It’s midnight in Hobart, Indiana, and the auditorium of the cozy, single-screen Art Theatre is echoing with the sounds of cartoon characters in the throes of passion. One orgasm follows another — first Marge Simpson then Bugs Bunny — as first-time attendees to the Art’s weekly screening of The Rocky Horror Picture Show go through a variation on a ritual that’s taken place for nearly 40 years: the hazing of the “virgins.” (Virgins under 18 get off the hook this particular night, due to the risqué nature of the initiation.) It’s not a packed house, but it’s lively as about 35 attendees, many of them regulars, file in and wait for the show to start, both on the screen and off.

The Art, built in 1941, is a lovely art deco venue with a mural-covered auditorium, but like many small-town theaters it has struggled in recent years. In 2012, it benefitted from a campaign designed to keep independent theaters alive by helping them to convert to digital projection. Since then, it’s stayed afloat with second-run offerings that focus on family-friendly fare. (I’m told that posters advertising Magic Mike XXL when I visit don’t represent the usual sort of film that plays there.) But each Saturday, the Art offers an added attraction: Help Me Mommy, which bills itself as Northwest Indiana’s Rocky Horror Picture Show Floorshow Cast, performs for an enthusiastic crowd as the 1975 Rocky Horror Picture Show plays behind them.

First released in September of 1975, Rocky Horror has never left theaters, and the film is more famous now than ever. But fame has also changed it, turning it from a fringe item to an institution famous enough to provide fodder for an episode of Glee and a possible TV remake for Fox. And as it’s grown in fame, the cult has appeared to lose some of its outsider fervor over the years, with weekly gatherings like this one becoming increasingly less common across the country.

You wouldn’t know that sitting in the Art Theatre in Hobart, though. Each week the cast and crowd goes through a ritual familiar to anyone who’s seen Rocky Horror at a midnight show: A pre-show featuring clips involving the Rocky Horror cast and other bits of weirdness. A “virgin auction” in which first-time attendees are subjected to good-natured, if obscene, initiation rituals. And then the show itself, in which the venerable movie plays out behind a live-action reenactment — known in Rocky fandom as a “shadow cast” — as its dialogue gets interrupted by a mix of crude, witty retorts from the audience, some nearly as old as the film itself.

That also makes them older than most of the attendees — who skew toward teens and 20-somethings. “I was always very socially awkward. Never felt like I fit in anywhere,” a man who asks to be identified only as T.R. tells me in the Art Theatre lobby a few hours before the show. These are odd sentiments to hear from a handsome 31-year-old union plumber wearing a Captain America T-shirt, with a physique and unerring gaze to match, but there’s not a hint of insincerity to them. “I did sports, I did theater, I did a bunch of stuff and never fit in.”

T.R. first saw The Rocky Horror Picture Show in 2002 and soon joined The Pink Invaders, the Northwest Indiana Rocky cast that would eventually evolve into Help Me Mommy. “Rocky Horror was the first place I ever came to where it was just like, ‘Yeah, we’re all a bunch of weirdos and we’re all accepting of each other. You can come hang with us and we’re not going to make fun of you for it.’”

If the ranks of Rocky Horror fandom are thinner than at their height in the ’70s and ’80s, they still remain a self-replenishing resource, and Rocky Horror continues as an important rite of passage for misfits of all stripes, in Hobart and elsewhere. But the larger midnight movie tradition from which it emerged is changing, and facing an uncertain future.

Prior to the ’70s, midnight shows were the realm of the occasional horror release and exploitation distributors who used the slot to attract night owls to seedy fare. But the midnight movie as we know it — as a Friday- and Saturday-night staple featuring cult films — came into its own as the ’60s turned into the ’70s.

The ‘60s saw a flurry of activity in underground film as Jack Smith, Kenneth Anger, Andy Warhol, and others made movies way outside the Hollywood system, films that took avant-garde forms and featured content too extreme for the mainstream. That didn’t mean there wasn’t an audience for them, though. Warhol’s 1966 film Chelsea Girls played New York for months, for instance, and in the latter part of that year, Mike Getz of Los Angeles’ Cinema Theatre — after having success in Los Angeles playing experimental films at midnight — hit upon the notion of sending a package of films on the road under the name “Psychedelic Film Trips #1.” They played at midnight across the country in theaters owned by Getz’s uncle Louis Sher, and they did well, making Getz something like the Johnny Appleseed of the midnight movie. As Jonathan Rosenbaum and J. Hoberman put it in their invaluable 1983 book Midnight Movies:

“In thus popularizing the underground, Getz provided the spadework for the midnight-movie explosion of the 1970s. … [T]he underground invented a grass-roots alternative to ‘straight’ movies, television, and theater. The underground demonstrated that, in America anyway, anyone could make some sort of movie and get it shown somewhere. Accordingly, all manner of long-repressed sexual content burst scandalously upon the screen. There was even money to be made doing it.”

The moment had arrived, in other words, for the underground to go overground, and a few key titles pushed it there. George Romero’s groundbreaking zombie film Night of the Living Dead found a second life, appropriately enough, at midnight. Alejandro Jodorowsky’s violent western fantasia El Topo played for appreciative, often stoned crowds. So did early John Waters outrages like Pink Flamingos and Female Trouble. Tod Browning’s 1932 film Freaks found new life in midnight revival screenings, and 1973’s The Harder They Come helped popularize reggae outside Jamaica. In some respects, the films had little in common. The blackly comic sensibility at the heart of Waters’ films shares little with the outré stylishness of El Topo. Yet an anti-authoritarian streak runs through all of them, one that resonated with counterculture audiences as one decade gave way to the next.

Not everyone saw this as an entirely healthy development. In a November 1971 New Yorker piece titled “El Poto – Head Comics,” Pauline Kael likened the crowds turning out for El Topo and “head” films like it to a “Black Mass,” criticized its extreme violence, and called it out for pretension and incoherence. “Jodorowsky employs anything that can give the audience a charge even if the charges are drawn from different systems of thought that are — as thought — incompatible.” “Head movies don’t have to be works of art,” she concludes, “they just have to be sensational.”

Yet even if the charges stuck, the popularity of El Topo and other midnight fixtures confirmed that there were viewers hungry for images and ideas they’d never see in respectable hours — and the communities that formed around them. In the heady buzz of a packed midnight house, these films reassured their audiences that because they were outsiders didn’t mean they were alone.

The midnight landscape started to shift in the 1980s with the rise of home video, and as with every movie-related technological advance, something was lost as something else was gained. It’s no coincidence that the golden age of midnight movies started to fade as VHS ascended, but it also allowed their weirdness to spread. Growing up in a tiny Ohio suburb, I doubt I would have had access to most of those films if not for a well-stocked video store, and I’m sure I’m not alone.

And as the films that once occupied it began to disperse, the midnight slot started to go mainstream. Warner Bros. capitalized on the anticipation around Tim Burton’s 1989 version of Batman by playing it at midnight on June 23rd, its opening day. The following year, Disney went even further, devising an elaborate, troubled promotion by which moviegoers could buy “I Was There First” Dick Tracy T-shirts that would double as admission to midnight screenings.

Over the next two decades, midnight screenings of mainstream films became commonplace, particularly in the years after the debut of Star Wars — Episode I: The Phantom Menace, which attracted considerable press for the long lines of fans waiting to see the earliest possible screening. Midnight premieres of films in the Lord of the Rings, Harry Potter, and Twilight franchises followed — franchises with rabid, passionate fan bases and subcultures. But so did midnight screenings of virtually every movie released: feel the need to see The Proposal the very minute the clock strikes 12:01 on Friday? Now you can.

Now midnight has become just another spot on the schedule in many theaters, with even the see-it-first element dropping out as more and more films started hitting theaters on the Thursday evening before their official release. Does a slot once reserved for the weird and the unexpected have the same meaning when you can see The Proposal or the remake of The Karate Kid at the same time?

As technology has changed, theaters have met even more challenges in attracting midnight audiences. At Chicago’s Music Box Theatre, assistant general manager Buck Le Pard points to increased competition from Netflix as a particular problem.”You want to do something, you’re hanging out with your friends,” Le Pard says, “like let’s go see a movie, and then it just became let’s see what’s on Netflix.” Music Box has experimented with “eventizing” midnight screenings, partnering with the website Consequence Of Sound and finding success with a midnight showing of Blue Velvet preceded by a celebration of all things David Lynch, complete with actors dressed as famous Lynch characters. The theater’s midnight selection also leans heavily on nostalgic offerings from the ’80s and ’90s, though sometimes it’s hard to figure out what films fall into the sweet spot. A recent run dedicated to ’90s blockbusters got off to a good start with Jurassic Park then faltered with Independence Day, Mission: Impossible, and Speed.

The distribution wing of Austin-based Alamo Drafthouse Cinema, Drafthouse Films, has the rare distinction of surfacing new, or at least newly discovered midnight movies: gems like the gloriously amateurish ’80s action film Miami Connection (set and filmed in Orlando) and Roar, a bizarre nature adventure film produced by and starring Tippi Hedren. Yet as fun as those movies are, there’s a notable shortage of new cult classics finding their audience at midnight.

The midnight tradition has always had a corner dedicated to films that fall under the so-bad-they’re-good heading. (It was midnight screenings, after all, that gave Reefer Madness its second life.) But these days, if a new film has any hope of joining the ranks of midnight regulars, it most likely falls under this heading. Films like Birdemic: Shock and Terror (2010) and The Room (2004) have attracted cults dedicated to their incompetence, which is compelling for reasons their creators could never have imagined. Irony has always been one mode for midnight appreciation; it’s now threatening to become the dominant one.

The rare exceptions are those films that get a second chance thanks to midnight audiences. A combination of home video and midnight screenings made The Big Lebowksi into the cult object it is today. Donnie Darko and Wet Hot American Summer both did meager-to-embarrassing business on their first release only to be rediscovered by appreciative, after-dark crowds. Considered a commercial disappointment on its 2010 release, Edgar Wright’s Scott Pilgrim vs. The World has since become a midnight staple. “I didn’t expect it,” Wright told me in an email exchange about the film:

“…But I had somewhat designed the movie to be the type of film that would have blown my head off as a teenager. So when it quickly started building up a passionate following even on its initial release, that connection back to the films that inspired me came full circle. I like movies that reward repeat viewings and it’s a very dense movie so people come back again and also drag their newbie friends too. When the movie underperformed on its opening weekend, the head of marketing at Universal, Michael Moses, sent me an email that simply said ‘Years, not days.’ I think he too knew that it could be a sleeper.”

“Give it another 20 years,” Wright concludes, “and we might be in the black.” If that second life offers cold comfort to studio bookkeepers, it at least assures filmmakers like Wright — and Donnie Darko’s Richard Kelly, and Wet Hot American Summer’s David Wain — that their work is finding the audience that missed it the first time around, however belatedly and in the late-night hours.

The fall and rise of Scott Pilgrim illustrates how difficult it is to predict the success, immediate or eventual, of any film — in prime time or midnight. A few titles, like the self-consciously cult-y 2008 musical Repo: The Genetic Opera have made deliberate attempts to stake their claim in the midnight market, with varying degrees of success.

Which brings us back to The Rocky Horror Picture Show, and back to Hobart, Indiana. While other key films from the golden age of midnight movies still get brought out for the occasional showing, most are solely home-video fixtures. Rocky Horror remains a theatrical draw even after appearing on VHS, DVD, and Blu-ray, being broadcast on cable networks from VH1 to Logo, and Glee. If a proposed TV remake ever happens, it will almost certainly survive that, too. Why? The harder I’ve looked for a simple answer, the more that answer has eluded me. Alexander Degman, a member of the Help Me Mommy cast, even confessed to me he didn’t like the movie. “I like everything else.” Degman says of the film itself, “Whether it’s good or bad or indifferent, it’s people putting dedication into something. It’s people having a passion for something and just getting people to do something else other than sitting in front of the computer playing video games all night long.”

The Rocky Horror Picture Show is, in some ways, the safest form of rebellion, in Hobart or elsewhere. “We have a big audience of teenagers,” cast member Sarah Denton (who shares a last name with the hometown of Rocky Horror’s heroes) tells me, “and I like to think that we’re more of a safe place than anything. Because it’s a Saturday night, it’s a bunch of teenagers; they could be going to parties, out drinking, out doing drugs, but we’re not doing that here.” Instead, they’re watching and interacting with a movie, one formed from glam rock, bondage gear, and the cheap, lasting stuff of B-movie fantasies and dirty rock and roll. It’s a movie that takes a buttoned-up couple from Denton, Ohio and drops them into a fantastic, dangerous, arousing world they didn’t even know was within their reach. And if two kids from Denton can find it, then maybe anyone can.

Yet if Rocky Horror seems in little danger of fading away, it’s unlikely we’ll ever see its likes again. The midnight slot has become a different sort of space in the years since its 1970s heyday, and watching movies a different sort of experience. For better or worse, some of the conditions that helped draw in midnight crowds — namely a passive tolerance of drugs, alcohol, and rowdiness — are hard to find these days. And midnights aren’t the only route by which films now find their cults. Leos Carax’s Holy Motors — a film very much in the tradition of golden-age midnight movies, may have only played in a handful of theaters for a few weeks, but now I can rent it in a click and watch it on my laptop, at Starbucks in the middle of the day. I can even link out to its amazing “Entracte,” on YouTube, a rousing musical showstopper that shifts the film into another gear as it moves toward its second act. You can watch it alone, if you like. But it will never quite be the same as its natural setting: in the dark, away from the rest of the world, after midnight.

Steve Williams couldn’t breathe. The former athlete had cardiomyopathy, which occasionally choked his lungs with fluid, making him gasp for air. But this felt different; Williams felt like he was dying. He was raced to an Orange County hospital, and shortly after checking in, his heart stopped. For 30 minutes, ER workers compressed his chest in an attempt to revive him. At one point, his wife Mary remembers being called into his room to say goodbye to her husband of 24 years. It seemed Williams was a dead man.

Incredibly, doctors rebooted Williams’ heart — but for three days, he was in an induced coma, his body packed in ice to minimize brain damage. When he woke up, his mental facilities were intact, but his body was ravaged. His liver was congested, fluid reappeared in his lungs, and his heart’s right and left ventricles were practically destroyed, making it hard for blood to circulate throughout his body. Without a heart transplant, he would soon die.

Donor hearts, however, are scarce; each year, only about 1 in 10 patients that need a transplant worldwide receives the life-saving surgery. Williams’ best shot at survival was to have his heart almost completely removed and replaced with a total artificial heart — a synthetic machine that stands in for a living, beating organ. Williams would be among a handful of patients who receive artificial hearts every year. Right now, they serve as life support for people who are about to have transplants. But if the device’s manufacturers have their way, one day soon mechanical hearts will replace our faulty biological ones forever.

Luckily, Williams had landed in one of the few places in the world that could perform the extreme procedure — he’d been transferred to the Cedars-Sinai Heart Institute in Los Angeles. But the proposal was harrowing: his chest cavity would be cracked, critical arteries severed, and the heart he’d lived with for 53 years pulled out of his body.

“Scared is such a simple word; we were totally blindsided,” Mary told me.

“But there was really no other choice for me,” Williams says.

The mechanical heart offered by Cedars-Sinai is manufactured by the Arizona company SynCardia. It has a simple design: a pair of fist-sized artificial ventricles that attach to the atria, the heart’s blood collection chambers; the aorta, the body’s largest artery; and the pulmonary artery, which connects the heart and lungs. Two tubes — with the circumference of garden hoses — attach the polyurethane ventricles to an air pump outside the body, which sends roughly 120 short bursts of air per minute into the machine. That air forces blood in and out, replicating a rapid heartbeat. The device must be constantly plugged into a power outlet or connected to a battery pack. Without electricity, the heart quickly shuts down; unless power is restored, the patient dies.

Williams agreed to have a mechanical heart installed as a temporary measure while waiting for a donor organ. In April of 2014, the day before his surgery, doctors introduced Williams to someone recovering from the same operation. The patient was tethered to an air compressor that sat at his bedside and made a perpetual thump-thump-thump-thump sound so loud it could be heard from across the room. Two tubes protruded from an incision on his stomach. The scene was unsettling, but Williams wasn’t fixated on the tubes, the pounding sound, or the man’s pulsating chest. He mainly noticed how healthy he looked.

Almost two-thirds of SynCardia total artificial heart recipients can get out of bed and walk two weeks after surgery, and the majority survive long enough to get a transplant. As odd as it might seem, Williams thought temporarily replacing his central organ with plastic looked promising.

Williams’ surgery went smoothly. He was discharged with a “Freedom Driver,” a toaster-sized battery and air pump designed by SynCardia that fits snuggly into a backpack and lets people return home. (Until 2010, most mechanical heart recipients, no matter how healthy they were, stayed connected to large, relatively immobile batteries and were confined to hospitals.)

The operation and the driver allowed Williams to resume a somewhat normal life. He saw his son graduate from high school, for instance. He takes daily walks on the beach, gardens in his backyard, and more than a year after he nearly died, remains in relatively good health.

“I was a 50-year-old guy that typically would be dead, and my wife would be the sad woman with two kids living off my insurance policy,” Williams says, as his “heart” labors away at his feet outside an Orange County Starbucks. “That’s what I have to say to myself to remember how lucky I am.”

But living with an artificial heart has its downsides: Williams is constantly accompanied by a caretaker, typically his wife, in case the device malfunctions; he is forever connected to a battery or plugged into a power outlet; the large doses of blood thinners he takes make his body run cold; and he can’t submerge himself in water. Then, there are the more impalpable challenges: he hasn’t heard silence in over a year — just the constantly beating sound of the air pump — and he draws stares in public. (Williams stopped seeing movies in theaters out of respect for other viewers; the noise from his artificial heart would bother most audiences.)

“It’s a love-hate relationship you have with these devices; you’re thankful you’re alive, but it’s a compromise to have them,” Williams says. “It’s human nature to want things to go back to the way they used to be. Sometimes they never do.”

There are few people alive in the world who have no natural heartbeat. Fewer than 2,000 patients have received an entirely artificial heart in the device’s three decades of existence, and most patients haven’t used the machines for long. As with Williams, mechanical hearts are typically just a bridge to an eventual transplant.

But that may change. A handful of companies, including SynCardia, are trying to get regulatory approval to market the first permanent mechanical heart for wide-scale use, which would replace a patient’s biological organ over the remaining course of their lifetime. “Think along the lines of knee replacements and hearing aids,” says Piet Jansen, chief medical officer for the French company Carmat, which is running an early clinical trial for a long-term mechanical heart. “At some point we will have an implantable blood pump that [permanently] replaces a sick heart.”

If it works, the technology could save a lot of lives. More than 5 million people in the US have heart failure, according to the CDC, and many need a new organ to survive. The lucky few that get a transplant often face complications such as rejection of the donated heart, infection, and vital organ failure. Mechanical hearts can be more reliable, and some predict they are the future of treating total heart failure.

But the technology raises also practical and ethical questions. It’s unclear whether plastic and metal hearts can ever truly replicate their biological counterparts, which pump 2,000 gallons of blood every day, service 60,000 miles of blood vessels (more than double the circumference of the world), and work without a hitch year after year.

Living with an artificial heart is also hard on patients, and once the surgery begins, it’s irreversible. “You’ve cut out a big chunk of the heart, and you’ve cut out the parts that are necessary to sustain life,” says Stuart Finder, a health care ethicist at Cedars-Sinai. “Once you start the intervention, you can never go back.”

The heart is just a pump, and for the past half-century bioengineers have doggedly tried to replicate the biological machine with a mechanical one. Success has always been just beyond their grasp.

In the 1960s, scientists promised to develop a total artificial heart by the end of the decade, to cure end-stage biventricular heart failure. The proposal intrigued Congress, which funneled hundreds of thousands of dollars into the newly created Artificial Heart Program at the National Heart, Lung, and Blood Institute. While it seemed relatively easy to craft an effective blood pump, designing one that agreed with the human body and worked for decades without fail was much harder.

One of the first prototypes, built in the 1970s, had an unlikely power source: plutonium-238 — a byproduct of nuclear weapons manufacturing. The element generates thermal energy for close to a century through radioactive decay; in theory, it could keep recipients alive well into old age.

The “atomic heart,” however, had some obvious drawbacks. Even when functioning properly, patients would be exposed to high levels of radiation. And what if it malfunctioned? Would enough radiation seep into the patient or harm their loved ones? Co-workers? Neighbors? A member of a federal assessment panel even wondered if terrorists could kidnap mechanical heart recipients, dissect them, and turn them into weapons.

Despite protest from researchers, the atomic heart was ultimately killed before it was implanted in anyone. But it did pave the way for the first clinically tested permanent mechanical heart: the Jarvik-7. The Jarvik’s design was simple: two man-made ventricles, connected to an external air pump, expand and contract the artificial organ, propelling blood throughout the body — much like Williams’ artificial heart today.

In 1982, Barney Clark, a 61-year-old dentist with end-stage congestive heart failure, agreed to have a Jarvik-7 installed. He didn’t expect to live long, but was willing to endure the operation for the sake of science. On December 2nd, after nine hours of surgery, Clark woke up with two plastic ventricles and titanium in his chest. He was the first human to live without a natural heartbeat.

Clark’s body had a hard time accepting the implant. He bled internally and suffered multiple strokes; his mental health deteriorated. In TV interviews, he stared off-camera with a distant look in his eyes, a washing machine-sized air compressor called “big blue” working tirelessly beside him. On a few occasions he supposedly asked doctors to unplug the device and let him die.

Then, on a snowy day in March of 1982, after multiple organs failed, Clark experienced brain death. His artificial heart, which beat nearly 13 million times over 112 days, was shut down. Four more Jarvik hearts were implanted after Clark’s, and each produced equally grim results. The device caused kidney failure, hemorrhaging, stroke, and seizure. Sociologist Renee Fox documented the Jarvik experiments in her 1992 book Spare Parts. One Jarvik recipient endured such heavy internal bleeding after his surgery, Fox writes, that half of his total blood volume had to be replaced. The patient later compared the device to a “threshing machine” in his chest cavity.

In 1988, after nearly a quarter of a billion dollars, the National Heart, Lung, and Blood Institute pulled funding from the Artificial Heart Program. The plastic-and-metal machines, it seemed, were incompatible with the human body, and the pursuit of the bionic man appeared to be over.

But not everyone gave up. While the device was impractical as a permanent replacement for the heart, a young heart surgeon from Arizona named Jack Copeland saw the potential to use artificial hearts as a bridge to transplant. In the mid-1980s he implanted a machine commonly referred to as the “Phoenix Heart” in a dying man. Controversially, Copeland didn’t have FDA approval or the person’s consent for the operation, but he saw no other options. “We didn’t have time to sit down and consider all of the implications,” he says. “We were trying to save his life, and that in the end that was the most important consideration.”

The urgency gave Copeland some legal protection, and to the chagrin of some medical ethicists, the FDA did not punish him. After the mechanical heart was installed, Copeland was impressed by its efficacy. “The guy did well on the device,” he recalls. “It worked a hell of a lot better than the heart that had been inside him.” The patient survived for 11 hours on the machine, but died shortly after getting a biological transplant.

A few months after implanting the Phoenix Heart, Copeland began using the Jarvik-7 as a bridge — this time with federal approval. In 1991 he co-founded CardioWest, which later became SynCardia. Its researchers obtained rights to the Jarvik heart and tweaked it. By 1993, after making minor changes to the old machine, Copeland and his colleagues had created a bridge-to-transplant device that was approved for a 10-year-long human trial.

Of the 81 people in the trial, 17 died after the artificial heart was installed. But the rest survived long enough to get a transplant — promising results considering that all the participants were considered at high risk of imminent death. Soon thereafter, the CardioWest device became the first and only FDA-approved artificial heart for a bridge to transplant.

Once widely rejected by the medical community, mechanical hearts have since become a viable temporary treatment for end-stage biventricular heart failure. “It’s like chemotherapy in the 1960s: people used to think it was crazy to give poison to someone who was dying, and now it’s routine,” says Francisco Arabia, a biological and mechanical heart surgeon, and surgical director of Cedars-Sinai’s mechanical circulatory support program. “There’s no doubt life is not the same with [a total artificial heart], but the human body can adjust to incredible things.”

On the sixth floor of Cedars-Sinai — where heart patients recover after surgery — a room hums with what sounds like a loud clock, each tick separated by exactly one half of one second. Kerrie Cancel De La Cruz, a 42-year-old mother from Texas with an ear-to-ear smile and a neck-to-chest scar, sits in the middle of a din coming from an air pump that keeps blood flowing through her 5-foot, 3-inch frame.

A large sticker that reads “Tink” in bold red letters is plastered across the pump. “I thought, you’re a ticker, but I’m not going to call you ticker,” she says to the machine. “I like Tinkerbell, so I named you Tink.” On a visit last April her walls were covered in family photos and a stuffed teddy bear hung off her bed. Flowers from her husband and a fellow artificial heart recipient wilted on her bedside table.

De La Cruz was born with a weak heart. At 18, she had a defibrillator implanted, and at 32, after giving birth to twins, developed atrial fibrillation, a disorder that causes an irregular heartbeat. Her skin turned pale grey, her stomach bloated with fluid, and she experienced such severe pain that it was hard for her to get out of bed. De La Cruz needed a new heart, but her small body and high level of antibodies made her an unlikely transplant candidate.

She was transferred to Cedars-Sinai in the spring of 2014, where Arabia explained that her best treatment option was a mechanical heart. The operation would increase her odds of survival seven-fold, while also helping her regain strength — which she would need in order to receive a donor transplant, Arabia said. Still, De La Cruz was reluctant. “At first I didn’t want it,” she recalls, tears welling up in her husky-blue eyes. “I was tired, I fought a lot. It felt like it was time for me to go.”

Patients like De La Cruz, who are out of options, often end up at Cedars-Sinai; the plush medical center adjacent to Beverly Hills performs more total artificial heart transplants than anywhere else in the country. When Arabia describes the operation, reactions are mixed, he says. A few patients — mostly men — are excited at the prospect of getting a “super heart,” but the majority don’t want the implant. It isn’t until they’re dying in the ICU that most people agree to it.

As Christiaan Barnard, the South African surgeon who performed the first human-to-human heart transplant in 1967 put it: “patients on the brink of death are between a lion and a crocodile.”

If the lion is a mechanical heart, a biological transplant is the crocodile: while it might be a better option, it still poses a grave threat. Taking a living heart from one person’s body and putting it into another person is a high-risk procedure with ample room for error. The donor heart first has to be methodically extracted from a very recently deceased body, transported hundreds, sometimes thousands of miles on ice, and installed in an entirely foreign vessel.

Occasionally a flaw in the donor heart will cause it to fail shortly after the transplant. Other times the transplant beats for a few months or years before giving out. “Human hearts are like a used car, they can have flaws, disease,” Arabia points out. Sometimes the patient’s body rejects the transplant — a painful process where the immune system attacks and kills the foreign organ. Nearly a quarter of heart transplant recipients show some signs of rejection within a year of surgery. To battle rejection, patients must take expensive and risky immunosuppressant medication. While the pills safeguard the new organ, they also increase the odds of infection, cancer, liver disease, and kidney failure.

Artificial hearts are different. Since they aren’t made of human tissue, rejection isn’t an issue, and defects are rare. “They work 100 percent of the time,” Arabia says. “You know what you’re getting.”

But they’re not without complications and downsides — both physical and emotional, as De La Cruz knows.

“This doesn’t only take from your body, it takes from your mind, spirit, and soul,” De La Cruz says.

Despite De La Cruz’s initial reservations, she had the operation. But in January, after a short stay at a temporary home in Los Angeles, she returned to Cedars-Sinai. While De La Cruz’s body accepted the mechanical implant, the Freedom Driver wasn’t strong enough to support her. She had to be hooked up to a more powerful air pump, forcing her to spend months living in a hospital room.

Each day she dreamt of resuming her old life: being home with her family; swimming in the Gulf of Mexico; eating and drinking freely (De La Cruz could only consume 1.5 liters of fluids each day, and had to stay away from sodium).

The ordeal was hard on her two children, particularly her youngest daughter, who’s in elementary school. “I shared the same room with her when I was home, and when I had to go back to the hospital, she said, ‘Mom, I really miss you, I can’t sleep,’” De La Cruz recalls, sitting cross-legged in her hospital bed. “When I asked her why, she said ‘because I can’t hear your heart beating.’”

“If you haven’t been through this, you don’t understand.”

Despite its drawbacks, SynCardia has helped hundreds resume a somewhat normal life while waiting for a donor transplant. Now, the company is preparing to leap forward: pending FDA approval, it plans to offer its total artificial heart — the same model De La Cruz and Williams have — as a permanent implant for patients in the US. If approved, the device will become available to a wider set of patients, including those not eligible for biological transplants.

Creating an effective long-term mechanical heart has been cardiac surgeons’ dream for decades, but the endeavor has been incredibly difficult and expensive. One company, called Abiomed, received FDA approval to use a permanent mechanical heart in 2006. The device was expensive and hard to implant. After clinical tests, only one person was fitted with the machine before the company abandoned the device.

One of SynCardia’s challenges is what appears to be a precarious financial situation. The Arizona firm registered for an IPO in September, hoping to raise money to pay off tens of millions of dollars in debt and expand into Latin America and Asia. “To serve as many patients that need the device, the company has to have the wherewithal to reach and treat them,” SynCardia CEO and president Michael Garippa says, referring to the IPO plan.

SynCardia withdrew its public offering in October and has since laid off some of its employees. Garippa blames the fumble on volatile market conditions.

The French company Carmat is developing an artificial heart, but it also faces an uphill battle. Unlike SynCardia’s machine, with its driver in a backpack, Carmat’s heart is almost fully implantable and is powered by small lithium batteries that can be worn on a belt or a holster under the arm. A layer of chemically treated cow-heart tissue — or “pseudo-skin” — coats the heart, wherever it touches blood, making it more biocompatible and reducing the need for blood thinners. It is still in early clinical trials.

Three people have been fitted with Carmat’s heart. The first patient survived two and a half months before the machine short-circuited — nearly three times longer than his projected survival. The second patient, who had the device installed last August, was able to return home roughly five months after operation, telling the newspaper Le Journal du Dimanche in April: “I have never felt so good.” Shortly after the interview, the machine’s motor malfunctioned and the patient died. The third patient, who was implanted six months ago, is still alive and well. Once a fourth person is fixed with the device and survives at least 30 days, the company can move beyond its early clinical trial and begin a second set of tests.

It’s clear that Carmat’s heart is still far from ready for the market. Aside from the two malfunctions, it weighs nearly three times more than an average human heart and is incredibly expensive. Nonetheless, the company estimates around 100,000 people in the United States and Europe could eventually benefit from the machine, which may usher in the next generation of mechanical hearts.

“If we had a small, totally implantable artificial heart, all powered electromagnetically, that would be the standard, there’s no doubt,” Arabia says, noting that it may be a while before a reliable, self-contained artificial heart is available. “We want something we can implant and forget about, just like a normal human heart.”

Perhaps the biggest moral conundrum, though, has nothing to do with living with a mechanical heart, but dying with one. Unlike biological hearts, artificial ones can pump blood throughout the body and oxygen to the brain when the rest of the body breaks down. Death is dragged out — and may require the machine be shut off. “I’ve had patients where the liver and kidney failed, the skin turned yellow and died, and the only thing working normally was the brain,” Arabia says.

On a hot spring afternoon, Steve and Mary Williams drive a few miles from their modest Orange County home to the Pacific Ocean, where the couple regularly takes a two-mile walk on a meandering bike and pedestrian path. Steve misses surfing more than almost anything else, and coming to the beach keeps him grounded.

After pulling into a handicapped parking space, Williams unplugs his heart’s air-compressor from the car’s charger, straps on a backpack that weighs roughly 14 pounds, and steps into the Southern California sun. As he walks up a wheelchair ramp to the coastal path, a man in an army fatigue hat and board shorts yells: “What are you doing, you’re not handicapped!” Williams stops and locks eyes with the pedestrian. “I don’t have a heart,” he responds, noticeably flustered. “What’s your disability?”

Mary ushers him away before things can escalate. “He gets comments like that all of the time,” she tells me. A waitress once asked if Williams had monkeys inside of his backpack, due to the strange sound coming from it. Another person accused him of carrying a bomb.

Stares and comments are minor compared to other complications. Earlier this year Williams accidentally knocked the air pump onto the floor while preparing to take a sponge bath, and the battery stopped working. He took a few steps out of the bathroom, then crumpled onto the floor. A backup battery kicked in, jolting Williams awake. The fall and the jolt left him with superficial injuries: a chipped tooth and a large cut above his left cheekbone.

“I’m trying to live a normal life but it can be hard with this thing attached to me,” he says, his words slowed by a stroke he had after the operation. Williams keeps track of all the promising breakthroughs in the heart transplant field. We discuss scientists’ effort to create bioartificial hearts using stem cells, full body transplants, and bioengineers’ quest to develop a “continuous flow device,” which are small, implantable turbine-like pumps that send blood through the body at a steady stream, much like a hose or faucet. It’s controversial technology, but the machine has just one moving part, doesn’t experience much wear and tear, and is relatively quiet.

Today, Williams is stuck in a sort of limbo. If his heart had stopped beating 30 years ago, he probably would have died. If it stopped beating 30 years into the future, he might have a permanent and non-intrusive artificial heart. Or perhaps a perfect clone of his original organ. Instead, he has a mechanical heart that, while burdensome, keeps blood pumping to his organs and oxygen to his brain. It’s a hard life, yes, but Williams feels lucky to have it. “When people ask what it’s like I tell them, ‘Sure, I don’t have a heart, but I had a lot of problems with the heart I was born with. If I kept it I probably wouldn’t be here today.’”

At 6AM on April 9th, Williams was making coffee for Mary when his home phone rang. He answered tersely, bothered that someone was pestering him so early in the morning. The call, however, was one he had anxiously been awaiting for the past 51 weeks: a donor heart was being transported to Cedars-Sinai for him.

Williams and Mary raced to the hospital where he was prepped for surgery. At 1:30 the following morning, Williams had his artificial heart removed, was placed on a heart-lung bypass machine, and had a stranger’s biological organ installed in him.

The surgery went smoothly. He had crossed the bridge, so to speak.

“I’m just in shock still: I can’t believe I don’t have the artificial heart; I can’t believe there’s no noise; I can’t believe I don’t have to take the blood thinners,” he said a few weeks after the operation. “I should be dead at least three times from this whole ordeal, but I’m not. I’m just so grateful for everything.”

A few months later, the transplant was still holding up well. Williams plans on volunteering at Cedars-Sinai and helping artificial heart recipients adjust to their new lives.

And in July, after seven months of being monitored closely as an inpatient at Cedars-Sinai, De La Cruz also crossed the bridge. Her operation, though, didn’t go as smoothly as Williams’: De La Cruz’s body rejected the new organ and she went into kidney failure. She is now in and out of the hospital and may soon need a kidney transplant.

“This is exhausting; I was looking forward to getting the heart and trying to live a better life,” she said at the end of October. Despite the grim circumstances, De La Cruz maintains her optimism. “There are still things that can be done to give me the quality of life that I was hoping for… It will be better.”

The one thing you never hear about Khalif Diouf is how much he laughs.

In general, the native New York rapper — better known as Le1f — is portrayed as a soft-spoken, yet deadly serious provocateur wholly dedicated to his bawdy, politically charged party raps. He helps that image along, of course: confrontational onstage, always strictly composed and mean-mugging like a couture model in photoshoots, and ready to unleash a social media beatdown on anyone unwise enough to call him transphobic or make bigoted comments about his career.

But in person, that all melts away with the arrival of his big, endearing exclamation point of a chortle, one that bursts loudly and often over whatever space he happens to inhabit — for example, a beige booth in a Midtown Manhattan diner, talking to a reporter about X-Men.

“I was just into mutant… situations,” he says of his nerdy childhood, laughing over a plate of plain French toast. (They were out of cinnamon-raisin bread.) He’s ordered a pitcher of water, too, worried he’s getting sick before his CMJ show at Williamsburg’s Cameo Gallery later this week. “I always liked Iceman, but he just came out of the closet. He said he didn’t want to before because he didn’t want to deal with the double-discrimination. I’m like, ‘Whatever, I’ve been a gay black dude my whole life. Get over it.’”

Le1f’s own breakout story plays out like an underground rap superhero saga. In 2012, having lost the production files in the eleventh hour for a song from his debut mixtape Dark York, Diouf found himself in need of a beat. He called on his friends, the production duo 5kinAndBone5, and when they sent over a track, he says, “I just drunkenly rapped a new song.”

That song was “Wut,” still his most well-known track to date. Despite its slapdash origins, it was a polished, major-label-quality banger that confidently showcased both his blackness and homosexuality in such a nonchalant manner that his stardom, in all its colorful glory, seemed inevitable. In the video for “Wut” video Le1f dances against an all-white background like the star of an animated iPod commercial made flesh, narrating the moves of his brightly dressed backup dancers; even when the lights go out, he keeps going, spinning in place while running an Afro pick through his close-cut, dyed-violet hair. The clip doesn’t even break the three-minute mark, but it only took that long for Le1f to be officially bookmarked as One To Watch.

The song and its video, coupled with a few fiery comments about homophobia in mainstream rap, went viral, and he quickly became a minor iconoclast, a talented artist with a fresh perspective. Two years later, he was performing with choreographed dancers on The Late Show with David Letterman. Plenty of independent artists have hit late-night TV stages in recent years, but for Le1f, it was a major moment: here was a radical queer black artist performing a song about being radical, queer, and black, on national television. Middle America had rarely been so visibly confronted by such a marginalized talent; it was an electric appearance that the press immediately lauded as a sign of big things to come.

But three years after “Wut” blew up, the inevitability of Le1f’s success seems far less certain. His second-most popular video, 2013’s “Spa Day,” has under half a million views; while “Wut” has been streamed 1.7 million times on Spotify, his second most popular “Boom,” has only made about 353,000. Le1f’s biggest hit, now three years old, has begun weighing on him.

“I kind of want to destroy it,” he says now at the diner, grimacing. “I think there are a lot of people who just see the ‘Wut’ video, but they’ll never hear any of the ridiculous, loud, and destructive music that I was [also] making at the same time.”

And nothing pigeonholes an artist faster than becoming a one-hit wonder — a fact of which he’s painfully aware.

“I appreciate that I can have a life from [‘Wut,’]” he says. But the fact that he’s continued to release new material since then, none of which has garnered the same degree of attention, is frustrating to him. “It makes me think that it really wasn’t about artistry at all,” he says. “It was about the hashtag: ‘gay black rapper.’”

Being oneself often has that sort of price, especially in pop music. But with his debut full-length album Riot Boi, which arrives this week on Terrible / XL Records, he’s making a serious attempt to get that fee waived.

At six-foot-three and dressed in all black, with one long, gold earring and a leather bucket hat, Le1f stands out in a restaurant full of tired, gray-haired white men in taupe suits. But like anyone who grew up having to make the best of their misfit status, he’s used to it.

Diouf comes from a line of classical vocalists — his mother and grandmother are professional opera singers who have performed at Carnegie Hall and the Metropolitan Opera House. “I’m like the only person in my family who can’t sing, actually,” he says. Instead, he found his niche in dance, a talent that eventually paved the way to boarding school at the prestigious (and predominantly white and wealthy) Concord Academy. As a teen he studied dance, design, and production — frequently showing up to class “wearing white plastic trench coats and Uggs” — and used his high school’s computer lab to record his own electronic and juke beats.

Still, the idea of performing vocally in any capacity seemed out of the question. “Having this nasally, gay voice was just something totally disgusting to me as a teenager,” he says. “I didn’t understand how to use it or what to do at all.” But at some point during his time at Concord — “a time when I was listening to Animal Collective in dance class, and crying to Björk at night” — Diouf first came across an inkling that even if he would never make it in the opera, maybe a rap career wasn’t so farfetched.

“M.I.A.’s ‘Galang’ video had just come out, and I saw how she incorporated her printmaking art into that music video, and I was like, ‘Yes, this is awesome!’” he remembers. “Her voice sounds crazy, but I love it, and she is summing up all her mediums [sic] into one awesome thing.’” He was also inspired by the equally idiosyncratic vocal stylings of British rapper Dizzee Rascal. “That was definitely an inspiring moment for me, to at least start working towards something.”

It wasn’t until 2007, the summer before he started college at Wesleyan, that the hours spent fooling around on school software started to pay off, courtesy of his DJ’s roommate and their friends — Himanshu “Heems” Suri, Victor “Kool A.D.” Vasquez, and Ashok “Dapwell” Kondabolu. The trio was looking to get a rap project off the ground, and Le1f had been selling beats to make extra money.

“I basically gave them this beat,” he says. “I really thought I wanted to rap over it, but at the time, I wasn’t feeling confident enough about rapping over anything that wasn’t electro or juke.” That beat became the foundation for Das Racist’s goofy 2008 viral earworm “Combination Pizza Hut and Taco Bell.” The favor paid off: the artists remained friends, and four years later, not long after Diouf graduated and returned to New York, Heems released Le1f’s debut mixtape Dark York via his label, Greedhead.

As the buzz around Le1f’s mixtapes grew (Fly Zone was released in January of 2013, Tree House that fall), the press developed its own narrative, observing similar releases from the likes of Mykki Blanco and Zebra Katz and declaring a burgeoning “queer New York rap scene.” It was news to Le1f.

“I didn’t even realize I was a ‘gay rapper’ until the Pitchfork article,” he says earnestly, referring to an article on the music site that linked the careers of Le1f and a handful of other subversive New York artists. For a while, he says, it was frustrating, but he soon accepted the label — “if that means screaming and vogueing in a dress in the dark onstage, then great.”

“I’m already a politicized figure, just by nature of how I’ve been in the press, being an out musician,” he says. And he has made a bit of a name for himself as an outspoken figure on social media: after Macklemore and Ryan Lewis won a VMA in 2013 for their pro-LGBT cut “Same Love,” Le1f tweeted — a lot. “It saddens me out that a straight man is the voice pop music has chosen for gay rights,” he wrote. (In the same incident, he also called out the Seattle duo’s horn-centric “Thrift Shop” for its resemblance to “Wut”: “At the end of the day, I openly have no respect for any artists ripping from other artists, so blatantly.”)

The way he sees it, at this point all he can do is throw his whole weight into the parts of his public reputation that really matter to him. “I wanted to make a record that addresses the politics that I wanted to speak about, and not what others [expected me to say],” he says of Riot Boi. “It feels like I’m breaking out of being a meme and convincing people I’m a musician.”

Still, he admits, while his subject matter may be subversive, he can’t help but aspire to mainstream acceptance, which of course comes with pressures to conform to certain cultural expectations. He spends a lot of time thinking about where he would fit in the pop world, should he get there. He admires Nina Simone, a pianist who began singing out of necessity and ended up a massively celebrated vocalist and activist, but also likes Beyoncé, a lifelong singer and meticulous performer. They’re nearly opposite aspirations, artistically speaking, but he wants both.

“But like, the cyborg version, post-human,” he says, adding, “That’s why I put the ‘1’ one in my name, you know: I wanted to be slightly a computer.”

But those aspirations are a secondary concern for now. More pressing is simply maintaining basic quality control: the loss of momentum he’s faced over the past couple years has limited how much he’s able to do as an artist, even in keeping the elements of his show that made people fall in love with him in the first place. “I love being able to work with dancers and a choreographer,” he tells me later over the phone. “[With Letterman,] I just wanted to have that moment. But for the most part, I can’t afford to do that all the time. I’m not exactly filling up rooms, and it’s hard to get promoters to want to spend that kind of money, and [hiring dancers] doesn’t really make money for my label.” He’s considering upgrading his live show by paying out-of-pocket for dancers, in which case, he says, “I’ll just be broke.”

Playing by the rules and breaking them all in the name of self-expression is a never-ending struggle, one that visibly frustrates him. “Sometimes I edit down too much, and I think I should go the opposite way, and just be a crazy person, see what happens,” he says, with another laugh, this one tinged with a note of cynicism. “Like, what if I actually just wrote a song from the voice of Mystique talking to Steve Jobs? What if I actually went to those levels?”

The day after lunch, we’re supposed to meet again at Dr. Wu’s, the low-profile Williamsburg studio where he and Jake Aron, his co-producer and former Wesleyan classmate, have been piecing together the record for the last year. He’s about an hour and a half late, but Aron says that’s nothing in rapper time. “Once we were supposed to have a session with Joey Bada$$ at 1PM,” he says, “and the guy didn’t show up ‘til 10.”

When he finally shows up, blunt and grapefruit slices in hand, the pair set to tweaking the backing track for “Grace, Alek or Naomi.” The engineer hasn’t sent the stems for any of his new songs and the show is tomorrow. They’ll have to use what they have on file instead of making a new version specifically mixed for a live venue.

The songs on Riot Boi are the result of three years, three mixtapes, and two EPs’ worth of sonic and lyrical experimentation; together they comprise a compact, meticulously crafted 43-minute mission statement that turns Le1f’s complicated existence as a black, openly queer, private school-educated rapper into a candy-coated, hedonistic, 8-bit rave. Each song is a new level in a ‘90s Super Nintendo game, teasing and condemning the establishment — homophobic, racist white people, in particular — into confronting the realities of his daily existence.

Riot Boi calls on the spiritual aid of an array of women, from Grace Jones to model Alek Wek to the damn Mona Lisa, as well as the literal aid of others, like compatriots Junglepussy and House of LaDosha (particularly in “Swirl,” a lush, wry track about the fetishization of blackness). Even his mom, credited as “Miss Geri,” contributed her talents, alongside Devonté Hynes, on the crisis-of-faith album closer “Change.” On songs like “Cheap” and “Taxi,” he explores complicated race and class politics disguised as diary entries (“Hippie hoodrat, product of my environment … N*ggas look at me must be thinkin’ I’m mega-rich”). Slather his signature ballroom-culture vibe heavily on the front end, and not even Kathleen Hanna made smashing the Patriarchy (and White Supremacy) such a jam.

And then there’s that album title — a direct nod to riot grrrl, another radical music tradition, albeit one that had its own blindspots to racial and socioeconomic issues. “That aggressive ideology, that [spoke] to the music of the time, and still invoked all these political, progressive, necessary topics, was so interesting to me,” he says. “But really, in my mind, there are two lanes for [cool] political music: that, and conscious rap. And I hate conscious rap. So the working title, to keep it in my head to not make conscious rap, was Riot Boi. And it just stuck.” The point being that in order to be truly subversive, the music has to be, on some level, fun. “I want Republicans to play it, and be like, ‘Oh my God, that’s cute,’ you know, and not realize what just happened,” he said at the diner, with another endearing chuckle.

At the studio, after a few minutes’ banter and careful grapefruit consumption, Le1f becomes quiet. He seems stressed about having to make a placeholder track for the show instead of using a real one. It’s probably not helping that the illness he had been worrying about at the diner seems to have gotten worse.

But it might also have something to do with the fact that Riot Boi is about to put everything he is — artistically, politically, spiritually — on the front line. Yesterday, after an hour of more lighthearted conversation, he had casually mentioned that Riot Boi is dropping the same day as albums from Justin Bieber and One Direction. He mentions it again today.

“Yup, both of them,” he says grimly. For that reason, he adds, he has no expectations for the impact it might have on his career. “I’m happy it’s coming out. I’m interested to see if people like anything.” He’s already released two singles, “Koi” and “Rage,” but the reception hasn’t inspired confidence.

“My YouTube ratio is kind of… I don’t know,” he says. The video for “Koi” came out two months ago and has just over 242,000 views. “Some people kind of like it?”

I reminded him that earlier he had called himself “an acquired taste.”

On Friday afternoon before his Cameo Gallery gig, Le1f blows off his soundcheck at the last minute. Apparently he’s still sick, and his manager tells me he’s conserving his energy. He finally arrives sometime just before midnight, after three or four acts have come and gone and his Terrible / XL labelmate Empress Of is just kicking off her penultimate set.

Cameo’s backstage area — if you can call it that — is essentially a fluorescent-lit storage closet. It’s currently almost as sweltering as the main room, and made even smaller by too many people talking in a small space. It’s too crowded even for its original purpose — Empress Of’s band don’t even bother storing their gear when they finish, instead taking everything from the stage directly back out into their van. Half the room looks a little too pleased to be here, but Le1f and his friends look perfectly at home, draped across the staircase like it’s a stoop, laughing and poking fun at one another amidst the din.

When he finally takes the stage wearing a massive knee-length tunic open to his waist and the bucket hat from the other day, he clears his throat several times and wanders around for a few minutes. The Cameo crowd has thinned out a bit, but nevertheless cheers loudly when he purrs into the mic, “Are we readyyyyy?”

Claps and horns pepper the air — it’s “Wut.” Without dancers or any specific choreography, he performs it like a karaoke warm-up, hitting all the right syllables but obviously on autopilot. It’s not long before his sickness gives way to the groove, though; by the time he unleashes “Koi,” words like “chemistry” and “cunt” drip off his tongue with his hallmark zeal. By “Lisa,” a trap beat that’s been warped like a bad VHS tape, he’s in full rap posture, delivering raw lyrics like “Homie, you could die from a side-eye” with Nina Simone seriousness while maintaining his ball-winning strut.

The room has been tropically hot for hours, but Le1f still survives a few songs before repurposing his shirt as a headscarf and letting loose on “Grace, Alek or Naomi,” dripping with sweat. The rest of the set, he’s on fire, getting down in fans’ faces; the crowd doesn’t devolve into a full-blown rave, but it hops until the very end. The backing track might not be perfect, but no one can tell the difference.

“This is the last song I’m gonna put you through tonight,” he assures the crowd, and goes right into “Rage,” Riot Boi’s first and most electrifying single.

He closes his eyes and crouches at the edge of the stage during the singsong intro, but when the chorus bursts in — “RAGE! RUH-RUH-RUH-RUH-RAGE! RAGE!” — he explodes to his feet, his energy doubling to meet the glass-shattering, corrosive, Saul Williams-style beat. The song relents here and there, returning to nursery rhyme melodies that gently explain he’s just here to have a good time — but of course, it’s the fury of the choruses that really matters, the screams of frustration (“I’ve been dealing with too much shit / All these pricks is ignorant”): a mutant situation all his own. When he shouts, “Push me to the front! / Pull me up! / Give me what I want!” it’s obvious he means it.

Nobody was more surprised by Xbox’s fall from grace than Microsoft.

In its first eight years, the Xbox 360 established the company as a powerhouse in the games industry, and one of the best options for streaming apps like Netflix, Hulu, and Pandora. As entertainment trended away from physical media and cable subscriptions, the Xbox 360 felt like a test run for Microsoft’s plan to one day control the living room.

With each opulent press conference announcing a new product, Microsoft seemed more confident. When it launched the Kinect motion controller in 2010, Wired called the event “the most lavish product launch” in video game history — complete with Cirque du Soleil performers, an animatronic elephant, and an Xbox orb large enough to crush the front row. By the time the company announced the Xbox 360’s successor — Xbox One — three years later, the company was riding a tsunami of success.

In theory, the Xbox One was an improvement on everything fans loved about the Xbox 360: a more powerful Kinect, new hardware that merged the console with your cable box, and lots of talk about the cloud. Microsoft only forgot one thing: the games. During the hour-long kick-off presentation, six minutes were allotted to new game announcements.

Fans felt betrayed and poured their fury into forums and game blogs; critics dubbed the event a disaster.

Leadership tried to pivot, using subsequent press events to announce the return of beloved franchises like Halo and Killer Instinct. But the messaging remained muddied and unfocused, requiring constant backpedaling. Because of the Kinect hardware, the Xbox One cost $100 more than its closest competition, Sony’s PlayStation 4. When Microsoft finally debuted the system on November 22nd, 2013, it landed with a thud. It’s been trailing PlayStation ever since.

Microsoft hasn’t abandoned Xbox, though. If anything, it seems to be doubling down by investing hundreds of millions of dollars, and adopting an audacious strategy to turn hundreds of millions of Windows users into Xbox customers. But first, Microsoft has one priority: win gamers back, at all costs.

“This year really felt like I came home,” says Shannon Loftis. As general manager of Microsoft Studios, Loftis is in charge of producing the next crop of Microsoft video games.

We’re sitting in a break area on the third floor of Xbox’s headquarters in Redmond, Washington. A life-sized Halo statue, guns ready, stands in the lobby beneath us. It’s mid-morning and quiet: a couple dozen employees pass by, but all you hear is the mechanical clack of the coffee machines and the dull pop of a virtual grenade detonating in a testing room a few doors down.

In an industry dominated by men wearing ratty graphic tees and sneakers, Loftis is chic and professional. She embodies confidence. Xbox employees talk about Loftis the way soldiers brag about a takes-no-shit commander. She has a reputation for being fierce and protective, demanding a threshold of quality and timeliness, but shielding studios from forces that threaten to impugn creativity.

A 22-year Microsoft veteran, Loftis has seen the company through good and bad, multiple times through. She was there in the late 1990s when Microsoft used games to entice home computer owners to Windows. In the early 2000s, she produced games for the original Xbox, and was involved in that console’s efforts to connect console gamers via the internet for the first time. She oversaw European game development for Xbox 360, including beloved brands like Crackdown and Fable. And when Microsoft began work on the Kinect motion controller, Loftis headed an internal studio that produced a batch of games and core tech. Most recently, Loftis served as general manager of Xbox Entertainment Studios, a defunct group assembled around the Xbox One launch to produce documentaries, films, and television shows around gaming franchises.

Today, Loftis, like Microsoft at large, is refocusing on producing games — a large part of her job is finding the developers to create Microsoft’s next big franchises.

Not every game made by Microsoft is made by Microsoft. Instead of producing games entirely in-house (see Nintendo’s Super Mario Bros.) or banking on third-party providers (like Activision’s Call of Duty), Microsoft often relies on what some have come to call “second-party” partnerships — deep collaborations between console makers and outside studios. It’s a win-win relationship: Microsoft Studios taps ready-made, talented teams of developers, and developers get financing, marketing, and support from Microsoft’s internal team of producers and engineers.

The approach has proved remarkably successful: Dance Central; State of Decay; and Gears of War, one of Xbox’s biggest franchises, began as second-party games. Both parties walk away with big credits under their belt and bags of cash in their bank accounts.

To revive Xbox, Microsoft needs to land the next big franchise. To do that, the company must persuade game developers — with all of their talent and experience — to collaborate with them. And to do that, outside developers and potential second-party partners must believe in Microsoft.

That’s why Loftis’ first action as general manager of Microsoft Studios was so crucial: she immediately consolidated the small groups that made up the organization into a single entity, streamlining workflow.

“I see our primary role really as being the proof point,” Loftis says. “We prove to gamers why it’s worth playing on the Microsoft platforms. We prove to game developers why it’s worth [making games for our platform].”

Collaborating with Microsoft Studios needed to be faster, easier, and better. The company’s future depended on it.

Microsoft is betting big on three properties for 2016: ReCore, Crackdown 3, and Quantum Break. All three began under previous leadership and company battle plans; none has felt the changes of the last few years more acutely than Quantum Break.

A third-person shooter, Quantum Break puts players in the snazzy dress shoes of Jack Joyce, one of two people given unique time-warping abilities during an experiment gone sideways at a fictional, MIT-like university. As time collapses around him, Joyce must use his powers to stop the malevolent Paul Serene, a man who can see into the future.

The story lives in a high-brow, low-brow greyzone, somewhere between the films of Christopher Nolan and Roland Emmerich. Judging from teasers, it could break in either direction, but there’s no denying the sheer awesomeness of every raw gameplay demonstration. In a demo screened last year, a humongous boat crashes into a bridge, and time stops as the hero leaps across pieces of debris to safety.

But behind the scenes, Quantum Break has had a bumpy road.

One of the two new games announced at the Xbox One reveal in 2013, Quantum Break was initially a poster child for Microsoft’s push to make the Xbox a multimedia convergence point. Streaming television and playing games wouldn’t just happen in separate, discrete apps. In games like Quantum Break, they’d blend together. In Quantum Break, the third-person shooter would be intercut with a live-action video from the villain’s point of view.

That means the team behind Quantum Break had to learn the ins and outs of TV production with help of teams in Los Angeles and Redmond, thousands of miles from the developer’s office in Espoo, Finland.

Then in 2014, Microsoft shifted gears. Microsoft Entertainment Studios was shuttered, and the Kinect was no longer automatically bundled with Xbox One. Quantum Break creative director Sam Lake compares the iterative process of the console’s development to that of making a game.

“Step by step,” Lake says, “certain things that had been more important [for Xbox One] became less important, where others turned out bigger than in the original plan.” He says at times Microsoft’s adjustments affected the “kind of important and strategic benefits of certain features in the game.”

Quantum Break still includes a television show — four episodes, 20 minutes long, complemented by some shorter cut scenes — but much of its marketing has emphasized its place in the shooter genre. Players will even have the option to skip the TV episodes altogether.

Lake sounds happy with where the project ultimately landed: at its core, it’s always been a traditional third-person shooter. “First and foremost, it’s this big, story-driven, cinematic action game,” says Lake. “It is a core game, in many ways. With some trial and error and experiments along the way. I think that’s where Xbox One [as a console] has landed, too — where games matter.”

The evening before I meet Loftis, I sat down in a windowless meeting room with Chris Novak, Xbox’s design architect. He has two decks for me: a PowerPoint presentation and an actual, physical deck of cards. This, Novak tells me, is the presentation an outside developer gets when Microsoft is considering a partnership for a new game.

Novak is implacable in age, sporting a pressed T-shirt, a leather jacket, and gelled hair. He could be 20 or 50: he has the energy of a teenager with the tenacity of someone who works in an industry where million-dollar projects go belly-up in a month.

Novak cues up PowerPoint. “The goal [of this presentation],” Novak explains, “is to get the developer to really understand what we can bring to bear, to help them make the game they’ve always wanted to do.”

The meeting is equal parts sales pitch and bonding exercise. Novak emphasizes Microsoft’s sheer scale — “when we have a company meeting, it takes up a stadium” — and stresses that partners get to lean on a 100,000-person-strong tech and entertainment company with competitive knowledge and access to everything from dark fiber networks, to office complexes full of servers, to an extensive user research division that pushes through tens of thousands of people each year. “We are a huge company,” he tells prospective developers, “think about how to wield us.”

Novak then fans the deck of 100 cards across the table.

Written on each card is a cheeky headline followed by a short blurb. A lot of the cards seem like obvious ideas — “Gameplay – Gameplay is King” — while others point to merchandising campaigns and cross-media productions. A handful of cards are labeled “Showcase” — these represent the most expensive and expansive gambles, the pie-in-the-sky plans Xbox wants to emphasize with its new generation of games and hardware.

Novak asks me for a game idea, and I blurt a series of unrelated words: a massively multiplayer zombie-bunny adventure — like Destiny but cuter.

Unfazed, Novak immediately lifts a card labeled Cloud Computing.

“We have this thing called Azure,” Novak says. “It’s billions of dollars invested in data centers all around the planet.” Novak is off, explaining how cloud computing can enable the zombie-bunny adventure game to rethink the traditional MMO. How would the game look if it had a shelf life of one year, five years, a decade? What is it we really want to do? What is our big ambition?

Somehow, we start discussing cancer.

“Our vision is that this solves cancer,” says Novak. “Zombie Bunnies is going to solve cancer. Okay great, how can we do that? We need the cloud architecture to not only enable matchmaking in the MMO, but also to do the DNA supercomputing crunching in the background.”

Novak doesn’t actually believe Xbox or any of its games will solve cancer. What he’s really after is the core purpose of a game — any game. Novak wants to know why you want to make a game, and how Xbox can help.

“In my experience,” says Novak, “there’s a card in here for every team.”

On occasion, Microsoft Studios goes directly to a developer with a card in mind.

For instance, Novak says, “we wanted to take billions of dollars of risk in order to make [cloud gaming] happen. But we needed a dance partner to kick those things off at the game level. We were effectively waiting for that person to show up.”

That person eventually showed up, and his name was Dave Jones. He is the creative director at Reagent Games, the lead studio developing Crackdown 3.

Crackdown 3 is flabbergasting to watch. Players come together to fight crime as well as destroy an entire futuristic city piece by piece, from the newsstands to the skyscrapers. Architecture splits into thousands of brittle pieces, and entire skylines fall in catastrophic domino effect.

But what looks like an explosion-fest from the outside, is, against all odds, an investment in Microsoft’s global cloud computing ambitions — for real.

Dave Jones is a godfather of the now ubiquitous open-world genre. In 1988, he founded DMA Design — the developer of the first Grand Theft Auto. In 2007, he took the genre online with the original Crackdown. Jones has also co-founded numerous video game developers and tech companies. His latest startup, Cloudgine, intends to improve the responsiveness of cloud computing. If it works, video game developers won’t have to rely on customers to buy the latest expensive hardware — instead more and more computing will happen on far away servers.

An early demonstration of the tech shows two identical skyscrapers bursting into pieces: one, running on traditional hardware, slows and stutters until it becomes a glorified slide show. The other, taking advantage of additional cloud computing, runs smoothly as chunks of building break and scatter across the screen.

Because Crackdown 3 doubles as a research and development opportunity, Microsoft is able and willing to toss in its support. When Jones’ team needs to better understand cloud infrastructure, Microsoft provides access to data centers. When the Cloudgine team has questions about cloud infrastructure, Microsoft’s experts give answers. If Crackdown 3 sets a precedent, Jones tells me, it could revolutionize the entire industry.

And both Xbox One and Microsoft, with years of experience already under its belt, would be positioned to benefit immediately. “Frankly,” says Novak, “we are changing the way that [Microsoft cloud computing platform] Azure’s functionality works because of his game.”

Not every game that comes through Microsoft is a multi-million-dollar gamble spanning multiple studios across the globe, nor is every game expected to reach the masses. At least not at first.

“I actually love this notion of finding seed IPs,” says Loftis, “and helping game creators realize their very first vision. The more we can engage with that developer, the more releases we can put into the marketplace, the more likely that it might become one of these franchises.”

Loftis points to State of Decay.

Initially released as a downloadable title for Xbox 360 in 2013, the open-world zombie game sold over 250,000 copies in its first 48 hours, and broke the million mark within its first year. The game didn’t always run reliably, and bubbled over with bugs. But its core mechanic — the player embodied many survivors, all of whom could die permanently — was fresh to the genre.

“It was very much an experiment,” says Loftis. In 2014, Microsoft Studios signed a multi-year, multi-game contract with the studio behind State of Decay, Undead Labs. Loftis would not comment, but rumors around the project describe a substantially larger, more ambitious, and presumably more expensive production for a massively multiplayer online game.

The “seed” model encourages creativity while minimizing financial risk — similar to piloting a TV show before ordering a full series. Downloadable games have no cross-country shipping fees, no disc manufacturing, and no partnerships with any brick-and-mortar storefronts. Thousands of developers have left their jobs at thousand-person studios to make smaller, more personal projects, giving Microsoft a bounty of talent to partner with.

ReCore is the youngest of Microsoft’s big titles for 2016, announced just this summer. If Quantum Break represents Xbox One’s early ambitions, and Crackdown points toward the future, ReCore fits snugly within the company’s current “games above all else” ethos.

Little is known about the project developed by Armature Studios, an Austin-based developer formed by the heads of the Metroid reboot for Nintendo, and Comcept Inc., an Osaka- and Tokyo-based developer founded by Keiji Inafune (Mega Man, Onimusha, Dead Rising).

Pressed for details, Armature co-founder and game director Mark Pacini humors me:

“ReCore, at its heart, harks back to a different era in gaming,” Pacini says, “and trying to put a new spin on it, based on games that myself and Inafune-san have done in the past that aren’t really being done anymore.” He name-drops games like Metroid and Zelda as possible inspirations.

When Pacini first met with Microsoft to discuss collaboration, he was skeptical. He’d been disappointed by the Xbox One’s reveal, particularly the emphasis on film and television, and was unsure he had interest in the console. Big publishers, he explains, also have a stigma of being too controlling, gradually expunging creative freedom from their partners with endless demands.

Microsoft came with a clear idea: a game for gamers. If ReCore really does turn out to be the Metroid / Zelda-like throwback Pacini hints at, then Microsoft will get what it wanted.

Pacini says Microsoft Studios has been involved but not prescriptive, a relationship that benefits from a mutual understanding of what both parties wish to create and an abundance of resources.

Where the development of Quantum Break evolved with the changing strategies and leadership at Xbox and Microsoft Studios, and Crackdown constructed itself around its proprietary technology, Pacini describes a much more streamlined approach for ReCore. That means a pure gaming experience, free of distractions like Kinect and SmartGlass.

“Those things just kind of get brought up less often,” says Pacini, “and all of a sudden they’re never brought up in a conversation anymore, and it has nothing to do with anybody fighting for one thing or another. It’s just a natural part of consumers responding to certain things they like or dislike about the hardware.”

“It feels like, for us, the runway’s been kind of cleared. So we can just concentrate on making a really good game and putting features in that make sense for this game.”

Are things easier, I ask, with Xbox’s new leadership, like Loftis and Xbox head Phil Spencer?

“With a lot of publishers,” Pacini says, “the immediate group that you’re working with — the publisher development group — they’re always kind of trying to get it past marketing people, past the businessman, past whatever. Our group has to get it past [the team at Microsoft Studio]. That’s a huge difference.”

This team, Pacini explains, has a history of making games; they know what a successful one looks like.

Back in the quiet corner of Xbox HQ’s break area, I have a question for Loftis that is so obvious and so big that I’m a little embarrassed to ask: Why? Why does Microsoft bother making games?

The company makes billions off Windows and enterprise software, and Xbox is a comparably small portion of the profit pie that, as of late, has attracted a good deal of bad attention. Xbox headquarters in the Redmond campus is surrounded by countless other buildings, many larger and with more complex and lucrative-sounding names on their signage.

“I think one of the best reasons for Microsoft to continue to engage in games is just because gamers are also consumers,” says Loftis. “That loyalty, the passion that they bring to their gaming and entertainment experiences, the communities that they build.”

Loftis offers Kinect as an example. The motion controller began as a peripheral for games on the Xbox 360, but even before Microsoft released an official version for Windows, hackers and do-it-yourselfers began finding other uses for the device, from art installations, to medical applications, to interior design tools.

“These incredible experiences,” Loftis says, “were so far beyond anything that anybody that worked in the Kinect program had ever visualized. It created a fly wheel around a creative ecosystem. That’s just good for the world. What’s good for the world is obviously good for Microsoft.”

The hard truth is that Microsoft’s renewed focus on Xbox and games and the gamer is bunk if nobody owns the hardware to play them. But the solution isn’t and can’t be as simple as selling more Xbox Ones.

Microsoft’s big mistake with the launch of Xbox One was ignoring how the typical person consumes media. At launch, the hardware brought together our various forms of entertainment on a singular device. You could play a game, have a Skype call, and receive fantasy football stats all at once, and in the same place. This was a Jetsons-like fantasy 15 years ago, in the era of the original Xbox.

But our media habits are neither as static nor as constrained today as they were in 2001. When we read a Kindle ebook or stream a Netflix show, we may start on a laptop at work, switch to a smartphone for our commute, and finish on a tablet or television at home. In 2015, we expect media to follow us wherever we go.

Enter Minecraft.

A little over a year ago, Microsoft acquired Minecraft, along with its developer Mojang, for a reported $2.5 billion. Yes, billion. If you haven’t heard of the Lego-like online game in which players construct entire worlds with friends, ask the nearest child or parent. For them, the game is inescapable.

Minecraft has sold over 70 million copies, making it one of Microsoft Studio’s most popular brands. The most successful Call of Duty sold 26.5 million copies. The most successful Grand Theft Auto sold 54 million. Minecraft has sold more copies than Halo 1 through 5 combined.

Minecraft’s global domination stems from its extreme availability. You can play it on Xbox One, Xbox 360, Windows, and a dozen other platforms, including Microsoft’s competitor PlayStation 4, Apple’s iOS, Google’s Android, and even Amazon’s FireTV.

“An acquisition like Minecraft,” says Phil Spencer, the head of Xbox, “makes so much sense because Minecraft is so pervasive both as a service, as video, obviously as the game across so many different devices — it really is a manifestation of what I think gaming can be.”

You can’t help but wonder if Phil Spencer was created in some unmarked moonshot laboratory to be equally palatable to both board members and customers. At public events, the man who oversees the future of Xbox hardware and software wears indie game T-shirts under a blazer. His conversations seamlessly weave between corporate strategy and ornate opinions on forgotten video games of the 1980s.

He also sits on the board of Windows 10.

“It gives me a real good insight into what we’re doing with the service and how gaming can show up in Windows itself. I feel good about the early steps there. It opens up a ton of opportunities, I think most importantly for our development partners.”

In the months after Minecraft’s acquisition, the company has begun merging Xbox and Windows, not in half steps, but leaps.

A new Xbox app for Windows 10 lets Xbox One owners stream their games onto their computers, allowing them to play in a different room. Features like messaging, voice chat, and activity feeds are included. It’s a work in progress, Spencer says, with plenty of additions and updates to come.

Over a billion computers run Windows, and over 100 million devices run Windows 10. They’re about the only numbers in software that dwarf Minecraft’s. “As we embrace [Windows users] as an active part of the Xbox community, it opens up opportunities for our first-party games,” says Spencer. He calls Xbox’s proximity to Windows 10 “beachfront property.”

As of its fall 2015 system update, Xbox One now runs on Windows 10. Along with an improved user interface and the option to play over a hundred Xbox 360 games on the hardware, the shift also made the Xbox experience, in the words of my colleague Tom Warren, “so much faster.”

Both Spencer and Windows 10 leadership have talked about the streamlining of development the operating system provides, painting a future in which apps created for one Windows 10 device work on all other Windows 10 devices. In April, Microsoft CEO Satya Nadella told investors that Windows 10 “will be a service across an array of devices and will usher in a new era […] where the mobility of the experience, not the device, is paramount.”

The initiative has already begun to seep through Microsoft Studios.

Lionhead Studios, responsible for the Fable series, is developing the next Fable Legends for Xbox One and Windows 10. Rare Ltd. is developing its multiplayer adventure game Sea of Thieves the same way. Core franchises are making their way to PC, too. Albeit slowly. Gears of War: Ultimate Edition, a polished remake of the original, is scheduled for release on Windows 10 in the future.

When I ask why the latest Halo isn’t available on Windows 10, the head of 343 Industries and the Halo franchise, Bonnie Ross, points to the headway the franchise has already made away from consoles: Halo mobile games; the Halo Waypoint social platform; and the upcoming Halo Wars 2, a real-time strategy game for both Xbox One and PC.

“It’s being deliberate,” says Ross. “So yes. We will, of course, play a critical role [on Windows 10], as I think that we’re an important IP for the company. It’s doing the right thing at the right time.”

If and when the time comes for new Halo games to appear on Windows 10, the thinning line between console and computer will have all but disappeared.

If the Xbox One launched with a catastrophic lack of focus on games, it’s fair to ask whether this Windows 10 strategy doesn’t also presage yet more problems, another move to minimize console systems.

“I love that console experience,” says Spencer. “It’s obviously the core of what Xbox is about, and Xbox One is our most important gaming device inside the company. There’s no doubt about that. We also know there are millions and millions of gamers who access Microsoft [in other ways,] whether it’s Windows, their phone, Skype — it could be many different things every day.”

“As head of gaming inside of Microsoft, I think about how we make sure that all of those customers feel as supported and engaged as the Xbox fans do who own a game console.”

Loftis describes the expansion as broadening their demographics and making gaming “more inclusive across the board.”

“Most of the games that we’ve announced,” says Loftis, “and that we’ve talked about so far are obviously very controller-centric,” says Loftis, “but believe me we are exploring as many ways to make all of our experiences relevant across all of our platforms as we possibly can.”

As televisions, computers, and smart devices become the all-in-one-devices that seek to displace consoles, does Xbox just become an app or a storefront or a brand?

“No,” Spencer says, “I fully expect that you’ll see another console from us […] Our best customers are Xbox console customers, and I want to keep those people engaged both on the Xbox One and anything we might do in the future. I’m 100 percent committed to that.”

“I don’t want to dilute what the Xbox console customer feels,” says Spencer. “I want to expand what we’re able to do for more customers.”

Before we wrap our interview, I ask Loftis about making games in the pre-Xbox era.

“We were a little bit of a renegade group,” she says, “making games back in the mid-’90s for Windows. In those days Microsoft was aware that Windows was a huge phenomenon but that in order to keep Windows relevant we needed to make sure that we were continually pleasing consumers. Not just back offices, but to have Windows at home as well.”

“That was really our initial purpose. There was this emerging thing called gaming at home — let’s latch onto that and see what we can do.”

Two decades, multiple leadership shuffles, and billions of dollars later, Microsoft’s original game plan comes full circle with a twist: court gamers, regardless of where one finds them. Windows or Xbox, work or home, somewhere in between. Wherever you go, there Microsoft is.

A couple weeks after my visit to Redmond, Shannon Loftis got another promotion. Though technically her title remains the same, she now shares leadership of Microsoft Studios with Hanno Lemke, the general manager of Microsoft Studios in Europe — the same studio where Spencer and Loftis first worked together — along with the company’s first-party studio heads. Loftis now has even more say over bigger games.

Loftis will also become a public face for the company. In years past, former Microsoft Studios VP Kudo Tsunoda and Spencer made appearances on stage at Xbox’s biggest and most lavish events. Now Loftis will get her chance. Maybe she will bring back the swagger and spectacle of the past.

You can imagine it now:

Loftis is lowered to the stage on a literal Xbox-shaped cloud. She steps up to the podium, and with a confident grin, proudly announces the latest entry in a property she nurtured from a small downloadable game to AAA blockbusters.

“You can play it this fall,” she’ll say, “on Xbox One and all Windows 10 devices.”

In time, says April Underwood, Slack’s VP of product, large enterprise software companies are going to invest in building smart bots. And smaller developers will be able to rely on those companies’ expertise in machine learning, so that they don’t all have to build their own engines for interpreting our texts. “I predict by the end of 2016, we’re going to see more really great examples of household-name companies creating great bot experiences,” she says.

Others speculate that the promise of bots may be exaggerated: the most ambitious bots of the day, Operator and Facebook M, are still in closed beta. And for the most part, bots still depend on platforms they don’t control (iOS, SMS, Slack) — which some venture capitalists believe limits their growth potential.

Benedict Evans, a partner at Andreessen Horowitz, told me that bots face a big challenge as they grow: the number and volume of human desires might always outpace our ability to write software that can address them. “You’re writing recipes,” he says. “And how many recipes can you write?”

There’s an interface challenge, too. Give people a blank box, and they may not know what to ask for. “People try to bolt AI onto every new user interface model. But we don’t actually have HAL 9000, and may be 50 Nobel Prizes away from that,” Evans posted on Twitter a few days after we spoke. “The problem with ‘no user interface’ is that, since you don’t really have HAL 9000 behind it, it’s almost as opaque as a DOS prompt.”

But with Facebook and Google now investing in building virtual assistants, bot makers are betting that people will learn. “Inevitably, they’re going to get educated on this interface,” says Operator’s Chan. And in many ways, they’re educated already. “The messaging interface feels very familiar to anyone who’s ever sent a text message,” Slack’s Underwood says. You’ll text your boyfriend to pick up pizza on his way home — why wouldn’t you text the same thing to Facebook M?

The fact is, as investor Semil Shah has written, messaging has usurped the browser on mobile devices: it’s where most of our activity takes place. And once you’ve dethroned the browser, which empires will crumble? Could a new e-commerce channel rise to challenge Amazon? Could a bot outdo Google when it comes to understanding what you’re looking for?

Those companies are rich enough to buy up competitors before they become existential threats. But could you disrupt struggling Foursquare or Yelp with a conversational UI? (See Luka.) What about fitness apps? (See Lark.) What about banks? (See Digit, or Penny.) Which media companies are poised to succeed in a world where we consume more information through text messages? (In the bot era, Facebook’s Notify app looks a lot like a media company.) What does a media company look like when it optimizes for Slack?

Does Siri add a text interface? Does Alexa?

And the web? The web won’t die; mediums never do. That said: “It’s going to continue the erosion of the power of the home page, and even the power of the search bar,” Brown says of bots’ rise. “The more software can notice the signals we’re sending it constantly, and deep-link us all the way to the answer, the less I have to go browse or search.”

Of course, all this depends on a still-rickety infrastructure of services and machine learning actually coming together. It depends on an automated messaging interface that feels as trustworthy as a message to your boyfriend, rather than a mysterious DOS prompt. And it depends on bots living up to their billing as friendly, powerful, all-knowing assistants. “Building an AI that can be ‘everything to everyone’ is obviously appealing, yet we’ve seen pitches of this sort fail dozens of times over,” says Maran Nelson, founder of Clara Labs, which is building a virtual assistant that schedules meetings. “The technical challenges to delivering on this promise at scale make this almost inconceivable.”

For now, using Facebook M to order takeout burritos feels like a novelty. What will it take to feel like a superpower? Is it coming within reach, as so many entrepreneurs in Silicon Valley now believe? Or are we still several advances in computer science from getting there?

We’re about to find out. Existing bots are getting smarter every day, through the sheer volume of data passing through them. Across Silicon Valley, an armada of new intelligences is now under construction. If there’s a killer bot to be found on our phones, 2016 may be the year it says hello.

On May 6th, 2008, a package containing $68,000 in cash arrived at a FedEx store in Palo Alto, California. The bills had been washed in lantern fuel, as per instruction, then double-vacuum-sealed and placed inside the cavity of a stuffed animal, which was then gift wrapped. The store had been chosen carefully: it was open all night, and located just 500 feet from a Caltrain station. The package was general delivery, to be picked up at the store by a man named Patrick Stout.

The money was being closely watched. The package had been prepared by a criminal informant, working in cooperation with a joint task force of agents from the FBI, IRS, and US Postal Service, who were investigating a tax fraud scheme. The informant had been arrested and flipped months earlier, betrayed by yet another informant. Now they were after the mastermind.

Around 5 o’clock the next morning, the target appeared. A wiry white man in a dark hoodie came in through the back entrance, presented a driver’s license in the name of Patrick Stout, picked up the package, and left the same way he came in. He tore open the package near a dumpster behind the store, pocketing the stuffed animal, and headed toward the train. Two officers tried to follow, but he recognized the tail and slipped away. Agents rushed to the station but couldn’t find him among the early morning commuters. The trains could have taken him anywhere from San Francisco to San Jose, with connections to each city’s airport.

Just a few minutes after the pickup, Stout was gone. It was as close as the cops would get for months.

From there, every lead seemed to dry up. Stout’s driver’s license was fake: the address didn’t exist and the ID number belonged to a woman in Bakersfield. Stout kept talking to the informant, joking that he was becoming paranoid. He didn’t know who he’d spotted following him at the FedEx store, but he was shy about setting up another cash drop. A few weeks later, Stout had $18,000 in gold bars mailed to the same FedEx office, but by the time investigators found out, the pickup had come and gone.

The informant led the task force to a nest of bank accounts where he’d been instructed to deposit money, but they were all in false names — Sam Blat, Benjamin Cohan, Aaron Johnson. There was more than $400,000 spread across the accounts, all of which could have been shuttered and seized using evidence the task force already had — but that would just have spooked the target. They wanted to catch him, not to scare him off.

Their best lead was the IP address Stout used to file the fraudulent returns, which traced back to a Verizon Wireless AirCard registered under the name Travis Rupard. Rupard — or were Rupard and Stout the same person? — had bought it through another post office box with another fake ID, kept active just long enough to receive the device. The whole point of an AirCard is to provide internet access that’s not tied down to a fixed address, which made tracking down the owner tricky. When Rupard used the card, Verizon knew which cell tower he connected to — usually somewhere around San Jose — but they couldn’t tell much more than that.

Three months after the FedEx episode, on August 3rd, the task force descended on an apartment complex near the San Jose airport, rented in the name Steven Travis Brawner. Agents caught Rupard outside the complex, and served a search warrant on his apartment and storage unit later that day. They found $117,000 in US currency, 230 ounces of gold, and 588 ounces of silver, along with the dark gray hoodie tying him to the drop at the train station and a Verizon AirCard tying him to the bank accounts. By the time the case was over, the agents would recover more than $1.4 million.

The suspect was charged with 35 counts of wire fraud, 35 counts of aggravated identity theft, and three other miscellaneous charges — enough to keep him in jail for the rest of his life. Taking his fingerprints three days later, the police finally worked back to his name — not Rupard, or Stout, or Brawner, or Aldrich, or any of the others. His name was Daniel Rigmaiden.

But there was something else, something that wasn’t reported on the seizure affidavit, the complaint, or any of the documents that followed. To track Rigmaiden down, the investigators had used a secret device, one that allowed them to pinpoint their target with far more accuracy than Verizon could. They called it a cell-site simulator, or by its trade name, Stingray. Neither term was found in the court order that authorized its use. The device had to be kept secret, even from the courts.

The Stingray had worked perfectly. Agents traced the suspect’s AirCard back to his apartment and now had more than enough evidence for a conviction. But in the years that followed, that open-and-shut case would turn into something far more complex. Working from prison, Rigmaiden would unravel decades of secrecy, becoming the world’s foremost authority on the device that sent him to jail. By the time he was finished, a covert surveillance device and the system that kept it secret would be exposed to the public for the very first time.

This past October, I met with Rigmaiden in Phoenix, where he’s lived since he was released from a nearby federal prison in 2014. In person, he looks like a scrappier J.J. Abrams, with thick, black glasses and spiky, black hair. Intensely private, he declined to meet at his home, preferring a shopping center some blocks away. It was a mild day and the sun was low, so we decided to go on a hike in an expansive park south of town where the flat plain of the city rises into rocky hills. He doesn’t get out much these days, but the trail brought out his old outdoors instincts. As we climbed, he guided me away from brittle rock faces and possible snake pits. “Normally I wouldn’t even come out without a full pack,” he told me. “You’re not really supposed to.”

These days Rigmaiden is primarily concerned with the typical worries of a newly released prisoner — finding a good job and an apartment willing to rent to an ex-con. He likes his probation officer and seems to be adjusting well to life on the outside.

But the world has lots of ways of reminding Rigmaiden of his time locked up. As we came down from the hills, a City of Phoenix maintenance truck rolled toward us, the driver wearing a uniform we couldn’t quite place. When we crossed in front of the truck, the driver rolled down the window and pulled out a camera, training it up the mountain and above it, to the sky. Were we accidentally trespassing? Neither of us was sure. Rigmaiden has to make an official report for every interaction he has with the police, so even minor mistakes can become dangerous. But the official didn’t seem to care about us, and we walked past without being stopped.

It was a tense, awkward moment, but Rigmaiden is used to the feeling. A natural outsider, he has spent much of his life coming to terms with authority, legal and otherwise. “I tried the non-participation way for 10 years, then I spent five years in jail,” he told me. “Now I’ve made the transition to try to change things for the better.”

Born in Seaside, California, Rigmaiden left home just after high school, living in a string of college towns up and down the coast. He became an expert at forging fake IDs and did a good business selling them to beer-happy college kids online.

Eventually, he decided to get off the grid entirely. “I just didn’t want to be attached to the whole society system,” he recalls. “I needed to take a step back and take a break from it all.” He traded the college towns for seaside motels, or a tent and remote campsites. “It was peaceful. What I liked about it was, you had to do everything yourself,” he says.

He spent one summer in Big Sur, an isolated stretch of the California coast famous for its massive redwoods and dramatic precipices. He set up camp deep in the forest and developed a taste for free-climbing, scrambling over the area’s boulders and cliffs. He liked the exertion, the independence, but most of all the heightened sensitivity that came with danger. “It’s not panic,” he told me. “It’s almost like you’re glued to the side of the rock. It’s that calculated. Because you know if you fall off you’re going to die.”

Once, looking out over a 60-foot dropoff, the ground gave way beneath him. He found himself sliding down the face of a cliff. “I was grabbing saplings to slow myself down but I ended up just pulling them out. It was this burnt stump that saved me,” he told me. “Eventually, I got back by swinging sapling to sapling.”

Those adventures cost more money than he could make from selling fake IDs, so he hit on another scheme, filing tax returns for the recently deceased. He liked the scam, he told me, because he saw it as relatively victimless. He didn’t have to steal from anyone who was still alive. All he got was whatever refund money would be coming back to the deceased, often thousands of dollars for a single return. It proved incredibly lucrative, pulling in far more than he needed to survive. Rigmaiden developed a rhythm: working his ID and tax return businesses for six months, squirreling away the extra money, then taking the rest of the year off.

He was meticulous about hiding his identity. His name changed constantly, with a new driver’s license for each new storage unit or post office box. Each step of his scheme was painstakingly arranged: from the computers that filed the returns to the post office boxes where the refunds were delivered, to the couriers that picked them up and the bank accounts where they were deposited. He approached it like a mental puzzle. As far as he could tell, there were no links that could trace the money back to him.

Except for that AirCard. He didn’t think police had the expertise to trace it, but he knew enough about the physics of wireless signals to be sure it could be done. In radio terms, the AirCard was noisy, blasting out data in all directions like a barking dog on a busy street. By the time the signal reached the cell tower, it was mixed in with noise from thousands of other sources — but in theory, a person with the right equipment could always trace the bark back to the dog.

For years, the weakness of the AirCard was a hypothesis — but the moment federal agents arrived at his apartment in the summer of 2008, he knew it was real. When he was arrested, Rigmaiden’s first thought was that the AirCard had given him away. The task force had used some device, some device no one knew about. His second thought was that he didn’t want to take on all the work that would be necessary to prove it. But even then, he knew there was no other way. “I knew I was going to have to learn the legal system to get out of this,” he told me.

It was two months into his stay at the Florence Correctional Center that Rigmaiden discovered the law library. Since his arrest, he had been trying to plead his case while scrambling from prison to prison — first a local penitentiary in California, then to Florence in Arizona where the case was being tried. He did his best to keep a low profile at Florence, steering clear of mentally ill inmates and making do through random shortages of staples like toilet paper. He wrote note after note to his lawyer, scribbling on the back of Prisoner Request Forms with the 3-inch golf pencils issued by the prison, but it didn’t seem to change anything. He was facing a federal case, represented by a public defender. Making noise about invisible signals and secret surveillance devices was only going to get him in trouble.

Still, he knew the government was hiding something. He had seen the warrant for the raid on his apartment, but it gave no sense of how they had tracked him, attributing the information to “historical cell tower information and other investigative techniques.” But a single cell tower serves thousands of phones at once. Even detailed records couldn’t have been enough to pinpoint a single apartment complex. There had to be something else.

If he was right, it meant the agents had lied to the judge, which could be enough to get the case dismissed. But the key to his argument was buried in murky technical details that only he understood. His lawyer didn’t know the first thing about cell tower data. Neither did the judge. He found himself drawn to the prison library’s open hours, the three hours a week when he could actually get answers.

He fired his first lawyer, then a second, then finally got permission to represent himself

He made a valuable new friend at the library, a disbarred lawyer who was serving time for fraud. Rigmaiden learned about the rhythms of a trial, the motions and requests each side uses to stake out territory. Eventually, Rigmaiden opted for a strategy of total legal war, flooding the court with motions and proposals. “When you hire an attorney,” he says, “they have to pick and choose the things they’re going to challenge because of time and resources. But I just challenged everything.”

He fired his first lawyer, then a second, then finally got permission to represent himself, which let him bump up his library time to five hours a day. He worked six days a week, sometimes 15 hours a day. When he couldn’t print the motions, he wrote them out by hand, using the same nubby half-pencil.

Meanwhile, he was also poring through records for any evidence of the mysterious device that had caught him. In October, the court gave him access to his discovery file, 14,000 pages of documents that laid the groundwork for the prosecution’s case. In the second to last box, he saw the word “Stingray,” scribbled in one investigator’s notes. He thought it sounded like a brand name.

The prison library didn’t have internet access, but the prison’s case managers would Google Search for you if you asked them. Eventually, Rigmaiden found a Stingray brochure from Harris Corporation, advertising exactly the capabilities he’d suspected. Now he just needed to prove police were using it. He found what he was looking for in the minutes of a Maricopa County board meeting: a unanimous vote to buy police equipment paid for by a federal grant. Because there hadn’t been an open bidding process, the department had been required to submit its invoice for public consideration. The invoice was for a cell-site simulator device, built by the Harris Corporation.

In 1995, a hacker named Kevin Mitnick broke into a software company called Netcom, stealing email archives and security programs. He concealed his location with a modem-connected cell phone, a hacked-together version of Rigmaiden’s AirCard. To zero in on Mitnick’s cell phone, police used a passive cell-site simulator combined with a silent SMS from the phone company that forced Mitnick’s device to check in. It was primitive, but they were employing the same technique that would catch Rigmaiden more than a decade later.

In 1996, Rohde & Schwarz created a device called the GA 090, which bundled the ping and capture functions together, effectively masquerading as a cell tower. Driving through a neighborhood with your GA 090, you could see the unique subscriber ID for everyone within range, akin to seeing everyone’s cell phone number. If one number was particularly interesting, you could drive around the corner and take another reading to triangulate exactly which house or car it came from.

The device exploited a fundamental part of how cell networks are built. Mobile phones rely on constant communication with nearby cell towers, always listening for signals from a tower that might be closer or less congested than the current connection. As soon as the phone hears a signal, it spits back an identification number. But crucially, the signal doesn’t have to prove it’s coming from a tower. There’s no authentication for the first stage of the process, so a device like the GA 090 could slip right in.

Security researchers have been concerned about that flaw since the ‘90s, some even accusing phone companies of keeping it open just to allow law enforcement to exploit it. But by the time those flaws were being used against Rigmaiden, the topic had largely faded from view. Researchers knew the system was broken, but everybody else counted it as a theoretical attack and moved on.

Meanwhile, devices like the GA 090 were gaining popularity in the law enforcement community. In 2003, Harris Corporation unveiled the Stingray, a sleeker, smaller version of Rohde & Schwarz’s earlier model, and it came with an aggressive push into the US market. Intelligence agencies began using them overseas to surveil targets or identify their devices. US Marshals put Stingrays in planes and flew them over cities, collecting tens of thousands of phone numbers in search of a single fugitive’s phone. Over time, the devices trickled down to local police departments, where they could be used to track down anyone from murderers to purse-snatchers.

But purchasing the device came with a catch. Every time an agency bought a device from Harris, they signed an agreement to keep it out of public court records. If Stingray methods were ever entered into evidence, Harris argued, criminals would catch on, rendering the device useless. Agencies still got court orders to use the devices, but they usually looked like a vaguely worded request for phone records. In most cases, the judges never knew what they were signing and defendants never knew how they’d been caught.

He was a self-represented prisoner chasing a mythical surveillance device — sure signs of a crank

But they couldn’t hide every trace. Rigmaiden found signs of Harris’ Stingray device in random corners of the web. He searched through Harris’ patent filings, which gave him far more insight into how the devices worked. He filed requests for information on the devices under the Freedom of Information Act, but departments locked up, citing confidential methods. He studied The Fugitive Game, a 400-page study of the Mitnick case, scouring for clues on exactly how police had tracked the rogue hacker. There was no public evidence of the device at the federal level, no documents or statements indicating it was in use. He turned his attention to local departments, hoping they would be less careful. When his library time was up, he would take the documents back, poring over them in his cell. Over more than two years, he built a file that sprawled to hundreds of pages, including every last trace of the Stingray he could find.

Rigmaiden needed allies, so he sent his file to half a dozen different privacy organizations, but never got back more than a form letter. Not only was Rigmaiden a self-represented prisoner chasing a mythical surveillance device, but he had filed hundreds of motions — these were sure signs of a crank. A normal person would take the plea deal and resign himself to his sentence. Rigmaiden’s phonebook-sized file was testament to how unusual he really was.

Eventually, Rigmaiden sent his file to Christopher Soghoian, then a PhD student who had published work on wireless spectrum surveillance. There was something unhinged about the file, to be sure, but Soghoian knew that what Rigmaiden was proposing could very well exist. “My reaction wasn’t, what is this strange device. It was, oh I read about this in graduate school. But I read about it as a thing that was possible, not a thing that the police in Baltimore were using,” Soghoian says. He became convinced Rigmaiden was right.

Soghoian passed Rigmaiden’s file along to privacy groups like the ACLU and EFF, many of the same groups who had ignored the package the first time around. But the breakthrough came when he sent Rigmaiden’s sprawling file to Jennifer Valentino-Devries, a reporter for The Wall Street Journal’s Digits blog. The next month, her story hit the front page of the paper, revealing Stingrays to the public for the first time and presenting the unchecked, widespread adoption of the technology as a direct challenge to constitutional rights. When a copy of the paper made it to his cell, Rigmaiden was surprised. For the first time, he had reason to think his discovery would matter for more cases than just his own. “I knew that it wouldn’t be a front page story if it wasn’t such a big deal,” he told me.

Rigmaiden’s legal situation, however, hadn’t changed. Arizona prosecutors wouldn’t drop his case just because of a fudged warrant, and while the Stingray news weakened their case, they were determined to see it through. He kept up his barrage of motions: he filed 22 different motions to “continue trial,” forcing the prosecution to adapt to his unusual speed. One filing was entirely concerned with the use of the word “preliminary” in testimony describing a handwriting analysis. After the FBI started ignoring his FOIA requests, Rigmaiden launched a civil suit against the bureau, eventually filing almost two dozen motions on that case before it was dismissed. The criminal docket sprawled to more than 1,100 entries.

In 2013, prosecutors finally offered a plea deal on his criminal charges. Rigmaiden believes the deal had more to do with his persistence than the merits of the case. “The reason they wanted to get rid of the case wasn’t because they were worried the Stingray was going to get exposed more, because at that point it was pretty much already out there,” he says. “The reason they wanted to get rid of me was because I was doing all that work. I was giving them so much work to do and it was pushing their resource limit.” He was still reticent to take the plea — which waived his right to appeal — but prison was wearing him down, and he ultimately decided he could do more on the outside. In April of 2014, after nine months of deliberation, he took the deal and walked out of Florence a free man.

In the four years since the first Journal article, Stingrays have turned up everywhere. They’ve been documented at 53 agencies spread across 21 states, used in major cities like New York and Chicago as well as smaller departments in Memphis, Durham, and San Jose. In a 2015 review, Baltimore police admitted using the device more than 4,300 times, sometimes for crimes as minor as a stolen cell phone. The US Marshals have even lent the devices out to Mexican officials hunting down cartels. In each case, the departments had been using the devices for years, long before Rigmaiden came onto the scene.

Meanwhile, an informal network of defense attorneys has cropped up, with public defenders sharing notes on how to spot and deal with Stingray cases. Because of the intense non-disclosure agreements around the Stingray, most prosecutors will drop cases rather than defend the use of the device in the face of a well-versed lawyer. Rigmaiden has become a kind of in-house expert for those lawyers, tracking surveillance issues on Twitter and consulting behind the scenes. When Washington state started drafting a law to limit use of Stingray devices in late 2014, someone passed the bill along to Rigmaiden. For a month and a half, he spent each morning going over a draft of the legislation, demonstrating the same obsession to detail he’d shown in planning his tax fraud scheme and criminal defense. Working with a local ACLU chapter, he recommended tweaks to the bill, specifying the language and expanding the rules to cover passive surveillance devices and other variants that are still secret.

In September, the Justice Department issued a new policy on cell-site simulators, instructing all federal agencies to get warrants before using the devices. For a lot of privacy groups, it was a victory — the first clear federal policy on Stingray use, openly instructing agents to play straight with courts. Rigmaiden was less optimistic. “A lot of it is just putting a new face on what they’ve been doing all along,” he told me. “In my case, they had a warrant. The problem was the information on the warrant.”

Rigmaiden now lives a modest life. He gets no money from his legal work, and when I visited, he had just quit a telemarketing job in favor of a slightly friendlier web development gig. He has no car (his record also makes auto loans difficult), so the telemarketing job meant a long walk to a bus to a train. Last year, he spoke at a defense attorney conference at the University of Arizona, an experience he’d like to repeat if he can. For now, the terms of his probation stop him from leaving the Phoenix area.

After climbing down from the hills, we drove to another park, this one a flat stretch of lawn closer to the center of Phoenix. It was quiet, empty except for a few couples and about a dozen people gathered near a stage for a benefit event. It was, as Rigmaiden pointed out, exactly the kind of crowd where you might want to use a Stingray. He had an app on his phone called SnoopSnitch, which looks for telltale disturbances in the network, one of the open-source Stingray-hunting projects that’s sprung up since the device became common knowledge. That day, the data showed two disturbances, both silent pings from the phone company. Neither was necessarily suspicious, although we’ll never know why they went out.

Rigmaiden suspects police have moved on from Stingrays to passive receivers, which sniff signals out of the air without disrupting the network at all. It’s the kind of thing you could build yourself with open software and no FCC violations at all, although you’d need the phone company’s help to connect the signals to actual cell numbers.

Cell-site simulators are now at least 20 years old, a long time for any one trick to stay secret. Police had been using the devices in secret for 12 years by the time they were trained on Rigmaiden. From there, it took another eight years to drag them into the light. Even that was only possible because of the chance alignment of a stubborn defendant, a legal shortcut, and a sympathetic judge. “If we hadn’t picked up the scent on this, they could have gotten another five or 10 years out of it,” Soghoian says.

This is the logic of surveillance, an arms race between police and criminals, but also between police and the legal systems meant to keep them in check. After 10 years off the grid and five years in jail, Rigmaiden is now on the side of those systems — privacy groups, lawyers, judges. It’s a strange place to find himself. How did he make the turn from dodging surveillance to actually fighting it? He’s still not sure, although it probably has something to do with getting older.

“You have to realize when you’re in a situation where you can make a difference, and grab onto it,” he told me. “I don’t think those opportunities arise very often.”