Glass, Darkly

Google Glass shares much of its electronics and software with the smartphone, but it’s a very different machine.

You hold a smartphone in your hand. And we do—at restaurants, at the movies, walking across the street, and even in bed. We use smartphones to check our mail, update Facebook, get driving directions, search the Internet to settle bets, and, sometimes, even to make calls. But Glass you wear on your face, and that fundamentally transforms all these human-computer interactions, making them more intimate. Because you don’t use your hands, and because it projects an image onto a transparent screen suspended in front of your eye and uses a vibration to stimulate your inner ear, using Glass is like being naked with the machine: synapses and wires united.

Glass is a head-mounted computer with a camera and a microphone. It sees what you see, hears what you hear. But Glass is not for “life logging”; unlike cameras such as the Narrative Clip, it is not intended to document your day. The battery doesn’t have enough juice, and there’s no organizing software. Glass is an always-ready smart device that answers your questions, alerts you to messages, and gives you driving directions. The see-through display is just out of your direct line of sight. When you choose to consult the display, it looks like a smartphone screen held eight inches from your face. But when you are doing something else, Glass is easy to ignore. Google’s challenge in making the device a successful consumer product will be convincing the people around you to ignore it as well.

Yet for them, Glass is hard to ignore. It’s not clear whether that’s because it’s so new or whether there is something inherently intrusive about this kind of device. When others look at me wearing Glass, the first thing they see is a plastic boom over my right eye with a camera pointed straight at them. Are they being filmed? There’s a small translucent prism—the “glass”—with a tiny illuminated rectangle that no one else can read. Am I paying attention to them, or to it? What is it whispering? They can’t hear.

When I raise a conventional camera to my eyes and press a button, the people around me assume I’m taking a picture. With Glass, I can take a picture with a wink. If I hold down the button on the touch pad built into the side of the device, Glass starts recording a video and continues until I stop it, or until the device runs out of storage space or its battery dies.

Glass is a marvel of integration and miniaturization, even though the specs don’t seem that impressive. The camera records five-megapixel (2528x1856) stills and high-definition (720p) video. It has a dual-core OMAP4430 processor. All that sounds like 2011 technology. But the slower, older processor, combined with the low-power display, lets Glass provide five to eight hours of intermittent use on a smaller battery. The result is that Glass weighs only 42 grams, and since the titanium headband distributes that weight across my cranium, the “nose weight” is similar to that of my ultralight 14-gram rimless eyeglasses.

Making It Mini: Glass is a marvel of integration and miniaturization, even though the specs don’t seem that impressive.

The processor inside Glass is powerful enough, because most of the processing doesn’t happen on your face; it happens in the cloud. Glass reaches the Internet using its internal Wi-Fi radio or by sharing your cell phone’s data connection over Bluetooth, a process called “tethering.” Sound enters your head via a “bone conduction transducer” that presses just over the ear. Bone conduction offers surprisingly good bass and midrange, but the treble is mushy. Better sound can be had from the included “mono earbud” that plugs into the micro-USB port, or through a pair of stereo buds ($85).

The user interface is based on voice-activated menus and a timeline that’s filled with “cards,” similar to the Google Now app available for Android and iOS. You navigate the timeline by dragging your finger across the touch pad. Each photo you snap is pinned to the timeline, as are New York Times headlines, Google+ status updates, e-mail messages, and notices from other “Glassware” applications. There’s also a Web browser, but it’s practically useless on such a small screen.

The Glass community

Glass is not ready for general use, and Google knows it. Today’s device is too different from existing mobile computers (cell phones, laptops, or car navigation systems), its software too immature, and the whole concept too geeky for it to be a successful mass-market product.

Instead, Google spent much of 2013 gradually releasing Glass to increasing numbers of “Glass Explorers,” chosen because the company felt sure they would be relentlessly enthusiastic about the technology. The first 2,000 were attendees who signed up at the company’s June 2012 developer conference. Next were the winners of a microblogging contest: the people who wrote the 8,000 best tweets and Google+ status updates that contained the hashtag -#ifihadglass. They were permitted to pay $1,500 and travel to Google’s offices in New York, Los Angeles, or San Francisco to be fitted with the device, a process that included several hours of training. (Some likened the experience of being invited into the Explorer program to finding a golden ticket to Willy Wonka’s chocolate factory.) The price, travel, and time commitment ensured that Explorers would be personally invested in working with Google to improve the product. Explorers were paying test subjects, voluntarily turning their homes, workplaces, and friends into Google’s living laboratory.

A wave of positive publicity followed. A startup named CrowdOptic gave Glass to some Stanford University cheerleaders, players, and coaches for a few football and basketball games—the video is short and shaky, but it’s like hanging with the teams! Several surgeons have used Glass in the operating theater. iJustine, a pretty YouTube personality, uploaded her five-minute review in June; it attracted 1.2 million views within six months. The September 2013 issue of Vogue featured Glass in a 12-page photo shoot.

The campaign made wearing Glass seem open, fun, upbeat, hip—even wholesome. Explorers posted smiling selfies and videos: a first-person roller-coaster ride, a walk through a park, or a touching moment with a child (all shot -#throughglass). Looking at the images, even those of us who weren’t Explorers felt connected. It’s a very different feeling from actually standing next to an Explorer; then I feel that I’m being observed, monitored, and analyzed. Or worse, that I’m being ignored.

The Explorers, meanwhile, have become a community. Explorers give advice to newbies and provide trouble-shooting advice. Battery life? One Explorer developed a free app that tells how much power is left; another started a company that makes PWR glass, a combination lanyard and external battery that nearly triples battery life.

Most of the tech problems are easily solved. But the thornier questions about privacy—or, rather, privacy perceptions—aren’t going away. Glass users report being accosted by people who feel sure they are being recorded. The problem: how do you prove you’re not taking a picture? (A common suggestion in the forum—to add a little red light—has two obvious problems: it could be hacked so that it would not light up, and it would make Glass seem even more obnoxious when it actually is recording.)

Last December, Google opened up the Glass Explorer program to all subscribers to the company’s $9.99-a-month music streaming service. The price is still $1,500, but now Google will ship the product directly to anyone. (Well, almost anyone. An employee of the MIT Enterprise Forum, which is managed by MIT Technology Review, bought a Glass unit in this fashion and lent it to me. But Google says it is not yet lending Glass to journalists for conventional product reviews, on the grounds that the Explorer edition is not a fully formed consumer product. This review is therefore unauthorized.)

Presumably that high price will still filter out people who might be intimidated by stares. But for Glass to be successful, the broader public has to accept it as well. Otherwise, Glass will remain a tool for specialized users like those surgeons.

Evolutionary, not revolutionary

The first true wearable computer was a foot-controlled machine that the mathematicians Edward Thorp and Claude -Shannon invented to help them win at roulette. That computer was built in Shannon’s basement lab and then (it’s said) tested at a Las Vegas casino in 1961. Others developed shoe-based computers for counting cards in 1983. These clandestine computers gave their wearers superhuman capabilities, but they needed to remain secret to work as intended.

On-your-face wearables showed up a few years later. These machines were all about cognition: enhancing the wearer’s ability to remember, record, and retrieve information. The defining characteristics of machines from this era were their “headgear”—wearable displays that put a video image in front of the user’s eye—and one-handed “chording” keyboards that were hard to master. These machines drew stares when people walked around with them, and their users (most of them academic researchers) proudly called themselves “cyborgs.” They reveled in their outlandish appearance, between geek and punk.

Back in 2001, wearable-computing pioneer Thad Starner (a Media Lab alumnus, now a professor at Georgia Tech, a consultant to Google, and the intellectual powerhouse behind Glass) wrote about the challenges faced by wearables in IEEE Micro. Revisiting those articles 13 years later, one can see that three key problems remain: input, output, and software.

Input: Glass is loaded with input devices, including a camera capable of recording stills and videos, a touch pad, and a nine-axis sensor (three-axis acceleration, three-axis rotation, three-axis magnetometer). But getting words into a wearable remains a challenge. Glass does away with the chording keyboard, using speech recognition instead. Google’s voice recognition works extremely well for navigating menus, because it’s easy to match a second of speech with one of a dozen possible menu choices. The technology is not as reliable for captioning photos and replying to messages, but with practice and patience it can be astonishingly accurate—though you can’t mind people overhearing what you are saying.

Output: The tiny prism that is the Glass display may seem futuristic, but it’s the descendant of technology that’s more than two decades old. Back in the 1990s a company called Reflection Technology developed a similar system called the Private Eye. It was a plastic box the size of a modern cell phone that used fancy optics to put a virtual screen in front of the user’s face. These screens were indispensable to the wearable-computer inventors of the 1990s, but because they were big and obscured the eye, they did as much as anything to make the early users resemble cyborgs. Compared with the Private Eye, today’s Glass display is small and fairly unobtrusive. The other output is audio. Today’s Glass uses audio to read text, to play music, and for alerts, but not for much else.

Software: As with many pieces of information technology, the biggest challenge for users will not be the hardware but the software. Glass is clearly useful, but its voice-based interface relies on menus that must be spoken in a particular sequence, rather than free-formatted commands that can be issued in any order. For example, if you say “Okay, Glass … take a picture,” the system expects that if you wish to share the image you must next choose how to share it, then select a recipient, then add a caption, all in that precise sequence. Users will want to utter unformatted commands—for example, so you can take three pictures in a row.

Likewise, the system for alerts and notifications is in desperate need of a redesign with a heavy dose of artificial intelligence and user modeling. Instant notifications are an annoyance if you’re getting 200 e-mail messages a day. Glass is quite literally in your face. It knows not only where you are but whether you are walking or standing still, whether you are talking to someone, whether that person is talking back, and more. Yet messages from my predefined “Glass Contacts” interrupt no matter what I am doing, while others are silenced even if I’m bored. Knowing when it’s the right time to interrupt and when it is not will take more research, and it will probably require algorithms that adapt to each user.

The Glass hardware also has its critics, most notably Steve Mann, who has been developing wearable technology since he was in high school 35 years ago. Mann, a Media Lab cyborg in the 1990s and now a professor at the University of Toronto, says that the Glass display will force each eye to focus at a different distance when images are overlayed on the real world, which could cause eyestrain or dissociation in some people. Mann’s headgear employs a different design that pro-jects images into the eyeball with a double-sided mirror and a pinhole projector, supporting a range of “augmented reality” applications like one that automatically labels buildings when you look out at an unfamiliar city.

“It’s astounding to me that Google and other companies now seeking to market head-wearable computers with cameras and displays haven’t leapfrogged over my best design,” Mann wrote last year in IEEE Spectrum. (See Mann’s “View” to read more of his thoughts on wearable computing.)

Google says that Glass is designed the way it is because eye contact is important to humans. (Even Mann’s most discreet display prevents eye-to-eye contact.) Google probably made the right call for a mass-market, always-worn display. A Glass camera and screen that obstructed the pupil would have a hard time as a consumer product, at least in this decade.

Glass and the law

In October 2013, Glass Explorer -Cecilia Abadie was pulled over for speeding by the California Highway Patrol. When the officer noticed her Glass—on—she received an additional charge for having a TV screen visible while driving. “Is #GoogleGlass ilegal [sic] while driving or is this cop wrong?” she posted.

Abadie’s ticket was thrown out a month later for lack of evidence: the cop didn’t see Glass enabled while she was driving, and he couldn’t prove the radar gun had been properly calibrated. Nevertheless, California, at least 37 other states, and the District of Columbia all have laws prohibiting the operation of in-car TV screens that can be seen by the driver, says Russ Martin, manager for state relations at AAA, the motorists’ association. In case Google Glass falls into some kind of loophole, lawmakers in Illinois, Delaware, New Jersey, and West Virginia have proposed specific bans on using wearable computers with head-mounted displays while driving.

I would not use Glass while driving, but I did use the turn-by-turn directions app while my wife drove us to a party. I’m glad I was in the passenger seat. I spent so much time fiddling with Glass that I would almost certainly have veered off the road, driven through a stop sign, or worse. The issue isn’t taking my eyes off the road—it’s that Glass steals away my attention. AAA’s Martin backs me up, pointing to a growing body of research showing that “hands-free cell-phone use offers no safety benefits over handheld cell-phone use” and that “voice-activated texting or e-mailing is one of the most mentally distracting activities a driver can engage in.”

Google’s legalese-filled disclaimer that appears when the navigation app starts up seems to be in agreement with AAA but goes even further: “Please keep your eyes on the road and obey applicable laws. Do not manipulate this application while in motion. Directions may be inaccurate, incomplete, dangerous, not suitable, or prohibited. Data is not real-time, and location accuracy cannot be guaranteed. Tap to continue.”

Using Glass while driving is probably more dangerous than listening to the radio, but it’s probably less dangerous than texting with your iPhone. What’s more, Glass might have untold safety benefits—there’s an app called DriveSafe that uses the tilt sensor to determine if you are falling asleep at the wheel and directs you to a rest area. Perhaps apps like DriveSafe, combined with an updated navigation app that has a better understanding of human cognition, will one day make the benefits of driving with Glass outweigh the risks.

Even nondrivers are likely to find a growing number of places where they can’t use Glass. A hip restaurant in Seattle banned the device, for example. “We want our customers to feel comfortable, not like they’re being watched,” Jason -Lajeunesse, an owner of the Lost Lake Café and Lounge, told Forbes. Glass is banned at casinos, of course. Google banned Glass at one of its own press events. And if you are visiting Washington, D.C., you would be well advised to remove your Glass before visiting the White House, the Pentagon (including its Metro station), the House and Senate galleries, or the Supreme Court.

The other kinds of rules Glass users need to understand are Google’s—rules enforced by license agreements with Glass app developers, designed to preserve the device’s aura of wholesomeness. Google has so far banned the use of facial recognition or voice prints to identify a person or “present personal information identifying anyone other than the user.” Google also prohibits use of the camera when the display is turned off, perhaps because taking pictures when the machine is not visibly on might seem sneaky. Last summer, Google banned a sexually explicit app called Tits & Glass. But there is sure to be pressure for adult-themed software.

Developers have skirted Google’s rules, of course. A company called -FacialNetwork has developed an app called NameTag that looks through Glass, quickly queries a database, and tells you if the person you are looking at is one of 450,000 registered sex offenders (or a near likeness, anyway). Another app, called Sex with Google Glass, will allow to people to see themselves through their partner’s eyes, turn out the lights, and even have the computer suggest new ideas. These apps aren’t in Google’s store, but you can “side load” them by putting Glass in developer mode and transferring them by micro-USB.

But that would be creepy.

“All the world’s a stage”

Much of the initial excitement about Glass concerns users’ ability to blog with point-of-view photos and video. But as Glassware improves, we’re likely to see a proliferation of new applications that are more useful, like a Word Lens app that’s able to instantly translate text between English and five other languages.

For Google itself, Glass is compelling because it makes Google an irrevocable part of the wearer’s life. Glass provides Google with hugely personal, sensitive data—just the thing for a company that aims to organize the world’s information and gets about 90 percent of its revenue from advertising. I expect the cost of Glass to drop toward $300, because Google will subsidize its use. The platform will be licensed to other manufacturers, just like Android. Already Hyundai has announced that its 2015 Genesis luxury sedan will have a Glass app that will unlock and start the car or help the user find it in a parking lot, among other things.

So far, the biggest problems I’ve encountered with Glass involve the display. A friend who has no vision problems reported eyestrain and a mild headache after using Glass for half an hour. My left-handed wife reports total frustration from having to use the device with her right eye—apparently her left is the dominant one.

“Glass isn’t for everyone,” reads the Glass FAQ, adding that people who have had Lasik surgery should consult a doctor first. “Don’t let children under 13 use Glass as it could harm developing vision,” the FAQ continues, although I suspect that the magic number 13 is more about the Children’s Online Privacy Protection Act, which prevents companies from collecting data on youngsters without their parents’ consent, and less about a child’s developing ocular system. In any event, Google offers those who bought Glass and can’t use it a full refund.

Glass clearly has significant opportunities for industrial, scientific, and medical applications. I also think that it could revolutionize the lives of the disabled. Glass can be a head-mounted eye for those with poor vision. It can deliver clear, understandable instructions to those who are cognitively impaired. It can let those who are paralyzed navigate the Web and communicate. Although decades of experience have gone into developing assistive technology for the disabled, it has often been too expensive for the intended users. With Glass they will be able to get the hardware for the price of a few fancy dinners.

Maybe as Glass becomes more useful to more people, it will become commonplace. Future wearables won’t generate stares if a lot of people are wearing them. But for many, I think, Glass faces an insurmountable problem. It’s impossible to miss. One of my friends, an Android developer and techno–utopian, says he understands the concept—he’s even got an Android watch—but he’s not interested in Glass. He has perfect vision and doesn’t want to wear something on his head. He’s waiting for the implanted version.

Simson Garfinkel is a contributing editor to MIT Technology Review and a professor of computer science at the Naval Postgraduate School. His last article for the magazine was “Windows 8: Design over Usability” (March/April 2013).