quiet since the 1850s, and by late March it was venting steam and ash and had opened a crater at the summit. Thousands of small earthquakes followed, and then it started doing

the thing that made every geologist in the region deeply, specifically nervous. The north flank of the mountain began to swell, not a little, not subtly. That side of the volcano bulged outward at a rate of roughly five feet per day, pushing out a total of well over 400 feet. Magma was rising into the mountain and physically inflating it, and any structural engineer, geologist or frankly any person who has ever overfilled a balloon can tell you where this ends. One of the scientists monitoring it described the situation as a keg of dynamite with the fuselit, which is not the kind of phrasing government scientists used casually, and he meant it, and he was completely correct. The man who said it was David Johnston, 30 years old, a volcanologist with the United States

Geological Survey. He specialised in volcanic gases, which meant his job was to climb into the crater of an actively inflating volcano and take samples, an occupation that makes deep sea welding look like a desk roll. He was by every account energetic, funny and enormously popular with the press, because he was the rare scientist who could explain a hazard in plain language without either dumbing it down or putting the audience into a coma. He was also the loudest voice arguing that the area needed to stay closed and this is where the story stops being about geology and starts being about people, which is where it always goes wrong.

Because there was enormous pressure to reopen access. The mountain was surrounded by land that people used. There were vacation cabins around Spirit Lake. There was a working timber industry with crews and equipment. there were property owners who wanted their belongings, tourists who wanted photographs, and a general public sentiment that the mountain had been rumbling for two months without doing anything dramatic, so how bad could it be? The restricted zone that was eventually drawn around the volcano was shaped in part by land ownership boundaries,

rather than purely by hazard modelling, which is an approach to public safety roughly equivalent to deciding the height of a floodwall based on where the property lines are. As it turned out, most of the people who died that day were, outside the restricted zone entirely. The day before the eruption, authorities escorted a convoy of cabin owners in to retrieve possessions, and a second trip was scheduled for the morning of May 18th. The eruption arrived first, and the fact that it happened on a Sunday morning is the only reason the death toll is measured in dozens rather than hundreds, because the logging crews

were not at work. That is the entire margin, a calendar. The most famous holdout was Harry R. Truman, 83 years old, who ran a lodge on the shore of Spirit Lake and had lived there for over half a century. He had no relation to the president and enjoyed the confusion enormously. He refused, publicly and repeatedly to evacuate, on the grounds that the mountain was his, he knew it better than any scientist, and if it went, he would go with it. He gave interviews. He got fan mail from schoolchildren. He became a national folk hero in about six weeks, the charming old man thumbing his nose at authority, and he did it while living with 16 cats, which is a detail that would be unbelievable in fiction.

There was a genuine argument in the press about whether he was brave or foolish, and the honest answer is that he was an old man who did not want to leave his home and had accidentally acquired an audience that made backing down impossible. He was never seen again. His lodge is under roughly 150 feet of debris and water. At 8.32 in the morning, a magnitude 5.1 earthquake hit directly beneath the volcano, and the bulging north face, which had been held together by nothing but optimism for weeks, let go all at once. The entire flank of the mountain slid off, two-thirds of a cubic mile of rock,

moving downhill at well over 100 miles an hour. And because that rock had been acting as a lid on top of pressurized magma and superheated groundwater, removing it did not just cause a landslide, it uncorked the mountain sideways. The lateral blast came out of the north face at hundreds of miles per hour, carrying rock, ash and gas at temperatures around 600 degrees Fahrenheit. It did not go up first and then spread. It went out, low and fast, hugging the terrain, and it flattened 230 square miles of forest. Old growth trees three feet thick were snapped off at the base and laid down in neat parallel rows, all pointing away from the mountain, like grass combed by an enormous hand. The pressure wave outran its own sound in places, which produced the peculiar situation of the

glass being clearly audible, hundreds of miles away in other states while being nearly silent in a quiet zone close to the volcano. The summit lost 1,300 feet of elevation in a matter of minutes. The ash column reached 15 miles into the sky in under a quarter of an hour, and the debris that mixed with snow and ice became lahars, which are essentially rivers of wet concrete, and those poured down the valleys and took out bridges, roads, logging camps and hundreds of homes. David Johnston was at an observation post about six miles north of the south. summit, on a ridge, with a clear line of sight to the bulging face. He had taken that shift from a colleague who had to leave for a graduate school interview, which is one of those details that

reorganises your understanding of luck. That colleague survived that day and later died at a volcano in Japan in 1991, doing the same job. This field has a body count and everyone in it knows it. When the flank collapsed, Johnston got on the radio to the survey office in Vancouver, Washington, and transmitted one line. Vancouver, Vancouver, this is it. Then the connection ended. The blast reached him in seconds. His body was never recovered.

Remnants of his trailer were found more than a decade later during road work. There is a genuine scientific value to that transmission, by the way, and it is not sentimental. He was watching the initiating event from the ideal angle, and his call is the timestamp that anchors the sequence. He worked until the last available second, and the last thing he did professionally was tell the office that the model was correct.

Another man on a ridge above him. A volunteer with ham radio equipment who had been reporting positions all morning, watched the wave of debris cross Johnston's position, and described it over the air before it reached his own. He did not survive either. What he transmitted became part of the record. Then there is Reed Blackburn, 27, a photographer for a local newspaper who was working a camp assignment

involving remote cameras positioned to capture exactly the kind of event that was now happening. He was about eight miles out which everyone involved considered a comfortable distance, because the modelling of how far a lateral blast could travel was based on volcanoes that had erupted upward like polite, well-behaved mountains. He was found in his car, buried in ash. The film in his cameras was destroyed by the heat, which is its own small tragedy, because he had positioned himself perfectly and got nothing. And then there is Robert Landsberg, and this is the one I would like you to sit with.

He was 48, a photographer, and he was about four miles from the mountain that morning, closer than anyone else who left a record. When the North Face collapsed, he had a clear view of it, and he understood immediately what was coming and that he had no chance of outrunning it. There was nowhere to go. Four miles is nothing to a pyroclastic surge. So he stayed and photographed it. He shot the collapse and the approach.

cloud as it filled the frame and then filled the sky. And then, with the ashwall closing on him, he did something that required a completely different kind of thinking. He rewound the film into the canister. He put the camera into his backpack. He lay face down on the ground on top of the pack, using his own body as shielding so that the film would survive the heat and the abrasion even though he would not. He was found 17 days later buried in ash with a backpack beneath him. The film was intact. The photographs would be. developed and published, and they are exactly what he intended them to be, a usable visual record of the leading edge of the blast taken from inside it. That is not a man taking a picture. That is a man

performing an act of archiving with about 90 seconds of life remaining, on the assumption that somebody would eventually come looking, and that his work would matter to them. Everything about that decision is rational, and everything about it is devastating, and both of those are true at the same time. 57 people died on May 18th. Loggers, campers, scientists, photographers, a couple with their car, an old man with his cats. Most of them were beyond the line where they were supposed to be safe. The eruption rewrote the hazard models for lateral blast worldwide, which is a real and permanent benefit and also a phrase that means those models were paid for with a specific number of specific people. Here is the part that connects to everything else tonight. All of that material is public.

Johnston's transmission, Lansberg's photographs, the photo sequences shot by a survivor 11 miles northeast, who caught the flank collapsing frame by frame, and lived because of a lucky ridge, the survey logs, the radio recordings. It is in federal archives and university collections and for the most part on the open internet. Free, indexed, and about four clicks away from anyone with a phone. You can, right now, in the time it takes to microwave something, look at the final photographs taken by a man who was about to be killed by the subject of the photographs. No warning label, no ceremony, no gatekeeper, just a file in a public archive, sitting next to the metadata, exactly as available as a stock photo of a golden retriever. That is not a criticism of the archives. Those images exist because

Lansberg wanted them to exist, and paid the maximum price to make sure they would, and hiding them would be its own kind of insult. It is a comment on what it means that the last seconds of a stranger's life, are now simply content that loads instantly, with a download button, and that we have all completely stopped noticing how strange that is, which is the perfect place to leave the question of signals, and move to the question of pictures, because everything from here on begins with an image that somebody found, and nobody could explain. June 15, 1989, Port St. Joe, Florida, a small town on the Gulf Coast where the main industries are fishing, paper and heat. A woman pulls into the parking lot of a convenience store to do the most ordinary thing a human being can do,

which is buy something at a store, and while she is there, she notices a white cargo van parked nearby with a man inside, and what appear to be children with him. Something about it registers as slightly off, though not off enough to do anything about, because that is how these moments always work in real life. Nobody hears a musical sting. She goes about her business, the van pulls away and lying on the ground where it had been parked is a polaroid. She picks it up, and this is the moment where an unremarkable Thursday in Florida turns into one of the most argued about images in the history of American missing persons cases. The photograph shows two people lying on a pillow side by side in what looks like the cargo area of a van.

On the left, a young woman, maybe late teens or early 20s, dark hair, eyes open and squinting slightly at the camera. On the right, a younger boy. Both of them appear to have their hands bound behind them. Both appear to be gagged with tape. Neither of them looks like they are having a good time, which is possibly the most restrained sentence I will produce today. And next to the young woman, positioned in the frame in a way that photography nerds will tell you

is either meaningless or extremely meaningful depending on the day, is a paperback book. The title is legible enough to identify. It is My Sweet O'Drina by V.C. Andrews. that book in your mind for a moment, because it is the hinge of this entire story. Nine months earlier, and roughly 1,500 miles away, on the morning of September the 20th, 1988, a 19-year-old named Tara Calico left her home in Belen, New Mexico, to go for a bike ride. This was not unusual. It was her routine, a long out-and-back run along a state highway

that ran south from town, around 30-odd miles round trip, which is a serious ride, and tells you something about her fitness and her stubbornness. She was a university student, she had a boyfriend, a tennis game scheduled for that afternoon, and a completely normal life with completely normal plans. She was riding her mother's bike that day, a pink mountain bike, because her own had a problem. Before leaving, she told her mother that if she was not back by noon, come looking for her. This was not a premonition, and I want to be clear about that, because true crime storytelling telling loves to turn ordinary caution into prophecy. She had been followed and hassled by vehicles on that road before. It was a rural highway with fast traffic and a young woman riding alone on it,

and she had developed the same low-grade situational awareness that basically every woman who exercises outdoors develops by the age of 16. She was being sensible, not psychic. Noon came, she did not. Her mother drove the route. What she found was nothing, which in a search is the worst possible result, because nothing gives you nowhere to start, no bike, no bag, no sign of a crash, no skid marks, no blood. Later, pieces of a portable cassette player and a tape were recovered along the road, some distance out, which suggested a struggle or at least a fast departure, though even that has been debated for decades because roadsides in 1988 were absolutely littered with broken consumer electronics. Witnesses reported that a light-coloured pickup with a camper shell had

been seen on that stretch of road, and some reported it appeared to be following her. The vehicle was never conclusively identified. The bike was never found. Tara Calico simply stopped existing at some point along a straight, flat, highly visible rural highway on a bright morning, and the search that followed produced volume without producing information. So when a Polaroid turns up in Florida nine months later showing a bound young woman who looks broadly like her, lying next to a copy of her favourite novel, you can understand why the temperature in the room changed. Because that book detail is genuinely uncomfortable. My sweet o'drina was Tara's favourite book, not a book she owned, her favourite, her family confirmed it,

and if you are constructing a probability argument, the presence of a specific paperback next to a woman who resembles a specific missing person is not nothing. Except this is exactly where the entire case turns into a lesson about how images work on the human brain. and how badly we all overestimate our ability to read them. The photograph made its way to law enforcement, which took a while because the initial reception in Florida was not exactly a full mobilisation. This was a photo found in a parking lot in a town of a few thousand people,

and the local response was closer to filing it than to launching a manhunt. Eventually it reached investigators in New Mexico and then Tara's family, and once her mother saw it, the case acquired the fuel it has been running on ever since. She was certain. She looked at that image and saw her daughter, and she spent the rest of her life saying so. And here is where I have to introduce the most important technical fact in this entire chapter, and it is the one detail that keeps the photo alive as evidence rather than dismissing it outright. The Polaroid was shot on a specific type of instant film that did not exist in stores until roughly May of 1989. That is eight months after the disappearance, which means, whatever else is true, this photograph could not

have been taken before Tara Calico vanished. It was taken after. Somebody, somewhere, in the spring or early summer of 1989, taped up two people, arranged them on a pillow in the back of a vehicle, and took an instant photo of them, and then that photo ended up on the ground in a Florida parking lot. Sit with that for a second, because it is the part that tends to get lost in the debate over whether the woman is Tara. Even if it is not her, this is still a photograph of two bound people, there is no innocent interpretation of it that anyone has ever produced. No prank explanation has ever surfaced. No photographer has come forward to say it was a staged Halloween joke gone weird. No family has ever identified the pair as their own kids messing around.

In more than 35 years, with the image published nationally, in newspapers, on network television and eventually across the entire internet, nobody has ever identified either person in it. Two human faces distributed to hundreds of millions of people and zero confirmed names. Now, the analysis. This is the part I actually want to spend time on because it is where the chapter earns its place. The image was examined by multiple expert bodies and the results were not just different. They were opposite. Scotland Yard in London examined it and concluded that the woman in the photograph was probably Tara Calico.

Probably. That is a strong word coming from a police service that. does not enjoy being wrong in public. Los Alamos National Laboratory in New Mexico, an institution best known for physics of a considerably more dramatic kind, ran their own analysis. Their conclusion was that it could not be determined, inconclusive. The image did not contain enough usable information to say yes or no, and the FBI examined it and concluded that the woman was not Tara Calico. Three serious institutions, one photograph, three answers. Probably yes, cannot say, and no. If you have ever wondered why courtroom experts contradict each other with

such confidence, this is a beautiful little case study, and unusually nobody here was being paid to reach a preferred conclusion. So how does that happen? Because this is not incompetence. It is the nature of the material. First, resolution. A Polaroid is a small print, and the faces in this one occupy a tiny portion of it. Blow that up and you're not revealing hidden detail, you are magnifying grain, There is a persistent belief, installed in the public imagination by every crime show ever made, that images contain more information than they appear to, and that a technician can extract it by leaning forward and pressing a key. That is not how any of it works. An image contains a fixed amount of information. Enhancement redistributes what is there and makes it easier for a human eye to interpret. It does not add facts. When you sharpen a blurry face, you are not to discovering the face. You're making a guess look crisp, which is arguably worse than leaving it

blurry, because crisp guesses are persuasive. Second, and this is critical, the version everyone argued about was in many cases not the original print. It was a photograph of a photograph, reproduced, copied, faxed, printed in newspapers, refotographed off television screens, scanned, uploaded and re-uploaded. Each of those steps throws away detail and adds artifacts. By the time the image reached most of the people forming strong opinions about it, they were looking at a degraded copy of a degraded copy, and confidently identifying a human being from it. Third, the face itself is doing you no favours. The woman's eyes are squinting, likely from a flash, her mouth is covered, her head is at an angle, and her hair falls across

part of her face. Facial identification depends heavily on relationships between fixed landmarks, And a good half of the useful landmarks here are either obscured or distorted. Add the fact that human faces in that age range are statistically, not as unique as we like to believe, and you have a situation where an honest analyst can look at the same pixels as another honest analyst and reach a different conclusion without either of them doing anything wrong, and then there is the body of evidence that is not the face. The strongest physical argument for identification was a mark on the woman's leg

that appeared consistent with a scar Tara had from a childhood accident. That is the kind of detail that feels decisive and is not, because a low-resolution smudge on a leg in an instant photograph can be a scar, a shadow, a fold in fabric, dirt, or an artifact of the film itself. Everyone looking at it already knew what they hoped to see, which brings us to the actual invisible force running this entire case. Confirmation bias is not stupidity. It is the default operating system of the human visual cortex. Your brain does not passively receive images. It actively predicts. what should be there and fills gaps with the most plausible candidate. This is why you see faces in electrical outlets and animals in clouds. It is also why, if you show a grieving parent a blurry photograph and tell them it might be

their missing child, their perceptual system will do everything in its considerable power to make that identification succeed. That is not weakness. That is neurology. And it happens before conscious thought gets a vote, which means the entire national conversation about this photograph has been, from day one, hundreds of thousands of people running their pattern-matching hardware on an ambiguous stimulus and then reporting the output as an observation. The boy in the photograph deserves his own section, because his story is the clearest illustration of how this works. Almost immediately people connected him to another New Mexico case, a nine-year-old boy who had gone missing earlier that same year in the spring of 1988 in a completely different part of the state.

two missing children from New Mexico, one photograph containing two bound young people, obvious conclusion, and the theory took off like it had been fired from something. Then, in 1990, that boy's remains were found in the mountains where he had gone missing. He had become separated from his family group during a hunting trip, and had died lost in the wilderness. No abduction, no van, no connection to Florida or to Tara Calico or to anything else in this story. It had always been a coincidence of geography and timing, and it had been treated as a lead by serious people for the better part of two years.

That should have functioned as a permanent warning label on the whole case. It did not because nothing ever does. What happened next is the thing this chapter is really about, which is the photograph acquiring a life independent of the investigation. Tara's mother, Patty Dole, took the image and the case onto national television, unsolved mysteries, daytime talk shows, tabloid news programs, the full 1990s media service, circuit. And this worked exactly as intended in one respect, which is that tens of millions of

people learned Tara Calico's name and saw her face. That is not nothing. That is how cases stay alive. But it also did something nobody plans for. It converted the photograph into a piece of pop culture. It became one of those images that circulates permanently, detached from the case, showing up in listicles and compilations and late-night videos, right alongside blurry crypted photos and haunted paintings. It became content. And once an image becomes content, it stops being evidence in any practical sense. Evidence sits in a file with a chain of custody.

Content gets cropped, recolored, mirrored, captioned, memified, run through whatever enhancement filter is trending this year, and reposted by accounts whose entire relationship to the case is that the picture performs well. Every single one of those passes degrades the original further and introduces new artefacts that the next round of amateur analysts

will treat as meaningful detail. I have seen versions of this photo online where the contrast has been cranked so hard that features appear which are not in the original at all. And people analyse those, confidently, in the comments, at length. The family paid for that visibility in the way families always do,

the attention brought tips, and the tips brought hoaxes, and the hoaxes brought phone calls, and eventually the household was fielding an ongoing stream of sightings, psychics, theories, cruel pranks, and strangers with certainty. They eventually left New Mexico entirely, in part to get away from it. Patty Dole died in 2006 without ever knowing what happened to her daughter, having spent 18 years being told by people on television that a photograph either did or did not show her child. Two other Polaroids surfaced over the over the years, incidentally, both showing bound or gagged women, one found at a construction site in California and one reportedly found on a train in Texas. Both got a brief spike of attention.

Neither was ever connected to this case or to any case at all, and the most likely explanation is the least satisfying one, which is that a country of 300 million people generates a certain background rate of disturbing photographs, and once the public knows to look for them, they start getting found and reported. Pattern recognition again, running at national scale. Meanwhile, the actual investigation went somewhere the photograph completely obscured, because for years law enforcement in Valencia County, New Mexico, has held a very different theory of this case, one that does not involve Florida, vans, cross-country abductions or instant film at all. In that theory, Tara was struck by a vehicle on that highway that morning, by local men who knew

her and had been harassing her, and what followed was a panic and a cover-up rather than a planned abduction. A sheriff went public with a version of this, stating that he believed he knew who was responsible, and that the obstacle was evidence, not identity. No charges have ever been filed. The people involved have never been publicly named as suspects in a formal sense, and every element of it remains unproven. Whether or not that theory is correct, notice what it does to the photograph. If it is correct, then there is a matter of the fact that you are The picture is a horrifying but unrelated document of two entirely different people, and 35 years of public attention has been aimed at the wrong image in the wrong state.

All that analytical energy, all those television segments, all those threads, would have been spent on a coincidence, and the case would still be unsolved either way, which is the part I find hardest to get past. Periodically there are new developments, or claims of them. Enhanced versions of the image get released with requests for public help, and of Conversaries generate coverage. Podcasts revisit it. Sometimes with new interviews, sometimes with recycled speculation dressed as new interviews. Every few years someone announces a breakthrough, and every few years the breakthrough turns out to be a lead rather than a resolution. The case file remains open. Nobody has been charged with anything. Tarah Calico has never been

found. So what did the photograph actually accomplish in 35 years? It kept the name alive, which is genuinely valuable, and I will not be cynical about it. that. Cases die from silence, and this one never got the chance. It proved absolutely nothing forensically about Tara Calico, and there is a real argument that it consumed investigative attention and public bandwidth that a small rural sheriff's office did not have to spare. And it demonstrated, better than any lab experiment could, that a photograph is not a fact. We treat images as the highest form of proof, the thing that settles arguments, the reason people say picks or it did not happen.

But an image is just a distribution of light values captured at one moment, and the meaning we extract from it is manufactured almost entirely inside our own heads, using assumptions, hopes and prior beliefs as raw material. Three of the most capable institutions in the world looked at the same rectangle and produced three incompatible readings, and none of them were lying. The other thing worth stating plainly is that there are two people in that picture who have never been identified in either direction. Not confirmed as victims, not confirmed as safe, not confirmed as anything.

If the woman is not Tara, then she is someone and nobody has ever come looking for her by name. That is arguably the most disturbing possible outcome and it is the one that almost never gets discussed because the entire conversation has been organised around a single question with a single name attached to it. An image arrived in perfect condition carrying real information about real people. people, and after three and a half decades of maximum attention, it has yielded no names, no location, no timeline, and no charges. The picture got through. The meaning never did. Which raises an obvious next question, and it is the one the internet has been trying to answer ever since it acquired the ability to organise. What happens when an image arrives that is

deliberately built to be decoded, and thousands of people decide to solve it together, all night at scale? of 2015, a video appeared online with no explanation, no channel identity worth speaking of, and a title that looked like a serial number rather than a title, 1-1B-X-1371, just under two minutes long, black and white. And within about a week it had become the most intensively analysed piece of amateur video on the internet, which is a competitive category. The footage itself is exactly what you would design if your goal was to make a person's shoulders rise toward their ears. A figure stands in an abandoned building, filmed in grainy monochrome. The figure is wearing

a plague doctor mask, which is the long beaked leather thing that actual 17th century physicians wore stuffed with herbs because they believe disease travelled through bad smells. A theory that was wrong, but did accidentally result in the most terrifying piece of headwear ever produced by the medical profession. It has been doing heavy lifting and horror imagery ever since, and it works every single time, which is why the costume industry sells roughly 9 million of them a year to people who want to skip the effort of being scary and go straight to the result. Don't you wish you could just hit skip on the worst parts of your life? You know the same way you can skip an ad?

I get it. I'm Siaya and I live in Ice Cove. I've made some questionable decisions that didn't end up the way I planned. And today I'm still figuring it out. Somehow things usually get worse before they get better. Apparently, that's how I'm. roll. So bundle up and come along for the bumpy ride. Stream a new episode of North of North Tuesdays on CBC Gem.

When WestJet first took flight in 1996, the vibes were a bit different. People thought denim on denim was peak fashion, inline skates were everywhere, and two out of three women rocked, the Rachel. While those things stayed in the 90s, one thing that hasn't is that fuzzy feeling you get when WestJet welcomes you on board. Here's to WestJetting since 96. Travel back in time with us and actually travel with us at westjet.com slash 30 years. The figure moves through the building with that stuttering frame-dropped quality that reads as unnatural to the human eye. It holds up sheets of paper toward the camera. It gestures. It walks. Behind everything is a layer of harsh audio that sounds like ordinary interference,

the kind of hiss and buzz you would expect from a damaged recording, and most viewers would file that under bad equipment and move on. That noise was the actual content of the video. Everything on screen was the packaging. Here is what people found once they started pulling it apart and the sequence matters, because it explains how a normal internet weekend turned into a global forensic project. First, somebody looked at the video description and noticed it contained a long string of characters that were not random. They were hexadecimal, which is just a compact way of writing numbers that computers find convenient and humans find ugly, and every programmer on Earth recognizes it

on site. Convert it to text and you get readable English. That text included a short, pleasant message, and, more interestingly, a set of geographic coordinates. Those coordinates pointed at Washington, D.C., in the immediate vicinity of the White House. At that point, the video stopped being a curiosity, and started being the kind of thing people forward to each other with no comment. A masked figure, a hidden message, and a map reference to the most photographed building in American politics is not a combination that goes unnoticed, and reports circulated that federal attention had at least glanced in its direction, which, honestly, would have been negligent otherwise. Second, and this is the part that made the video famous, somebody ran the

audio through a spectrogram. I need to explain this properly because it is genuinely one of the most elegant tricks in digital media, and most people have no idea it exists. A spectrogram is a visual representation of sound. Time runs along the horizontal axis, frequency runs up the vertical axis, and brightness shows how much energy is present at each frequency at each moment. It is a standard tool. Every audio engineer has one open right now. Its normal use is diagnostic, letting you see a hum or a click that you can hear but cannot locate. The trick is that the process runs both ways. If a picture drawn in a spectrogram produces a specific pattern of sound, then you can start with the picture, work backwards and generate an audio file

that contains it. The result sounds like harsh a tonal noise to a human ear because it is not designed for a human ear. It is designed for a graph. Play it and it is garbage. Look at it. And there is an image sitting inside your speakers. This is not a new technique and it is not exotic. Musicians have been hiding things in spectrograms since the late 1990s, most famously a British electronic producer who buried a distorted image of his own face in a track, which is a flex of the highest order. It shows up in game soundtracks, in album Easter eggs, in marketing campaigns. So when people ran this particular audio through the analysis, the noise resolved into pictures, skulls, symbols, text, human figures, including an image of a woman who appeared to be bound,

which landed with a very specific weight given that this was a video already trading in implied violence. There were short passages of Morse code embedded in the track as well, which decoded to brief phrases and more coordinates and additional strings that pointed onward rather than concluding anything. Note what this means about the production. Somebody built this backwards. They started with images, encoded them into audio, laid that audio under a video, and shot the video in a location that would maximize the effect. That is not a random creepy upload. That is planning in multiple technical disciplines, executed by someone who understood exactly how the internet would react to each layer as it was peeled, and the internet did react at genuinely astonishing speed.

Within days there were collaborative documents, threads with hundreds of contributors, independent verification of decoded strings, people arguing about which spectrogram window size produced the cleanest rendering, people with actual signal processing backgrounds volunteering their time, people geolocating the building from the architecture and window frames, people cross-referencing the coordinates against maps. It was, in structure, a real investigation. It had methodology, it had peer review, in the sense that a bad decode got shredded within 20 minutes by three strangers. It had specialisation, with the audio people and the imagery people and the location people operating in parallel. The total labour

involved would cost a serious amount of money to purchase. It was performed for free at night by people who mostly had work in the morning. And what did all that effort produce? Not an answer, more layers. Every solved element pointed at another encoded element. Decode the hex, get coordinates, analyze the audio, get images and morse. Follow those, get further strings and references. The structure was not a lock with a key. It was a hallway of doors, each opening onto another hallway, which is the oldest and most effective design pattern in puzzle construction, because it converts curiosity into an unpaid subscription. Then, after the case had built up a substantial mythology, a man came forward and said it was his. He gave a name, produced supporting material,

explained the piece as an art project with themes he was willing to discuss at length, and released follow-up work, and the reaction from the audience was not relief or applause. It was disappointment, followed immediately by denial. Large numbers of people simply refuse to accept it. He must be a copycat claiming credit. He must be a front. The real thing must be underneath. Even now, if you go looking, you will find people arguing that the stated authorship is itself part of the piece, which is unfalsifiable and therefore extremely convenient. That reaction is the actual subject of this chapter, because it tells you something uncomfortable about what the audience wanted and it was not the truth. Consider the economics.

An unsolved mystery as an asset. It generates content indefinitely. It supports videos, threads, podcasts, reaction clips, deep dives, debunks and debunks of debunks, forever, at no cost to anyone, because there is no ending to spoil. A solved mystery is a depreciating asset

with a fixed expiration date. The moment a thing is explained, its traffic collapses. Every single person producing content around a mystery has a quiet, structural interest in that mystery remaining open, and I include myself in that,

which is why I am telling you rather than pretending otherwise. The audience has the same interest, just less consciously. Solving is not actually the goal. Solving is the story we tell ourselves about why we are up at 3 in the morning adjusting the contrast on a spectrogram. The real reward is the searching, which is why the announcement of an author felt like a theft rather than a gift.

Now the part I actually want to make sure lands, because it changes how you look at everything else tonight. That video was frightening in the way a haunted house is frightening. It was constructed from the ground up out of pre-tested components. A mask with four centuries of built in dread. An abandoned building, which is a location that produces unease automatically because human architecture without humans in it reads to our brains as evidence that something went wrong. Monochrome, because colour implies the present, and grey scale implies archive footage,

and archive footage implies this already happened and nobody stopped it. Dropped frames, because our visual system is exquisitely tuned to human movement, and any deviation from natural motion sets off a life. arms without conscious input. None of that is supernatural. It is a professional understanding of perception applied with skill, and the encoded layers work the same way. They exploit the fact that a hidden message reads to our brains as a message someone did not intend for us to find, which carries far more weight than the same content stated openly. Put a skull on screen, and it is a graphic. Bury it in a spectrogram and it becomes a secret, and secrets feel true

in a way that statements never do. The genuinely disturbing element in this story is not the plague doctor. It is the mechanism the video exposed, which is that thousands of technically capable people can be mobilized within hours at no cost by anyone who is willing to construct a puzzle and that this crowd will apply real expertise with real rigor to material whose meaning may be entirely arbitrary.

Because here is the trap sitting underneath all of it. A sufficiently determined group of humans will find structure in almost anything. Give a thousand smart people a random string of numbers and tell them it is a code and within a day you will have several elegant, internally consistent, completely wrong decryptions, each defended by someone who is not stupid

and has done actual work. Numbers can be read as coordinates, dates, ASCII, phone numbers, page and line references or frequencies. With enough interpretive freedom, every number is meaningful. This is the same faculty that reads shapes into blurry photographs, running on numbers instead of pixels, which means collective decoding has a specific failure mode. It cannot reliably distinguish between a message and an inkblot.

The process is identical from the inside. You look, you find pattern, you get the dopamine hit, you tell the thread. Nothing in the experience alerts you to whether the pattern was put there or generated by you. There is a much larger and stranger example of this that ran for years, in which a series of extraordinarily sophisticated puzzles appeared online involving image steganography, book ciphers, prime numbers and physical posters taped to walls in cities around the world. Thousands of people worked on it.

Some reportedly reached the end and were never heard from publicly again. Nobody has ever confirmed who ran it or why, and the leading theories range from a recruitment operation to a private community to an extremely committed group of enthusiasts. The point is that the puzzle worked exactly the same way, regardless of who made it. The engine does not care about the driver, and this is where I want to be careful and fair, because there is a version of this chapter that ends with a lecture about gullible people online, and that would be both smug and wrong.

The crowd that took apart this video was not gullible. It was, in the technical sense, correct. There genuinely were hidden images in that audio. There genuinely was encoded text in that description. Every major decode was real and reproducible, which is the actual test. The crowd did not hallucinate the content. It found what was there, quickly, and verified it. The error was not in the analysis.

It was in the assumption underneath the analysis, which was that a puzzle this well built must be attached to something real, that the effort implied a purpose beyond effect, that nobody would go to that much trouble just to make something unsettling. But people absolutely do go to that much trouble. artists do it, filmmakers do it, marketing departments do it aggressively and with budgets, and there is a whole tradition of promotional campaigns that hid clues in album artwork, on physical objects and in websites, sending fans on genuine hunts that ended in a product launch.

Individuals do it because building an elaborate thing and watching the world take it seriously is one of the most satisfying experiences available to a person with time and skills. The gap here, and it is the same gap as everything else in this episode, sits on between what a message contains and what a message means. The transmission was flawless. The content arrived, intact, and was correctly extracted by strangers with no instructions. What did not arrive was the context, and without context an audience will manufacture its own, and it will manufacture the most dramatic one available, because the dramatic interpretation is the one that feels proportional to the effort of decoding. So the honest answer to what scared us in 2015 is not

the bird mask. It is that a single person with a camera, a decent understanding of audio processing and a location with peeling paint, managed to command the sustained attention and specialised labour of a global crowd, and could have pointed that crowd anywhere, at a real address, at a real person, at a name. That capability did not go away when the video was explained. It is standing right there, permanently available, and the only reason it does not go badly more often, is that most people who possess it are fortunately, more interested in making art than in making trouble. The next case in this block is the exact inverse of that, and it is my favourite kind of story, an account that uploaded tens of thousands of videos over years,

watched by nobody, meaning nothing, with an explanation so boring that when it finally arrived, a significant portion of the internet refused to believe it out of sheer disappointment. The channel was called Webdriver Torso, which sounds like either a fitness product, and or a minor transformer, and it had been quietly operating since 2013 without a single human being paying attention to it, which was arguably the point. What it uploaded was this, 11 second videos, each one consisting of a plain white background with a blue rectangle and a red rectangle on it. The rectangles change size and position. Underneath, a single steady tone at a different pitch each time, that is the entire creative output, no titles that meant anything, just strings of

random characters, no descriptions, no thumbnails worth looking at, no tags, no commentary, no playlists, no attempt whatsoever to be found. And it did this tens of thousands of times. At its peak, the account was pushing out a new video every 20-odd seconds, around the clock for months. By the time anyone noticed, the archive ran into the tens of thousands of uploads, an enormous body of work in the literal sense and an absolute void in every other sense. Total viewership across most of those uploads was functionally zero, not low, zero. Videos sitting in public on the largest video platform on Earth that no human eye had ever landed on. Then, in 2014, someone stumbled onto it, and the internet did what the internet does when it finds a locked door in a hallway it did not know existed.

The theories came fast and they escalated even faster, and I want to walk through them in order of increasing ambition, because the progression is genuinely funny. The first and most popular theory was that this was a number station for the modern era. Number stations are real, and this comparison is not as unhinged as it sounds. For decades, shortwave radio has carried broadcast consisting of nothing but a voice-reading strings of digits, sometimes preceded by musical stings, transmitted on schedule to nobody in particular. They are almost certainly intelligence communications, using one-time pad encryption, which is unbreakable when done properly,

and requires only that a field agent own a cheap radio and a pad of paper. The beauty of the system is that anybody can listen and nobody can tell who the intended recipient is. Some of these stations have been running since the Cold War, and a few are still going, which means somewhere out there is a government employee whose entire job is reading numbers into a microphone in an empty room. So the logic went,

if a spy agency wanted a modern equivalent, a video platform would be perfect, free hosting, global distribution, encrypted delivery courtesy of the platform itself, and enough background noise that nobody would ever notice. Blue rectangle at position A, red rectangle at position B, tone at frequency C. That is three variables per frame. You could encode a serious amount of information in 11 seconds. It is a completely sensible theory, and it is wrong, but I respect it.

The second wave of theories went artistic. somebody was making a statement about machine aesthetics, or the meaninglessness of content, or the void at the heart of the attention economy, a generative art project running unattended. This one was popular with the sort of person who uses the word praxis in conversation. The third wave went to artificial intelligence, and specifically to the idea that a system was talking to itself, or teaching itself, or leaving notes for a future version of itself, which in 2014 was still a fairly exotic idea, and today is just Tuesday. And then the fourth wave, inevitably went cosmic.

This was a beacon, a signal designed for something that does not experience video the way we do, whatever it was, it was not for us, and the rectangles were the visible edge of a conversation happening at a level we could not access. Major newspapers covered this. Serious publications assigned actual journalists to the story of the blue and red rectangle, investigations were mounted, people built databases, people analysed the tone frequencies looking for patterns, people correlated upload timestamps against world events, which is the internet equivalent of reading tea leaves with a spreadsheet open, and in the middle

of it the channel produced two moments of genuine strangeness that poured gasoline on everything. One upload broke the pattern entirely, and showed a night-time shot of the Eiffel Tower. Another anomaly featured content that had nothing to do with the rectangles at all. In a body of tens of thousands of identical machine outputs, a couple of human fingerprints appeared, and that is the exact texture of an actual mystery, the moment where the wallpaper peels and you see brick. Then the answer arrived,

and it was so boring that it functioned as a physical insult. It was a test account, belonging to the platform, uploading automated quality assurance videos. Here is the mechanism, and it is genuinely useful to understand. When you upload a video, the platform does not simply store your file. It transcodes it into a whole family of versions at different resolutions and bit rates,

so it can serve the right one to a phone on a bad connection or a television on fibre. That transcoding pipeline is enormous, constantly changing, and full of things that can silently break. Colors can shift, aspect ratios can stretch, frames can drop, audio can desynchronize by a few milliseconds. And the worst kind of breakage is the kind that does not throw an error. It just quietly makes everything slightly worse for hundreds of millions of people. So you test it continuously with a known input. You generate a video where you know exactly what every pixel and every frequency should be. You push it through the real pipeline, you pull the output back down,

and you compare it against the source automatically. Blue rectangle should be here, at this size, in this exact shade. Tone should be at this frequency. Any deviation is a bug in the encoder. Simple geometric shapes and pure tones are perfect for this precisely because they are trivial to verify by machine and impossible to misinterpret. Which also explains the name. WebDriver is the standard term for software that operates a browser automatically, clicking and typing the way a person would, and it is the backbone of automated testing at basically every technology company on Earth.

The rest of the name was random. There was no meaning to decode because nobody involved was communicating with anybody. A script was talking to a pipeline about rectangles, and the reason those videos were public rather than private is the most human explanation available, because making them public was the default setting, and nobody bothered to change it. The greatest mystery of the mid-2010's internet existed, because an engineer did not click a drop-down, which, given where this episode started, feels like a running theme. The platform eventually confirmed it, with the particular tone of a company that has realized its internal plumbing has become a paranormal phenomenon and then leaned into the joke

by building the channel into an Easter egg in its own product. Search the name and the logo turns into coloured rectangles. That is a corporate communications department having the time of its life. Now, the part of this that I actually think matters beyond the comedy. The overwhelming majority of what exists on the internet was never intended for a human being to look at. Test files, monitoring pings, health checks, synthetic transactions that pretend to be a customer buying a product every 60 seconds so that a dashboard somewhere stays green. Placeholder assets, automated backups of automated backups. There are entire server farms whose full-time job is generating and consuming content

that no person will ever perceive, and this has been true for decades. Older viewers will recognise the ancestor of this, because broadcast television did the same thing openly for years. Before the day's programming started and after it ended, stations transmitted test cards, geometric patterns, colour bars, calibration targets, sometimes accompanied by a steady tone. Some of them became weirdly beloved, including a famous one involving a girl and chalkboard that a generation of British children found deeply unsettling because they were seeing it in an empty living room at four in the morning. Nobody thought those were coded messages because everybody understood what they were for. The pattern was for the engineer, not for you. The only difference now

is that the test patterns live in the same place as everything else, with the same interface, the same recommendation algorithm and the same comment box. There is no separate channel for machine content. The infrastructure and the entertainment share one address, and from the outside you cannot tell which is which without being told, which produces a specific permanent condition. you are constantly walking past machinery that looks exactly like a message. Your ability to distinguish the two depends entirely on whether somebody happens to explain it to you. And that is why this chapter is not just a palette cleanser, it is the control experiment, because right alongside it there is a case with the exact same shape and no ending whatsoever.

Around 2015, a channel called Unfavorable Semicircle began uploading at a scale that made the rectangle account look modest. Tens of thousands of videos again, this time consisting of a distorted half-visible image and a harsh tone, with durations that varied wildly from a fraction of a second to hours. The titles were fixed strings followed by numbers that climbed steadily, like a counter. There was an image of what appeared to be a person, degraded, and audio that sounded like a voice run through something unkind. People applied the whole toolkit to it. frequency analysis, frame extraction, sequence tracking, everything the crowd had learned from previous puzzles. It generated a substantial amount of speculation, some fairly rigorous documentation efforts and no resolution.

The channel was terminated by the platform at one point, which by itself is interesting because that is a human decision rather than an automated one, and then material appeared again under other names, on other platforms, which is the behaviour of a person rather than a process. Nobody has ever credibly claimed it. No company announced it was a test suite. No artist stepped forward. No researcher published a paper. The best available guesses are that it was a long-running art project, a personal obsession or an experiment in automated publishing,

and the honest answer is that we do not know, and there is no mechanism by which we are likely to find out. Compare the two situations and you get the actual lesson of this block which is not about aliens or spies at all. The rectangles got solved because they had an owner, and the owner was an enormous company with the communications department, a legal obligation to not look sinister, and a strong commercial interest in explaining itself. Somebody could be asked, and that somebody had a reason to answer. The other case has no such person. Or rather, it certainly has one, and that person

has chosen silence, and there is no lever anywhere in the world that can be pulled to change that. not attention, not journalism, not a crowd of skilled analysts working for years. The information exists completely intact inside one human skull and there is no protocol for extracting it, which is the same wall the broadcast hijacking ran into, arrived at from a completely different direction, and it is worth noticing that the internet has not actually changed this. We have more data than any generation in history, better tools, faster distribution, and a global volunteer workforce, and we still cannot make a person tell us something they have decided not to say.

So here is where the shape of this episode turns, because up to now the failures have all been on the receiving end, signals that arrived without meaning, images that arrived without names, machines talking to themselves in public. What is left is the opposite case, what happens when a person transmits perfectly to a real audience, and the audience transmits back.

when a comment gets read and answered and demanded of, and the person on the other end goes out into the desert to prove it. Westfield, New Jersey is the kind of town that gets used in commercials when a company wants to communicate the concept of stability. Large old trees, good schools, a train line into Manhattan, houses built when houses were built to intimidate. It is roughly 45 minutes from midtown and about 400 years from anything dramatic ever happening. Visit BetMGM casino and check out the newest exclusive. The Price is Right Fortune Pick. BetMGM and GameSense remind you to play responsibly. 19 plus to wager.

Ontario only. Please play responsibly. If you have questions or concerns about your gambling or someone close to you, please contact Connects Ontario at 1-866-531-2,600 to speak to an advisor. Free of charge. BetMGM operates pursuant to an operating agreement with Eye Gaming Ontario. In the summer of 2014, Derek and Maria Broaders closed on a house there. Six bedrooms, built at the start of the 20th century, sitting on a street called the Boulevard,

purchased for something over one and a third million dollars. They had three young children. Maria had grown up in the area. This was the Dreamhouse purchase, the one you plan around for a decade, and they immediately started renovation work before moving the family in, because the Dreamhouse had Dreamhouse plumbing. Three days after the closing, a letter arrived, addressed to the new owner. It was typed. The handwriting on the envelope was odd and deliberate, and the content of it turned an extremely expensive celebration into a permanent problem, because the sender explained, calmly and at length, that the house had been under observation by his family for generations. His grandfather had watched it in the 1920s. His father had watched it decades later. Now it was his turn, and the arrival of new owners had been a matter of anticipation. He wrote about the property as though it belonged to

him in some deeper sense, and had merely been passing through the hands of temporary occupants. Then he asked about the children. Specifically, he wanted to know their names. He wanted to know how many there were. He referred to them in terms that no reasonable person could file under neighbourly interest, referencing young blood and asking whether they had discovered what was inside the walls. He signed it as the watcher. Now, I want to be a little bit of the watcher. Now, I want to to be precise about why this is so effective, because on paper it is just paper. Nobody was threatened with a specific act. Nobody was physically approached. If you describe the contents to a lawyer in 2014, you would get a fairly discouraging answer about what could actually be charged.

But look at the structure of it. The sender knew the closing date, which is public record, but requires someone to go and check it. He knew the house was changing hands before the family had moved a single box in. He wrote with the confidence that the same thing. He wrote with the confidence of someone with a long relationship to that specific address, and crucially, he demonstrated proximity because everything he described required a line of sight to that property. The letters continued. Later ones referenced the renovation work in progress, complained about the noise of the construction, and mentioned details about the family's presence at the house, including what the children had been doing on the property during visits and where vehicles had been parked.

One referenced the layout of the house and how the rooms were being used, which is a detail that reorganises your understanding of what is happening, because it means either the sender had been inside at some point, or the sender was watching windows carefully enough to reconstruct interior arrangements from outside. At no point had the family spent a single night in that house. They never did, not once, in five years of ownership. Here is the detail that turns this from a horror story into an infuriating one. The previous owners had also received a letter from the same sender

shortly before they moved out, after living in the house for over two decades. They read it, found it strange and threw it away without telling anybody, and specifically without mentioning it during the sale, which raises a question that ended up in court, does a letter from an anonymous person who has been observing your house count as a material defect you have to disclose to a buyer? Property disclosure forms in the United States are extremely thorough about certain things, They will ask you about basement moisture, roof age, underground tanks, termite treatment,

whether the well has been tested and whether the property sits in a flood zone. There is no line on any standard form that says, has anyone been sending correspondence about this house from an unknown location? The broadnesses sued. The legal outcome did not deliver them a house they could live in, which is the only remedy that would have mattered. The investigation was, for a suburban police department, genuinely substantial. Westfield Police took it seriously and worked it.

for years. The family hired private investigators, including people with serious backgrounds. The letters were examined for the usual forensic markers. The envelopes produced DNA from the seal, which is the sort of evidence that solves cases on television in 11 minutes. The result indicated the sample likely came from a woman, and it matched nobody in any accessible database, which is the single most frustrating possible outcome, because it proves a person exists without narrowing them down by more than roughly half the planet. Neighbours were interviewed. Suspicion inevitably fell on people living nearby

because the letters demonstrated observation. An observation requires an observer somewhere with a view. There were people in the immediate vicinity who attracted attention and I'm not going to walk through them by name because nobody has ever been charged and being publicly named as a suspect in an unsolved stalking case is its own life sentence. Everyone who has looked at has denied involvement

the DNA has never been matched. The case sits open. Handwriting analysis was performed. The letters were printed rather than handwritten, which removes the richest source of individual identification. Postmarks narrowed the mailing area but not usefully, because a mailbox is a device specifically designed to accept objects from anyone without recording who put them in.

And then, in 2015, the story got out to the press, and the second half of this chapter begins, which is the part I actually think is the more disturbing one, because up to that moment, the family had a private nightmare involving one anonymous person. What they got afterward was a public nightmare involving everybody else. Local coverage jumped to national coverage within days, then international. A house in New Jersey with an anonymous stalker who wrote about generations of surveillance

is an irresistible package, and the media did what media does with irresistible packages. Photographers came, news trucks parked outside, and because a house has a fixed physical address that cannot be anonymised, edited or taken down, that address was in every article. The consequence was immediate and permanent. The property became a destination. People drove past to look. People got out and took pictures. People filmed themselves on the lawn. Teenagers dared each other onto the porch. Later, when short video platforms arrived and rewarded exactly this kind of content, the whole thing accelerated, because now there was an algorithmic incentive to go stand in front of a stranger's house at night with a phone.

Take a second to appreciate the specific cruelty of the sequence. A family was targeted by one person who watched their house. The response to that was an arrangement in which thousands of people came to watch their house. The problem was not solved, it was franchised. Meanwhile, a substantial online contingent decided the family had written the letters themselves. That theory circulated widely, and it deserves examination, because it is not entirely. crazy on its face. Insurance fraud and buyers remorse are real phenomena, and investigators are professionally obligated to consider the people closest to a case. The police did consider it. The theory ran into two structural problems. The first is that the previous owner had received a

letter before the sale, which the new family had no way of knowing about or arranging. The second is that the outcome for the family was catastrophic rather than profitable, which is generally not what a successful scheme looks like. None of that stopped. anything, because online theories are not evaluated by cost-benefit analysis. They are evaluated by whether they are fun to argue about. So a family that could not sleep in a house it owned also got to read comment sections explaining that it had done this to itself. They tried everything available. They rented the property out for a period, which produced its own grim moment when a letter arrived again while tenants were living there. They applied to have the house demolished

and the lot subdivided which was rejected by the town on planning grounds that are entirely defensible if you were a planning board, and entirely maddening if you are the person who owns an unlivable house. Eventually in 2019 they sold it, for substantially less than they had paid, taking a loss in the hundreds of thousands, after five years of owning a property they had never spent a night in. Then the story was adapted into a streaming series, which introduced the whole thing to an audience of tens of millions, and sent a new wave of visitors to a street where actual families with actual jobs are trying to live, including the people who bought the house and had nothing to do with any of it. That is where this case has settled, unsolved, unpunished,

and permanently monetised by everyone except the people it happened to. So what actually happened here? Let me lay out the possibilities, honestly, because there are only a few. The first is that it is exactly what it presents as, a person, likely local, with a fixation on that specific property, expressing it through the safest possible channel. Anonymous letters are attractive to this kind of fixation because they combine maximum psychological impact with minimum legal exposure. No contact, no trespass,

no threat that can be cleanly prosecuted, no digital trail, and postage costs less than a coffee. It is, in the coldest analytical terms, an extremely efficient way to destroy someone's life at almost no risk to yourself. The second is that it was instrumental, instrumental rather than obsessive.

Somebody wanted that family gone, or wanted that house to become cheap, or wanted the sale reversed for reasons connected to the neighbourhood rather than the building. The letters would then be a tool rather than an expression, and the generational grandfather narrative would be set dressing, designed to make it sound ancient and inevitable. The third is that it began as one thing and became another. Somebody wrote a strange letter once, watched what it did, and discovered they enjoyed it.

What all three have in common is that they require nothing supernatural, no hidden chamber, nothing in the walls, and no history of the house being anything other than a large, expensive building in a nice town. Investigators went looking for a dark past on that property, because everybody did, and what they found was an ordinary history of ordinary families, which is the least satisfying and most likely answer. And here is the connection back to everything else tonight, because this case is a perfect specimen of a communication that worked flawlessly in every direction, except the one that mattered. The letters were delivered. They arrived at the correct address, on schedule, through a functioning national postal system,

and they were read by exactly the intended recipients. The message was received with total clarity, and the emotional payload landed exactly as designed. In terms of transmission, this is the most successful communication in this entire episode. What never arrived was the return address, not the one printed on the envelope, but the identity behind it. A person delivered years of precise, specific, personal information about a family, and that same person remains, after more than a decade of police work, private investigation, forensic testing, and the sustained attention of the global internet, a complete blank. We have their DNA, we have their writing, we have their approximate location, their approximate

schedule, their fixation, and a rough sense of their voice. We have everything except the one field that would let anyone do anything about it. That is the actual horror of the modern information environment, and it is the opposite of what people usually complain about. We are told constantly that privacy is dead, that everything is tracked, that nothing can be hidden. And then a person mails paper into a mailbox and vanishes so completely that a wealthy family, a police department and millions of curious strangers cannot produce a name. The technology did not fail. There was simply nothing for it to grab. No account, no device, no login, no transaction, no metadata. Just a hand, an envelope, and the patience to stay quiet, which is where I want to leave the

unsolved cases, because there is one more story tonight, and it goes the other way entirely. Everything up to now has involved information arriving without an identity attached, a face behind a mask, a woman in a photograph, a figure in a spectrogram, a sender with no name. The last case starts with the identity missing and the pertin present, a young man who died in an accident and spent 20 years in official records under a placeholder because nobody knew what to call him. And the same internet that turns real addresses into tourist attractions is the reason he has a name on his headstone today. June 1995, Interstate 95 in southern Virginia, just outside a town called Emporia.

A car left the road at speed in the early hours and hit a tree. Two young men were inside, both died. The driver was identified within hours because he was driving his own vehicle and everything in the world that connects a person to a car was in place. Registration, license, family, a home address, people who knew where he had gone that night. The passenger had none of them. that. No wallet, no license, no identification of any kind, and because of the nature of the crash, his face could not be used to identify him either, which removed the single method that solves

the overwhelming majority of these cases in the first 24 hours. So investigators worked with what he had on him, which came down to a handful of objects, and those objects tell you almost everything about who he was without telling you anything about his name. He had ticket stubs from Grateful Dead concerts, from shows played in the Washington, D.C. area in the days immediately before the crash, and he had a couple of handwritten notes, the sort of casual scrawl that friends leave for each other, one of which referred to him by a first name, Jason. That is it. That is the entire archive of a human life, as it existed in an evidence bag in 1995. Now, if you are under 30, I need to explain the specific context here, because it is essential to understanding why this case went cold for two decades.

The Grateful Dead were not simply a band. They were functionally a mobile city. A substantial population followed that tour around the country full-time, living out of vans and cars, selling food and crafts in parking lots to fund the next stop, sleeping wherever, and moving on when the band moved on. Thousands of people, most of them young, most of them broke, most of them completely uninterested in being on any list anywhere. If you needed to design a lifestyle in which a person could vanish without anybody immediately noticing, you could not do better, and I say that with genuine affection for what that scene was, which meant a young man hitching a ride out of a show, with no documents and a first name, was in a category of person that the system of 1995 was

almost perfectly designed to lose. There was no national database anyone could query in an afternoon. Missing person's information lived in filing cabinets and local records, organised by jurisdiction, which meant a report filed in one state might as well have been filed on the moon. Photographs were physical. Fingerprints required a prior arrest to be useful. DNA analysis existed but was expensive, slow, and reserved for cases with an identified suspect and a prosecutor asking for it, not for an unidentified passenger in a single vehicle accident with nobody to charge.

So the county did the decent thing available to them. They buried him locally. They put a marker on the grave. and because the ticket stubs were the only real personal detail anyone had, the name that ended up attached to him was Grateful Doe, Doe being the standard placeholder in American records for a person whose identity is unknown. He stayed that way for 20 years, and I want to make sure the scale of this registers,

because it is the part almost nobody knows. There are, at any given moment in the United States, somewhere in the region of 40,000 sets of unidentified human remains held in medical examiner offices, county storage and cemeteries, 40,000. That is a stadium. Federal researchers have called it the nation's silent mass disaster, which is one of the very few pieces of government phrasing that is actually good, because that is exactly what it is. A catastrophe distributed across decades and jurisdictions so thinly that it never looks like one. Every one of those cases has a family on the other end who does not know, and in most of them, the missing person report and the unidentified remains

are both sitting in official records, in the same country, sometimes in adjacent states, having never been compared to each other, because until quite recently there was no mechanism that would compare them. That mechanism eventually arrived in the form of a public national database, launched in the mid-2000s, that did something obvious and revolutionary at the same time. It put unidentified remains cases and missing persons cases into the same searchable system and then let ordinary members of the public access it. Not just law enforcement, anyone. That decision is the hinge on which this entire chapter turns because it meant that for the first time, the enormous surplus of curiosity that the internet generates could be pointed at something

with an actual measurable output. Which brings us to the volunteers, somewhere around 2014 and 2015, this case got picked up by online communities that specialize in unidentified person's work. Forum users, a dedicated corner of Reddit, people who spend their evenings cross-referencing case files, comparing dental records against reports, mapping timelines and maintaining spreadsheets that nobody is paying them to maintain. When WestJat first took flight in 1996, the vibes were a bit different. People thought denim on denim was peak fashion, inline skates were everywhere, and two out of three women rocked, the Rachel. While those things stayed in the 90s, one thing that hasn't is that fuzzy feeling

you get when WestJet welcomes you on board. Here's to WestJetting since 96. Travel back in time with us and actually travel with us at westjet.com slash 30 years. I have spent this entire episode being fairly hard on collective internet behaviour, and this is the point where I get to balance the ledger, because these communities are the same phenomenon operating with a different target, same skills, same obsessiveness, same willingness to spend a Tuesday night at two in the morning comparing two images. The difference is that here the work has a defined success condition and a real deliverable,

and the people doing it have developed an actual ethic about not naming living suspects, not harassing families, and running tips through law enforcement rather than posting them for engagement. What they did with this case was, in essence, a marketing campaign for a dead stranger. A forensic facial reconstruction had been produced, an approximation of what the young man had looked like,

built from the structure of the skull, and volunteers took that image and pushed it everywhere. Social media, forums, groups devoted to the media, music scene he had been part of, communities of people who had toured in that era, thousands of shares. The reasoning was simple and correct. The only person who could solve this was somebody who had known him, and the only way to reach that person was to put his face in front of enough eyes. Reconstructions, by the way, are not portraits, and they are not as reliable as the confident ones look. They are an educated approximation of soft tissue

over a known bone structure, and the accuracy of things like no shape, hairstyle and expression is limited by the fact that skulls do not record any of that. The realistic goal is not to produce a picture someone recognises instantly. It is to produce a picture that makes someone stop scrolling and feel a flicker of familiarity, and then go check whether their old friend ever came home. That is exactly what happened. Somebody who had known a young man named Jason in that touring scene

saw the image, felt that flicker and made contact. From there, investigators had something they had never had before. A probable full name attached to a probable home state. The name was Jason Callahan. He was from South Carolina. He was 19. His mother had not filed a missing person report, and if that sounds strange, sit with it for a second, because it is one of the quietest and saddest details in this entire episode. He was an adult who had left home to travel with the music, which is a thing that young people did constantly in that era. He checked in irregularly by design. He had told her he was going off with a band.

When the checking in stopped, it did not look like a disappearance. It looked like a 19-year-old on the road, living exactly the life he had told her he was going to live. There is no moment where the alarm goes off in that scenario. There is only the slow accumulation of time, which is a much crueler way to lose somebody, because there is never a day on which you are entitled to panic. She had looked for him, for years,

without a case number, without a system to search, without any of the tools that a person in her position has today. The confirmation came through DNA, a comparison between a sample from the remains and a sample from his mother, and the match was made, and in late 2015 the case closed after 20 years and six months. The marker in that Virginia cemetery no longer reads as a placeholder. It has his name on it.

His family finally got to do the ordinary and necessary human things that had been suspended for two decades, which is the entire point of identification, and which is worth being clear about, identification does not bring anybody back. It does not deliver justice because there is no crime here to punish. It delivers one specific thing, which is the ability to grieve accurately.

And this is where I want to tie the whole episode together because it is not really about spooky things at all. Everything tonight has been a variation on one problem, a signal that travelled perfectly and carried nothing recoverable, an alert that carried perfect information that was full, arriving in seconds while the correction took over half an hour. A report that reached the right building and stopped 12 inches short of a radio. A photograph that has been examined by three world-class institutions and identified nobody.

A puzzle that thousands of people solved correctly without ever learning what it meant. A machine talking to itself in public because nobody changed a setting. A comment that reached the entire internet and cost the person who wrote it his life. Letters that arrived at exactly the right address with no. return address that anyone could find. Every one of those is the same failure wearing a different costume. Information exists and information does not arrive where it needs to be, and then there is this one, which is the same problem solved. The information existed the whole time. A young man was buried in Virginia, his mother was in South Carolina. Both facts were true simultaneously

for 20 years, sitting a few hundred miles apart in a country with functioning roads and telephones, and there was simply no connection between the two. What closed that gap was not a technology breakthrough, although the database and the DNA work were both necessary. What closed it was a large number of people voluntarily doing the tedious, un-glamorous labour of moving information from where it was to where it needed to go. That is the same network that hijacks broadcasts,

invents theories about grieving families, turns a real house into a tourist attraction, and demands proof from a man in the desert. It is not a different internet. It is the identical machine, with the identical capabilities, pointed somewhere useful. And the difference between those two outcomes is not technical. It is entirely a question of what the people using it decide to spend their attention on, which is a genuinely optimistic thing to realise.

Because attention is the one resource in this entire episode that any of us actually controls, there are still tens of thousands of those cases open right now, those communities are still working. The tools have improved enormously since 2015, and identifications that were impossible a decade ago now happen regularly, which means the number of families still waiting is going down for the first time in history. That is a good place to stop. If you have a story that belongs in a future volume, something strange, unresolved, or sitting in a corner of the internet where nobody has looked properly, there is a new mailbox for exactly that, and it is in the description

below, send them over. Half of what ends up in these episodes start as somebody's message until... Don't you wish you could just hit skip on the worst parts of your life? You know the same way you can skip an ad? I get it. I'm Siaiaia and I live in Ice Cove. I've made some questionable decisions that didn't end up the way I planned. And today I'm still figuring it out.

Somehow things usually get worse before they get better. Apparently, that's how I roll. So bundle up and come along for the bumpy ride. Stream a new episode of North of North Tuesdays on CBC Gem. the next one