Part IThe Story
Thirty minutes into Casino Royale, James Bond is driving a silver Ford Mondeo through the Bahamas, trailing the banker Le Chiffre, when he glances down at his Sony Ericsson phone. It is shameless product placement, and the handset already looks antique. On its screen, though, a small graphical icon of the Mondeo moves across a map toward its destination. In mid-2008 Garrett Camp was twenty-nine, recently and comfortably rich from the sale of his first company, and living in an apartment in San Francisco's South Park neighborhood. He liked to put the Daniel Craig DVD on when he was relaxing at home. He liked the movie fine. What stayed with him was that icon: a car, rendered as a glyph, approaching.
Most people who watched that scene saw a spy and a gadget. Camp saw an interface. He saw something closer to a solved problem, or the outline of one: a vehicle whose position was knowable, whose approach could be watched, whose arrival did not depend on standing at a curb in the cold with an arm raised while empty cabs went past. "I remember one day — after watching James Bond and seeing him tracking his car on his phone — I was looking at my iPhone, waiting for the cab company to call me," he told an audience at his alma mater years later, "and I thought, why can't I just push a button and they would come?"
The button became Uber. Uber became, by the end of 2025, a company with $52 billion in annual revenue, 202 million monthly users, 10 million drivers and couriers, and an average of 42 million trips and deliveries every day. It became a verb in several languages and a defendant in many more jurisdictions. It became the subject of a book called Super Pumped and a television series of the same name, in which the man who first imagined it was played by Jon Bass, a supporting actor, while Joseph Gordon-Levitt played the man who ran it.
That casting was accurate. "There's nothing particularly memorable about Garrett Camp," began the longest magazine profile anyone has written about him, in the Globe and Mail's Report on Business in 2016. No uniform, no turtleneck, a wide and genuine grin, long curly hair, "but he doesn't exactly ooze charisma." His personal website at the time consisted of a thumbnail headshot on a dark-grey background and thirteen words: I'm an entrepreneur based in San Francisco. Founder @ Expa, Uber and StumbleUpon. @gmc. He was, at that moment, the third-richest Canadian alive, behind only David Thomson and Galen Weston, heirs to fortunes built across generations. Camp had built his in about seven years, mostly by having an idea and then, with considerable deliberateness, handing it to someone else.
What the icon on Bond's phone did not show, because icons don't, was who was driving.
By the Numbers
The Inventor's Ledger
$75MeBay's price for StumbleUpon, May 2007
$800One New Year's Eve on a private driver, the bill that rankled
4%Camp's approximate stake in Uber at its 2019 IPO
$3.7BValue of that stake on the eve of trading
72BTrips and deliveries Uber has coordinated since 2010
$350MRaised across three Expa funds, 2013–2022
$72.5MPrice of the Beverly Hills house drivers picketed in 2019
No reason to leave the basement
The Camps of Calgary were builders. His father was trained as an economist and his mother as an artist, and at some point both of them abandoned those careers to design and construct houses together — which is to say that Garrett grew up in a family whose business was the deliberate manufacture of comfortable interiors. By the time he finished high school, his parents' own house had three storeys, a proper office, and a computer room in the basement. "There wasn't much reason to leave," he told Brad Stone.
He didn't, much. An uncle had given the family an early Macintosh, the floppy-disk kind, and through the long Alberta winters the boy sat with it, playing point-and-click adventure games, fiddling with primitive computer graphics, writing basic programs. He played sports and learned electric guitar and asked, by every account, a great many questions. He enrolled at the University of Calgary, a short commute from home, and kept living with his parents to save money. The one interruption was a year in Montreal, interning at Nortel Networks, then the most important technology company Canada had ever produced and soon to become one of the most spectacular corporate collapses in the country's history. He came back to Calgary. He took his bachelor's degree in electrical engineering in 2001 and stayed for a master's in software engineering, and only after turning twenty-two did he finally leave the nest — for a campus apartment, with classmates, a few kilometers away.
It is tempting to read the basement as a deficiency, an absence of hunger. Put it next to the childhood of the man who would become his partner and it looks more like a different temperature of ambition. Travis Kalanick grew up in Northridge, in the San Fernando Valley, the son of a civil engineer who worked for the City of Los Angeles — a man whose job, literally, was the public roads — and a mother who sold retail advertising for the Los Angeles Daily News. As a teenager Kalanick sold Cutco knives door to door. At eighteen he started a test-prep company with a classmate's father. He dropped out of UCLA. Where Camp's parents built private houses, Kalanick's father built the municipal infrastructure that Kalanick would later spend a decade fighting.
Camp's master's research concerned collaborative systems, evolutionary algorithms, and information retrieval — the problem, roughly, of how a machine finds what a person wants when many people are helping it learn. The thesis was not abstract for long. It was already running, on a few thousand computers, under a name he had come up with himself.
What if you don't know what you want
The idea for a company came from Geoff Smith, whom Camp had met through a childhood friend. "It was actually Geoff who first suggested we start a company," Camp said in 2008, back on campus in Calgary, a year after the sale. "I came up with the name soon after, and then we figured out how to make it work while we lived together that first year." Two of Smith's co-workers, Justin LaFrance and Eric Boyd, made four. They batted around ideas for about a month. The one they kept was a homepage that would be different every time you opened it — "a new page, just for you."
StumbleUpon launched in November 2001. It took the form of a browser toolbar, a row of about fifteen buttons along the top of the screen. You told it, broadly, what you liked. You pressed Stumble, and it took you somewhere. You pressed I like it or a thumbs-down, and those verdicts went back into the system, which compared your taste to the tastes of people who resembled you and refined what it sent next. Camp's framing of it, repeated in interview after interview for a decade, became the cleanest statement of the product thesis he has ever produced:
"Google is cool if you know what you want. But what if you don't know what you want?"
It is worth sitting with how odd a proposition that was in 2001, when Google was still a fringe search engine and the dominant metaphor of the web was the library catalog. Camp was building for the person with no query. Within a year, with no marketing and no outside money, StumbleUpon had 50,000 users. "In dorm rooms, people would ask, 'What are you doing?'" Camp remembered. "And people would say, 'Oh, I'm Stumbling.' And that's kind of how it spread."
He kept going to class. For the first few years the company was a part-time project, built "out of our bedrooms," as he told MIT Technology Review. By the time he was finishing his thesis in late 2005, StumbleUpon had somewhere between 500,000 and 600,000 users, and the money still hadn't come. Then it came to him — through the product. A heavy user named Brad O'Neill, an angel investor in Silicon Valley, sent Camp a note through StumbleUpon's own messaging system: If you ever want to raise some money, I could really help you. They met in December 2005. "I wasn't actively seeking investment," Camp said. "I was still in school." He went down for what he thought would be a few weeks. The investors O'Neill introduced him to "liked the fact that we'd achieved 500,000 or 600,000 users with zero funding and no help. We were just hacking away." The round came together in two weeks: $1.5 million, from a syndicate that included Mitch Kapor, of Lotus and the Mozilla Foundation, and Ram Shriram, who had written one of the first checks into Google and sat on its board. Google's investor backing Google's philosophical opposite was the kind of hedge that Silicon Valley considers elementary and Calgary, at the time, might have considered a contradiction.
The company moved to San Francisco in 2006. "We were a couple of guys with no attachments and I was just done my thesis so I was looking for a change," Camp said. Over the next twelve months StumbleUpon went from 500,000 users to 2 million. The offers started. In May 2007, eBay bought it for $75 million.
Why eBay? An auction marketplace and a serendipity engine have nothing to say to each other, which was precisely the attraction. "Some people were like, 'I don't get it,'" Camp told Canadian Business. "And I kind of liked the idea of not being integrated." Under eBay, StumbleUpon stayed in San Francisco, kept its team, and ran as what Camp called "pretty much still an independent subsidiary." He liked having someone else do the payroll and HR. He got the money and kept the product. For about a year, this was perfect.
Then it wasn't. "In the first year, I liked it, and then I got a little frustrated with how things weren't moving quite as fast," he said. "Once you get to a certain size, you have to have rules and regulations, just to protect yourself. When you're small, you're just five people, you just trust everybody."
Set that sentence aside. It comes back.
In May 2008 the Calgary Herald photographed him in Mac Hall, the student center at the University of Calgary, holding his BlackBerry. He was twenty-nine and, the paper noted, "could easily be mistaken for a student." He gave full credit to his co-founder: "Geoff is definitely the unsung hero here. He doesn't talk to press much but I never could have built SU without him." A local blogger who watched him in an interview around then wrote, with Canadian approval, that Camp didn't come across like "the hyper-active, self-absorbed, 22 year old assholes" the industry produced in bulk: "He's a guy you could have a beer with."
Seven months later, in Paris, he would go out for a drink with exactly the other kind of guy.
The price of a night out
The origin story Uber told about itself for years went like this: on a snowy night in Paris in 2008, Travis Kalanick and Garrett Camp couldn't get a cab, and they vowed on the spot to fix it. Vanity Fair's Kara Swisher called it "a tasty blame-the-French origin story" that was "only partially true." It is also, in a minor way, a story about a different city.
The more prosaic catalyst was a bill. On a New Year's Eve, Camp and his friends — by some accounts including Kalanick — had spent $800 hiring a private driver to shuttle them between parties. The night had been fun and the number had been absurd. Camp was by then a millionaire, which is the detail that makes the anecdote interesting rather than banal: this was not a man who needed to save $800. He was a man offended by inefficiency. A driver sat idle between pickups. A car had capacity nobody used. Why couldn't that cost be shared across many people who each needed a ride at a different moment?
And then there were the cabs in San Francisco, which in 2008 were notorious. "I was frustrated that they would never show up," he said. "I would be standing there watching these cabs go by me and think, 'these cabs are empty, why are they not stopping? They are not coordinated at all!'"
They are not coordinated at all. For a man whose graduate work concerned collaborative systems, a fleet of empty cars cruising past waiting customers was not an inconvenience. It was a design error.
In December 2008, Camp and Kalanick were both in Paris for LeWeb, the annual European tech conference. Both were flush and idle. Kalanick had just sold Red Swoosh to Akamai for roughly $19 million; Camp's StumbleUpon money was a year old. Back in a shared apartment on the outskirts of the city, in one of the brainstorming sessions Kalanick called the JamPad, a group of entrepreneurs traded start-up ideas. An on-demand car service, inspired by the snow, was one of them. According to people in the room, it did not stand out. Kalanick, back in San Francisco, more or less moved on.
Camp did not. He obsessed. He bought the domain UberCab.com. He carried a notebook. Scott Belsky, the co-founder of Behance, was working with Camp on a partnership with StumbleUpon around then and recalls him producing it across the table: "I'm working on this side project that helps people summon a limo car." Belsky's reaction was the reasonable one — "You just bought your company back from eBay! ... Aren't we supposed to be focused on one company at a time?" — and also, it would turn out, the wrong one.
The notebook contained an unresolved question of positioning. "He was thinking about whether the brand narrative should be 'Everyone's taxi' or 'Everyone's private driver,'" Belsky remembered. Camp chose the second. UberCab would be black cars only, priced at roughly one and a half times a taxi, sold as a small luxury — the private-driver experience Camp had paid $800 for, sliced into individual rides. The Fast Company version of the pitch, as he gave it in 2012, was less elevated: "[The idea for Uber] was a lifestyle thing. Me, my co-founder, and our 100 friends could roll around San Francisco like ballers."
There was also a deck. In his original presentation to investors in 2008, Camp described the business as "profitable by design." It would take fifteen years, and losses that the Financial Times totaled at $31.5 billion, before Uber posted its first operating profit, in 2023. Camp later admitted, without much evident embarrassment, that he "didn't really think about the scale that it would reach."
He funded it himself — $250,000, by one account — and built it with friends. Two of them did most of the engineering. Oscar Salazar, the lead developer, was not an American citizen and took his compensation entirely in equity, a stake reported to have been worth around $250 million when Uber went public. Conrad Whelan was a fellow University of Calgary engineer, a bachelor's in 2002 and a master's in 2005, a year behind Camp at every stage, and he would stay on at Uber as an engineering manager. The first Uber, in other words, was partly a Calgary product: two Schulich School of Engineering graduates translating a Bond film into a dispatch system.
By the middle of 2009 they had a prototype. In the prototype you could see the car approaching on a map.
Handing over the keys
Neither founder wanted to run it.
Camp was still CEO of StumbleUpon, which he and a group of investors had bought back from eBay in 2009, and he had no intention of leaving it. Kalanick had agreed to serve as what both men called a "mega advisor" — part-time, temporary, in his own description there to "temporarily run the company," get the product to prototype, and find someone to operate it full-time. The someone was Ryan Graves, who answered a Kalanick tweet, became Uber's first employee in February 2010, and was named CEO that spring. In May 2010 UberCab gave its first public rides at the SF AppShow, ferrying attendees around in limos summoned from a phone. In October 2010 the San Francisco Municipal Transportation Agency served it with a cease-and-desist order for operating as an unlicensed taxi company.
Kalanick told the company to ignore the order. He did change the name — dropping "Cab" so no one could accuse Uber of falsely advertising itself as a taxi service — and he kept the cars on the road. Two months later, in December 2010, Graves stepped down to become chief operating officer and Kalanick became CEO. "I was really involved at the very beginning," Camp said later. "I handed him the keys in mid 2010."
The keys came with equity. Camp and Graves each signed over a large portion of their shares to Kalanick when he took the job; by one detailed account his stake went from 12 percent to 23 percent. This was not a coup. It was, by every indication, exactly what Camp wanted: a person with the appetite to operate the thing he had designed, compensated enough to care. "My goal is to create something that can live beyond me," Camp told Bloomberg's Eric Newcomer in 2019. "Any one person is going to hold a company back if it depends on them."
— Travis Kalanick, at an early Uber event in San FranciscoWhen you open up that app and you get that experience of like, 'I am living in the future. I pushed a button and a car rolled' … Garrett is the guy who invented that shit.
Kalanick, it should be said, never stopped crediting him. But Kalanick also arrived at Uber carrying a decade of scar tissue that Camp, the bootstrapper whose angel round had found him through his own product, did not share. At Scour, the peer-to-peer search company Kalanick joined out of UCLA, investors led by Michael Ovitz had fought the founders over terms and ended up with majority control; then the movie and music industries sued the company for $250 billion and drove it into Chapter 11. Red Swoosh, the company Kalanick started next, he called his "revenge business." It nearly died repeatedly. In September 2001 it used about $110,000 in payroll tax withholdings to stay alive, a decision whose authorship has been disputed in every subsequent telling. Kalanick went more than three years without a salary. At one point the company was down to two employees. He moved back in with his parents to save money and later complained that "it sucked," because he "wasn't getting ladies."
Both founders, in other words, had lived at home in their twenties. One had done it because there was no reason to leave. The other had done it because there was nowhere else to go.
The lesson Kalanick drew from Scour and Red Swoosh was that investors were adversaries and control was oxygen. When Bill Gurley of Benchmark put $11 million into Uber in February 2011, for 20 percent of a company valued at $50 million, Kalanick structured the terms to favor himself and the company. As subsequent rounds came in, he limited what financial information investors could see, and he made sure that shares sold to new investors carried one-tenth the voting power of the shares held by the three insiders — Kalanick, Graves, and Camp.
Camp, then, was not merely a co-founder emeritus. He was one of three people holding the super-voting stock that made Uber's governance a fortress, and from 2010 he was chairman of its board. Years later Kalanick would describe that board to Apple's Eddie Cue, during one of his standing meetings with Cue and Tim Cook, in terms that Mike Isaac recorded in Super Pumped: "The board is irrelevant. I hand pick all of these guys. They do what I tell them, and the way I've structured things, I do what I want."
Camp, asked about Uber in those years, liked to say he was "really just more of an adviser."
The quiet partner and the loud one
There is a version of the Uber story in which Camp is the angel on the shoulder — the gentle Canadian inventor whose beautiful idea was coarsened by a Californian brawler. It is a satisfying version. It is not quite available on the evidence.
What the evidence shows instead is a division of labor that each man seems to have found congenial. From 2011 on, Kalanick ran Uber the way he had survived Red Swoosh: as a war. Uber launched in New York in May 2011, Paris that December, Toronto in March 2012, London that July. In July 2012 it introduced UberX, cheaper cars; in April 2013, after Lyft, Sidecar and Wingz had won licenses to let ordinary people drive their own vehicles, Uber copied the model it had publicly called illegal. By December 2013 it was in 65 cities. Everyone's private driver became, in practice, everyone's taxi, the other option in Camp's notebook, which turned out to be the bigger business. Kalanick entered cities first and argued with regulators later. Uber's strategy, as critics summarized it, was to "seek forgiveness rather than permission."
Camp, meanwhile, ran StumbleUpon. He took it from a Firefox add-on to a cross-device product with apps for iPhone, Android and iPad and a staff of more than ninety. He talked to Mashable about international expansion: "We kind of know what the rules are here." In May 2012, with StumbleUpon at 25 million registered users, he stepped down as its CEO after ten years. "It was my first job and my first love," he said. Asked whether he might have preferred to leave earlier and join the real-world fight at Uber, he said no. He had a few new ideas, he said, but they would be "virtual" products. Kalanick was better suited to the operational role. Camp wanted to work on products.
The thing about the operational role is that it is where the decisions get made.
Between 2012 and 2017 Uber's operators made a long series of them. In August 2014 a rival counted 5,560 rides ordered and canceled by people tied to Uber. In November 2014 a senior vice president, Emil Michael, floated at a dinner the idea of spending "a million dollars" on opposition researchers to dig into the personal lives of hostile journalists, specifically Sarah Lacy of PandoDaily, who had accused the company of sexism. That same year the company's "God View," which let insiders track individual riders, drew a letter from Senator Al Franken. A data breach exposed 50,000 drivers' information; Uber waited more than five months to tell them. Uber built a tool called Greyball that served fake versions of its app to suspected law enforcement officers in places where its service was illegal — Portland, Oregon; Australia; South Korea; China; later, documents would show, across Europe. It built another, codenamed Ripley, to lock and power down office computers during government raids, and used it at least two dozen times between spring 2015 and late 2016. In 2016 hackers stole data on 57 million riders and 600,000 drivers; Uber paid them $100,000 and concealed the breach for a year.
When 124,000 internal documents from those years leaked in 2022 — the Uber Files, handed to The Guardian by the company's former European lobbyist Mark MacGann — the most quoted line came from a message thread about sending Uber drivers into a French taxi protest that colleagues feared would turn violent. "I think it's worth it," Kalanick wrote. "Violence guarantees success."
The names in that coverage were Kalanick's and his lieutenants'. Camp's appears nowhere in it that I can find, which is consistent both with his account of his role and with the critique of it. The Globe and Mail put the question to him in 2016, at the height of all this. "It's pretty simple," he said. "Uber is great. Uber is bringing efficiency to a market that wasn't efficient." And beyond that? "I can talk about it generally, but when you're on the board of a company, you don't have the daily exposure to these sorts of issues. I just don't think it's your place to comment on something unless you're well-informed."
He was, at the time, chairman of the board.
Newcomer's 2019 Bloomberg portrait, assembled from interviews with more than ten current and former employees and investors, called Camp "a factory of ideas, with a preternatural aversion to confrontation," and reported that some former colleagues considered "his most severe lapse" to be "allowing Kalanick's power to go unchecked. At pivotal moments in the boardroom, Camp was indecisive or unreachable." The same article delivered the line that has trailed him ever since: of everyone made rich by Uber, Camp "holds the distinction of reaping the largest windfall for doing the least amount of work."
He did not dispute it. "This was by design," he said.
— Garrett Camp, to Bloomberg, May 2019My goal is to create something that can live beyond me. Any one person is going to hold a company back if it depends on them.
Read generously, this is the creed of a systems designer: a well-designed system should not depend on any single node, including the one who drew it. Read less generously, it is the creed of a man who has arranged never to be the node that fails. Both readings are true. The trouble is that the company he designed so as not to depend on him came to depend almost entirely on someone else.
Growing pains
On February 19, 2017, a former Uber engineer named Susan Fowler published a blog post titled "Reflecting on one very, very strange year at Uber." She described being propositioned by a manager on her first day on a new team, reporting it, and being told by HR that the man was a "high performer" — and later being threatened with termination for continuing to complain. The post went everywhere. Eight days later Uber's new senior vice president of engineering, Amit Singhal, resigned after it emerged he had not disclosed a sexual harassment complaint from his time at Google. Two days after that, Bloomberg published a dashcam video of Kalanick, in the back of an UberBlack, berating his own driver, who had complained about falling fares. Kalanick apologized in a memo to staff and said he needed to "grow up." In March came Greyball, the Seoul escort-bar visit, a criminal inquiry. Former Attorney General Eric Holder and his firm began an investigation of the company's culture.
At the end of May, Kalanick's parents were in a boating accident on Pine Flat Lake in California. His father was seriously injured. His mother was killed.
When Uber's board met that spring to discuss whether Kalanick should take a leave of absence, Camp was traveling abroad. He later called it a "stressful time" and declined to say much more about it.
The Holder report was presented on June 11 — "hundreds of pages" long, with a dozen pages of recommendations, the most consequential being that Kalanick step back and have his responsibilities reallocated. On June 13 Kalanick announced an indefinite leave. At the all-hands meeting where Arianna Huffington, a board member, described the findings and noted that research showed one woman on a board made it likelier there would be more, another director quipped that it would mean more talking; he resigned the same day.
On June 20, Camp published a post on Medium titled "Uber's path forward." It was the most he had said publicly about the company in years.
"Like many of you, recent events have left me upset and deeply reflective," he wrote. "When we designed the Uber prototype app back in 2009, we were just focused on creating an app that would make it easier to get around." He went on: "All companies have growing pains. And because Uber grew so fast these growing pains are much more serious. Over the years we have neglected parts of our culture as we have focused on growth. We have failed to build some of the systems that every company needs to scale successfully."
A friend, he wrote, had recently asked him what went wrong, "and the answer is that we had not listened well enough to those who got us here … our team and especially our drivers. In a highly competitive market it is easy to become obsessed with growth, instead of taking the time to ensure you're on the right path."
He closed with an anecdote: "Last week I heard a story that really stuck with me. A college professor made a comment that 'Uber has liberated grandmothers worldwide' based off the mobility Uber has brought to her mother."
The post did not mention Travis Kalanick. It did not mention sexual harassment. It did not use the words "women," "harassment," "Fowler," or "Holder." The reception was brutal. Popsugar called it "tone-deaf." The consultant Josh Bernoff ran it through a highlighter, flagging platitudes in bold and weasel words in italic, and called it "mild and meaningless." Recode's headline captured the half-step it had taken: "Uber co-founder: We've been too obsessed with growth." TechCrunch noted dryly that Camp "has not been involved in Uber's day-to-day operations for quite some time."
The next day, after five major investors including Benchmark delivered a letter demanding it, Kalanick resigned as CEO. Uber added in-app tipping — which drivers had requested for years and Kalanick had refused as a source of "friction" — shortly afterward.
It is striking, reading the post now, how precisely its diagnosis matches its author's vocabulary. We have failed to build some of the systems. Not a person failed; a system was missing. When you're small, you're just five people, you just trust everybody — Camp had said that about eBay in 2011, as a complaint about big-company procedure. Uber had become enormous while running on the trust appropriate to five people, and the systems that might have protected people from it had been treated, all along, as somebody else's operational problem.
That summer, Camp did something more pointed than the post. Kalanick, out of the job but still on the board, was telling people he was "Steve Jobs-ing it" and maneuvering to influence the choice of his successor. On a Monday in early August, Camp sent a memo to employees after fielding questions at a product leadership all-hands. "Despite rumors I'm sure you've seen in the news," he wrote, "Travis is not returning as CEO. We are committed to hiring a new world-class CEO to lead Uber." It was signed simply Garrett Camp, Co-founder. Three weeks later the board chose Dara Khosrowshahi of Expedia, whose interview presentation had included a slide reading There cannot be two CEOs.
At some point in that tumult Camp gave up the chairman's title. In March 2020 he left Uber's board entirely, staying on as an observer. Kalanick had already sold roughly 90 percent of his shares, about $2.5 billion worth, and walked out on December 31, 2019. Of the two men who had been in the apartment in Paris, only the one who had never really been there remained.
Buying back the past
Camp's relationship with his own creations has a recursive quality, like a recommendation engine feeding on its own history.
He sold StumbleUpon to eBay in 2007 and bought it back, with a group of investors, in 2009 — in a spin-out TechCrunch headlined "StumbleUpon Beats Skype In Escaping EBay's Clutches." He talked about being a "born-again startup." He stepped down as its CEO in 2012 and handed it to Mark Bartels. In August 2015 he announced he was becoming its majority shareholder again. "Some difficult changes to the product and company will be needed, and these changes will take time," he wrote. "But I strongly believe that systems like StumbleUpon play an important role in helping people discover what matters most to them." TechCrunch, enjoying itself, wrote: "He just can't quit SU."
In May 2018, with a Medium post titled "SU is moving to Mix," he shut StumbleUpon down and moved its users to a new discovery product built partly through his studio. Forbes's headline that August read, with something like affection: "Uber Cofounder Garrett Camp Is Back To An Old Problem: Finding Interesting Things On The Internet."
The studio was Expa. He had set it up in 2013, self-funded at first, deriving the name from "experience apps," and explained it to TechCrunch with the candor of someone thinking aloud. "I've come up with 10 ideas over the past year that I'm interested in," he said. "But realistically only three or four will really come to fruition." He did not want to be an angel investor. "Instead of doing angel investing, I'm trying to take that same amount of capital and put it into something where I can be involved in the branding and design." He did not know what to call it. "I've asked a lot of people what should we call it, and we don't have an answer." And he described a habit that recasts his entire career as a running experiment: "Every time I make a mistake with a company, I write it out and try to figure out why it happened." He does it, he said, for the things he gets right too. "I've come up with a good set of guidelines, and if I apply these guidelines through Expa, I can hopefully end up with a couple of cool companies."
His model, he said, was Uber. "When I think of all the things you can do right or wrong, Uber definitely got most of them right. There's a few that they got wrong, but none of them are fatal."
That was 2013. Expa raised $50 million at launch, $100 million in 2016, and $200 million in 2022, more than half of that last fund his own money; the rest came from multifamily offices and wealthy friends, among them Ram Shriram — the angel from the 2006 StumbleUpon round, still writing Camp checks sixteen years later. Its portfolio includes the challenger bank Current, valued at $2.2 billion in 2021; the open-source analytics tool Metabase; Collective, a back office for the self-employed; Reserve, eventually sold to Resy. Its managing partner, Roberto Sanabria, first worked for Camp at StumbleUpon. By its own count Expa has helped 72 companies to a cumulative value of $8 billion — real money, and also about a twelfth of what Uber alone was worth on its first day of trading.
Some of the bets have the unmistakable silhouette of the first big one. In 2013 Camp was coaching BlackJet, a seat-sharing service for private jets that the press called "the Uber for private planes," from his home. Fortune's 2016 headline: "Celeb-backed BlackJet Is Grounded. Again." Later Expa backed Aero, which sells regular people seats on private aircraft. Everyone's private driver, then everyone's private pilot. The notebook question never quite closes.
The ideas don't stop either. In 2022 he told TechCrunch's Connie Loizos that he'd recently discovered he had "like, 3,500 notes" relating to company-building in his iCloud account, and that about 10 percent of them were ideas for new things — "a solid 10" of which could be big companies. He can't help himself, he suggested.
— Garrett Camp, to the Globe and Mail, 2016It seems visionary, but really you were just strategically hacking away at something for a year.
The windfall and the gate
On November 3, 2017 — four months after the Medium post — Camp published a different kind of letter. He had just turned thirty-nine, and he had just come back from his first trip to Africa: two weeks, six locations in Kenya, some on safari and some in rural towns. "In one small town, two young girls playing in the street waved at me saying 'Jambo', and I took this photo," he wrote. "This was my first time visiting small towns in a developing country, and meeting people who struggle just to find clean water or enough to eat."
He had met Bill Gates for the first time earlier that year and had been impressed that Gates started his foundation in his forties and took "a very active and analytical approach." Camp announced he was joining the Giving Pledge, committing to leave half his assets to charity. And then he framed the decision, unmistakably, in his own idiom: "My trip to Kenya made me realize that beyond just making charitable donations, I should also begin spending part of my time on projects where hands-on systems design can create more efficient solutions, and help people who need it most." He would read the research on effective philanthropy. He invited people to email him suggestions.
The vehicle became Camp.org, an umbrella that has funded Every.org — a "frictionless" platform that lets nonprofits accept donations in stock, crypto, donor-advised funds and cash — and built free software to help detect deepfakes. Frictionless was the word Kalanick had used to block driver tipping. Camp applied it to giving. The vocabulary of the button travels.
Uber went public on May 10, 2019, at $45 a share, raising $8.1 billion and valuing the company at $82 billion. The debut was a dud; the stock closed below the offering price. On the eve of trading, Bloomberg put the value of Camp's stake at $3.7 billion and offered some comparisons: with that money he could buy his hometown Calgary Flames and probably the seven other teams in their division, or give a new Toyota Camry to nearly a fifth of the population of San Francisco. He held about 81.5 million shares that April, roughly 4 percent of the company. After the IPO he sold $210 million worth and gave other shares to charity.
That June, he and his longtime partner Eliza Nguyen paid $72.5 million for a mansion in Trousdale Estates, the hilltop enclave of Beverly Hills. It was briefly the most expensive home ever sold in the city — briefly, because the next day the Spelling Manor went for $120 million. It was not, however, too briefly to become a story. That summer, Uber drivers were in open revolt: in March and May they had struck across California, organized by Rideshare Drivers United, over pay, hours, and the lack of benefits; Uber, newly public, was posting billions in losses and laying people off. The Guardian ran the purchase under the headline "Uber co-founder buys record-breaking LA mansion for $72.5m as drivers fight for wages." And drivers went to Trousdale and picketed his house.
They had found him. The man who had stepped back from the operational role, who didn't think it was his place to comment, whose website contained thirteen words — the people at the far end of his system had located the node at the top of it, and they were standing at his gate.
The house was the kind of thing his parents built, scaled past recognition.
Where cabs go
On November 17, 2015, Camp went home. The Hunter Centre for Entrepreneurship at the University of Calgary had invited him for Global Entrepreneurship Week, and about 140 students and community members came to hear him in conversation with the university's president, Elizabeth Cannon. He did not come alone. Beside him on the stage was Conrad Whelan — BSc '02, MSc '05, the classmate-of-a-sort who had helped him code the first UberCab, by then an Uber engineering manager. Two engineers from the same school, a few years and one gigantic company later, back on the campus where one of them used to be mistaken for an undergraduate.
A month earlier, on October 15, Uber had launched UberX in Calgary, into the usual municipal fight.
Camp's advice to the students was what it always is. "Fundamentally, I want to build a product that I want to use myself," he said. "Will it make my life better and will other people want to use it? If I don't want to use it myself, then I'm not going to be excited about it, and then work becomes work and it's not fun." And on persistence: "If you're really excited to solve that problem, you're going to stick through that multi-year period where you're not going to get any traction, and half the people say it's a stupid idea."
That spring, the Globe and Mail had asked him how he expected to be remembered. "In general," he said, at his syllable-skipping pace, "people sort of remember me as a good system designer who worked on interesting products that ended up getting good traction." It is the least grandiose summary of a multibillion-dollar career on record, and also a perfectly exact one. He had designed systems. He had worked on products. They had gotten traction. Everything after the traction — the 70 countries and 15,000 cities, the $11.4 million Pennsylvania fine and the AU$271.8 million settlement with Australian taxi drivers, Greyball and Ripley and the 57 million stolen records, the strikes, the Supreme Court of the United Kingdom ruling that drivers were workers, the €290 million and €825 million Dutch fines, the first operating profit, the $10 billion net income, the robotaxi partnerships with Nvidia and Rivian, the $14.8 billion offer for Delivery Hero — all of that had been other people's daily exposure.
There is a document from before any of it. In the fall of 2007, when he was twenty-eight and StumbleUpon had just been bought by eBay, MIT Technology Review named him to its list of top innovators under thirty-five and asked him for a few of the best things he had personally stumbled upon — the interesting sites he'd found that he never would have thought to search for. He gave them five. A visual history of audio sampling in hip-hop and electronic music. A site that told you how to reach a human operator when you called customer support. Twittervision. A gallery of computational art built in Flash.
The first entry on the list was a site called cabspotting.org, built by Stamen Design right there in San Francisco, which he had found through a journalist he'd friended on StumbleUpon. His description of it ran to seven words.
"Where cabs go in San Francisco."
Part IIThe Playbook
What follows is an attempt to extract the operating principles from Camp's career — including the ones he demonstrated by omission. He is unusual among founders of his stature in that the most instructive parts of his record are not only what he built but what he declined to do: run Uber, speak about Uber, govern Uber in any forceful way. Several of these principles are worth copying. At least two are worth studying as warnings. The difference is the point.
The principles
- 1.Treat your own irritation as a dataset.
- 2.Collapse the workflow into one visible action.
- 3.Get proof before you get money.
- 4.Let your best users recruit your investors.
- 5.Launch aspirational, then descend.
- 6.Separate the inventor from the operator — and pay for it.
- 7.You cannot delegate governance the way you delegate operations.
- 8.In a crisis, only specificity counts as speech.
- 9.Keep a mistake ledger.
- 10.Run ideas as a portfolio, not a vocation.
- 11.Buy back what you understand.
- 12.Carry your design language into your giving.
G
The Camp Arc
Invention, hand-off, absence, return — a career organized around stepping back from things he started.
2001
StumbleUpon launches in November as a part-time project among four University of Calgary friends.
2005
With ~600,000 users and no funding, Camp meets angel Brad O'Neill, who found him through the product itself.
2007
eBay buys StumbleUpon for $75 million; Camp keeps it as an un-integrated subsidiary.
2008
The $800 New Year's Eve, the Bond icon, LeWeb in Paris; Camp buys UberCab.com and pitches "profitable by design."
2009
Camp and investors buy StumbleUpon back from eBay; UberCab prototype built with Oscar Salazar and Conrad Whelan.
2010
Camp "hands him the keys": Kalanick becomes CEO in December; Camp becomes chairman.
2012
Camp steps down as StumbleUpon CEO after ten years and 25 million registered users.
2013
Expa founded to build, not merely fund, new companies.
2017
Medium post "Uber's path forward"; August memo stating Kalanick won't return; Giving Pledge in November.
2018
StumbleUpon shuts down; users move to Mix.
2019
Uber IPO; Camp's ~4% stake worth roughly $3.7 billion on the eve of trading.
2020
Camp leaves Uber's board, remains an observer.
2022
Expa raises a $200 million fund, more than half of it Camp's own capital.
Principle 1
Treat your own irritation as a dataset
Camp's two great products began as personal grievances with an engineer's precision attached. StumbleUpon came from the observation that search assumed you already had a query. Uber came from three distinct annoyances layered on top of each other: an $800 bill for a driver who sat idle most of the night, empty San Francisco cabs that cruised past waiting customers, and the absurdity, in an age of GPS-enabled phones, of not knowing where your ride was. "They are not coordinated at all!" is not the complaint of a commuter. It is a bug report.
The distinction matters. Plenty of founders build for their own pain; fewer decompose that pain into its mechanical causes. Camp's irritation was useful because it was specific — idle capacity, missing coordination, invisible position — and each specific cause pointed at a design feature: shared cost, dispatch, a map. His own formulation, delivered to students in Calgary, is that the test is personal use: "If I don't want to use it myself, then I'm not going to be excited about it, and then work becomes work." The deeper lesson is that personal use is only the entry point. The value comes from asking, every time something annoys you, which component of this system is failing?
Tactic: For the next frustration you encounter twice in a week, write down not the complaint but the three mechanical reasons it happens — then ask which one a single piece of software could remove.
Principle 2
Collapse the workflow into one visible action
Every Camp product reduces to one button. StumbleUpon was literally a button labeled Stumble. Uber was, in his recurring phrase, "push a button and they would come." What made the Uber version transformative was not only that it removed steps — call, wait on hold, describe your location, hope, flag down, negotiate, pay — but that it made the remaining wait legible. The Bond icon was the insight: an approaching car on a map converts anxiety into anticipation. The waiting was still there. The uncertainty was gone.
This is a subtler design principle than "simplify." Simplification removes; Camp's move was to remove everything except the one thing the user actually cares about and then make that thing visible. A taxi dispatch that silently worked faster would have been an improvement. A tiny car crawling toward you on a screen was a revolution in how the experience felt — and, not incidentally, in how much people trusted it.
Tactic: Map every step your user takes, delete every one that exists only for your convenience, and for the step you cannot delete, find a way to show its progress in real time.
Principle 3
Get proof before you get money
StumbleUpon reached 50,000 users in its first year with no marketing and no outside capital, and between 500,000 and 600,000 by the time anyone offered to invest. Camp was still in graduate school, still attending classes, still treating it as part-time. When investors finally saw it, what they "liked," in his words, "was the fact that we'd achieved 500,000 or 600,000 users with zero funding and no help." The round closed in two weeks. Uber's first stretch was similar: Camp put in his own money — about $250,000 by one account — and built a working prototype with two friends before anyone else was asked to believe.
The advantage is not thrift. It is negotiating position and clarity. A founder who arrives with traction does not have to sell a vision; he has to decide which of several interested parties to accept. And the period without money forces a team to discover whether the product spreads on its own — whether, in the StumbleUpon case, people in dorm rooms will start describing what they're doing with your verb.
Tactic: Before your first pitch, set a usage milestone you can reach with your own resources and refuse to raise until you hit it.
Principle 4
Let your best users recruit your investors
Brad O'Neill was a heavy StumbleUpon user before he was an investor. He found Camp through the product's own messaging system and offered to help him raise money. That introduction led to Mitch Kapor and Ram Shriram — and Shriram was still writing Camp checks into Expa sixteen years later. The best capital Camp ever raised arrived through a channel he had built for an entirely different purpose.
There is a general mechanism here. Investors who discover you as users have already done the most important diligence — they believe the product works because it works on them — and they bring a quality of conviction no deck can manufacture. Camp didn't engineer this; he was, by his own account, "not actively seeking investment." But a founder can make it more likely by building products that sophisticated people use and by keeping the door open to them inside the product itself.
Tactic: Identify your ten most engaged users, find out what they do for a living, and talk to any who invest or advise before you talk to anyone else.
Principle 5
Launch aspirational, then descend
The question in Camp's notebook was "Everyone's taxi" or "Everyone's private driver." He chose the second. UberCab launched with black luxury cars at roughly 1.5 times taxi fares, sold to Camp and his "100 friends" who wanted to "roll around San Francisco like ballers." Only after the brand meant something — in July 2012 with UberX, and in April 2013 with ordinary drivers in personal cars — did Uber come down-market and become, in practice, everyone's taxi, the bigger business all along.
Belsky drew the lesson that consumers "always want special treatment," and that is part of it. The other part is strategic sequencing. A premium launch gives you better unit economics while volume is low, a smaller set of more forgiving early adopters, and a brand position that survives the eventual move down. Starting cheap rarely lets you go up; starting premium usually lets you go down. The InfoTable below sets out how the two positioning options played out.
U
The Notebook Question
The two framings Camp weighed for UberCab, and what each became.
| Framing | Launch expression | What it became | Phase |
|---|---|---|---|
| "Everyone's private driver" | Black cars only, ~1.5× taxi fare, San Francisco, 2010 | UberBlack; the brand's original aura | Launch |
| "Everyone's taxi" | Chosen against in 2008 | UberX (2012), personal-car drivers (2013), 72 billion trips | Scale |
| "Private planes for regular folks" | BlackJet, coached from Camp's home, 2013 | Grounded, twice; later echoed by Expa-backed Aero | Risk |
Tactic: Write two positioning statements for your product — one premium, one mass — and launch with the premium version unless you can show the mass version is the only one with enough demand.
Principle 6
Separate the inventor from the operator — and pay for it
Camp knew what he was and wasn't. "I'm a system designer at heart, and wanted to be directly involved in the creation of new products," he said. Running a regulatory insurgency in 70 countries was not that. In mid-2010 he "handed [Kalanick] the keys," and he and Ryan Graves signed over a large portion of their shares to make the new CEO's commitment total — by one account Kalanick's stake nearly doubled, from 12 to 23 percent. It was the most important business decision of Camp's life and it was, financially, an extraordinary success: he reaped what Bloomberg called "the largest windfall for doing the least amount of work," and said, "This was by design."
The lesson is not merely "hire an operator." It is that a founder who recognizes he is the wrong person to scale something should make the right person an owner, not an employee — and should do it early, when equity is cheap and the decision is reversible. Camp's restraint here was a genuine form of discipline. Few founders of anything that big are willing to give away the steering wheel at all, let alone generously.
Tactic: Write a one-paragraph job description for the person who will run your company at 1,000 employees; if it doesn't describe you, start recruiting that person while you still have equity to offer.
Principle 7
You cannot delegate governance the way you delegate operations
This is the warning. Camp handed Kalanick operations, which was wise. He also, in effect, handed him the board, which was not. Camp was chairman. He held super-voting shares — the insiders' stock carried ten times the votes of new investors'. Yet he described himself as "really just more of an adviser" and said it wasn't his "place to comment on something unless you're well-informed." Former colleagues told Bloomberg his most severe lapse was "allowing Kalanick's power to go unchecked," and that at pivotal moments he was "indecisive or unreachable." When the board met in 2017 to discuss Kalanick's leave, Camp was abroad. Kalanick, for his part, told Apple's executives that "the board is irrelevant."
A chairman holding super-voting stock is not an adviser. He is one of the few people in the structure who can say no. The voting power Camp shared with Kalanick and Graves was designed to protect founders from interfering investors — and it worked, which meant that the only checks remaining on the CEO were the founders themselves. Camp's own Medium post supplied the diagnosis: "We have failed to build some of the systems that every company needs to scale successfully." The board is one of those systems, and governance is the one responsibility a founder cannot hand to the operator, because the operator is the thing being governed.
Tactic: If you hold control rights in a company you don't run, put three things in writing — what behaviors would make you act, how you'll stay informed enough to notice them, and the date you'll next review both.
Principle 8
In a crisis, only specificity counts as speech
Camp's June 2017 post said Uber had "neglected parts of our culture," failed to "build some of the systems," and needed to "listen better." It did not name the CEO, the harassment, the investigation, or the women who had come forward. It was received as "mild and meaningless" and "tone-deaf." His August memo, by contrast, said one concrete thing — "Travis is not returning as CEO" — and it landed, because it was a commitment rather than a mood.
The asymmetry is instructive. A person who rarely speaks publicly has a reservoir of credibility, and the first words out of it get weighed heavily. Abstraction spends that credibility on nothing. "Growing pains" is a category; "a manager propositioned an engineer and HR protected him" is a fact. In a crisis, audiences read vagueness as either ignorance or evasion, and neither is survivable for a chairman. The InfoTable compares the two statements.
M
Two Statements, Summer 2017
Same author, same crisis, opposite reception.
| Dimension | "Uber's path forward" (June 20) | Memo to employees (early August) |
|---|---|---|
| Named the CEO? | No | Yes — "Travis is not returning as CEO" |
| Named the problem? | "Growing pains," "neglected parts of our culture" | "The CEO search and confusion around the process" |
| Made a commitment? | Values to be "updated" | Hire "a new world-class CEO" |
| Reception | Called tone-deaf | Settled a rumor |
Tactic: Before publishing any crisis statement, delete every sentence that would be equally true of another company in another crisis, and publish only if something remains.
Principle 9
Keep a mistake ledger
"Every time I make a mistake with a company, I write it out and try to figure out why it happened," Camp told TechCrunch in 2013 — and he does it for the things that go right, too. Over a decade this became, in his phrase, "a good set of guidelines," which he built Expa to apply. Nine years later he mentioned in passing that he had about 3,500 notes on company-building in iCloud.
This is the discipline of an engineer applied to a career. It explains how the same person could build a web toolbar in 2001, a dispatch system in 2009, and a fintech portfolio in 2020: the domain changes, the ledger carries over. It also explains Camp's habit of evaluating companies by tallying decisions — "Uber definitely got most of them right... there's a few that they got wrong, but none of them are fatal." That last judgment, made in 2013, is a useful caution about the ledger's limits: a mistake that isn't fatal at year four can still be the one that nearly ends you at year eight.
Tactic: After every significant decision, write three sentences — what you decided, what you expected, and what happened — and reread the whole file once a quarter, looking for the mistake you keep repeating.
Principle 10
Run ideas as a portfolio, not a vocation
Camp sketched Uber in a notebook while running StumbleUpon, which Scott Belsky found faintly scandalous. He handed Uber off while keeping StumbleUpon, then stepped away from StumbleUpon to build Expa. He arrived at Expa with "10 ideas" and the expectation that "realistically only three or four will really come to fruition." He considers about 10 percent of his 3,500 notes to be ideas, "a solid 10" of them potentially big.
The conventional advice — focus on one company — is right for operators and wrong for people like Camp, whose comparative advantage is the generation and early shaping of ideas rather than their execution at scale. Expa formalized this: it's a studio rather than a fund, because Camp wanted to put capital "into something where I can be involved in the branding and design" instead of writing angel checks. The portfolio approach also takes the pressure off any single idea, which may be why he could abandon BlackJet and StumbleUpon without visible agony. Knowing which kind of founder you are determines which advice applies.
Tactic: If you consistently have more good ideas than you can execute, stop treating that as a focus problem and build a structure — a studio, a co-founder pipeline, a partner program — that lets other people operate the ideas you originate.
Principle 11
Buy back what you understand
Camp sold StumbleUpon in 2007 because the price and terms were right and integration wasn't required. He bought it back in 2009 when corporate procedure had slowed it down. He took majority control again in 2015, when it needed "difficult changes," and in 2018 he shut it down himself and moved its users to Mix. Each move was made with information no outside buyer had: he knew the product, the team, and the thesis better than anyone alive.
There is a sentimental reading of this ("he just can't quit SU"), and that may be partly true. But there's a sound strategic one as well. An exit is not necessarily permanent, and the person who built something is often the best-informed buyer when its new owner loses interest. The 2009 spin-out happened at a time when eBay was shedding non-core assets; Camp's re-entry put the company back in the hands of someone who still believed in the original question. That StumbleUpon eventually closed doesn't make the buybacks errors. It makes them a founder's prerogative to decide how his own experiment ends.
Tactic: When you sell a company, negotiate a right of first refusal on any future divestiture, and keep a relationship with the acquirer's corporate development team for at least three years.
Principle 12
Carry your design language into your giving
Camp's Giving Pledge letter is, among other things, a product brief. He wanted to spend time "on projects where hands-on systems design can create more efficient solutions." He resolved to research effective philanthropy before deploying capital. The result, Camp.org, backs Every.org, which removes the friction from how nonprofits receive stock, crypto and other complex gifts, and has built tools against deepfakes — infrastructure, in other words, rather than single grants.
The principle generalizes beyond philanthropy: whatever problem you turn to next, bring the method that made you good at the last one. For Camp that method is finding the uncoordinated system and adding the coordination layer. It worked for empty cabs. It works plausibly for donations, which are full of forms, delays and dropped intentions. The method is the asset; the domain is replaceable.
Tactic: Before you write your first significant charitable check, spend a month identifying the friction in how money currently reaches the cause you care about — then fund the tool that removes it.
Free playbook
Get The Business Model Playbook
58 business models, one visual page each: how the money flows, the metrics that matter, and who runs it. Free when you join the Faster Than Normal email.
Free. No spam. Unsubscribe anytime.
Part IIIQuotes and Maxims
In their words
— Garrett Camp, on StumbleUponGoogle is cool if you know what you want. But what if you don't know what you want?
— Garrett Camp, University of Calgary, November 17, 2015Fundamentally, I want to build a product that I want to use myself. Will it make my life better and will other people want to use it? If I don't want to use it myself, then I'm not going to be excited about it, and then work becomes work and it's not fun.
— Garrett Camp, to TechCrunch, May 2013Every time I make a mistake with a company, I write it out and try to figure out why it happened.
— Garrett Camp, 'Uber's path forward,' June 20, 2017In a highly competitive market it is easy to become obsessed with growth, instead of taking the time to ensure you're on the right path.
— Garrett Camp, Giving Pledge letter, November 2017Philanthropy isn't just about donating money, but also sharing your advice or spending time solving important problems.
For the founding mythology in its most breathless form, there is The Founding of Uber and The Real Story Behind the Founders Travis Kalanick and Garrett Camp, which promises to reveal "the secrets that will make you a Billionaire" — a promise Camp himself, who once described his career as "strategically hacking away at something for a year," has never made. The fuller accounts remain Mike Isaac's Super Pumped and Brad Stone's The Upstarts.
Maxims
- Debug your annoyances. A frustration you can decompose into mechanical causes is a product spec in disguise.
- One button, one visible wait. Remove every step but one, then make the progress of that step something the user can watch.
- Traction before terms. Six hundred thousand users with no funding buys a two-week round and your choice of investors.
- Your users are a cap table. The person who loves the product before investing will believe in it longer than the person who read the deck.
- Start as the private driver. Premium positioning gives you room to descend; mass positioning rarely lets you climb.
- Give the operator real ownership. If someone else is better suited to run your company, make them an owner early, while equity is cheap.
- A chairman is not an adviser. Control rights you don't exercise are responsibilities you've abandoned, not ones you've delegated.
- Name the thing. In a crisis, a statement that could describe any company describes nobody, least of all you.
- Write down the mistake. A decade of logged errors becomes a playbook; a decade of forgotten ones becomes a pattern.
- Know which kind of founder you are. Inventors should build portfolios of ideas; operators should build one company — and both should stop taking each other's advice.
