Tesla, The New York Times And The Truth About 'Truth' In Data

Elon Musk, CEO of Tesla Motors, reacts to a reporter's question following the electric automaker’s initial public offering on
Elon Musk, CEO of Tesla Motors, reacts to a reporter's question following the electric automaker’s initial public offering on Nasdaq, Tuesday, June, 29, 2010 in New York. (AP Photo/Mark Lennihan)

elon musk tesla motors new york times Tesla CEO Elon Musk.

The fierce dispute between The New York Times and the CEO of Tesla Motors over the merits of its electric car is remarkable for what it says about how we argue in a time of seemingly limitless information.

Tesla co-founder Elon Musk and New York Times writer John Broder are not arguing over interpretations or subjective characterization about the experience of the new car. Rather, they are arguing over basic facts -- things that are supposed to be easily established in an age when cameras, sensors and instantly searchable databases are ubiquitous.

After Broder wrote a critical review of Tesla's Model S, Musk declared the article a "fake" and promised data from the vehicle's logs would "tell [the] true story" about the mischaracterized mishaps on Broder's route.

But since then, Musk and Broder have each produced their own measurements, taken from the car and reporter's notebook, respectively, and still no "true story" has emerged. The conflicting accounts -- both quantified, both recorded in detail -- are as much a lesson in the perils of drawing conclusions from high-tech tracking as they are a guide to the optimal conditions for electric cars.

Data is supposed to be the authoritative alternative to selective anecdotal recollection, though more data seems in some cases to only make our disagreements more heated, with every party able to marshal a seemingly stronger and tailored case. In Tesla-gate, Big Data hasn't made good on its promise to deliver a Big Truth. It's only fueled a Big Fight.

The problem is that the data shows what happened, but not why, argued David Weinberger, author of Too Big to Know and a senior researcher at Harvard University's Berkman Center for Internet and Society. While the stats and facts capture what occurred inside the car, they tell us precious little about the people who made those things happen and why they behaved the way they did.

Observers aren't interested in the Tesla tiff because they care deeply about whether Broder turned down the heat when he said he did, or set cruise control when he claims, explained Weinberger. Rather, the data is useful to us only as it helps us shape a clear narrative from the data -- the story of the smug CEO attacking an innocent journalist, or the tale of the corrupt reporter trying to take down an innocent entrepreneur. Yet hard numbers capture action, not intention, and the conflicting accounts offer no more insight into the "true story" onlookers want to assemble.

"The only reason we care about the data is because it fits into a story that was interesting for reasons that have nothing to do with the data itself," Weinberger said. "We're not trying to understand the data or even what happened. We're trying to understand what the humans in the story were doing. Was Broder driving around to drain the battery, as Musk says? Or was he circling around trying to find the badly-marked power source the way Broder says? What we want to know is human intentionality, and data doesn't settle that question."

At least not yet.

Smaller sensors, more astute algorithms, dwindling data-storage costs and more detailed readings on our every move are leading us to a world where everything is tracked and analyzed. Machines are increasingly adept at capturing not only what we did, but more precisely, what we meant by it.

Massachusetts-based startup Affectiva has developed emotion-recognition software that allows the company to track audience reaction to different stimuli -- advertising, in particular. Solariat, which applies artificial intelligence to social media marketing, can pick up on people's intentions based on what they share to a social network. We're also closer than ever to being able to turn back the clock to reconstruct any experience and analyze what really took place. Rick Smolan, author of The Human Face of Big Data, predicted we're not far from a world where every conversation will be recorded and indexed: "For almost every moment," Smolan noted in an interview last fall, "it will start being like there really is a time machine where you can step back in time to any moment."

But for the time being, we're stuck dealing -- and arguing over -- data that provides an incomplete view of human behavior.

"Facts and data do a much better job in domains, such as science, where there is no human intentionality," noted Weinberger. "As soon as you involve a human doing things for some purpose, data is helpful, but not conclusive."