<TitleType>01</TitleType> <TitleText>Margaret Preston</TitleText> <Subtitle>Recipes for Food and Art</Subtitle>

chapter one community
Some people just look the part of what they do. There’s no universe in which Beyoncé doesn’t become a pop star. The Rock appears to have been hewn from solid granite. Nicholas Braun, aka Cousin Greg on the HBO show Succession, was born to play that exact kind of guy. As far as people who’ve worked at Twitch go, this is Marcus “DJWheat” Graham. Whether he’s broadcasting from his office or chilling on his couch, the now-forty-five-year-old, salt-and-pepper-haired livestreaming pioneer is all wide expressions and big gestures. He moves like a cartoon character, as though he was animated to captivate. His voice booms no matter what he’s discussing, the volume knob on his enthusiasm level evidently stuck at 11, never to so much as glance at 10. You can just see, almost immediately, why this person in particular was one of the first successful video game streamers, predating even Twitch itself. Graham can talk for hours and hours and hours, and then more hours after that. He will; it’s just who he is. But you’ll find it impossible to look away, like a snake under the spell of a master charmer.

Right now, he’s talking about pinball machines, which he’s turned into the centerpiece of the lower floor of his Nebraska home.

“You never have the same game twice. So the adage of ‘Once you get a pinball machine, they multiply’ is totally true,” he says, gesturing toward favorites based on Teenage Mutant Ninja Turtles and The Mandalorian. “They’re amazing, and they’re actually better investments than, say, an NFT or something. They have been skyrocketing in price over the pandemic, partially because manufacturing costs have gone up and they produce less machines, but also people are like, ‘I want things to do.’?”

Graham’s appreciation of these machines is infectious. The craftsmanship of a good pinball machine, he explains, is a thing of beauty—a clockwork assemblage of flippers, knobs, springs, switches, lights, and speakers meant to spirit you away to some fantastical realm (or a sewer where giant turtles eat pizza) amid the din of a crowded arcade or bar.

This passion, this attention to detail, is what turned him into one of the first prominent esports commentators, one of the first well-known gaming livestreamers, and eventually, one of the first Twitch employees. Where Graham goes, people gather to listen—a quality that came in handy back when he began his career. In the late nineties, when Graham first started commentating over competitive games, there was barely an American esports scene to speak of. There certainly wasn’t software designed with the idea in mind or an easy means by which to broadcast online to more than a small handful of people. This eventually led him to streaming, which he also got into when the medium was in a larval form.

Graham’s career was only possible because of the people he was able to draw to his early, hacked-together broadcasts. He owes everything to the community he helped build, which in turn built him. This, more than anything else, is the lesson he’s taken to heart: Community is key. It’s the soul of his operation, the gasoline that fuels his passion.

It’s why he joined Twitch. It’s also why he left.

THANKS TO HIS dad, Graham grew up with a keen interest in radio. This led to a broadcast journalism major in college, which in turn led to a realization.

“I realized that radio was fucking dumb,” Graham says. “And what I mean by that is, radio has such a limited audience. And I’m discovering that as I’m realizing the internet is the future, right?”

So in 1997, after he’d graduated and begun working an IT job, Graham went to his local Best Buy and picked up a gadget that claimed it would let him start his own radio station online. Unfortunately, its audience was significantly more limited than radio’s.

“One of the bullets on the back of the box was ‘Start your own radio station’ with an asterisk next to it,” Graham says. “Like, you know, you need to have this stuff to run this, and of course I’m like, ‘Well, fuck the asterisk.’ I took it home and realized I could have five, at max ten people listen. But getting those first ten people for esports stuff was amazing.”

Graham’s game of choice at the time was Quake III, a multiplayer-focused entry in a classic sci-fi first-person shooter series without which modern genre pillars like Halo, Destiny, and Apex Legends wouldn’t exist. But at the time, online gaming was still relatively new, and competitions with money up for grabs were a niche within a niche, especially in North America. Big matches were infrequent and inconsistent, and central sources for news about them were difficult—sometimes impossible—to come by.

Initially, Graham commentated largely for the benefit of a Quake team he was coaching; he’d watch their games and record notes for them about how they were playing. It was a teammate who gave him the idea to start broadcasting to the (relative) masses.

“[My teammate] was like, ‘Hey, you should do this live like you’re covering sports,’?” says Graham. “And I was like, ‘Yeah, why not?’?”

With the help of an old friend (who’s since gone on to work at Meta), Graham managed to set up a server on a media service called Shoutcast that allowed him to reach more people than his previous, severely limited approach. At the time, this was about as good as it got.

“We filled up that five-hundred-person Shoutcast server almost right away,” says Graham. “And it was, like, the most atrocious thing ever. I had no idea what I was doing. I wasn’t even really thinking about it, right? It took me a little while to go, ‘Oh, now let me take my knowledge of radio and apply that to making an opening intro, [explaining] why you should care about this game, post-game interviews—you know, getting it all together actually.’?”

Radio expertise proved valuable on the tech side of things as well. Though Graham’s setup at the time was rudimentary, the bones of it were not so different from hardware utilized by modern streamers.

“It was hard as hell back then,” he says. “Even trying to get mics to work and stuff was unbelievable. Nowadays everyone’s got a [studio-quality microphone], and back then we were getting stick mics from Walmart…. But because of all the radio experience, I basically built a radio console that I could plug into a computer, which is what a lot of streamers still do today.”

What Graham and his friends didn’t recognize was that by broadcasting to nascent gaming audiences, they were also tapping into a force that would become essential to the DNA of livestreaming as we now know it: interactivity. Quake fans would gather in text-based internet relay chat (IRC) rooms and react to what was happening in matches, offer their own opinions, and of course, tell Graham when he was screwing up. (The latter desire, it should be noted, has powered far more pivotal instances of human ingenuity than anybody is willing to admit.)

“Someone could be like, ‘Wheat, you’re an idiot,’ and I could respond to them immediately,” says Graham. “We had a little bit more delay back then, but I was blown away by this instant interaction that we saw take place. So it wasn’t just the commentary part of it. Once it got tied to IRC, suddenly it became the interactive experience that I think people really love today.”

It’s this alchemic reaction of broadcasters and viewers/listeners creating collectively—cheering, jeering, and building on each other’s ideas in real time—that would later go on to define livestreaming as a medium. No longer was a broadcaster like Graham simply talking at audiences and occasionally taking listener calls or running contests. The two sides of the equation had become connected, tethered by a moment in time regardless of space. Graham knew he was onto something. But the road to fully realizing his vision of the future would prove longer than expected. And much, much bumpier.

IN THE EARLY 2000s, Graham took his first crack at making a career out of this whole “talking over video games” thing. He’d been commentating on larger and larger tournaments, and people with deeper pockets had begun to take notice. In 2004, Graham hit the jackpot: an American esports upstart called the Global Gaming League. Its owners, flush with investment cash, convinced Graham and his wife to uproot from Nebraska and move to Los Angeles for the purposes of taking esports and video game content to the next level.

“GGL let me experiment with livestreaming like crazy,” says Graham. “They were helping me pay bandwidth bills that were absurd. I went to them and said, ‘I want a livestream from [annual video game news conference] E3,’ and they were like, ‘Great, what do we need to do to make that happen?’ I said, ‘You just gotta get a $25,000 internet pipe to the booth on top of [what you’d already normally pay to be there].’ They’d be like, ‘Awesome!’?”

Even mainstream celebrities were beginning to suspect there might be gold in gaming’s still-untapped hills.

“GGL let us put together a hip-hop gaming league that had Snoop Dogg as a commissioner [and] Method Man and Eric V. playing Madden against each other,” says Graham. “We were doing shit way, way, way ahead of its time.”

However, this also marked the beginning of a pattern that would ultimately echo forward to Graham’s time at Twitch: a company making good, forward-thinking decisions, followed by bad decisions, followed by decent decisions, followed by worse decisions. For Graham, it was whiplash inducing.

“They would do these great esports events,” he says. “They were paying players. They were putting out good prize money. They were getting good sponsors. But then they’d do something stupid like say, ‘We want to rename esports to vsports so then we can own the name.’ So you can imagine what it was like, especially for me—knowing what you know about me and Twitch—working for a company that has these moments of brilliance and then the dumbest fucking ideas you could ever imagine.”

Needless to say, vsports, as a name, did not catch on.

Graham got an even bigger break in 2005, when a television producer named Mike Burks approached him at an esports event he was casting.

“I think I was casting in a Hawaiian shirt,” says Graham. “It was all sorts of goofy, awkward gaming nerd geekiness happening here. Occasionally you’d see a mom that came with their kid or whatever.”

Burks, not from the world of gaming at all, stood out—his twenties a distant memory, very tall and rocking a hairstyle older (both in fashion and literal age) than any of the competitors at the event. Burks was impressed by Graham’s ability to commentate effectively all by his lonesome without a large behind-the-scenes crew feeding data into his ear like traditional sports broadcasters typically had. Burks revealed to Graham that he was working with DirecTV on a pilot for what would become the company’s esports offering: the Championship Gaming Series (CGS). Graham leaped at the opportunity to ascend beyond his small online niche and grasp at televised legitimacy.

“I wouldn’t give up the experience ever,” says Graham. “It was years of live television. I don’t know where I could have ever gotten better experience. It was brutally hard.”

But CGS was vanishingly short-lived, beginning with a pilot in 2006 and then folding in 2008 after just two regular seasons. Graham estimates that DirecTV spent tens of millions of dollars on CGS, but it failed to find the kind of audience the company was looking for. This, too, served as a lesson for Graham, albeit one of a different, more sobering sort.

“I think working for DirecTV is the first time I started seeing, ‘Oh shit, this is how companies exploit gaming for advertising dollars, profit, etc.,’?” he says. “On one hand, I was OK with that because they were at least pushing the envelope. But the reality of it is, two years later when they shut everything down, you feel the hangover of it all. Gaming is also the first thing [companies] are willing to jettison.”

This left Graham at a loss. His grand hope for gaming had failed, and the traditional entertainment industry rejected him like white blood cells fending off an infectious disease.

“I started realizing there was no landscape for people like me [in LA],” he says. “I had conversations with William Morris [now WME] where I basically pitched them the idea of Twitch, and they were like, ‘This shit is never going to be big,’ and laughed me out of their offices. It was absolutely crushing.”

Faced with a lack of direction and mounting financial uncertainty, Graham and his wife moved back to Nebraska, and Graham resumed working in IT. It was, as he describes it, a “long fall” from his lofty days in live television, but it was a living. But when the content creation bug first bit him, it had burrowed deep. Despite the near-fatal damage prior years had done to his spirits, Graham resumed creating shows on a website of his own in 2009. To support his hobby, he made his PayPal info public. After just forty-eight hours, viewers had kicked in $5,000.

“At the time I was like, ‘Oh my fucking god. That’s huge. I don’t even know what to do with this,’?” says Graham, echoing a sentiment Twitch streamers—supported in large part by viewer donations and subscriptions—would later espouse.

Ensuing success led to interest from Ustream, an early livestreaming platform originally created in 2007 to aid deployed soldiers in communicating with their families, but which went on to host broadcasts from politicians, artists, and video game streamers. Ustream’s founders sought out popular personalities, including Graham and a whole host of esports luminaries, to provide content. Graham believed in Ustream’s tech and vision, so he went to bat for the company and persuaded others in the esports space to start streaming there. But the relationship soured quickly. While Graham and other gaming streamers brought a consistent audience to the site, Ustream’s eyes often wandered elsewhere.

“We would come to them and be like, ‘We’re gonna do this huge Quake thing, and we’re gonna advertise it and put it all over the place,” says Graham. “?‘Will you give us [placement on your] front page?’ And they’d be like, ‘Well, we would, but we’re gonna have Snoop Dogg do a fifteen-minute chat on his channel, and we want to keep that up [on the front page] the whole time….’ Yes, they’d get an explosion of users for Snoop Dogg, but that didn’t make anyone want to come back. Snoop Dogg wasn’t coming back and streaming.”

Snoop Dogg was a household name, but his presence on Ustream was functionally a stunt, a gimmick. To build a real audience, the site needed homegrown personalities who were consistent.

“[Ustream] were like, ‘Hey, we need to focus on these people who we spent so much money to bring on board,’ but there was never really any stickiness,” concurs Justin “Gunrun” Ignacio, an early esports streamer and commentator who came over to Ustream because of Graham. “The viewer base would just go off platform.”

Moreover, money was quickly becoming an issue.

“[Ustream] gave me, like, no money,” says Ignacio. “My two thousand viewers would make [me], like, a couple dollars.”

It was around this time that the founders of a competing streaming platform called Justin.tv, also created in 2007, began to realize they’d need more than cutting-edge livestreaming tech to succeed. The site began as a single channel on which founder Justin Kan livestreamed his life 24/7, an alluring gimmick that attracted attention from mainstream publications and TV shows—and which, for better or worse, led to some of the first-ever instances of swatting during a livestream. After a handful of months, however, Kan and his compatriots decided to broaden their horizons. With plans for expansion came a need for more content creators with different styles of broadcast. Back then, though, there was no template, no guide on what a good stream looked like. People just threw whatever they had at a camera to see what stuck. And then sometimes, they immediately thought better of it and gave up.

Initially, Graham did not like what he saw.

“[At first] Justin.tv was kind of a joke because most of what was on there was garbage,” he says. “It was, like, someone with bad video and audio going live for thirty seconds and then not being live. There were a few gems in there, but it was mostly garbage or pirated content.”

But Justin.tv kept studying the space and expanding. Before long, the company started trying to poach Ustream creators with the goal of building out a vertical dedicated to gaming. This initial ambition was followed by a period of second guessing, explains then-COO of Justin.tv Kevin Lin. Lin and Kan had watched esports titans like Graham get sent packing by traditional television behemoths and came away thinking it was just too early to bet the farm on video games.

“A lot of those TV attempts basically died,” says Lin. “The industry was reeling from [the fact that] DirecTV and those other TV companies had promised all this wealth and money to the esports industry, nascent as it was, and then it got canceled…. And so we felt it was too early to really deep dive.”

Yet, it was hard to ignore the obvious potential of games to become a truly mainstream force somewhere down the line. Meanwhile, something resembling today’s creator economy was beginning to take shape on other platforms—to gain a silhouette and spine, if not a recognizable face or body.

“YouTube was pretty new,” says Lin. “Creators were struggling to understand how advertising works and [were] feeling like, ‘Wait, I get paid 80 cents per one thousand views? That doesn’t seem like it’s actually going to sustain me.’ But it was starting to work. You were starting to see creators that were making maybe tens of thousands of dollars per year…. We started doing research and found that gaming was the third largest category on YouTube at the time.”

And then a watershed moment happened: Lin, Kan, and numerous other Justin.tv employees got really, really into StarCraft II, the 2010 sequel to one of the greatest strategy games of all time.

“[The original StarCraft] was pretty much the only content I would watch on Justin.tv,” says Lin. “Flash-forward a bit, the StarCraft II beta test comes out in 2010. A bunch of our employees got access to the beta and just started playing nonstop…. Just noticing our own behavior: We were going home and watching StarCraft videos on YouTube just to get better at the game. We were playing in the office every single day. And we had that sort of proverbial light bulb moment of ‘Wait a second, should we just lean into this?’?”

In calls and meetings with creators, Justin.tv talked a big game. Graham also found that, on a more basic level, the company’s executive team actually gave him the time of day, a far cry from the dearth of attention he felt his concerns were receiving at Ustream. But eventually, for Graham and others in his scene, what it came down to was money.

“Justin.tv started talking to people like me,” Graham says, “[esports commentator Sean ‘Day9’ Plott], [Ignacio], and other people in the gaming space, and they’re like, ‘What would get them to use Justin.tv over Ustream?’… And I was like, ‘Money. I hate to say it: It’s because none of us are making money. We have been kind of unsupported as gamers over time, and all we want to be able to do is support our hobbies. The company that first helps support gaming and second helps us put money in our pockets to leverage this passion we have for gaming—or esports or StarCraft or whatever—that’s what we want.’?”

In response, Justin.tv’s team said they were interested in spinning up a partner program for streamers that would offer better pay to Graham, Ignacio, and others. And so, Graham made up his mind: He’d paddle his operation—now a small network with multiple recurring shows featuring himself and other hosts—over from Ustream to Justin.tv, and he’d encourage other gaming streamers in his scene to follow suit. But almost immediately, he found himself in choppy waters.

GRAHAM, OF ALL people, knows how streamers feel when Twitch’s missteps leave them in a giant, foot-shaped crater. He knows this because he was perhaps the first person to whom it ever happened. It was 2011, and Twitch wasn’t even Twitch yet.

“We signed with Justin.tv,” says Graham. “It was a pretty big deal for us because at that point we were running basically a production company. We had all of these lower thirds and intro videos and links everywhere. And I said to Kevin Lin, ‘You know, you guys have been great, but please don’t fuck us.’?”

About a month and a half later, Justin.tv announced that its gaming efforts would be spun off into a new company and site called Twitch while the Justin.tv mothership would continue playing home to other types of broadcasts. A last-minute name change might not seem like the biggest deal, but if you’re a creator dealing with graphics, advertising, and a fundamental rebrand to suit your new online home, it’s the kind of thing that throws a jumbo-sized wrench into your plans.

“I remember that being a very heated telephone call with Kevin,” says Graham. “Basically I called him and said, ‘Kevin, you fucked us! We just spent all this money to redo all of our graphics and all this shit. Now we’re going to have to put another $5,000 into changing all this stuff and hiring the right people to do it. And this is exactly what we asked you not to do.’?”

On Graham’s end, the principle of the thing mattered more than the material challenges. He’d jumped ship from Ustream in part because he felt like Justin.tv leadership had the creator community’s best interests at heart. But then, suddenly, it seemed like he was dealing with the same set of problems as before stacked on top of each other, disguised in a slightly different trench coat.

Lin, to Graham’s surprise, took it well. He apologized and offered to pay the costs.

“We weren’t really creatives,” Lin explains of Justin.tv’s executive team at the time. “None of us actually streamed, and the thought of graphical assets didn’t come to mind. We just didn’t think about it. It was an honest mistake not informing people more ahead of time. We didn’t realize, ‘Oh crap, creators have all these assets?’ But of course they do. We see it every single day with our own eyes…. [Running Twitch] was a constant, iterative learning process, and that was a key learning.”

The incident, minor though it was in the grand scheme of things, is in Graham’s mind emblematic of what made Twitch work in its formative days.

“If one thing was true from the very beginning, it’s that Justin.tv’s original founders knew what they didn’t know,” says Graham. “They respected the opinions of others who had a perspective, a lens, experience in the space. That went a really, really long way.”

All the way, it turns out, to Lin offering Graham a job just a few months after that tense phone call.

“I think what Kevin saw was someone who understood livestreaming—who really cared about not fucking the creators,” says Graham. “This is the lifeblood. This is what needs to be the most important thing.”

Twitch’s community-focused hiring approach neither began nor ended with Graham. In its early days, Twitch also hired Ignacio and numerous other esports luminaries like Ben “FishStix” Goldhaber, who’d previously created a website that served as a centralized resource for esports streams. This quickly snowballed into a creator-first approach not just to hiring but to everything else.

“We had this very specific rubric of ‘How do we drive love, fame, and money for streamers?’?” says Goldhaber, Twitch founding team member and now-former director of content marketing. “When you put it that way, it sounds kind of cold and calculating. But what that meant was that we spent all of our time talking to the people who were creating gaming content and asking, ‘What do you want? What do you need to succeed?’ The first couple years of product development at Twitch was really just interviewing streamers, figuring out what they wanted, and building it for them.”

Twitch wasn’t just calling streamers on the phone, either. It quickly gained a reputation for its presence at countless video game conventions, tournaments, and events, many of which attracted thousands or tens of thousands of people. During this time, video game events were experiencing a rapid growth spurt in tandem with a widening embrace of video games and geek culture at large. PAX, which grew from 3,300 attendees in 2004 to 70,000 in 2011, did not carry the same sort of nerdy stigma as the (allegedly) B.O.-soaked Star Trek conventions of yore. Video games were becoming cool. Twitch, though intended to create connections that would bypass hills, seas, and other harsh realities of our mortal plane, grew at first through boots-on-the-ground word of mouth.

“I don’t think you could possibly overstate how important it was for us to be at these events and tournaments,” says Goldhaber. “At a lot of these, there was a Twitch partner lounge. At others, there were at least a few reps from Twitch there. I think that made an incredible difference, just to show the community that we cared, show the streamers that we cared. It also elevated streamers…. The Twitch parties became legendary, right? Having these exclusive parties made them feel very special. It built a sense of community around Twitch and was an aspirational goal for prospective streamers.”

The strategy was a conscious one on the part of Twitch leadership. It was intended, Lin explains, to work both ways, bringing the gaming community into the Twitch fold and ensuring that employees were hearing suggestions, questions, and concerns directly from creators.

“We encouraged people—everyone in the company, whether you were a brand-new junior employee all the way to a senior leader—to talk to creators,” says Lin. “We allowed people—and paid for them—to go to one event a year. Could be [Penny Arcade Expo], could be E3, could be an esports tournament. And while there, we’d set them up with creators to have a deep dive or casual conversation. This would help people understand how to talk to creators, how to extract good questions, good ideas, good problems and then turn them into product or feature ideas.”

The goal was to foster a sense of community with all Twitch creators, even if they weren’t on the verge of bursting through the stratosphere and achieving superstar status.

“People would come along and set up these relationships with creators,” says Ignacio, “even if they never became, like, huge, huge top streamers.”

Graham remembers the moment that solidified this approach: a get-together with partnered streamers at one of the first Boston-based “East” versions of the annual Penny Arcade Expo (PAX).

“We just rented out a sushi bar and invited all these esports people that were streaming on Twitch and Justin.tv,” says Graham. “What every attendee walked away with was exactly what I was searching for for so long: [that feeling of] ‘I love that Twitch cares about us, what we do, what we say, what we’d like them to build,’ etc. It became very clear to everyone at Twitch that we needed this to be our marketing.”

Graham quickly found himself on the front lines of this effort. His job, in effect, was to persuade esports players and teams to stream on Twitch, a prospect they initially found bewildering. Why, after all, would they want to spend months with their fingers on keyboards and heads in the clouds, devising galaxy-brained strategies, only to give everything away in broadcasts rivals could watch?

“I spent a year literally explaining what Twitch was,” says Graham. “I spent six months of that arguing with players and teams [about whether] streaming on Twitch would be bad for them, because their opponents would tune in and watch them and understand their strategies. I was up against so much resistance. But I was just like, ‘Let me show you how successful [Plott] is being, and all he’s doing is talking about games.’?”

Little by little, Graham eroded the esports community’s suspicion. It helped that he was, in many ways, the ideal person for the job. He was a bona fide livestreaming pioneer, and he knew in his heart—from countless hours of experience—that Twitch was onto something.

“I’ve never enjoyed sales, but to me this was the best version of sales, which is evangelism,” he says. “I truly believed livestreaming was the future.”

The idea that Twitch cared—that not just the company but the entire streaming community was a family—became central to Twitch’s identity. Even Twitch’s slogan, “Bleed Purple,” was initially a community creation. Shannon Plante, an eventual Twitch employee who was, at the time, just a fan and moderator for streamers and the Twitch subreddit, came up with it in 2014 after attending a video game convention in the UK. On Twitter, she decided to thank Twitch for, essentially, giving a shit. She hashtagged her tweet #BleedPurple, riffing on the slogan of a local hockey team.

“I was thanking [Twitch staff members] for being involved not just for streamers or large video game studios and trying to build those relationships, but for caring about indie developers,” says Plante. “I was just basically being a fan of how everything felt so tight-knit back at that time. I really wanted to inject myself in that because it gave me energy. It was just such a wonderful buzz.”

Before long, Twitch staff let her know that employees had begun using the slogan internally, as well as at parties: “There was a party in 2014 where they had started to use it printed out on regular paper, pasted on the walls, because that was the budget they had for parties back then,” says Plante.

Twitch hired Plante for a community-focused role in 2015, and it used the slogan to announce that she was joining the company. “That was really special for me,” she says. “And then it became something I started to see on every banner and billboard that they made. They made it their official thing after that.”

“Bleed Purple” became Twitch’s official slogan, Plante believes, because the underlying sentiment resonated not just with Twitch employees, but everybody who used the platform.

“It was something that took off because of a strong sense of community,” she says. “It wasn’t just me or Twitch staff who said it. It became something that streamers, smaller and larger, began to use constantly. Everyone felt that way. It was a very contagious sense of identity and inclusion.”

In some cases, Twitch wound up stopping major event streams from falling apart at the seams. Ignacio proved instrumental in this regard. Before Twitch hired him, he was a community college student living with his parents, with short black hair and an unassuming, blue jeans-clad style, which disguised a person with boundless expertise in a field that had just barely begun to exist. Streaming was just a hobby, but Ignacio had a preternatural knack for squeezing blood from stones, hacking together a livestreaming setup involving obscure Russian software to ensure his video game streams looked better than anybody else’s without his modest computer going up in flames. Before long, he gained a reputation for using his whiz-kid know-how to save esports events whose livestreams were collapsing under the weight of high-end graphics and rickety internet connections. That piqued the interest of folks like Lin, which led to a job at Justin.tv. Before long, Ignacio, whose irrepressible enthusiasm made him talk so fast that it seemed his mouth couldn’t keep up with his brain, was helping Twitch solve tech problems at its San Francisco home base and in the field.

For Ignacio, working at Twitch was a dream come true, but as Twitch came to both embed itself in the gaming community and rely on esports for viewership, he got a front-row seat to the most precarious elements of the platform’s business.

“I feel like I have all this personal PTSD from handling these tumultuous moments in the field of ‘Oh wow, this is something that’s never happened in existence. I need to solve it immediately for this event to not be a crippling failure,’?” he says.

Some events, like the debut of livestreaming on Xbox 360 from Germany, were pure pressure. Others, like the 2012 Battle.net World Championship in China, were surreal.

“I just remember being in the server room trying to solve this issue, and then [former NBA basketball player] Yao Ming comes in, and everyone’s like, ‘Yao Ming really wants to see production right now!’?” says Ignacio. “I was like, ‘Oh cool, I’m just trying to do my work’ while being a crazy twenty-one-year-old wearing a bunch of StarCraft swag and trying to fix crazy things.”

Even though Twitch had made strides in low-latency video tech, livestreaming remained—as in Graham’s mid-2000s Global Gaming League days—expensive. At first, Twitch had to pay another company called Akamai to handle especially large viewership spikes and avoid overloading its own servers.

“Whenever there was a major event, we’d call Akamai and tell them, ‘Hey, we’re thinking we’re gonna get, like, thirty thousand viewers for this, so let’s pay you some money and get this set up.’ But if it broke, like, eighty thousand [viewers], we’d have to pay a much higher cost for it,” says Ignacio. “We’d have to forecast esports events so we wouldn’t die from the cost of [overflow]. I had one of the keys for that. So did [CEO] Emmett Shear. I liked using the nuclear missile analogy when it came to that, like ‘OK well, there goes however much money that was.’ We had a giant red thermometer of ‘Here’s how much money was being lost, and we need to make this money up somehow.’?”

Lin, perhaps because he had no other option, came to view even these moments through an optimistic lens.

“We had so many failure points, whether it was at ingest or on our systems,” he says. “It was constant, constant breaking. But it was also a great signal of how much people wanted us to exist. They would suffer through all these pain points that were out of anyone’s control in many cases.”

Graham always maintained hope that Twitch would eventually stop bleeding green.

“[Twitch’s founders] were thinking about ‘What does our infrastructure need to look like in three to five years in order for us to be a sustainable business?’?” he says. “They were spending money like a startup, but in my opinion they were hyper-focused on the two things that were going to determine their success: creators and infrastructure.”

IN THE BEGINNING, Twitch didn’t have many ironclad rules, at least compared to today’s time-honed terms of service that outline who stays and who goes when it comes to granular forms of racism, sexism, harassment, and numerous other varieties of inappropriate conduct. Of the few rules in place back in 2011, one towered above all the rest: Video games only. No non-gaming streams allowed. Those, upon detection, would be banished to Justin.tv, forced to roam the wastes with 24/7 livestreams, truck driving streams, farm streams, and whatever else people could think of to broadcast that wasn’t gaming.

“Anytime we found anyone, any content, even if it was a talking head show,” says Lin, “we would message them and say, ‘Hey, you can’t do this here. You have to go to Justin.tv. Twitch is really about video games.’?”

As far as standard-setting edicts go, it was a pretty simple one—at least, on the surface.

“We felt like it was pretty obvious: The majority of the screen real estate should be a video game,” says Lin. “You can place yourself in the corner or whatever. That was totally fine for a while, but of course our community loves to push the envelope. So eventually, people would just put a tiny little corner of video game, to test the line.”

Once early streamers started getting in trouble for that, they flipped the dynamic on its head, giving games center stage but broadcasting even more audacious main events down in the corner of the screen.

“You have the camera where all the [actual] content would happen in the lower third,” says Ignacio. “So they’re ‘playing’ World of Warcraft, but… they’re having a crazy party or doing a dating game or something over Skype.”

This taught Lin, Graham, Goldhaber, Ignacio, and numerous others a lesson that seems obvious in hindsight, but that flew in the face of conventional wisdom at the time: Livestreaming is powered by personalities. Everything else—even God-tier video game skill—plays second fiddle. It was the beginning of a seismic shift that would ultimately send shock waves to every corner of the internet. Now there’s a personality for everything, whether it’s politics or detecting scam artists. Younger people, in fact, primarily get their news from creators rather than traditional sources. Twitch learned early the power of a friendly face.

“While a lot of the impression at the time was that people wanted to watch who are skilled [gamers], it really is about personality,” says Lin. “Esports was just starting to be understood by the industry and consumers, and so we had this concept that it was all about skilled play, when the reality is, people were watching for the streamer, the personality…. You don’t have to be that good. It helps if you can do both.”

But even though Justin.tv would seem a more natural fit for streams that emphasized personality above all else, Twitch’s focus on gaming provided glue for an otherwise formless medium. Games—often lengthy, or in the case of competitive hits like StarCraft and League of Legends, potentially endless—gave streamers something to react to, to talk about, every day. Consistency allowed viewers to form habits and build schedules around their favorite creators. They could expect streamers to show up with a daily dose of comfort food, as opposed to the dissonant blend of gimmicks and stunts that had come to characterize other livestreaming services. Technology provided the dance beat, but gaming taught everybody the steps.

“No one knew what their script was going to be if they were just going to sit in front of the camera and talk all day,” says Graham. “You did have guys like [Justin.tv streamer] Marijuana Man where literally he smoked marijuana and then would talk and listen to music, and that’s why some people showed up. But it really came down to: Were you building community around your content? And the gamers were building a community.”

Community led to a consistent audience, and it did not take long for prospective streamers to realize that Twitch was where the viewers were. Twitch cemented this community dynamic by working in collaboration with Plott—one of the most popular streamers on the platform at the time—to create a button that allowed viewers to subscribe to their favorite streamers for $4.99 per month, granting perks like exclusive chat emotes. Subscriptions became badges of honor to viewers, a sign of commitment, bolstering group cohesion and boosting streamers’ (and Twitch’s) revenue. Twitch’s user base, meanwhile, grew and grew, from 3.2 million in 2011, to 20 million in 2012, to 55 million in 2014. It did not take long for Justin.tv’s founders to realize which arm of their operation was buttering their bread: In 2014, Justin.tv and Twitch’s parent company rebranded to Twitch Interactive, and it shut down the Justin.tv website to focus on Twitch.

Still, it was, to some extent, that first Garden of Eden-like moment that came to define Twitch’s culture: A single rule decreed from on high, almost immediately defied. Twitch streamers would spend the following years trying to skirt the outer edges of rules and norms, while Twitch would try to define what rules and norms even were. Some of this outside-the-box thinking led to pioneering innovations: While early stars like Plott were by-products of the esports world, others like Ben “CohhCarnage” Cassell and Saqib “Lirik” Zahid ascended to new levels of stardom by playing a variety of different games—rather than esports standards like StarCraft and League—and anchoring broadcasts with affable personalities. Those streamers, more so than Twitch, also set the bar for what was an advisable amount of time to spend streaming. They set it high.

“[Zahid], back in the day, was one of the first streamers who was ever just like, ‘I’m streaming eight to ten hours a day,’?” says Goldhaber. “That’s part of why he blew up. That became the meta of how you grow an audience on Twitch.”

Upping time spent streaming conferred a multitude of benefits. Twitch spent much of its lifetime structured as a series of categories organized by number of viewers, and the more hours a streamer spent in a category, the more chances they had to scrape viewers from the underbellies of bigger names who are logging off for the day. Moreover, there was the spectacle of it all: Lengthy streams represented a sort of endurance feat, whether they came in the form of a single marathon broadcast that lasted for twenty-four-plus hours or multiple broadcasts that lasted ten-plus hours, perhaps every day, indefinitely. Viewers came to revere that level of dedication, as they did with YouTubers who’d post multiple videos per week, or even per day. Eventually, they started to rely on it. At that point, for streamers, there was no turning back.

Nowadays, burnout is a major topic of discussion among top streamers, and it’s natural to wonder if early Twitch employees thought to pump the brakes and encourage, well, breaks. At the time, however, most weren’t thinking of what the Twitch landscape might look like in a decade.

“At the time it was so novel, the idea of playing video games for money or for a living,” says Goldhaber. “It’s easy to take for granted, but the concept of [a] creator—someone who makes their entire full-time living on the internet—barely existed at the time. So I can’t recall a lot of conversations where [burnout] was a big concern. I think it was mostly elation that people were making a full-time career doing what they loved, and we were facilitating that.”

Twitch’s own work culture was also not the healthiest, which might have contributed to that particular blind spot. As with many startups, passion could turn into toxic work habits.

“I had no concept of work-life balance,” says Ignacio of his early days at Twitch. “Grind culture was the meme of the era.”

But while Twitch might not have contemplated the long-term impacts of streamer labor at the time, those norms, once solidified, certainly informed the company’s future thinking; it’s not uncommon for modern-day Twitch contracts to demand upward of 150 or 200 hours per month of live time from big-name creators.

Other creator-born innovations were more straightforwardly beneficial to both the creators and the company. In 2014, a channel called Twitch Plays Pokemon moved beyond the idea of a streamer entirely, allowing thousands of viewers to simultaneously input commands into chat to collectively—and chaotically—pilot a player character through the 1998 Game Boy hit Pokemon Red. The concept went viral, with millions ultimately tuning in and creating memes based on the broadcast’s most absurd triumphs and failures. Twitch staff came to view it as a watershed moment, one that—thanks to the resulting news coverage—showed mainstream media that Twitch had bottled millions of Pikachus’ worth of lightning. Its community was doing the unthinkable, in large part because nobody had ever thought of anything like this before. Twitch users were dreaming up entirely new forms of entertainment.

“I think [Twitch Plays Pokemon] completely changed the way people looked at Twitch,” says Graham. “They finally went, ‘Oh shit this whole interactive format and medium is the future. It will define this industry for a long time to come.’?”

Other early streamers, like Steven “Destiny” Bonnell, innovated by pairing gaming prowess with offensive language and shock humor, echoing back to edgy pockets of early YouTube and forward to the perpetual drama engines that power large portions of Twitch today. Bonnell has since been indefinitely banned from Twitch for “hateful conduct,” according to Twitch, but other streamers of the era prone to harboring questionable beliefs or casually employing inadvisable language—Chance “Sodapoppin” Morris, Jaryd “Summit1g” Lazar, and Zack “Asmongold” (who has not disclosed his last name) among them—are now about as popular as they’ve ever been.

Following in this grand tradition, every era of Twitch has had its crop of edgelords, some of whom took things too far, others of whom are still around today. In 2016, Tyler “Tyler1” Steinkamp began a reign of bellowing, profanity-laden toxicity that got him banned from his game of choice, League of Legends, for two years. In 2017, Paul “Ice Poseidon” Denino became famous for repeatedly courting real-life controversy until somebody swatted him on an airplane, at which point Twitch banned him. In 2023, Adin Ross—who’d already committed countless other infractions—finally got the boot after moving to another platform entirely and allowing antisemitism in his chat. Félix “xQc” Lengyel, once Twitch’s biggest North American star and still a top streamer, has been suspended for everything from porn to airing copyrighted content to cheating in a Twitch-hosted tournament. The list goes on, as it does on other platforms, each of which have adopted their own content moderation policies over the years—some, like YouTube, gradually clamping down, while others, like Twitter (now X), have come to allow all manner of toxicity, even as advertisers have fled. Despite Twitch’s wholesome modern-day marketing, it was founded on edgelords.

“It was the Wild West,” says Graham of livestreaming’s earliest days. “People think there’s edgelords now—there are—but everyone was an edgelord back in the beginning.”

Graham, to some degree, counts himself as part of that lineage. As an example, he points to a 2007 pre-Twitch stream in which he hosted an event called “Strip Halo,” which pitted adult actresses Mia and Ava Rose against viewers of Graham’s show in games of the popular sci-fi first-person shooter, and if viewers won, the Rose sisters had to remove articles of clothing. This was near the peak of the era in which video games were considered a male pastime, at least as far as marketers and gatekeeping subsets of the fandom were concerned. Graham feels like he contributed to a status quo of objectifying women, one that informed the Twitch community’s future hostility toward female streamers.

“Some of us were like, ‘We can all be Howard Sterns on the internet,’?” says Graham. “I’m not overly proud of all that.”

Early Twitch was characterized by a male-dominated culture among streamers and their communities, as well as in the halls of Twitch’s offices.

“We did attract a certain kind of bro-y gamer guy,” says Goldhaber. “The majority of the partnerships team was a bro-y gamer guy [type]. A lot of bro-y gamer guys.”

Partnerships, the team on which Graham spent his first several years at Twitch, were instrumental in Twitch’s early success. At the time, Twitch hand-selected which streamers achieved partner status, granting them additional moneymaking features on the site, a direct line of communication into Twitch, and the ability to attend exclusive parties. This gave Twitch’s partnership team outsized influence and led to accusations of favoritism, especially where moderation—what happened when streamers broke the rules—was concerned.

“There was such a lack of structure,” says Graham. “On one hand, that was great for creators, because as partnerships [team members], we want to look out for other partners…. But [partners] really only got banned for the most egregious shit.”

“In the early days, everyone did everything,” says Goldhaber. “We all kind of tapped in for moderation. I don’t think we had a lot of consistency.”

While Twitch standardized its moderation practices more over the years and eventually put together a dedicated trust and safety team, Graham believes that too few people doing too much work led to a reactionary moderation approach rather than a proactive one.

“A lot of [moderation] was ‘Just let it keep going until we really need to change it.’ It was forced reaction. ‘This happened, now we need to do something about it.’… Did the [rules] evolve gracefully over time? No. Did support evolve gracefully over time? No.”

These growing pains left scars. Racism and sexism—never explicitly allowed on Twitch, but addressed slowly by staff (or not at all) in the platform’s early days—were able to snarl their roots around the bases of many communities. Long-standing issues sprouted from this tainted soil: regular harassment of female streamers both on Twitch and off, by streamers and viewers, as well as the practice of spamming chat with an emote of a Black streamer named Mychal “Trihex” Jefferson anytime a Black person appeared on-screen, sometimes as a dog whistle for slurs. That’s just the tip of the iceberg of old-school Twitch toxicity, and though the company has worked hard to curtail those issues with significantly better enforcement of its own rules and built-in moderation tools streamers can deploy at will, the platform will never be able to fully erase the stain on its culture.

“If you’ve seen [Netflix mystery movie] Glass Onion, one of the main characters is a Twitch streamer, and he’s an awful fucking human being,” says Cristina Amaya, a former customer service rep at Twitch who went on to become a director at esports event company DreamHack. “That’s the joke, right? All these awful people are representative. That’s what people think of our platform.”

This insufficient approach to self-moderation and self-regulation also revealed the downsides of Twitch’s cross-pollination within its own community. Twitch employees regularly mixed and mingled with streamers, who they regarded as peers despite a power imbalance in which employees could grant streamers status both socially and—when it came to site features—at the press of a button.

“Not to downplay it or say it’s acceptable, but first of all, everyone’s in their early to mid-twenties drinking a lot at conventions,” says Goldhaber. “Twitch employees and streamers were all partying together, and they were all young…. So yeah, [streamers and employees getting together] definitely happened. That kind of thing was not all that surprising.”

“There was a bit of a bro-y culture,” Goldhaber reiterates. “There was a lot of drinking culture.”

In a few cases, these dynamics led to unsavory outcomes, though they would not come to light until much later. In 2020, a streamer who went by the handle Vio accused Hassan Bokhari, then accounts director of strategic partnerships at Twitch, of abuse of power and sexual assault that took place both before and after the two began dating in 2015. She said Bokhari would use his status as a Twitch employee—one with special access to Twitch partners—to lavish her with perks like a username change, a special holiday package meant for Twitch partners (she was not one at the time), and eventually partnership. When the two met up in person at a 2015 video game convention, Vio said Bokhari immediately began pressuring her with unwanted sexual advances. This turned into a pattern across multiple in-person encounters, until she, in her own words, “gave in.”

“I was preyed on, manipulated, gaslighted, violated, and sexually assaulted,” Vio wrote in a 2020 post about her experiences with Bokhari.

According to a report by Kotaku, this led Twitch to hire an external investigator and ultimately dismiss Bokhari from his role.

A wave of similar stories came out around the same time as part of a #MeToo movement in the Twitch and content creation space. These focused not just on Twitch employees but other empowered individuals: managers and agents, some of whom had already begun to draw flak for negotiating predatory contracts and secretly skimming money off the top of deals between creators and brands. Omeed Dariani, then CEO of Online Performers Group, one of the biggest management firms in livestreaming, faced accusations not unlike those leveled at Bokhari, with a streamer saying he engaged in predatory behavior at a 2014 video game convention. After word of his alleged actions got out in 2020, numerous streamers dropped the firm, and it ultimately shut down.

Even if justice was eventually served in some cases, the impact of the male-dominated, drinking-heavy culture that sprung up around Twitch shaped the environment. In 2020 alone, over one hundred people, mostly women, came forward to accuse industry figures and streamers of abuse and exploitation. At the time, Twitch issued a statement saying it had banned several streamers and would “continue to assess accusations against people affiliated with Twitch and explore ways Twitch can collaborate with other industry leaders on this important issue.”

Both prior to and following this reckoning, Twitch employees who were not white men found themselves in protracted uphill battles within the company.

“[Early] Twitch was gaming nerdy white dudes,” says Amaya, who worked at Twitch from 2017 to 2018. “They never brought in people that did makeup or [other non-gaming types of content creation]. And even when those people were at Twitch, when I was there, they never really got a voice in the room.”

“It was always a fight,” says a former Twitch employee, who chose to remain anonymous out of concern for future job prospects, who worked on diversity and charity programs at Twitch. “It was never, ‘Hey, let’s just do this.’… Everybody was fighting over the limited opportunities to get in front and be seen as somebody leading a project because it was the only way to get promoted. That means people who excel at things that are important to the company that are administrative—which tend to more often be women or marginalized identities—don’t get promoted as often.”

One example of this, according to the former Twitch employee, was Twitch leadership’s response to the idea of a Black History Month celebration on the platform—now a successful annual tradition that spotlights Black creators every February.

“[Leadership] fought me like hell over the first Black History Month,” the former Twitch employee says. “They didn’t want to tweet it. They said it wouldn’t do well. A guy on the team at the time was like, ‘I don’t think this tweet will even get five hundred retweets.’ That Black History Month tweet from [Twitch’s] first Black History Month was their best-performing tweet for almost a whole year.”

A 2020 report from GamesIndustry.biz painted an even more severe picture of Twitch’s diversity-related lapses over the years. In the piece, former Twitch employees from all prior eras of the company’s history spoke of multiple instances of sexual abuse, incidents where female streamers were scrutinized and punished for non-rule-breaking attire, and dismissals of concerns from employees of marginalized backgrounds, such as racist language both in the workplace and in Twitch chat. At the time, Twitch issued a statement saying that it “takes allegations of this nature extremely seriously,” but also, “many of these allegations are years old, and we’ve taken numerous steps to better protect and support our employees and community, and will continue to invest time and resources in this area.”

Lin, who was part of Twitch’s C-suite from its inception until 2020, acknowledges that Twitch leaders made mistakes when they were young and “flying by the seat of our pants,” but they also tried to course correct later on.

“Consciously we tried to change this,” Lin says. “A topic we spoke with a bunch of our friends in the industry about was ‘How do we increase diversity? How do we proactively do this?’ The hard part there is, we were running a business, and we had investors. When we talked to investors, it was constantly ‘Well, why aren’t you hiring experienced people?’ And the reality of what came before us was, most of the experienced people were male and white.”

Studies even back in 2011, however, showed that the lack of women and minorities in tech wasn’t merely a pipeline problem. People from non-white, non-male backgrounds reported disproportionately worse experiences in the workplace, which led to turnover. In other words, we’ve known for quite some time that companies play an active role in growing (or thinning) that part of the workforce. Still, Lin insists he and other Twitch leaders did their best, improving hiring practices, creating a streamer ambassador program with a focus on diversity, and implementing weekly meetings where anybody in the company could ask questions.

“We actually cared about this stuff,” says Lin. “No one will believe it, but we did. Could we have done better? One hundred percent. But a lot of this feedback [from the GI.biz article] is, like, any company is like this. Oh, an exec doesn’t listen to you? Yeah, guess what? There’s two thousand people in the company. Sorry, it’s impossible to talk to two thousand people every single month. It’s not possible. Maybe technology can solve this one day.”

The former Twitch employee who chose to remain anonymous views that potential solution as part of the problem: Twitch leaders like Lin and former CEO Emmett Shear, she says, had a tendency to view Twitch through a lens of technological neutrality—to believe that they’d somehow created a platform that was unaffected by the prejudices of those who made it.

“Nobody wanted to admit that users are racist,” says the former Twitch employee. “Every time I would talk about how the top users on Twitch’s platform for a long time were mostly white and men, people would just say, ‘That’s just what people feel is the best content on our platform.’ I’m like, ‘So you think Twitch is completely neutral on this?’ And they’re like, ‘Absolutely.’?”

Graham believes he grew as a person over the course of his time at Twitch. He came to recognize the double standard that impacted colleagues like the aforementioned former Twitch employee who opted remain anonymous, whom he ended up mentoring. “We could literally say the same thing in a room,” says Graham, “and if I spoke it people would say, ‘Oh that’s [Graham] for you,’ but if [she] spoke the same thing it would be ‘That’s insubordinate and [she] is hard to work with.’… I think she was constantly labeled incorrectly because she cared about community and was passionate.”

However, despite early ripples that would turn into future tidal waves, Graham retains a forgiving view of the first few years, before Twitch—and its problems—grew several times in size.

“I think where Twitch was falling short, or where we had problems, was just having a lot of work and not enough people to do it,” Graham says of Twitch’s startup days. “The growing pains were the hardest part. It wasn’t that the ideas were bad. It wasn’t that the vision and the mission weren’t right. It was always that we needed more people to do more things so we could accomplish more.”

MORE PEOPLE WOULD come in 2014 and the years following. Many more.

When Twitch got big enough, the major tech giants came sniffing around. Amazon emerged as the early favorite, its main competition in this clash of tech titans being Google, already the owner of YouTube. Upon hearing rumors that an acquisition was nigh, Graham feared a future in which Google would strip Twitch for its (by this point global) livestreaming infrastructure and roll the rest into YouTube. Effectively, this would have meant no more Twitch—and by extension, no more Twitch community.

“I didn’t want Twitch to get sucked into YouTube,” he says.

Twitch leadership, it turns out, felt the same way.

“Our biggest hesitation around Google was, we were certain to just become another cog in the YouTube machine,” says Lin. “By comparative scales to YouTube, we were still quite small. Would we ever really get the attention and resources we needed? Or would we just get absorbed entirely… when we still had all these big ideas, products, initiatives?”

Google also pitched Twitch on its broader strategy around games, but Lin “wouldn’t say that was very convincing” at the time. Amazon, on the other hand, seemed more sold on Twitch’s vision of the future.

In 2014, Amazon acquired Twitch, spurring long-term change that would shake the company to its core.

“We had this dream of ten years down the road, what Twitch’s place in the games industry could be, and that eventually became a project internally called ‘developer success,’ which was about building these audience-interactive games,” says Lin. “The easy example we’d use was Hunger Games. Imagine you’re playing Hunger Games—like games and audiences buy stuff while the game is live—or, at minimum, you could easily pull your viewers [from a stream] into a game.”

Lin and others at Twitch wanted to keep exploring those ideas, but they’d hit a wall: cost. Twitch did not publicly talk numbers on this front until 2022, by which point it was a very different company than it had been in 2014, but the outline of the problem has been present since day one: As Twitch gains more viewers, it grows more expensive to run.

“Bandwidth is expensive,” says former Twitch software engineer Theo Browne. “The more data you have and the further the distance it has to travel, the more expensive it gets.”

“We were profitable up until 2012, but as Twitch started to really take off, the cost just outpaced our ability to continue to improve monetization,” says Lin. “Twitch is an expensive site to run, and [then-CEO of Amazon Web Services] Andy Jassy gave us a lot of freedom to continue to build. As long as we kept growing, he was cool with it.”

With the Amazon acquisition came some pretty immediate benefits, chief among them integration with the mega-corporation’s popular Prime program. Amazon decided to use Twitch to sweeten the Prime pot: If a user signed up for Prime, they’d receive a free monthly subscription to a Twitch streamer of their choosing—meaning Amazon was effectively paying for millions of Twitch subscriptions per month. This put cash in streamers’ pockets and allowed Twitch to keep building.

“We sunk a lot more money into Twitch’s continued growth and development,” says Lin. “Amazon had a lot of patience there, I think, because things like [Twitch] Prime were working.”

This also meant resources for rapid expansion, the kind that would have been inconceivable before Twitch became a beneficiary of Amazon’s mighty money spigot.

“We ended up ballooning from when we got bought,” says Lin. “[At the time] we were at two hundred and forty people across the globe. Just one year after we grew to a little over four hundred, because we had all these resources and were encouraged to do so. Our [Amazon] integration team was like, ‘Yeah, you should go faster. Just grow!’?”

But that’s also when the troubles began.

“It became somewhat unnatural,” says Lin. “We’d been historically very careful about hiring. We talked about it all the time. Myself, Emmett, [and others] would ask, ‘Do we really want to grow this much? They’re telling us to, but it feels unnatural. What’s right? Do we stay at this size?’?”

After a little over a year, Graham and others on the more community-oriented side of the company began to feel the impacts of Twitch’s accelerated growth.

“The biggest downhill we saw after the Amazon acquisition is when we started hiring people that did not believe in the things that made Twitch successful in the first place,” says Graham. “I think Twitch hired a lot of people who had good Silicon Valley résumés, that could make a PR splash—but at the end of the day [who] were total fucking boomers who could not grasp what Twitch was, how it worked.”

During this time, Graham transitioned between different director-level roles that focused on everything from features to Twitch-produced live shows and events. As he almost always had, he continued to stream as well, something that was becoming rarer and rarer among Twitch employees. With that in mind, he took it upon himself to advocate for what had come before—or at least a more robust understanding of it. But a new layer of management made that difficult.

“The new guard created this sense of—and it was supported by individuals like Emmett, I’m not sure why, maybe he was too passive—everything that needed to be done needed to be new,” says Graham. “One of the reasons I think people felt I was a disruption within Twitch is because I would force the function of history. I’d be like, ‘You know, we used to do this,’ or ‘We tried this, and it failed. Why are we trying it again?’ [They’d say], ‘Oh, well it worked over at Twitter or Pandora.’ But we weren’t Twitter or Pandora. We’re not Facebook. We are Twitch.”

Ignacio also moved around within Twitch during this time period, switching from working under Lin to working under Graham. He realized he enjoyed being involved in more creatively oriented projects within the company, but he, too, could feel the company’s foundations shifting beneath his feet.

“People were brought on who’d never used the platform,” says Ignacio. “They focused on social media, too. They launched a system similar to Twitter likes and a friend system because they were trying to compete with Facebook for a while…. We all agreed it was a weird move, and even if you liked it, we didn’t do it right. And we gave up too fast.”

From 2015 onward, Twitch slowly expanded beyond the realm of video games, catering to a wider audience than in its pre-Amazon days and growing commensurately. Between 2014 and 2015 alone, Twitch went from 16 billion minutes watched per month to 20 billion, with a peak viewer count of 1 million concurrent viewers in 2014 and over 2 million in 2015. Some of Twitch’s non-gaming efforts bore fruit that sustains the platform to this day: “Just Chatting” streams—in which creators interact with their audiences sans a game—are some of the most popular on the platform, and many creators now focus on artistic pursuits, outdoor activities, or trawling the internet and gawking at other people’s content instead of ranking ever upward in games. But these broadened horizons also turned Twitch into a graveyard of half-baked projects: a karaoke game for streamers called Twitch Sings that, despite proving costly, never caught on and was eventually shut down; a merchandise program that split revenue with streamers; a store that also gave streamers a small cut of resulting sales; and countless site features and Twitch-produced shows. Meanwhile, teams—numerous and diffuse as a result of Twitch’s continued expansion well past the one-thousand-employee mark—competed as often as they cooperated, resulting in a plethora of projects that never cleared the runway. Twitch wanted to expand, but it failed to pick a direction and stick to it. And all the while, it struggled to synergize with Amazon, whose game development efforts hit snag after snag.

“Amazon Game Studios finally hit it big with [2022 role-playing game] Lost Ark, but the first, like, seven years of Amazon Game Studios was a complete disaster,” says Goldhaber. “[For Amazon] the buying of Twitch was supposed to be like, ‘We have a game studio, and then we market it on Twitch.’?”

Twitch continued to grow in popularity, though arguably more because of what was happening around it than inside it. In 2018, Fortnite—a battle royale shooter in which each match sees a hundred players compete to be the last one standing—took the platform by storm, with the game’s fusion of candy-colored cuteness and high-stakes blood sportsmanship catapulting creators like Tyler “Ninja” Blevins to previously unseen heights and hooking mainstream stars like Drake in the process. Before long, streamers were appearing in television advertisements and at the Super Bowl. It was as though the world suddenly awoke to a fact Graham and many others had known for quite some time: Everybody plays video games. Just as the Fortnite craze began to simmer down, 2020 brought another colossal influx of new Twitch users as the result of a pandemic-stricken populace starved for human connection. Every year brought with it a new all-time viewership high. Twitch had officially arrived.

But sheer numbers—by this point, Twitch had millions of streamers and over 100 million monthly active users—presented challenges that even Twitch’s most community-oriented employees couldn’t surmount. The company was forced to automate the partner program, resulting in a system that served many (over fifty thousand partners, compared to the hundreds of Twitch’s early days) but satisfied few. It became impossible to provide the lion’s share of big streamers with personalized attention, and many came to feel that Twitch failed to beef up its communication chops elsewhere to compen-sate. Streamers grew increasingly irate over suspensions they felt were under-explained, rules that seemed inconsistently applied, Twitch’s unwillingness to pay licensing fees to allow streamers to broadcast music and other copyrighted material (despite other platforms like Facebook and TikTok having done so to varying degrees), and slow responses to issues like a 2021 epidemic of “hate raids,” in which trolls overwhelmed streamers’ chats with fake accounts that spammed hateful messages. All the while, there was a looming sense that instead of providing streamers with the tools they wanted—features that would aid with pernicious issues like small streamers’ inability to be discovered by viewers—Twitch began to prioritize turning a profit. This, to a degree, was an accurate read: After years of expansion, Amazon had come to collect. It wanted Twitch to prove it could be a profitable enterprise.

Twitch didn’t feel like Twitch anymore, and Graham began to feel hopeless.

“The number of dumb ideas I had to shoot down in my time at Twitch, I wish I had a nickel,” he sighs, outlining an instance in which a decision from on high functionally ruined Twitch’s official channel on its own website by encouraging viewers to spam chat indefinitely. “These particular employees couldn’t truly show empathy for a creator. I was trying to explain to them why it would ruin a chat if I can’t have a conversation with the people in it. If Twitch can’t do it [on its own channel], how can it be an example for anyone else?”

When it came to decisions large and small, Graham was dismayed with his new compatriots’ lack of experience in the trenches of streaming, their inability to see how all the pieces of the site fit together.

“How does [a decision] affect a moderator?” he says. “How does it affect a creator? How does it affect an audience member? How does it affect the trifecta of how these things work together? The more and more Twitch goes away from that, the worse things seem to get. We started with Amazon purchasing Twitch. This is the long-term consequence.”

Longtime streamers like Ben “CohhCarnage” Cassell, who came up alongside Graham in Twitch’s early days, could feel the change. And hear it, because Graham has never been one to stay silent.

“Some of what [Graham] said on his stream were the kind of things where when you hear them, you know exactly what he’s talking about,” says Cassell. “When he can say something like ‘I stood up in a meeting full of people and asked if anyone was thinking of the people creating on this platform,’ when you hear something like that and get a vision in your head of literally a room full of people, and one person is standing up and asking if anyone is thinking of this? That’s a good example of where things are now [at Twitch]…. I think we all kind of hope that feeling of community, of shared passion, lasts as long as it can. For Twitch, we are past that time. We are not in that realm anymore.”

Other former Twitch employees with whom Graham worked have slightly different takes. While the Twitch employee who chose to remain anonymous also ultimately found herself dismayed at the extent to which Twitch took its community for granted (and discouraged her from advocating for it), she feels like Amazon helped professionalize a company where not everybody swapped out their halcyon-day beer goggles for rose-tinted glasses.

“A lot of early employees, Twitch made them, and it caused a lot of people to be ego-driven,” says the former Twitch employee. “[They would] speak out of turn, but also people who were friends with creators would advocate for them to get things, and it made it really messy. I think that a lot of what [Graham], for example, looked at from the olden days [as better]—and I’ve talked about this with creators who say, ‘Oh, Twitch was better,’ too—is people saying they liked it better when they had more influence and control.”

Goldhaber concurs that, even if he doesn’t love the exact form it took, Twitch needed to move past its rowdy early days.

“You have to grow up a little bit at some point,” he says. “As the company scaled, we’re talking about going from twenty employees when I joined to one thousand employees five years later. You need some growth in terms of management, in terms of leadership. And I do think some was needed.”

But as Twitch grew larger and larger, continental drift set in where employees’ objectives were concerned. Many wanted to see Twitch succeed, but according to Graham and Lin, others were mostly there to climb the corporate ladder.

“By nature of these big companies, these large groups of people, people aren’t aligned,” says Lin. “I think we were younger and much more naive, and we believed that all the other Amazon people that were coming at us to say, ‘Hey, let’s do something together,’ were sincere in their interest in working with Twitch and not so much driven by their own selfish objectives. Ninety percent of the time, that is what was going on, and so we became somewhat jaded. We became a little less interested in exploring anything too much deeper because a lot of these conversations led to dead ends. And wasted time, debates, and arguments.”

“There were people I worked with for three years, and it was not uncommon for me to hear [them say], ‘I still don’t understand why anyone would watch something like this,’?” says Graham. “I don’t know how to respond to something like that, because that’s not much different than [them] coming into someone’s house, taking a big shit on the living room floor, and then leaving and thinking they’ve done nothing wrong. I felt so hurt when something I worked so hard on was treated like it was a fad—like it was just a stop on a long road or whatever corporate ladder somebody was trying to climb in Silicon Valley.”

IN HIS FINAL days at Twitch before departing in 2022, Graham attempted to sell Twitch on the idea of a program meant to foster greater empathy between Twitch employees and creators.

“Honestly, when you start at Twitch, you should be able to fill out a form of what your interests are and be matched up with a creator or set of creators, or have a list of creators,” he says. “And your fucking job for two weeks should just be to watch Twitch, to chat, to subscribe to someone, to look at their Twitter. If you do that, you’re immediately gonna have some sort of connection, you’re going to immediately be able to build some sort of empathy, and then you have to reinforce that over time.”

He is unsure if anybody picked up that ball and ran with it, but he assumes it now lies on the floor of a darkened gymnasium, gathering dust.

Graham describes his departure from Twitch as the result of “death by a thousand cuts,” but ultimately it came down to a guilt-laden sense of having failed the community. Controversies over copyright issues on the platform, he says, played a big role, because he feels like he could’ve done more in Twitch’s early days to head them off at the pass. But the hate raid epidemic—and the headline-making #TwitchDoBetter hashtag it spawned that begged Twitch for some kind of response—was the last straw.

“That was an incredibly deep cut, not only because of what it represented—a group of marginalized individuals getting raided and harassed—[but] what really pissed me off was that we didn’t have a conversation with RekItRaven, the creator behind [the hashtag], for two fucking weeks,” Graham says. “As someone who spent five years pre-Amazon building Twitch, all based off of creators, [that was unacceptable]. There were times in 2012 and 2013 where a creator would be like, ‘Oh, I cut myself in the kitchen,’ and I’d be like, ‘Are you OK? Do you need anything? Can I help you find an urgent care?’ We would pick up the phone and we would fucking call them because they were people, and we cared about what they were doing.”

Kaitlyn “Amouranth” Siragusa, one of the most popular female streamers on Twitch, has found that a lack of consistent communication has strained her relationship with the platform over the years.

“Twitch is a weird platform because no one can really ever get direct communication from the people on the inside,” says Siragusa, whose risqué content has gotten her suspended from Twitch on several occasions. “I don’t really have a clear thought of whether I’m good with them or if I’m on thin ice.”

But the Twitch of now is not the Twitch of the good old days. While the anonymous former Twitch employee—who worked alongside Graham and advocated for many of the same causes—understands where he’s coming from, she believes Twitch has stranded itself in an awkward middle ground: It brands itself as a community leader, but in reality, the community has been leading itself for a long time.

“It puts Twitch in a really weird position because their marketing is ‘We’re a community, we’re a family,’?” she says. “But in those moments, it’s very much not. It’s a corporation.”

The solution, she believes, would be for Twitch to drop the facade entirely.

“I think Twitch would benefit from stopping trying to sound like they’re a community leader and instead positioning themselves as the best tool and the best place for communities to organize and meet and grow,” she says. “For people from the old days, like [Graham], that’s really hard because when your role, or your whole identity or success or popularity, has been wrapped up in being a voice of the community, it definitely feels like things are being taken away from you.”

Graham is fully aware of the limits of direct one-to-one communication, especially on a site that now services hundreds of millions of people per month. But he, much like the users behind the viral hashtag, still believes Twitch can do better.

“We handled #TwitchDoBetter like a corporation. And great, but we should’ve handled it like a corporation that actually cares about the people that are using its product,” he says. “I came to a point in 2021 where I realized that there is a complete lockdown on communication from Twitch…. So here I was realizing that the glory days will never return. And that’s fine; companies exist. But that’s not an excuse for giving up on the values the company was built on.”

Graham, Lin, Ignacio, Goldhaber, and Amaya are no longer working at Twitch. Hundreds of others have also departed the company in recent years. In 2021 alone, over three hundred employees left. Since Graham’s departure, the company has changed even more. In 2022, Twitch removed a contract option that allowed for a 70/30 revenue split on subscriptions in streamers’ favor. This had previously only been offered to a small percentage of creators, but for years it functioned as an aspirational goal and a sign that Twitch valued the faces of its platform. The decision was met with outrage, especially when paired with increased pressure to run ads that disrupted broadcasts and turned away new viewers.

Even creators who understood from a business perspective why Twitch made these moves, like politics juggernaut Hasan “HasanAbi” Piker, struggled to bear the weight of Twitch’s heavy-handedness.

“When the announcement came that the 70/30 split was no longer [available], or when I learn that I’m going to have to run more ads no matter what,” says Piker, “those are things that remind me of my place, ultimately.”

In early 2023, Emmett Shear—CEO of Twitch since the beginning—stepped down following the birth of his first child. New CEO Dan Clancy’s first major public announcement, just a week later, was a sobering one: As part of a round of nine thousand job cuts at Amazon, four hundred Twitch employees (out of a little over two thousand) would be laid off. Some were recent hires. Others had been with the company for over half its existence.

Devin Nash, who rose out of the esports community, represented top streamers publicly and behind the scenes, and eventually came to run an agency dedicated to connecting brands with creators, was at various points about as connected to Twitch’s inner workings as one could be without being an executive there. He believes that Twitch’s drift off course and into a sea of corporate BS was unfortunate but inevitable.

“The story of Twitch—I was so dramatic about it in the past—is just the classic mergers and acquisitions story,” says Nash. “Executives that love the company, that have the vision, that really want to build something special and understand the culture really well, they cash out. They have incredible opportunities, they’re ready to move on to the next thing. No moral, ethical judgment. That’s just how business works. They move on, the big corporate giant brings in people that understand [the company] a little bit less, and at the same time, they’re getting so much budget to hire people at the management level that understand less, too. Before you know it, you’ve come to a place four or five years later where no one really understands why the whole thing worked.”

“The culture of Bleed Purple,” he adds, “there just wasn’t anybody left at the company to tell that tale.”

Plante, who first coined “Bleed Purple,” left Twitch in 2017. She’s still part of the community—now the director of business operations for longtime streamer Chris “Sacriel” Ball, who is also her husband—but she regrets the slogan.

“It has been weaponized in my mind,” Plante says. “It’s become something that late Twitch, third-gen Twitch—whatever you want to call it—has definitely used against the community. Not for the community or with the community. That’s why I feel like it really doesn’t fit anymore.”

“Twitch is not a dead platform. The concurrent viewership is still absolutely astronomical,” she adds. “It’s the sense of community culture that maybe is in a coma, and someone’s about to pull the plug.”

Lin cannot help but ponder what would’ve happened if he and other Twitch executives had kept the company independent. No Amazon, no Google. Just Twitch.

“If we stayed independent, would we have been able to grow quite as fast? Maybe not,” he says. “We would have had to invent additional [means of] monetization. We would’ve had to raise a lot more [investment] money. But I believe we would’ve been much more disciplined on costs and monetization building. The platform might look a little bit different than it looks today, [but also] from a policy perspective and a fun perspective.”

“I’d still be there,” he adds. “I never wanted to leave. I love the work. I love the people. I think we could have stayed smaller as a team…. Would we have survived? I don’t know. I really don’t. We’ll never know.”

GRAHAM SITS ON his luxuriously lengthy couch, but he looks tapped. By this point, he’s nearing the end of a long, emotionally taxing journey through his professional history. The day has played host to no small number of unexpected cameos. Not his former coworkers—those conversations, to give you a peek behind the curtain of the journalism factory, would end up taking place later over the phone—but a flock of wild turkeys outside Graham’s house as well as Graham’s wife and teenage son.

The latter, who grunts a hello and then recedes into his bedroom, briefly becomes the focal point of conversation. From behind his closed door, he’s audibly talking to others while playing a video game. Graham explains that his pride and joy—despite inheriting his father’s formidable height—is not following in his footsteps. Instead of streaming to an audience of hundreds or thousands, Graham’s son is broadcasting to just a small handful of personal friends over Discord, a popular communications app.

“A lot of kids today are streamers; they just don’t do it publicly,” says Graham. “They hang out in a group of eight, and it’s like, ‘It’s this person’s turn this week to stream some game’ that they all sit there and [joke about] in real time, right? It’s basically the same as streaming, except you’re saying, ‘We’re doing this because we like the social aspect of it, not because we feel the need to make money or want to make a video that we’re going to upload.’?”

It is, in a sense, closer to the platonic ideal of livestreaming than what Twitch has made it into: the feeling of sitting on a couch with friends while one person plays a video game and everybody else hoots and hollers and talks shit. Just like the old days, but through new technology. For a generation raised in the constant blaring noise of social media, there’s a clear desire to curate more personal spaces—free from the prying eyes and ears of companies and the general public. These communities, vastly smaller than the ones Graham has dedicated his life to serving, seem at first glance more capable of building themselves, of sustaining themselves, sans the need for outside interference or advocacy. In a time when global squares like Twitter collapse and algorithms wall off platforms like Instagram to the degree that people exist in their own little pockets, it’s not difficult to imagine a future in which atomized communities become the norm. Perhaps, in that future, there will not be a need for someone like Graham to perform what ultimately proved an impossible task—at least, at the scale Twitch forced him to operate.

But no matter what shape gaming communities end up taking, Graham believes that in the face of corporate interests and incentive structures, they’ll always need an advocate. He’s getting back to it, he explains with the wide eyes of somebody half his age, at a company called Fortis, where he hopes to create a process by which community can directly influence the creation of video games. After years of putting out fires, he feels relieved to just help make something again. Maybe it’ll pan out, maybe it won’t, but he’s glad to be back in his community-first wheelhouse.

“I want to keep fighting that fight as much as I can,” he says.

1

CHAPTER 1 Why You Should Care About Chronic Inflammation