The Uber Achievers

In 1997, the computer scientist Niklaus Wirth, best known as the creator of Pascal, conducted a fascinating interview with the magazine Software Development, which I’ve quoted here before. When asked if it would be better to design programming languages with “human issues” in mind, Wirth replied:

Software development is technical activity conducted by human beings. It is no secret that human beings suffer from imperfection, limited reliability, and impatience—among other things. Add to it that they have become demanding, which leads to the request for rapid, high performance in return for the requested high salaries. Work under constant time pressure, however, results in unsatisfactory, faulty products.

When I read this quotation now, I think of Uber. As a recent story by Caroline O’Donovan and Priya Anand of Buzzfeed makes clear, the company that seems to have alienated just about everyone in the world didn’t draw the line at its own staff: “Working seven days a week, sometimes until 1 or 2 a.m., was considered normal, said one employee. Another recalled her manager telling her that spending seventy to eighty hours a week in the office was simply ‘how Uber works.’ Someone else recalled working eighty to one hundred hours a week.” One engineer, who is now in therapy, recalled: “It’s pretty clear that giving that much of yourself to any one thing is not healthy. There were days where I’d wake up, shower, go to work, work until midnight or so, get a free ride home, sleep six hours, and go back to work. And I’d do that for a whole week.”

“I feel so broken and dead,” one employee concluded. But while Uber’s internal culture was undoubtedly bad for morale, it might seem hard at first to make the case that the result was an “unsatisfactory, faulty” product. As a source quoted in the article notes, stress at the company led to occasional errors: “If you’ve been woken up at 3 a.m. for the last five days, and you’re only sleeping three to four hours a day, and you make a mistake, how much at fault are you, really?” Yet the Uber app itself is undeniably elegant and reliable, and the service that it provided is astonishingly useful—if it weren’t, we probably wouldn’t even be talking about it now. When we look at what else Wirth says, though, the picture becomes more complicated. All italics in the following are mine:

Generally, the hope is that corrections will not only be easy, because software is immaterial, but that the customers will be willing to share the cost. We know of much better ways to design software than is common practice, but they are rarely followed. I know of a particular, very large software producer that explicitly assumes that design takes twenty percent of developers’ time, and debugging takes eighty percent. Although internal advocates of an eighty percent design time versus twenty percent debugging time have not only proven that their ratio is realistic, but also that it would improve the company’s tarnished image. Why, then, is the twenty-percent design time approach preferred? Because with twenty-percent design time your product is on the market earlier than that of a competitor consuming eighty-percent design time. And surveys show that the customer at large considers a shaky but early product as more attractive than a later product, even if it is stable and mature.

This description applies perfectly to Uber, as long as we remember that its “product” isn’t bounded by its app alone, but extends to its impact on drivers, employees, competitors, and the larger community in which it exists—or what an economist would call its externalities. Taken as a closed system, the Uber experience is perfect, but only because it pushes its problems outside the bounds of the ride itself. When you look at the long list of individuals and groups that its policies have harmed, you discern the outlines of its true product, which can be described as the system of interactions between the Uber app and the world. You could say this of most kinds of software, but it’s particularly stark for a service that is tied to the problem of physically moving its customers from one point to another on the earth’s surface. By that standard, “shaky but early” describes Uber beautifully. It certainly isn’t “stable and mature.” The company expanded to monstrous proportions before basic logistical, political, and legal matters had been resolved, and it acted as if it could simply bull its way through any obstacles. (Its core values, let’s not forget, included “stepping on toes” and “principled confrontation.”) Up to a point, it worked, but something had to give, and economic logic dictated that the stress fall on the human factor, which was presumably resilient enough to absorb punishment from the design and technology sides. One of the most striking quotes in the Buzzfeed article comes from Uber’s chief human resources officer: “Many employees are very tired from working very, very hard as the company grew. Resources were tight and the growth was such that we could never hire sufficiently, quickly enough, in order to keep up with the growth.” To assert that “resources were tight” at the most valuable startup on the planet seems like a contradiction in terms, and it would be more accurate to say that Uber decided to channel massive amounts of capital in certain directions while neglecting those that it cynically thought could take it.

But it was also right, until it wasn’t. Human beings are extraordinarily resilient, as long as you can convince them to push themselves past the limits of their ability, or at least to do work at rates that you can afford. In the end, they burn out, but there are ways to postpone that moment or render it irrelevant. When it came to its drivers, Uber benefited from a huge pool of potential contractors, which made turnover a statistical, rather than an individual, problem. With its corporate staff and engineers, there was always the power of money, in the form of equity in the company, to persuade people to stay long past the point where they would have otherwise quit. The firm gambled that it would lure in plenty of qualified hires willing to trade away their twenties for the possibility of future wealth, and it did. (As the Buzzfeed article reveals, Uber seems to have approached compensation for its contractors and employees in basically the same way: “Uber acknowledges that it pays less than some of its top competitors for talent…The Information reported that Uber uses an algorithm to estimate the lowest possible compensation employees will take in order to keep labor costs down.”) When the whole system finally failed, it collapsed spectacularly, and it might help to think of Uber’s implosion, which unfolded over less than six months, as a software crash, with bugs that were ignored or patched cascading in a chain reaction that brings down the entire program. And the underlying factor wasn’t just a poisonous corporate culture or the personality of its founder, but the sensibility that Wirth identified two decades ago, as a company rushed to get a flawed idea to market on the assumption that consumers—or society as a whole—would bear the costs of correcting it. As Wirth asks: “Who is to blame for this state of affairs? The programmer turned hacker; the manager under time pressure; the business man compelled to extol profit wherever possible; or the customer believing in promised miracles?”