The phone call that changed Jini Kim’s life came at 2 a.m. in November 2013. The White House needed the former Google product manager’s help with Healthcare.gov, which had been meant to help people buy health insurance but was riddled with embarrassing glitches. She hopped on a plane that day and worked marathon hours to fix the site, giving up Thanksgiving, Christmas, and her birthday. By the time she left, six months later, the site had enrolled eight million people in insurance plans—and Kim had gained insight that would be crucial for her own health-care analytics company, NunaHealth.
Founded in 2010, Nuna helps companies shape their health-insurance benefits and wellness programs. It analyzes anonymized data about employees’ behavior to determine the answers to questions such as “Are there differences in how people in certain demographic groups seek health care?” or “Can more generous health insurance help improve the productivity of someone with a seriously ill family member?”
Before she bailed out Healthcare.gov, Kim viewed the government the way many people in Silicon Valley do: as a hindrance to innovation.Accordingly, Nuna originally sold its services only to corporations. But during her stint working for the Obama administration, she saw the enormous potential the government had to effect change. “You can touch millions of people so easily,” says Kim, recalling a day at a Healthcare.gov call center when she overheard desperate people crying because they were unable to sign up for insurance.
Upon her return to San Francisco, Kim expanded Nuna so that it now also works with local, state, and national governments. For example, the company helps the Centers for Medicare and Medicaid Services find patterns in their vast amounts of data.
For Kim, reforming health care is not a theoretical issue. Her 33-year-old brother, Kimong,has severe autism. She has been involved in his care since she was nine years old and had to sign him up for Medicaid on behalf of her immigrant parents. She still lives at home to help out. Nuna’s meeting rooms are named after Kimong’s favorite Sesame Street characters, and she brings him to work regularly to give her parents a break. The name “Nuna” comes from the Korean word for “big sister,” one of three words he knows.
—Yukari Iwatani Kane
I grew up in very rural Ireland. The Internet was a connection to the greater world. It was very clear just how potent a force the Internet was and could be. While my brother John and I were tinkering with some new apps in Ireland and then in Boston and Silicon Valley, we experienced firsthand the difficulty of accepting online payments. We were just baffled at how convoluted and awkward the process appeared to be. The ecosystem seemed designed to reduce the number of Internet businesses.
“The same way Google exists as a foundational component of the Internet around information retrieval, it felt like there should be a developer-focused, instant-setup payment platform. Many people in financial services told us it couldn’t work.
“Stripe now processes billions of dollars a year for thousands of businesses, from startups to publicly traded companies. There’s a ton of database and distributed-system work that has to be done to make that experience possible. We have a 10-person machine-learning team that works on compliance, risk, fraud, identity verification, all of those things behind the scenes.
“Making it so easy to participate in the online economy has a far larger effect than one might imagine. We’re enabling new business models, like crowdfunding. And mobile marketplaces, like Lyft, Postmates, and Instacart. That enables more people in society to take advantage of these services. My youngest brother is disabled, and for him it’s not just a convenience. He can now do grocery shopping in a way that he could not before.”
—as told to Robert D. Hof
Tumor cells that metastasize through the blood are generally very difficult to detect until they have spread to the point of being deadly.
Dena Marrinucci cofounded Epic Sciences in 2008 to commercialize a cell detection and analysis technology that she developed to find cancer earlier. It can find and profile nearly all the tumor cells in two tablespoons of blood taken from a patient. On average, a sample that size has 50 billion red blood cells, 50 million white blood cells, and only a few circulating tumor cells. “You’re basically looking for needles in a haystack,” says Marrinucci.
Other technologies miss some circulating tumor cells because they are scanning for only one biological marker or are filtering cells by size. Epic says it finds more because it detects not only genomic abnormalities but also other biological markers, such as protein expression in cells. That should be useful in tracking the progress of a patient’s cancer over time, so that treatments can be adjusted as the disease evolves. Twenty-six pharmaceutical companies are using Epic’s technology in clinical trials of cancer drugs.
Marrinucci had just begun graduate school at the Scripps Research Institute in San Diego in 2004 when her grandmother was diagnosed with advanced melanoma. Less than a year earlier, however, doctors had given her grandmother an all-clear after a PET scan. “By the time you see cancer cells on a PET or CT scan, there are thousands of them,” she says. “And that’s what we’re trying to change.”
One of the most audacious projects funded last year under the Obama administration’s BRAIN initiative aims to intervene in mental disorders using an electrical brain interface. The plan is to develop a system that both senses and modulates abnormal electrical activity, in hope of helping patients with conditions ranging from severe anxiety to post-traumatic stress disorder. Rikky Muller, an Israeli-born entrepreneur and the cofounder of Cortera Neurotechnologies, is designing the implantable hardware intended to interact directly with the brain.
Muller has long been interested in brain interfaces with clinical potential. After training as an electrical engineer and then designing chips for digital cameras, she gravitated toward neuroscience. In graduate school at Berkeley, she worked on neural implants that might decode human thought to control robotic prostheses. She also built a wireless device that could interpret brain signals in detail while resting on the surface of the cortex, rather than deeper in the brain. That work led to the founding of Cortera, in 2013, during the final year of her PhD studies. “We thought it could change patients’ lives,” she says.
Devices that record electrical activity directly from the surface of the brain—similar to Cortera’s founding work—are already used clinically to map the cortex during surgery and to pinpoint the location of seizures. In theory, these devices could also monitor severe neurological or psychiatric conditions on an ongoing basis. Muller is cagey, however, when it comes to Cortera’s plans in the growing neuromodulation market. “We do have a specific application in mind,” she says, “but we are not disclosing what it is.”
Watch this Innovator at EmTech 2015
Meet the Innovators Under 35
Ben Rubin is talking about the key qualities of Meerkat, an app that helped fuel a live-streaming craze this year. Type in a subject, press a button to start filming with your smartphone camera, and Meerkat sends out a tweet with a link that your friends can click to watch—and comment on if they want. That’s all it had to be, he says: “The medium is new, and if you make a complicated product in a medium that already makes people uncomfortable, you end up with zero adopters.”
One thing Rubin couldn’t control, however: after Meerkat got popular, Twitter began offering a similar app, Periscope, and cut off Meerkat’s access to its network. That made it harder for new users to find friends who also use Meerkat. The company has since let users connect Meerkat to their Facebook profiles.
Rubin envisions live-streaming eventually giving rise to a new form of entertainment: “an ongoing live show that is taking place in real time and involves the audience and everyone. Something where you’re no longer the couch potato; you’re part of the script.”
Kevin Systrom started Instagram in 2010, when he was 26, with a guy he’d befriended in a San Francisco coffee bar. Eighteen months later, when the company was just 13 people and still without a business plan, Mark Zuckerberg came calling with an offer of $300 million in cash and $700 million in Facebook’s pre-IPO stock. Systrom said yes only after he persuaded Zuckerberg to keep the Instagram brand alive and to let him and cofounder Mike Krieger run it.
Three years later, it’s clear that the creation of Instagram was remarkably well timed and well executed. The service is like Twitter, but with pictures and videos primary rather than text. It works because people like to tell stories with pictures: it’s easy, and it has impact across languages and cultures. Instagram has more than 300 million users, who post more than 70 million photos and videos every day.
One big question still faces Systrom, though: can he turn all this attention into a real business? He started rolling out an advertising program last fall and remains coy about how it’s doing. Systrom says he just has to find a way to present the ads without upsetting his users, the vast majority of whom are younger than 30.
Systrom himself is something of a model for an emerging kind of high-tech entrepreneur, at the intersection of technology and the liberal arts. He’s a jock, having been captain of his high school lacrosse team. He’s also artistic, having effectively minored in photography while getting an engineering and management degree at Stanford. He knows the corporate world: he’s on the board of Walmart. And he’s an extrovert, as comfortable with runway models in New York and movie stars in Hollywood as he is with coders in Silicon Valley. As mobile applications and social networking permeate more of our economy, people who understand how these technologies make the physical world more interesting or productive will become as important as the hard-core engineers.
Melonee Wise imagines that all homes will have autonomous robots—something like The Jetsons’ Rosie the robot maid, minus the apron and Brooklyn accent. Just one problem: Wise, chief executive of the year-old startup Fetch Robotics, thinks it won’t happen in her lifetime, because the challenges in hardware and software are too big. “I’m probably one of the most pessimistic roboticists you’ll ever meet,” she admits.
Nonetheless, Wise still thinks smaller and more powerful computers, affordable sensors, more adept machine vision, and better artificial intelligence are coming together to make robots capable of a wide range of tasks—if not yet all in a single machine. That’s why Fetch Robotics is going after one promising area: warehouses and e-commerce fulfillment centers, which are plagued with high turnover, injuries, employee theft, and a chronic shortage of workers, who, of course, also have a biological need to sleep.
Although dedicated robots are common in giant distribution centers (see “Inside Amazon”), Wise thinks there’s a bigger market for more flexible “mobile manipulation” robots that can help smaller companies ease into automation. In a simulated warehouse set up in a corner of Fetch’s San Jose headquarters, a knee-high, cylindrical rolling robot called Freight smoothly follows Wise like a very attentive dog as she picks up boxes of crackers and cereal from shelves. She drops them in a plastic crate atop the robot, and when she’s done with the fake order, it zips off to a mock shipping area.
Another robot, Fetch, is intended not to aid but to replace warehouse workers. It has one jointed arm with a gripper on the end, along with a “head” that uses a depth camera similar to a Microsoft Kinect game controller, so it can identify and pluck items from a shelf and place them in Freight’s crate. Both robots are taught to navigate by leading them around the warehouse to create a map. They’re even trained to recognize people’s legs, so they can follow particular individuals. Unlike some robots that navigate using radio beacons or bar codes on the floor, Fetch’s robots use 3-D laser scanners to get around and avoid obstacles, expected or otherwise.
Wise won’t disclose the exact price of Fetch robots, but she says they will be in the tens of thousands—much less than the cost of an employee. The company has sold some of its initial run of 40 robots to unnamed pilot customers, with plans for a much larger run if the automated workers can do the job.
—Robert D. Hof
Watch this Innovator at EmTech 2015
Meet the Innovators Under 35