Theranos Diagnosed With Apple Disease


Steve Jurvetson | Flickr

The tragedy of Theranos. As the media’s attention today turns to the latest iPhone unveiling, here’s a morality tale that suggests the perils in over-imitating Apple. A Wall Street Journal investigation was the proximate cause of the downfall of med-tech startup Theranos, which once promised to revolutionize the blood testing business. But a new look at the story of the company’s implosion (Vanity Fair) suggests that Theranos founder Elizabeth Holmes built the firm on sand from the very start by choosing a path of secrecy rather than transparency. Holmes’ emulation of Steve Jobs extended to more than the superficial shared preference for black turtlenecks; like Jobs, she applied control-freak tactics to prevent rivals, neutral third parties, and even Theranos employees themselves from obtaining information about the company’s products. That meant that, when the Journal finally raised questions about its work, Theranos had no defenders. If Theranos was as bogus as it now appears, maybe there was never much to defend. Either way, the lesson for the rest of us couldn’t be clearer: Don’t trust anyone’s revolution unless the data is shared, the science is reviewed, and the conversation is open.

The Clinton campaign’s strength in numbers. If this year’s two presidential campaigns were startups, Trump’s would be the one that prefers winging it to wonking out. Clinton’s, meanwhile, would be the one that lives and dies by data. A profile of her campaign’s director of analytics, Elan Kriegel (Politico), shows how thoroughly statistics drive the Democratic candidate’s efforts — the timing of email campaigns, the houses that on-the-ground volunteers visit, the targeting of postal mailers and online advertising and TV campaigns. Trump famously scorns deep-diving into data, while Clinton has organized her entire effort around a “culture of testing.” In a few weeks we’ll know which candidate bet right. (If you’re betting against data smarts, though, hope you’re wealthy enough to take the loss.)

When data models go wrong. Putting your faith in numbers only makes sense if you trust how those numbers were derived. In a new book, Weapons of Math Destruction (Boing Boing), Cathy O’Neil looks at the dark side of statistical prediction. Mathematical models shape more and more of our business choices and public-policy decisions. But these models are only as good as the assumptions that drive them and the quality of the sample data that trains them. Models are “opinions embedded in mathematics.” To use them responsibly requires that we shed our reverence for the authority of numbers and subject them to rigorous interrogation before we put our trust in them. The worst kind of model, O’Neil writes, is one that is opaque to its subjects yet can harm them: like, say, the data-driven marketing blitz of for-profit educational mills, or the secretive statistical screens applied to your resume in a job-application algorithm.

See what’s on the slab. Plenty of companies dream of making the food system more sustainable by satisfying our hunger for meat in new ways — by, for instance, growing it in labs rather than on animals in the field. One problem this industry faces: “Lab-grown meat” sounds icky. “Cultured meat” doesn’t sound much better. How about “clean food”? That’s the new name for your petri-dish-grown steak, if a new trade group succeeds in its rebranding effort (Quartz). On the one hand, it’s a pretty brazen rebranding effort. On the other, the sense of taste is so susceptible to the power of suggestion that “clean food” just might sell.

The next bus will be an Uber. The Verge has a great, balanced read on what happens in small communities when they turn over public services to private NewCos. After the failure of a mass transit plan in Altamonte Springs, a Florida suburb, local officials asked Uber to, essentially, become part of the local transit system for a year-long pilot test. Altamonte Springs subsidizes the rides, and Uber gets citizens to their destinations. The upside: Trips are faster and more convenient. The downside: They cost a bit more. Also, you can’t travel without a smartphone and a credit card. You’re out of luck if you’re disabled. And forget about government transparency, since Uber likes to keep its numbers to itself. That’s plenty of negatives, yet Altamonte’s experiment has been successful enough that other towns near and far are flocking to follow suit. This train is rolling already — we need to keep it on the right track.

Featured in NewCo Shift: Outside the Box. In our latest Shift Dialog, NewCo editor-in-chief John Battelle talks with Box founder Aaron Levie about why he’s betting on independence.

Want to follow the biggest story in business? Get our NewCo Daily Weekly newsletters!

Leave a Reply