The First Death
In the summer of 2004, a small team of programmers and artists in Vancouver was quietly burying a dream. They had spent nearly two years building something called Game Neverending — a sprawling, whimsical online world where players could craft, trade, and simply exist together, long before anyone had coined the term "social gaming." The game was strange and beautiful and, crucially, unfinished. The money was running out. The team, led by a philosophy graduate named Stewart Butterfield, had to make a decision that would define the next two decades of his life, though none of them knew it yet.
Buried inside the failing game was a small internal feature — a simple photo-sharing tool the team had built so players could post and swap pictures inside the world. It was never meant to be the product. It was scaffolding, a side door, something they'd hammered together to solve a minor problem. But as the game died, that scaffolding was the only thing still standing.
They stripped it out, rebuilt it, and gave it a name that sounded almost like a dare: Flickr.
Within a year, Flickr had become one of the defining products of the early social web. In 2005, Yahoo bought it for a sum reported to be in the tens of millions of dollars — a life-changing outcome for a team that had, by every conventional measure, just failed. Butterfield became something of a minor legend in Silicon Valley: the guy who built a game nobody finished playing and accidentally built one of the most important products of the decade instead.
It would have made a fine story if it had ended there. A neat little parable about failure and reinvention, the kind that gets trotted out at graduation speeches. But Butterfield wasn't finished making the same mistake — or, as it would turn out, the same miracle — twice.
The Second Death
By 2009, Butterfield had spent four restless years inside Yahoo, watching a company that felt increasingly allergic to the scrappy instincts that made Flickr work. He left, and did the only thing that made sense to a man who had just watched a broken game turn into gold: he tried to build another game.
The new company was called Tiny Speck. The new world was called Glitch — an ambitious, gorgeously illustrated, almost defiantly uncommercial browser game where players wandered a surreal universe milking giant butterflies and cultivating vegetables shaped like human faces. It had no combat, no violence, no clear win condition. It had heart, and it had a devoted cult following. What it did not have, after years of development and millions of dollars in venture funding, was enough players to sustain a company.
In November 2012, Tiny Speck announced it was shutting Glitch down. For a team that had already lived through one public death, this one landed differently — harder, in some ways, because they'd believed they'd learned the lessons the first time. Employees who had poured years into painting virtual skies and scripting talking vegetables now had to explain to family and friends that the game was closing, the servers were going dark, and nobody quite knew what came next.
Butterfield would later describe the particular grief of watching a community dissolve, not just a company. Reflecting on the shutdown years afterwards, he said that "every death of an online community is its own kind of tragedy" — because the people inside it believe, right up until the end, that the relationships they built will simply carry on somewhere else. Most of the time, once the servers go dark, they don't.
What came next was sitting in plain view the entire time, exactly the way it had with Flickr. Buried inside the guts of Glitch's development process was a piece of internal infrastructure nobody outside the company had ever seen: a messaging system the team had built purely to talk to each other. It let them post updates in shared channels, search old conversations, and drag-and-drop files into threads instead of drowning in an endless chain of emails. It was unglamorous. It was internal. It was, again, scaffolding.
But something about it kept nagging at Butterfield. The team had built dozens of tools over the years to manage the chaos of remote and distributed work — engineers in one city, artists in another, testers scattered everywhere. Most of those tools were forgettable. This one, people didn't want to let go of.
The Tool That Wouldn't Die
Here is the detail that separates this story from a simple "failure teaches success" fable: Butterfield didn't have a flash of inspiration so much as he had the discipline to notice a pattern he'd already lived through once. He had already watched a side-project outlive its parent. He knew what that felt like from the inside. So when Glitch's servers went dark, instead of mourning and starting from a blank page, he turned to his shrinking, exhausted team and asked them to look again at the internal chat tool they'd built almost by accident.
They rebuilt it as a standalone product. They gave it a name built from an acronym that doubled as a piece of quiet self-deprecation: Searchable Log of All Conversation and Knowledge. Slack.
Butterfield has since pointed to that process as unlike anything he'd built before. Most software gets shaped by committee, by argument, by whoever fights hardest for their feature — but this wasn't that. As he put it later, building Slack "was totally different from the normal process of software development," a rare case where the product had already proven itself before anyone had to sell the idea internally. There was nothing left to argue about. The tool had already won the argument by surviving.
It launched to the public in August 2013, and the growth that followed startled even the people who built it. Within its first day, thousands of teams signed up. Within a year, it was being used by hundreds of thousands of people daily, spreading almost entirely through word of mouth inside tech companies that were sick of email threads and fragmented tools. Slack didn't need to convince people that communication was broken — everyone already knew that. It just had to be the tool that didn't feel like a chore to use.
What made Slack different wasn't really the technology. Group chat had existed for decades before it. What Butterfield's team understood, because they had lived it as exhausted developers trying to ship a dying game, was that workplace tools fail when they demand effort from tired people. Slack was built by a team that was burned out, under-resourced, and racing against a shutting company — which meant every feature that survived had to justify its existence to people with no patience left for friction. That constraint, born entirely out of failure, became the product's greatest strength.
The Arithmetic of Falling Upward
By the time Slack went public on the New York Stock Exchange in June 2019, it did so through a direct listing rather than a traditional IPO — itself a quiet rebellion against convention, fitting for a company whose entire existence was an accident of a failed video game. Its opening valuation placed it among the most valuable enterprise software companies to ever reach the public markets, built by a founder who, eight years earlier, had been personally emailing apologies to a community of devoted players whose imaginary world he was about to switch off.
The story reached its formal conclusion in 2021, when Salesforce completed its acquisition of Slack in a deal worth tens of billions of dollars — one of the largest software acquisitions in history. The company that began as a footnote inside a game about giant butterflies had become core infrastructure for how millions of people organize their working lives, from hospitals coordinating shift changes to newsrooms breaking stories to small businesses two employees deep.
What the Story Actually Teaches
It would be easy, and lazy, to reduce this to a motivational poster: fail twice, win once. That flattens what actually happened, and it lets the real lesson slip through the cracks.
The truth is more uncomfortable and more useful. Butterfield didn't succeed by refusing to give up on his dream. He succeeded by being willing to let the dream die completely — twice — and then paying close enough attention to notice what quietly survived the wreckage both times. Most people who fail either abandon everything, dream included, or cling so hard to the original vision that they can't see the smaller, stranger thing standing in the wreckage that actually works. Butterfield's real skill wasn't game design or coding. It was the rarer discipline of grief followed immediately by clear-eyed observation — the ability to bury a project he loved and, in the same week, ask what part of it was still breathing.
There's a quieter detail in this story that rarely makes it into the retellings: both times, the thing that survived wasn't the ambitious, publicly celebrated product. It was the invisible, unglamorous tool built to solve an internal, unsexy problem — how do we share photos while playtesting, how do we talk to each other while we're scattered across time zones? Nobody set out to build Flickr or Slack. They set out to fix a small annoyance while trying to save something bigger, and the fix outlasted the thing it was built to save.
For anyone who has watched a business, a project, or a carefully tended dream collapse, there's something genuinely steadying in that. Failure, in this story, isn't a wall. It's more like a controlled demolition — violent, humiliating, and expensive, but occasionally revealing a foundation underneath that was stronger than anything built on top of it. The game died. Twice. What was underneath it is now how a large fraction of the working world says good morning to each other.