The 'Mechanical Turk' provides the illusion of technological advance, but humans are still doing the work.
(image by mandiberg) DMCA
Data gathering and crunching by glorified cab company Uber has been back in the news lately in a big way, and not in a good way. A senior executive at Uber notoriously suggested that Uber use this data to conduct "oppo research" on journalists who are critical of the company; in the midst of the ensuing hullaballoo, revelations emerged that Uber management had misused such data on multiple occasions, spooking at least one journalist with a comment that "I was tracking you," on Uber's "God View" screen.
One question has been little raised in the controversy since: exactly what use is the data analytics that Uber engages in? I mean, besides intimidating reporters and creating ethically questionable blog posts about the one night stands of passengers?
The answer, according to Uber, is that this data manipulation is central to Uber's ability to deliver cars to passengers quickly and reliably. Impressive algorithms, real-time tracking, and "Big Data" analytics are what makes Uber different from all those other, old-fashioned taxi dispatching services. This, claims Uber, is what makes them key to the future of urban logistics, and justifies their insane-sounding valuation of up to $30 billion.
The problem? It's just not true. The data tricks, in truth, do little or nothing.
Uber is really nothing more than a mechanical Turk.
The original "mechanical Turk" was an 18th Century scam of Uber-esque proportions -- a mechanical figure, dressed as a "Turk," sitting at a table and playing chess against live opponents. The public was amazed as this apparent automaton won game after game. How was it possible? What miracle of modern technology was this?
The secret, however, was that there was no miracle. Instead, a chess master hid inside the cabinet, operating the controls of the "automaton." This was not advanced technology, only the illusion.
You may have heard of Amazon's "Mechanical Turk," the crowdsourcing website that calls itself "a marketplace for work that requires human intelligence." Amazon's Mechanical Turk farms out tiny increments of work, tedious to humans but too complex for computers, which are largely taken on by a desperate, third world workforce willing to work for pennies at a time.
That, at heart, is Uber's model as well. How do Uber's cars appear in the neighborhood when you want them? The drivers are already waiting there. Uber relies, not on data analysis, but on the willingness of an unsustainably large and underpaid driver workforce to spend hours waiting for business, hoping to make the payments on the vehicle they bought through Uber's predatory lending partnerships.
It's not that the algorithm team do absolutely nothing. For one thing, they are always trying to find new ways to entice their underpaid workforce to spend more hours on the road, for instance through bonuses promised to drivers who stay online, waiting for calls, for as many as 400 hours a month (do the math on that!)
Beyond that, Uber's data analytics team really contributes little except for deciding when to turn on the ever-unpopular "surge pricing" to "lure" drivers to busy areas of town, and designing a computer simulated "Uberg" city to model which way drivers should drive after dropping off business. Laughably, none of this does anything more than recreate behaviors and knowledge which experienced taxi drivers already possess.
So -- assuming you haven't joined the #deleteuber movement yet -- next time you take a ride with them congratulate your driver on his game. He is, after all, the man hidden in the box: the chessmaster who operates Uber's illusionary mechanical Turk.
(Article changed on November 24, 2014 at 18:13)
(Article changed on November 24, 2014 at 19:18)