Study Shows Software Can Predict Psychosis Better Than Psychiatrists
Mental health issues manifest in a number of ways, and they're not all behavioral. Increasingly, scientists are using speech analysis software to detect subtle changes in voice acoustics and patterns to detect or even predict potentially problematic conditions.
A study published Wednesday in NPG-Schizophrenia by researchers at Columbia University Medical Center, the New York State Psychiatric Institute, and IBM's T. J. Watson Research Center found that digital speech analysis correctly predicted whether 34 youths at risk for mental illness (11 female, 23 male) would develop psychosis within 2.5 years. The system, which evaluated the study participants quarterly, correctly predicted all of their outcomes; five became psychotic.
The algorithm evaluated transcripts for predictive "semantic and syntactic features" like coherence and phrase length. "These speech features predicted later psychosis development with 100% accuracy, outperforming classification from clinical interviews," the researchers wrote.
Clinicians are able to accurately categorize patients as "at-risk," but within that subpopulation it is difficult to determine who will actually experience psychosis and potentially develop schizophrenia. If voice recognition software can help identify these individuals, they may be able to receive more effective care. "Computerized analysis of complex human behaviors such as speech may present an opportunity to move psychiatry beyond reliance on self-report and clinical observation toward more objective measures of health and illness in the individual patient," the researchers wrote.
Guillermo Cecchi, one of the investigators from IBM Research, told the Atlantic, "What this means is that over 45 minutes of interviewing, these young people had at least one occasion of a jarring disruption in meaning from one sentence to the next. As an interviewer, if my mind wandered briefly, I might miss it. But a computer would pick it up.”
Appeals Court Reverses Injunction on NSA Bulk Phone Record Collection
On Friday, a Washington, D.C., appeals court reversed a lower court's decision that the National Security Agency couldn't conduct bulk phone-record collection. The panel of three judges said that plaintiff Larry Klayman (a conservative activist) hadn't proved that his personal phone calls had been collected by the NSA's wide-cast net and therefore lacked standing. But privacy advocates don't need to panic.
Congress decided to shut down NSA bulk data-collection programs on June 1 but allowed for a transition period during which the NSA can do ongoing surveillance. As the Washington Post points out, the appeals court's decision does not speak to the question of what surveillance is constitutional, or the legality of what the NSA previously did. The only appeals court that has weighed in on that is the 2nd U.S. Circuit Court of Appeals in New York in a ruling from May. That court said that NSA bulk collection was not in keeping with the Patriot Act and was "unwarranted."
This new ruling from the D.C. appeals court doesn't affect that other decision. The Guardian explains, "The ruling reversed an injunction from a lower court on the phone records surveillance program—but only in a technical sense, as the injunction never actually went into force."
Klayman says he will add other plaintiffs whose data was collected to the complaint so it can move forward again. He told the Guardian, "It’s outrageous this court would allow the constitutional rights of Americans to be trampled upon."
Even beyond these legal battles, privacy advocates are also concerned about how much surveillance the NSA and other government intelligence bodies can still do using the existing troves of data, malware campaigns, and overseas channels they have access to.
Instagram Will No Longer Limit You to Square Photos. Too Bad.
Instagram has decreed an end to the tyranny of the square, a longstanding policy that restricted users to posting photos with sides of even length. The company announced the end of its reign with some wistfulness. “Square always has been and will be part of who we are,” it tweeted. And yet freedom marches on, toward the manifest destiny of portrait and landscape formats.
To what end? The landscape format will be more accommodating to certain Instagram genres, like photos of friends posing arm-in-arm, against the backdrop of some fun locale. But at a time when technology has freed us from almost all boundaries of what content we can record and publish and when, there is something to be said for restrictions. Twitter’s 140-character limit is one example. Vine’s six-second rule is another. Like Haiku or iambic pentameter, these limitations create economies that force authors to consider what each word or image is worth.
If you're making art, restrictions can be helpful. Brian Eno, the British music producer and artist, uses them to fight the chaos of choice. “In modern recording one of the biggest problems is that you're in a world of endless possibilities,” he told the Telegraph in 2009. “So I try to close down possibilities early on. I limit choices. I confine people to a small area of maneuver.”
The square confines Instagram users to a small area of maneuver. It forces us to consider what details are essential, and which can be cropped out. It spares us from indulgence of the landscape and the false promise of the panorama.
But Instagram, which is owned by Facebook, is in the business of accommodating its users, not challenging them. One of the problems with the square, the company explained in its announcement, is that “you can’t capture the Golden Gate Bridge from end to end.” This example speaks to the needs of a certain kind of Instagram user who enjoys planting his flag on settled territory. Like an iPhone videographer at a Taylor Swift concert, the guy Instagramming the Golden Gate Bridge is not creating a rare or essential document, only proof that he saw it with his own eyes.
And why did he bother doing that, anyway? Clearly, because photographs cannot really capture the scope of the Golden Gate Bridge, or St. Peter’s Basilica, or the view from your car window as you drive up the Pacific Coast Highway. The impulse to capture these moments on camera is shaded by the knowledge that the moment, in all its immediacy, is too large to fit in a frame of any size.
This is not to discount the medium altogether. Instagram photos, like memoirs, give us a good way to preserve and share moments from our lives. They allow us to retouch and stylize those moments to reflect how they felt, or how we want them to feel in retrospect. A couple of new framing options will not fundamentally change that.
Still, photos are not moments, and the square, in its inadequacy, forces us to acknowledge the distinction. It helps remind us that the feeling of standing at a vista with the breeze in your face and salt on your tongue and looking out at orange steel draped over nearly two miles of causeway cannot be boxed and shipped. Broadening the lens only adds dimensions to an illusion.
1 Billion People Visited Facebook on Monday
Zuckerberg wrote, "Our community stands for giving every person a voice, for promoting understanding and for including everyone in the opportunities of our modern world," and Cox noted, "I couldn’t be more excited about connecting the next billion." Their comments are in line with Facebook's ongoing, but controversial, effort to bring Web connectivity to everyone through its Internet.org initiative.
Facebook hit 1 billion monthly active users in October 2012, but it has taken until now for 1 billion people to visit the site in one day. As of June 30, the service has 1.49 billion monthly active users.
It would be pretty surreal to be able to say that you invented something that was used by "1 in 7 people on Earth" in a single day. There weren't even 1 billion cars in the world until 2010.
Hurricane Forecasts Have Become Much, Much Better Since Katrina
With a potentially strong hurricane bearing down on the United States the same week as Hurricane Katrina’s 10-year anniversary, it feels like a good time to take a step back and think about what’s different now.
As far as meteorology is concerned, Katrina may as well have been a century ago.
After the disastrous 2005 hurricane season, the National Oceanic and Atmospheric Administration began to plan for a crash course in greatly boosting the accuracy of hurricane forecasts. The Bush administration approved the Hurricane Forecast Improvement Project in 2008, and it has since exceeded even its own lofty goals.
In a statement Wednesday, NOAA said: “Since the 2005 hurricane season, NOAA has launched 5 new satellites, deployed new coastal observing systems and made major breakthroughs in oceanic and atmospheric research, all of which has resulted in a remarkable *40% reduction* in the margin of error of a hurricane’s expected track.”
Seen graphically, the result is stunning:
The cone from Katrina vs what the cone would look like for Katrina in 2015--big improvement in track skill in 10 yrs pic.twitter.com/mcSfDZfqw9— Eric Blake (@EricBlake12) March 27, 2015
In a tweet earlier this year, Eric Blake, a hurricane specialist at the National Hurricane Center, called the stunning improvement in hurricane track forecast accuracy over the last decade “one of the most incredible success stories of our lifetimes.” Five-day forecasts today are just as accurate, on average, as three-day forecasts were the year of Katrina. That means two extra days for people in the path to prepare.
Forecasting hurricane strength days in advance has historically proven more challenging than track forecasting, but there’s been vast recent improvement there, too. The U.S. flagship high-resolution hurricane model, the Hurricane Weather Research and Forecasting Model, has improved its accuracy at a rate of 10 to 15 percent per year since 2011.
Earlier this year, HFIP fell victim to its own stunning success, and the budget took a big cut. But the program should still be able to benefit from a massive new NOAA investment in faster supercomputers.
You’d be forgiven if you haven’t noticed much benefit from the vastly improved hurricane forecasts. That’s mostly because, with the possible exception of Hurricane Sandy, there’ve been (thankfully) very few high-profile opportunities to test new forecast systems in the last 10 years. The U.S. is in the midst of a record-breaking drought of hurricane landfalls with winds of 111 mph or higher—“major” hurricanes. We’ve grown complacent, and sooner or later, our luck will run out.
I can guess what will happen then: There’ll be an ominous forecast cone, worryingly camped over a major coastal city for a few days. Officials may wait until the day before to order mandatory evacuations, and many locals might choose to stay. After the stormwaters recede, the damage will be measured in the tens of billions. People on the evening news will say, “We never saw it coming.”
Just this week, on Twitter and in weather message boards, I’ve noticed Floridians confidently quip something like: “If we’re in the cone at five days, I know I can breathe easy. We never get hit when the storm is pointed at us that far out.” NOAA, to its credit, is gently challenging that narrative this month.
Still, at this point, there’s reason to believe that better forecasting isn’t the most important thing in minimizing American losses to hurricanes. Providing two or three extra days of warning may not mean much for low-income families whose evacuation options are limited, as Katrina painfully showed. In a recent op-ed, Peter Neilley, the scientist in charge of forecasting operations at the Weather Channel, said that when preparing for the next major hurricane, psychology is now as important as meteorology. In Katrina, “there was a gap between the perceived accuracy of the forecast and the real accuracy,” Neilley wrote. “Society’s perspective on forecast accuracy lagged behind the true gains that our science had made up until that point.”
The same is true now. Even with perfect forecasts, society can never be perfectly prepared for extreme weather. Only when meteorologists and emergency managers place the “why” of improving forecasts above the “how” will society truly benefit. This is a lesson that the meteorological community is still struggling to learn. That’s why after Katrina, after the horrible tornadoes of 2011, after Hurricane Sandy, we all asked, “How could this happen?” At some point, improving society requires a re-think of why people become vulnerable in the first place, and then taking action to ensure those vulnerabilities are addressed. Better weather forecasts help, but what we need is a better society that prevents those vulnerabilities from reaching potentially disastrous levels in the first place. Many meteorologists are already thinking this way, but we’ll need a whole lot more before we can say we’ve made progress since Katrina.
North Dakota Police Drones Can Be Weaponized If They’re Not Lethal. Wait, What?
A bill passed by North Dakota's legislative assembly that was meant to require warrants for drone searches evolved into something entirely different, thanks to an amendment from a lobbyist.
The Daily Beast reports that House Bill 1328, sponsored by Rep. Rick Becker, R-Bismarck, aimed to forbid all weapons on police drones. But Bruce Burkett from the North Dakota Peace Officer’s Association amended the bill to prohibit only lethal weapons, leaving the door open for “less than lethal” weapons like Tasers, pepper spray, and rubber bullets. (Let's not even get into the fact that “less than lethal” weapons actually have killed people.)
As the Verge points out, North Dakota is one of six Federal Aviation Administration pilot programs for trying out commercial drone use in civilian airspace, and drones are allowed to fly at up to 1,200 feet in the state instead of the usual 400-foot limit.
Becker told the Daily Beast of the amendment, “This is one I’m not in full agreement with. I wish it was any weapon. ... In my opinion there should be a nice, red line: Drones should not be weaponized. Period.”
Report: A Lot of People Don’t Bother Using Fancy Car Tech
Cars are supposed to be able to do basically anything these days. They listen to you and try to answer your questions, they know all your favorite music, they self-park. Soon they'll be doing all the driving for us. But the 2015 Driver Interactive Vehicle Experience Report from J.D. Power shows something unexpected: A lot of people don't seem to care about any of this.
The Los Angeles Times notes that 43 percent of people surveyed don't use their cars' voice recognition to call up things like GPS directions. Thirty-five percent never tried automatic parking, 32 percent avoided apps like Yelp, and 20 percent didn't even use half of the tech features in their cars. (The survey asked about 33 tech features that seemed to be available in all respondents’ cars.)
The report polled 4,200 people between April and June who had bought or leased cars no more than three months before taking the survey. Research indicates that people are unlikely to explore car features and start using new ones after the first three months of owning a car, Reuters reports.
It seems that most people, especially those in the 21- to 38-year-old range simply used their smartphones instead of attempting to engage with their cars' tech features. For all ages the lack of engagement seemed to be a combination of active avoidance and not knowing all of the things the cars could do.
Kristin Kolodge, the executive director of driver interaction at J.D. Power, told Reuters, "Customers say, 'I have a competing technology that's easier to use, or I've already paid for it—so why do I need it again?' ... Is it really making it easier? That's where some of the value is being challenged." She noted in a statement that the tech features people seem to like the most are more related to actual driving mechanics—things like maintenance diagnostics, cruise control, and blind spot monitoring—than entertainment or connectivity.
Though these results could have implications for the (supposed) impending rise of self-driving cars, it also could play out that people have more time to figure out all the features in their cars when they're not actually driving them. And when people are already so practiced on their smartphones, it's hard to see how plugging in some directions at a red light is more difficult than trying to get the voice of your car's "personal assistant" to stop blasting out of the speakers about alternate route options and Top 40.
Voting Selfies Violate the Sanctity of the Ballot Box. That’s a Good Thing.
In one of his more controversial essays, E.B. White writes about politicians, religion, and the dogma of democracy. In 1956’s "Bedfellows," White aptly observing that the “Democracy is itself a religious faith. For some it comes to being the only formal religion they have.”
Prescient and wise though he was on certain matters, the famously reclusive White could not have foreseen the appeal of selfies and the slavish devotion of those who snap them, nor their place in debates about suffrage. Selfie takers may be radiant, but they’re not terribly humble and can be a bit piggish. The advent of the selfie ballot may be causing a spiritual crisis for those who worship democracy.
To produce a ballot-box selfie, complete the following steps: Take a photo of yourself with your marked ballot. Use flash only as needed, i.e., when your district’s booth is relegated to the dark corner of the high school gym. Collect your “I voted” sticker, and Photoshop image thereof into original photo. (This part is optional but encouraged.) Post to social media platform of choice. Rocking of vote complete. Watch the likes roll in—that is, of course, as long as you’re not in Indiana.
On Monday, the New York Times reported on a newly enacted Indiana law, in which a voter may not take a photo of her ballot in the polling place to distribute or share using social media. The photos may be shared, however, to document a problem or malfunction with voting system. Indiana is hardly the first state to adopt such a law. White’s beloved Maine amended its code in 2011 to remove a pre-existing (pre–selfie heyday!) ban on sharing of photos of one’s ballot. Research by the Digital Media Law Project indicates that while only a handful of states explicitly ban recordings inside the polling place, there are actually plenty of laws on the books that, if fully enforced, would stifle ballot selfie expression. And until recently, New Hampshire voters were also bound by a law that left them free to live, but not to post. Earlier this month, the U.S. District Court in New Hampshire overturned the law on the grounds that it amounted to a content-based restriction on speech, giving Kardashian lovers everywhere reason to celebrate and bow down to the altar of the “gram.”
In 2014, New Hampshire updated a decades-old election law to extend the general ban on a voter sharing her marked ballot to specifically prohibit social distribution. The rationale? Selfies and the information they contain could lead to coercion and vote buying. Boss Tweed political-machine types and corporate overlords with a cause, the argument goes, will demand voting booth selfies as proof that smartphone-wielding employees are toeing the party and company line. The increasing political influence of corporations is real and a source of concern. The prospect of a resurgence of vote-buying by powerful interests is scary. It is also already illegal.
The Indiana and New Hampshire legislatures’ motivations in banning the ballot selfie may or may not be noble. Underlying the laws is the notion of a pure voting experience unsullied by external forces, a celebration of political expression, of participation, of the power of the one, of the power of the many. Voting is the cornerstone of our democracy, so this is appealing stuff. I admire the instinct to protect the vote. Despite whiffs of paternalism, it’s a welcome balancing influence to the unfailing belief in transparency as a cure to all political woes. But the doctrinaire approach—protect the sanctity the booth—misses an important point: It’s hard to celebrate by yourself. One need look no further than the Iowa caucuses for proof that politics is a team sport. Democracy (when it’s functioning as intended) is a religion open to all. So it is with the selfie. Selfies are meant to be shared for all to see and comment upon. And I expect the tweeting masses won’t give up on the right to frame without a fight, or, at the very least, a rousing debate.
This Might Be the Best Internet Service Provider in the U.S.
When we moved to our current Bay Area home in 2010, it seemed wise—given our reliance on the Internet for our work—to get both cable Internet and DSL service, to ensure a connection in all but the worst of circumstances. We had one choice for cable: Comcast, which is under no obligation to share its lines or central facilities. But because of different rules governing copper-line connections, we had more than one choice for DSL. And rather than send money to AT&T, a company that has earned widespread contempt—notably, most recently, for its eagerness to turn over its customers’ data to the government’s pervasive-surveillance maw—we opted for a more trustworthy alternative.
Our DSL comes from a small company called Sonic, based north of San Francisco. It’s an independent in an industry dominated by a cable-phone cartel notorious for greed, customer disservice, and control-freakery. Sonic is innovative and aggressive in good ways, expanding its footprint by providing excellent service at a fair price. It has (from first-hand experience) a genuine commitment to customer service. And, reflecting the civil-libertarian beliefs of its founder and CEO, Dane Jasper, it is the anti-AT&T when it comes to privacy and security.
Sonic consistently gets a perfect score from the Electronic Frontier Foundation for protecting users’ privacy. Unlike most Internet service providers, which hang on to users’ data for months or years, Sonic retains customer data for just two weeks, long enough to troubleshoot network problems and provide law enforcement help in emergencies but not long enough to fuel copyright trolls or government fishing expeditions. “We’re not trying to help people evade the law,” Jasper says. “We're trying to protect the privacy of our lawful customers.”
Sonic also fought back when the government subpoenaed a user’s information during its ongoing investigation of WikiLeaks—“so he would have an opportunity to fight it,” says Jasper. The court let Sonic tell the customer, who was identified in press accounts as Jacob Appelbaum, but the ISP was required to hand over the data.
Like other DSL competitors, Sonic co-locates its own equipment inside the phone company's facilities, effectively renting the copper lines. For the small percentage of Sonic customers whose setup requires the use of AT&T's central-office electronics—most customers are using Sonic gear on both ends of the conection—Sonic provides a virtual private network service at no extra charge. AT&T may be saving and handing over its own customers’ information willy-nilly, but Sonic customers’ Internet traffic (at least, for people taking advantage of the VPN) is subject to Sonic’s data-retention policies, not AT&T’s, a major improvement.**
Jasper says Sonic has business and residential customers—he won’t say how many, and he doesn’t have to since he runs a privately held company—in 125 California communities. The vast majority have some variant of DSL, which uses the copper wires from the traditional phone system. But advances in technology have given companies like Sonic a way to offer significantly more connection speed than was available on DSL a decade ago.
Sonic’s speed offerings, for DSL customers who live close enough to phone company facilities (we don’t, unfortunately), are amazingly fast: up to 100 megabits per second, and occasionally more, for a dual-line setup that connects two phones. The high-speed data comes with a wired phone line—a kind of “double play” in telecom parlance that is only $60 a month plus various fees and taxes. A single-line setup is $40 a month but still, in a good location, is plenty fast. (Sonic has a deal with DirecTV to offer a small discount on the satellite provider’s service). Now Sonic is deploying fiber, so far in two northern California communities with more to come; fiber is Sonic’s most important new frontier. Jasper’s goal is to offer super-fast fiber, gigabit-level connections, plus phone for the $40 base price as widely as possible.*
Jasper and Sonic aren’t the only noncartel competition in the country, though the collective independent footprint is dwarfed by the telecom giants’ installations. There are many wireless ISPs, though their data speeds can’t generally keep up with wired connections. A Toronto-based company, Tucows (disclosure: the CEO is a friend), has moved into the ISP business as well: It’s taken over several small independent systems (one built by a city) in the eastern U.S., with plans to move further into the business. And, of course, Google is a huge competitor to the entrenched telecoms in the few cities it’s been wiring for gigabit service.
Even if Sonic never brings its fastest connections to our address, I’ll stick with them. It feels important to do business with companies that believe in doing the right thing. From what I’ve seen, Sonic is one of them.
*Update, Aug. 26, 2015: This post was updated to clarify that the fastest DSL service requires two linked phone lines, while most customers get a single line at a lower cost.
**Update, Aug. 27, 2015: This post was updated to reflect that while Sonic's DSL customers are using AT&T's lines, only a small percentage of them are using AT&T electronics.
Beating Extinction Will Take More Than Freezing Animal Sperm
Zoo animals are giving humans a run for their money in the assisted reproduction department. Mei Xiang, a giant panda at the National Zoo, gave birth to twin babies this past Saturday, thanks to artificial insemination. And earlier this month, scientists announced the birth of a bouncing baby black-footed ferret, conceived with cryogenically preserved sperm from a father who had died twenty years ago.
This is great, you’re thinking. Why can’t we just artificially inseminate all the endangered animals? Game over, extinction … right?
Well, no. But that line of reasoning isn’t too far off from what scientists were thinking back in the ’70s and ’80s. That’s when, inspired by successes in the livestock industry, researchers started freezing and stockpiling the sperm, eggs, and other tissues of endangered animals in “frozen zoos.” Farmers have successfully artificially inseminated cows for decades, and it’s common practice for prize horses and dogs. (And humans, of course.)
But it turns out wild animal insemination is a whole different beast.
Scientists, first off, simply don’t know how reproduction works for the vast majority of species. “We had to go back first to basic reproductive biology, because, of course, a cheetah is not a dairy cow,” says Pierre Comizzoli, a biologist at the Smithsonian Conservation Biology Institute.
Artificial insemination for cows has been successful because the livestock industry has poured millions of dollars and decades of research into studying them. But that’s just one species worth of research. With 6,264 endangered or critically endangered species on the IUCN Red List, scientists are spread thin trying to tease out the intricacies of their reproductive systems.
On top of that, science-mediated sex is really complicated (even more complicated than run-of-the-mill sex). “Artificial insemination is still an experimental model,” Comizzoli says. It works sometimes, but the odds of success aren’t great. So vials and vials of endangered animal sperm are just sitting around in liquid nitrogen in zoos across the country, waiting for the day the science catches up to the promise of their gooey contents.
Take the big cats. They’re especially challenging to artificially inseminate, says Adrienne Crosier, who’s been studying cheetah reproduction at the National Zoo for more than a decade. That’s because cheetahs don’t tend to produce good quality sperm samples. “A lot of cells in fresh ejaculate are motile and of good quality,” Crosier says, but what samples scientists can get from the small cheetah population are plagued by low sperm count and abnormally shaped cells. Mostly, that’s because cheetahs have an unusual lack of genetic diversity, thanks to a population bottleneck 10,000 years ago.
Female cheetahs aren’t helping out much either: Their estrous cycles are so irregular that Crosier has to administer hormones to reset them, and her team relies on male cheetahs to let them know when the females are in heat. Otherwise, she says, “it’s impossible for us to tell.”
Once scientists have collected a semen sample (a logistically complex procedure that may involve anesthetizing the animal), the sperm needs to be frozen and, when the time comes, thawed. “It’s asking a lot of these cells,” says Budhan Pukazhenthi, a Smithsonian scientist who works with ungulates like zebras, antelopes, and deer. Scientists basically pickle the sperm cells with a cocktail of antifreeze chemicals, which involves a delicate dance between lowering the temperature of the sperm and introducing the solution to keep sharp ice crystals from forming within the cell. Reanimating the sample involves reversing the process while controlling for temperature and thawing rates. “It’s almost like cooking,” Pukazhenthi says.
But, again, not all sperm is created equal. Giant pandas owe their human-orchestrated reproductive success to their hardy sperm: It’s resistant to cold, not fragile, and relatively easy to freeze and reanimate. Elephants, though? Scientists are still struggling to freeze the sperm and keep it alive after thawing. Cheetahs are somewhere in between—researchers have worked out a protocol to preserve sperm well, but insemination itself hasn’t gone so well. Out of 50 attempts or so, inseminated cheetahs have given birth to 11 litters. “You may have a perfect-quality sperm sample that’s just absolutely beautiful, and then your female doesn’t respond to the hormone, or vice versa,” Crosier says. “So some of it is all the stars aligning.” She and her team are working to increase those odds.
Still, artificial insemination has a ton of promise: Instead of transporting animals to breed between zoos and stressing them out in the process, it would be much simpler for everyone involved to just move their genes instead. Scientists can keep populations from becoming too inbred by pairing up genetically favorable couples. And a single sample can have a longer shelf life—and better staying power—than a single stud.
Using artificial insemination would be especially useful for cheetahs, who are extremely picky about their mates and rarely breed in captivity, says Crosier. “We need to be patient, but we also have species that need to be saved as soon as possible,” Comizzoli says. “This is kind of a race we have to win.”
Also in Wired: