There's a specific kind of lie that tech bosses tell: the "premature truth". That's a lie that they excuse by saying it'll come true soon.
Uber has always lied about everything, from driver pay to the company's finances.
Now, lying about driver pay is relatively low-risk for a company like Uber. After all, the company created the app-based worker misclassification playbook, designed to make vulnerable and precarious workers even more vulnerable and precarious. The lies Uber tells about its workforce only have to survive contact with credulous municipal governments so that they keep up a supply of get-out-of-jail-free cards for Uber's many labour, licensing and consumer rights violations.
But lying about your finances is another matter altogether. When Uber lies about its finances, it is lying to rich people, and (unlike workers) rich people have many ways to punish people who lie to them. Despite this, Uber spent more than a decade lying flagrantly and lavishly about its balance sheet.
The most shameless financial lie Uber told is that it was profitable due to the increased valuation of the many ride-hailing companies around the world that Uber owns large shares in. How did Uber come to own large shares in these companies? It happened after Uber's competing ride-hailing companies in countries around the world failed.
Uber's subsidiaries lost billions of dollars, and eventually Uber sold them to its (also money-losing) domestic rivals in all-stock deals that saw Uber acquiring vast amounts of worthless, illiquid shares in the world's failing ride-hailing companies. Then, for years, Uber would announce that its stake in these companies had appreciated in value. These revaluations were purely vibes-based. Uber's former rivals were privately held, and while small volumes of shares traded hands in private deals, Uber's pronouncements that, say, the Chinese company DiDi was now worth vastly more than it had been worth a few months before was sheer fantasy dressed up as profits.
Uber did this over and over (and over and over), relentlessly dogged by Hubert Horan, an aviation analyst who became the world's foremost close-reader of Uber's financial statements.
Despite these lies, Uber eventually found a way for its early investors to cash out, through a very successful IPO that saw the House of Saud made whole (and then some) on the $3.5bn they ploughed into the business so that it could lose $31bn in the preceding 13 years.
The company did achieve actual profitability after its IPO. It got there through the simple expedient of paying drivers less and charging riders more. The company spies on its drivers to calculate the lowest wage they'll accept, based on their inferred economic desperation, a process that Veena Dubal calls "algorithmic wage discrimination".
They also use surveillance data about riders to calculate the highest price you'll pay for a journey. Investors call this "personalised pricing", while everyone else calls it "surveillance pricing". It's a way for Uber to reach into your bank account and reduce the value of your dollars, by charging you more than the next person would pay for the same service.
This is something AI turns out to be extremely good at. Businesses can program chatbots to automatically segment their customers and employees based on surveillance data and then perform a series of automated experiments to find the highest prices and lowest wages they'll accept.
When it comes to wage theft and price gouging, Uber leads the pack. Of course, it helps that they're willing to lie like crazy. Uber routinely claims that it pays its drivers hourly rates that are far in excess of minimum wage, but it helpfully omits that drivers only get paid while they have a passenger in the car, and that the company punishes drivers who turn down too many unprofitable jobs.
Drivers only get paid for a fraction of the time they spend working for Uber, and on top of that, they have to pay for fuel, insurance, depreciation, loan interest, and the cost of getting vomit out of their back seats. Add it all up and Uber drivers make peanuts. In Toronto, Uber pay nets out to something like US$6.37 (£4.80) an hour.
Uber can only be profitable for so long as drivers continue to subsidise its business, and so long as riders are willing to pay more than they ever did under the traditional taxi system (much of which was seen off by Uber's Saudi-subsidised decade of predatory pricing, which also induced a lost decade in global transit investment).
Inflation (supercharged by AI pricing) has hit all workers, but it's been especially hard on Uber drivers, who didn't have any financial cushion to start with. Eventually, drivers were bound to turn their backs on Uber. "Eventually" is now, according to Uber's third-quarter earnings call, where chief executive Dara Khosrowshahi told investors that "driver supply [was] not keeping pace with strong consumer demand in several international markets".

I can't think of any way in which replacing starvation-wage drivers who provide their own cars with $200,000 robotaxis that Uber must operate, maintain and insure will work out better
As The Gig Economy Project writes, the driver shortage is especially sharp in India and Brazil, two gigantic countries where (according to Uber’s chief financial officer, Balaji Krishnamurthy), there is too much consumer "price sensitivity".
Uber is also having trouble getting drivers to accept the rock-bottom wages they're offering for food delivery. Despite this, Krishnamurthy assured investors that the company's measures to "improve driver availability" wouldn't affect their bottom line. Instead, Krishnamurthy promised to raise prices on riders elsewhere, something he called "rebalancing incentives":
“That is part of the art that we’ve got and then there is a lot of algorithmic science that goes into, if you have a dollar, what is the maximum predicted return for that dollar and should I invest that dollar in a driver in Brazil or should I invest that dollar in a rider in India, for example. The algorithm can structurally make these choices based on training on billions of trips and allocate these investments in the way that humans just can’t.”
The Gig Economy Project goes on to say that "robotaxis will fundamentally alter the unit economics of taxis, but we are still some time away from the point where robotaxis are price competitive with human-driven vehicles”.
Boy is this an understatement. Remember, Uber is already paying drivers sub-minimum, sub-starvation wages, which drivers must spend to cover all the capital, depreciation, maintenance and insurance costs, along with (skyrocketing) fuel costs. I can't think of any way in which replacing starvation-wage drivers who provide their own cars with $200,000 robotaxis that the company must operate, maintain and insure will work out better for the company than the current situation.
This is as good as it gets for Uber, in other words. Being able to tap an army of desperate workers who supply 99% of its business's capital and variable costs out of their own pockets is vastly superior to operating a fleet of expensive robots that Uber has to purchase and service.
Profitable "driverless" cars have been around the corner for more than 15 years now. Waymo claims to have delivered this through the brute-force expedient of using the most expensive sensors and vast amounts of incredibly expensive cloud computing, but their balance sheet remains stubbornly opaque. We don't know how often a "driverless" Waymo is just a remote control car being piloted by a low-wage worker in an overseas call centre (there's a reason Indian technologists joke that AI stands for "Absent Indians").
We also don't know how often a Waymo is being overseen by a high-waged engineer. What we do know is that when General Motors was forced to reveal this number for its driverless car company, Cruise, it turned out that they'd replaced a single, low-waged cab driver with 1.5 engineers, each of whom earned an order of magnitude more than the cabbie they'd displaced.
Uber's claims that they were profitable (based on accounting gimmicks) were lies, but they were also premature truths: the company did become profitable (through price-gouging and wage theft). But there's a limit to how low you can push your workforce's wages, and even if the AI crash delivers a fresh cohort of desperate workers who subsidise Uber's operations for a few more years, they, too, will eventually find that they can't live on $6 an hour.
When that day comes, robotaxis won't save the business model. Robots may cost less per hour than Uber drivers (though not by much), but owning a robot is always going to be more expensive than getting drivers to supply free cars to Uber. If Uber can't sustain profitability with that massive subsidy, it is going to hit a wall, and it won't be pretty.
This is an edited version of a Cory Doctorow post from pluralistic.net. It is published under a CC BY 4.0 creative commons licence
Cory Doctorow, who was born in Toronto and now lives in Los Angeles and London, is the Nerve’s tech columnist. His new book The Reverse Centaur's Guide to Life After AI is published by Verso

