Humans are a pesky problem for Uber—one that it deals with on a massive scale.

“We’re in the hundreds of thousands of partners connected to our system,” Uber CEO Travis Kalanick said at the TechCrunch Disrupt conference in San Francisco Monday morning. The transportation company is adding “tens of thousands” every month, he said. 

In May, Kalanick waxed effusive about the potential for self-driving cars to cut his costs. After all, the biggest expense his transportation company’s customers pay for is “the other dude in the car.” (The “other dude” is also known as “the driver.”)

In fact, Uber is desperate to get more dudes. A Verge report recently revealed that Uber is paying its recruiters up to $750 per driver poached from rival companies like Lyft and Sidecar. The demand for cheap transportation is there. The supply of drivers, not so much. 

That’s not a technology problem. It’s an economic problem: how to charge passengers enough to court drivers, while keeping rides cheap enough to attract passengers. At the same time, Uber must fend off competitors who are doing pretty much the exact same thing.

“You do something good for the drivers and the riders are upset, you do something good for the riders and the drivers are upset,” Kalanick observed.

(Uber might cost-effectively expand its labor pool by having its CEbro refer to its drivers as “people” and doing other things to court women behind the wheel. There aren’t any official statistics, but many people have noted that Uber drivers are predominantly male.)

Uber likes to talk about the sophisticated software it uses to match riders and drivers. But the ground war with its rivals is more spy-tech than high-tech. Burner phones, throwaway credit cards, and chat rooms to discuss which Lyft drivers to target for recruiting are all part of Uber’s strategy. In August, the company hired one of President Barack Obama’s political strategists, David Plouffe, as its “campaign manager.” His mission: Lobby for Uber-friendlier regulations in the U.S. and abroad.

UberPool—which lets multiple riders share a car—is one example of where technology could help Uber. (Despite the vogue for calling Uber and Lyft “ride-sharing” services, until very recently, they’ve only offered rides for single passengers or groups riding together.) Matching rides with separate passengers is a very complex, real-time problem.

The bigger problem with making UberPool work is simply having enough riders going in the same direction, an issue Kalanick calls “liquidity.” That’s supply and demand—sophisticated pricing and clever marketing—more than technology.

When Uber starts using driverless cars, the game changes. Soon after Kalanick uttered his quip about getting rid of “the other dude in the car” in May, he backtracked, with Uber spokespeople offering statements about how self-driving cars are decades away. (In fact, Google, which has invested hundreds of millions of dollars in Uber, has a small fleet of robot cars testing roads today.)

So Uber faces a contradiction: Ultimately, it sees its destiny as getting rid of drivers and replacing them with technology. Between now and then, the company will need hundreds of thousands, if not millions, of drivers to trust it with their livelihoods. And it will constantly be experimenting with those drivers’ paychecks, testing new products like UberPool—which might earn drivers more money, or it might just let Uber lower the rates it charges passengers while keeping driver net pay the same.

Along the way, we’ll learn whether Uber is a technology company—or just a very large marketing operation that happens to supply iPhones to drivers.

Photo by Owen Thomas for ReadWrite