In Shanghai, an Uber driver logs into an online forum. He is not looking for a passenger. He is looking for someone to pretend to be a passenger.
He finds a willing taker. He pretends to collect the customer and drop her at the airport. In fact, they never meet. Then he goes online and sends her the agreed fee: about $1.60 (£1.20).
Or perhaps the driver goes a step further, making up the other person as well as the journey. He goes to the online marketplace Taobao, and buys a hacked smartphone.
That lets him create multiple fake identities. He uses one to arrange a ride with himself.
Why? Because someone's willing to pay him to give people lifts in his car.
Rational economic incentives
Investors in Uber and their rivals have run up billions of dollars of losses - in China, and elsewhere - paying people to share car journeys.
Naturally, they're trying to stamp out the imaginary journeys, but they're convinced subsidising genuine rides is a smart idea.
This all seems bizarre, perverse, even. But everyone involved is rationally pursuing economic incentives.
To see what's going on, we have to understand a phenomenon that's spawned many buzzwords: "crowd-based capitalism", "collaborative consumption", "the sharing economy", "the trust economy".
50 Things That Made the Modern Economy highlights the inventions, ideas and innovations which have helped create the economic world in which we live.
Suppose I'm about to drive myself from downtown Shanghai to the airport. I occupy only one seat in my car.
Now suppose you live a block away, and you also need to catch a flight. Why don't I give you a lift? You could pay me a modest sum, less than you'd pay for some other mode of transport. You're better off. I'm better off.
There are two big reasons why this might not happen. The first, and most obvious, is if neither of us knows the other exists.
Until recently, the only way you could advertise your desire for a lift would be to stand at a junction, holding up a sign saying "airport". It's not very practical - especially since the plane won't wait.
This function of matching people who have coincidental wants is among the most powerful ways the internet is reshaping the economy.
Traditional markets work perfectly well for some goods and services, but they're less useful when the goods and services are urgent or obscure.
Consider the plight of Mark Fraser.
It was 1995. Mark Fraser gave lots of presentations, and he really wanted a laser pointer - they were new, and cool, but also forbiddingly expensive.
Fraser, however, was an electronics geek. He was confident that if he could get his hands on a broken laser pointer, he could fix it.
But where on earth would he find a broken laser pointer? The answer, now, is obvious - try Taobao, or eBay, or some other online marketplace.
Back then, eBay had only just started. Its very first sale: Mark Fraser bought a broken laser pointer.
Mark Fraser was taking a bit of a risk. He didn't know the seller. He simply had to trust that they wouldn't simply pocket his $14.83 (£11.20) then disappear. For other transactions, the stakes are higher.
That's the second reason I might not give you a lift to Shanghai airport. I've no idea who you are. Perhaps you're planning to attack me and steal my car? You might doubt my motives, too - perhaps I'm a serial killer.
Enablers of trust
After all, hitch-hiking was a popular pursuit a few decades ago, but fell out of fashion after some well-publicised murders.
Trust is an essential component of markets - so essential that we often don't even notice it, like a fish doesn't notice water.
In developed economies, enablers of trust are everywhere: brands, money-back guarantees, and of course repeat transactions with a seller who can be easily located.
But the new sharing economy lacks those enablers. Why should we get into a stranger's car - or buy a stranger's laser pointer?
In 1997, eBay introduced a feature that helped solve the problem: Seller Feedback.
According to Jim Griffith, eBay's first customer service representative, "no-one had ever seen anything like [it]".
The idea of both parties rating each other after a transaction has become ubiquitous.
You buy something online - you rate the seller, the seller rates you. You use a lift-sharing service, like Uber - you rate the driver, the driver rates you. You stay in an AirBnB - you rate the host, the host rates you.
A few positive reviews set our mind at ease about a stranger. Jim Griffith says he's "not so sure [eBay] would have grown without [seller feedback]".
Online matching platforms would still exist, of course - eBay already did - but perhaps they'd be more like hitch-hiking today: a niche pursuit for the unusually adventurous, not a mainstream activity that's transforming whole sectors of the economy.
Platforms like Uber and AirBnB, eBay and TaskRabbit create real value.
But there are losers.
For all the touchy-feeliness of the buzzwords - "collaborative", "sharing", "trust" - these models aren't just heart-warming stories of neighbours coming together to borrow each other's power drills.
They can easily lead to cut-throat capitalism.
More from Tim Harford:
Established hotels and taxi companies are aghast at competition from AirBnB and Uber.
Is that just an incumbent trying to suppress competition? Or are they right when they complain that the new platforms are ignoring important regulations?
Risk of discrimination?
Many countries have rules to protect workers, like guaranteed hours or working conditions or a minimum wage. And many people on platforms like Uber aren't just monetising spare capacity.
They're trying to make a living, without the protections of a formal job - perhaps because a company liked Uber competed them out of a job.
Some regulations protect customers, too - for example, from discrimination. Hotels can't legally refuse you a room if you're, say, a same-sex couple. But hosts on AirBnb can choose to turn down guests after seeing not just your feedback but your photos.
AirBnb builds trust by emphasising the personal connection, and that means showing people prominent pictures of who they're dealing with.
That also invites people to act on their personal prejudices, consciously or otherwise. People from ethnic minorities have been shown to suffer as a result.
What should be the 51st Thing?
Tim Harford has discussed 50 things which he argues have made the modern economy. Help choose the 51st thing by voting for one of these listener suggestions:
- The credit card
- Global Positioning System (GPS)
- The pencil
- The spreadsheet
You can vote on the 50 Things That Made the Modern Economy programme website. Voting closes at 1200 GMT Friday 6 October 2017, and the winning 51st Thing will be announced in a special podcast on 28 October 2017.
How online matching platforms should be regulated is a dilemma causing lawmakers around the world to scratch their heads.
It matters because it's potentially huge business, especially in emerging markets where there isn't yet a culture of owning things like cars.
And it's a business with network effects: the more people use a platform, the more attractive it becomes.
That's why Uber and its rivals - Didi Chuxing in China, Grab in southeast Asia, Ola in India - have invested massively in subsidising rides and giving credits to new customers: they wanted to get big first.
And, naturally, some drivers have been tempted to defraud them.
Remember how they did it? By using an online forum to find a willing fake customer, or an online marketplace to buy a hacked smartphone. Matching people with the particular things they want really is useful.
Tim Harford writes the Financial Times's Undercover Economist column. 50 Things That Made the Modern Economy is broadcast on the BBC World Service. You can find more information about the programme's sources and listen online or subscribe to the programme podcast.