I Am Robot Boss
Popular Science's editor in chief conducts a terrifying experiment on his employees, from 3,000 miles away.
I roll out of my office and immediately crash into a box. Across the hall, three colleagues in a group stop talking and turn to watch me. I feel intensely self-conscious. My appearance, I know, must be off-putting, but I don’t actually know what I look like. I’m a continent away, steering a semi-robotic avatar via my laptop. My instinct is to drive to the restroom and look in the mirror, like a newly unwrapped plastic surgery patient, but I don’t know how I’ll open the door. So I turn and roll away in embarrassment.
The projection of myself that’s out there in the world, 3,000 miles away, is burning up my valuable reserves of professional dignity. I am formless, cut off. I am literally out of control.Seeking refuge, I steer unsteadily, like a drunk, into the cubicle of the articles editor. Normally, she’d greet me in a friendly, professional way. Instead, she laughs in terror and defensively turns her monitor toward me to prove that she’s working. I try to say something funny, hoping to retrieve our rapport, but she can’t seem to hear me. As other members of the Popular Science staff gather in her cube, I notice a chat window at the bottom of my screen. “Isn’t there audio?” I type. A pleasant but stilted female voice asks the question for me. There should be, I hear a colleague say. “Low battery,” the screen reads. I’ve evidently been left unplugged too long, and it’s clear that I’m not going to be able to make it back to my charging station. I type a string of expletives and a promise to fire everyone. The woman’s polite monotone is halfway through my tirade when the screen freezes. My avatar is now dead, stranded awkwardly in my colleague’s cube. At my desk in California, the screen goes blank. I’m seized by panic. The projection of myself that’s out there in the world, 3,000 miles away, is burning up my valuable reserves of professional dignity. I am formless, cut off. I am literally out of control.
* * *
For nearly two decades, telepresence robots, or, more properly, remote-presence devices (RPDs), have been something of a white whale for the tech industry. Engineers didn’t have the processors, the miniature microphones, cameras and sensors, or the cheap, fast broadband necessary to support them. But now they do, and in the last five years, a number of companies have sprung up to introduce the first truly functional devices. In the last 18 months alone, at least five companies launched new products, which range from tiny remote-controlled cradles for an iPhone to large rolling platforms that cost as much as a car. As the value of skilled labor rises, these companies are beginning to see a way to eliminate the barrier of geography between offices.
As the editor in chief of Popular Science, I am in the unique and somewhat complicated position of leading a magazine based in New York from a remote office in California. Every few weeks, I fly back and forth between my home in the San Francisco Bay Area and the office in Manhattan to get some face time with the magazine’s two dozen editors and designers. The rest of the time I communicate via email and phone. I’ve also built a makeshift teleconferencing rig in each of my two offices. But running a magazine is a dynamic and constant job that requires a fair amount of improvisation. That’s hard to do via Skype.
After working this way for a year and a half, I was ready to try something new. My wife was about to have our second child, and I needed a way to travel less. I figured that a test of the various RPDs would at least yield a few laughs and at most a new perspective on my job. So when I began calling remote-presence companies to introduce myself, it didn’t occur to me that there was a larger, more important question to explore: Is it possible to work together, to connect as colleagues must, both intellectually and emotionally, without being in the same building?
I wasn’t the first to consider this issue. In 1995, UC Berkeley engineer and computer scientist Eric Paulos flew a small blimp into a conference and tried to talk to people through it. Designed to take up no more room than a standing human, the blimp carried a 600-gram payload (video camera, microphone, remote-control systems) that provided the flyer a steerable means of speaking with someone in a remote location, although no means of projecting his or her face. In 1997, Paulos and his adviser, professor John Canny, published a paper on “tele-embodiment.” It was one of the first widely recognized papers on remote presence, detailing an “inexpensive, simple, networked, tele-operated mobile robot.” They were way ahead of their time, and they admitted as much in the paper: “Realistically, it will be a long time before the majority of people feel comfortable passing seamlessly between their immediate real world and the reality of tele-embodied worlds.”
Sixteen years later, the comfort they described is still incredibly difficult to engineer.
* * *
Though they represent a new category of electronics, RPDs are really a collection of common parts assembled in new and interesting ways. They require a brain, a mode of locomotion (whether wheels or treads or balls), a means to see and steer, and, of course, a screen, a camera, and some microphones. Among the dozen or so products on the market right now, I wanted to test a range of products, from affordable to expensive, from tricked out to bare bones. I settled on four.
The first to arrive was the VGo, which went on sale in 2011 and made waves last year in a Verizon commercial during the NFL playoffs. Designed in white, with the elegant plastic lines of an early Apple product, the VGo has a sort of evil-robot-emperor styling very much at odds with its somewhat diminutive stature. At four feet tall (an optional 12-inch height extension adds eight pounds and $1,690), it put me, at six-foot-seven, in the unfamiliar position of looking up at everyone, and I took to flashing its LED headlights at people like the laser weaponry of some thwarted Napoleonic alien. The small dimensions of the screen on which my face appeared also led several people to lean down, hands on their knees, the way they might speak to a toddler. I did not feel particularly boss-like in those moments.
Next was Anybots’ QB, a very friendly-looking and much taller device in which I felt far more comfortable. It’s dirt-resistant, with large wheels, and carries a cellular transmitter, which makes it ready for the out-of-doors. I briefly considered steering it into one of our elevators and out onto the street, but I was wary of being stopped by security and terrified of being stolen. The QB includes a downward camera that helps it automatically adjust its path to avoid certain obstacles, which is fantastically helpful. But the lackluster audio and video quality meant everyone instinctively leaned in to hear and be heard. Though they occurred eye to eye, conversations felt like a strain.
To try something entirely different, I visited the offices of Romotive. The company had just launched its first product, a partially Kickstarter-funded rolling iPhone mount that users can drive around. It’s something a grandparent might use with a distant grandchild, chasing him to and fro. It’s cute and packs a lot of features, and the price is right (just $149), but at a height of only a few inches, I wasn’t going to roll around on some conference table, peering up into everyone’s faces.
After three trials, I was starting to feel that although each RPD had its virtues, none imparted the flexibility or dignity I needed as a boss. Then I called Scott Hassan.
* * *
Hassan was an engineer on the original Google technology at Stanford and the founder of seminal robotics maker Willow Garage and is now CEO of Suitable Technologies, whose remote-presence device is called the Beam. When I arrived at his office in Palo Alto, he seemed bewildered as to why I was visiting in meat form, as he called it. Why didn’t I just visit via one of his devices? At first, it sounded like a transparent shtick—the sort of cultish Silicon Valley propaganda CEOs often use to dignify useless cameras, pens, or cars. But as we walked the factory floor, a stream of co-workers rolled past in their own Beams, tossing off casual greetings. I met one young engineer, seated in what looked like a kitchen, who has been traveling the world for a couple of years without missing a day of work. He currently codes from Santa Fe, where he spends his lunch hours skiing. Before that he worked from Hawaii, where he spent his lunch hours surfing.
Hassan uses his product’s name as a verb. He beams to his parent-teacher conferences so he doesn’t have to drive across town. He beams into his kids’ rooms at bedtime when he’s traveling. He beams into his dad’s house to say hello.
He won’t tell me how many companies are testing Beam, and he admits that no one has “bitten big” on his product. But the potential of remote presence devices is, in theory, enormous. New and remarkable efficiencies are possible in almost any industry you can think of. Anybots CEO David Rogan told me that a retail chain was experimenting with a central sales staff that “moves” among different stores via the QB depending on customer flow.
At the office.
But mastering the social stuff is what’s most prickly, Hassan tells me. And the Beam is engineered to create trust. It shuts down the moment it loses Wi-Fi so no one inadvertently rolls away. (In the VGo, I often had to reconnect and repeatedly found myself against a wall far from where I lost vision—the remote-presence equivalent of a blackout drinking episode.) The Beam doesn’t offer a way to record what’s going on, so colleagues won’t censor themselves when it’s around. And it doesn’t include a sensor system to detect and navigate obstacles, the way the QB and VGo do, because Hassan says he believes that to create collegial trust the human operator must be obviously and completely in charge.
Hassan insists that the Beam isn’t a robot. “The word misses the point,” he says. And that’s what helps to make his creation—a vessel for humans—the most functional approximation of remote presence I’ve seen.
My time as a robot boss taught me many things about a successful RPD. First, design and engineering matter. It’s not enough to have a screen and a mic and a speaker and a battery and a motor. The screen must be big and clear and at eye level. The mic must pick up conversation from the most distant person in a meeting. The speaker mustn’t require raising one’s voice. The battery must last a workday. The motor mustn’t sound like the vacuum of an approaching janitor.
When a device meets all those standards—as I found only the Beam did—it allows users to forget (or at least ignore) that it’s there. And that’s the trick. If the thing forces adjustments on the participants, the conversation is at best a novelty and at worst a nuisance. If the thing offers no obvious reason to speak up or bend down, it begins to blend in. The conversation evens out, and eventually, the exchange becomes something close to normal.
And that’s when the true advantages of remote presence begin to crop up. Rather than wait for my colleagues to teleconference me, I can roll off and join them. When a designer wants to show me the layout he’s put on the wall, I can survey it with him in his office. When a casual conversation with a guest yields a great story idea, an editor can grab me and make an introduction. I have my trial Beam for only a few months more, and I’m beginning to dread life without it.
My managing editor, an otherwise fearless person, physically recoils when I approach. “Get it away from me,” she says without humor.The second thing I learned about remote presence is that smart design and engineering won’t make it work for everyone. There are several people in my office who simply cannot handle the Beam. My managing editor, an otherwise fearless person, physically recoils when I approach. “Get it away from me,” she says without humor. (I roll into her cube whenever I can.) One of the Web editors instinctively photographs me whenever she sees me rolling past but won’t actually engage. She’d rather place another screen between us, it seems.
If you ask whether I think remote presence can eliminate physical distance between people, the answer is no. The Beam can’t provide the intimacy of touch, for instance, and you can’t hand anyone a drink. And that’s why remote presence makes sense in a workplace. People who work together are inherently constrained, thank God, by the rules of professional decorum. You shouldn’t be doing any touching at work anyway.
But for most of the Popular Science staff, the Beam provides a tolerable and even useful extension of their boss. Colleagues approach to show me their latest Nerf gun. The high-resolution camera allows me to zoom in on cover sketches. People say hi.
In all-staff meetings, which we typically hold by conference call, it eliminates the awkward anticipatory silences and simultaneous speaking so common with that lower form of communication. Colleagues can see whether I’m about to speak, so conversations flow more naturally. Sure, occasionally our “robot only” Wi-Fi network goes down, but that’s expected. The Beam drops my calls less frequently than AT&T does.
And recently, my job tested the Beam as only an office can. An editor asked whether she could speak with me privately for a few minutes. Normally, she’d have ducked into an office and called me from her cell phone. Instead, I asked whether the Beam was okay.
I followed her down the hall and through an open door and asked her to close it behind us. I quietly rotated to face her and slid forward and back to find a comfortable distance. She sat and began to speak.
I won’t tell you what we discussed, but it was one of the dozens of private conversations a supervisor and an employee might have. Perhaps she was asking for help with a difficult colleague. Maybe she wanted a raise. Or she’d been offered another job. The point is, she could see my expression of interest and concern and could presumably hear the puff I blew from my cheeks in dismay as she laid things out. We talked for a while. It was a subtle and complicated conversation, and yet somehow the Beam conveyed the sentiments we each entrusted it to convey. In less than 10 minutes, we had resolved things. She thanked me and rose. “Was this weird, doing this by robot?” I asked.
“Not at all,” she said after a moment. “I guess I’ve gotten used to it.”
Jacob Ward is the magazine’s editor in chief. This article originally appeared in the November 2013 issue of Popular Science.
An impressive combination of sensors and motors for under six grand, but the small screen at stomach level is the fatal flaw, as the audience leans in awkwardly to see the operator. (A height extension improves things but raises the price.) Also, it’s impossible to go anywhere unannounced: Its locomotion is audible from 15 feet away. $5,995 (plus annual subscription fee)
The QB is easy to drive. Its wheels can handle bumps up to 2.5 inches tall, and assisted steering self-guides it through doorways and around obstacles. Its telescoping neck ranges from 3′ to 6’5”, and it can handle moisture and dirt. Sadly, poor image quality and mixed audio can stifle conversation, and its somewhat goofy appearance robs it of dignity. $9,700 (plus connection fees)
A toy rather than an office solution and as much about the software as it is about the mechanical package, the Romo is for entertaining a child or connecting that child with a distant friend or family. That said, it’s impressive at this price. Next, the company is working toward giving it computer vision so it can navigate obstacles on its own. $149 (iPhone not included)
With a monitor big enough to show a life-size face, dual-band radios for seamless handoffs between Wi-Fi stations, and truly lifelike audio at both ends of the conversation, the Beam is the best (and most expensive) of the bunch. It’s the little things: a quiet, fast motor, an eight-hour battery life, and reminder emails if you park it somewhere off-charger and forget. $16,000 (plus connection fees)