ON THE THIRD FLOOR of a generic brick office building in Boston’s Fort Point, Baxter nods, a quick up-and-down motion to acknowledge that he understands the task being asked of him. He needs to perform a simple movement: Pick up a cardboard square from the table, then put it down a few feet away, the kind of mind-numbing chore a human can stand to repeat only so many times. For Baxter, however, it’s something entirely different.
That’s because Baxter isn’t a human — he’s a 300-pound, 6-foot-tall, two-armed robot created by Rethink Robotics, a Boston startup founded by one of the people behind the Roomba vacuum cleaner. Roomba was designed to free its users from domestic drudgery, tirelessly cleaning the floors so people had a bit more free time at home. Rethink has designed Baxter for the industrial workplace, but the goal isn’t all that different. The so-called collaborative robot is meant to improve productivity on assembly lines while freeing human co-workers for more involved tasks.
Factory robots have been around for more than half a century, but new collaborative robots — also known as “cobots” — are a different breed. Rather than a massive, dangerous machine cordoned off by safety cages, Baxter, released in 2012, is stationed out in the open working alongside people. There, the cobot can do anything from moving plastic pieces into milling machines to loading six-packs of beer into crates.
On top of that, programming Baxter doesn’t require a computer-science degree. To get the cobot to pick up the cardboard square, Rethink marketing executive Eric Foellmer literally takes Baxter by the arm and guides him through the sequence of motions. Baxter nods to acknowledge the new program, then starts in, the expressive eyes — programmed into an iPad-like display that functions as his face — betraying not the slightest hint of boredom. The process takes less than a minute. Whenever he wants, Foellmer can reprogram Baxter just as easily.
Baxter’s $25,000 price tag and ease of use make the cobot an affordable plug-and-play solution for many small- to medium-sized businesses, the kind that haven’t been able to use robots before. The promise of that market has helped attract more than $100 million in outside investment to Rethink. In April, the company closed on a $40 million funding round that included GE Ventures, Goldman Sachs, and Amazon founder Jeff Bezos’s Bezos Expeditions, one of Rethink’s earliest investors.
And now Rethink has unveiled a little brother for Baxter designed to capitalize on the growing market. His name is Sawyer, and he’s a one-armed cobot weighing just 42 pounds. Able to work on items like circuit boards that require finer motor movements, he’s sleeker, quicker, and more dexterous than his big sibling. Rethink’s cobots are growing up quickly. But as in any adolescence, things might not be easy.
ALTHOUGH THE WORD “ROBOT” wasn’t introduced until 1921 by a Czech playwright adapting a word for forced labor, the idea of automatic humans goes back centuries. Leonardo da Vinci even sketched out designs for one — complete with waving arms and an articulated head — around 1495. It took until 1961 for the world to see its first industrial robot, named Unimate, which made its debut in a General Motors plant in New Jersey. That type of robot would go on to displace untold workers in the auto industry and elsewhere.
But Rodney Brooks, Rethink’s 60-year-old founder and chief technology officer, has advocated for a kind of robot that can actually help workers. In 2003, Brooks, who once led the MIT artificial intelligence lab and cofounded Bedford-based robot maker iRobot, gave a TED Talk titled “How Robots Will Invade Our Lives.” In it, he talked about robots that deliver medicine and take away dirty dishes and bots that let soldiers safely disable roadside bombs in Iraq and Afghanistan. Then there was Roomba, which had just come on the market at the time of Brooks’s talk. Few people could have imagined wanting a vacuuming robot at home back then, but Roomba has gone on to sell more than 10 million units.
Brooks has long envisioned robots and humans cooperating in the workplace, and his future is becoming the present. At Praxis Packaging in Grand Rapids, Michigan, Baxter might use his arm, custom fit with four suction cups, to unpack cans of infant formula for a couple of hours, then switch in minutes to working with boxes of vitamins for Costco. “We run multiple orders in a given shift, so it’s not uncommon that Baxter could be changed over between a few tasks in a given day,” senior vice president Scott Hanmer says. The company has four Rethink robots and more on the way.
At Rapid-Line Inc., also in Grand Rapids, one of the nation’s largest metal fabricators, the focus is on quick-response manufacturing — speedily meeting the needs of clients, whatever their order. The company now uses its two Baxter robots to feed and remove parts from a computer-controlled milling machine that fabricates steel components for office furniture and panels.
Rethink designed its cobots, especially the smaller, one-armed Sawyer, to fit into human-sized work spaces, which limits the alterations a company needs to make when it purchases one.
They can adjust to changing conditions as well. Let’s say Sawyer’s job is to take a circuit board from one spot and insert it into a casing a few feet away. If one casing came down the manufacturing line a few degrees off, a traditional robot wouldn’t be able to adjust, causing a costly delay. A Rethink robot, however, uses feedback mechanisms in its gripper to “feel” its way into the casing, successfully placing the circuit board in the correct spot.
Both robots have animated “eyes” on their faces. They not only give the cobots a little personality, but also provide directional clues to the humans working alongside them. The eyes look to the left, for instance, before the arms move that way. Baxter also has built-in sonar and camera sensors to detect humans when they enter his space. Unlike industrial robots, he stops if he bumps into a person, helping to avert injury. As Brooks has said, Baxter’s “got a little common sense.”
Baxter and Sawyer, however, aren’t the only collaborative options around. In fact, they face stiff competition. In May, Massachusetts-based Teradyne, a maker of computer equipment, agreed to buy the Danish company Universal Robots for $285 million. Universal released its first robot for this kind of work in 2009 and became profitable by late 2010. It had $38 million in revenue in 2014, more than 70 percent higher than 2013. According to Scott Mabie, Universal’s general manager of the Americas, his company has shipped 5,000 cobots to firms around the world, including more than 1,000 in North America.
For its part, Rethink has sold somewhere north of 1,000 robots. And that’s a problem, says Frank Tobe, editor and publisher of The Robot Report, a leading website on the robotics industry. Tobe acknowledges that Rodney Brooks has done more than anyone to create the category of collaborative robots, but he believes Rethink’s products haven’t yet lived up to Brooks’s vision. “Universal — whose robots work flawlessly and are the easiest to train — has been the principal beneficiary of his marketing efforts,” Tobe says.
Rethink has a lot riding on Sawyer, its smaller, faster unit. Set to ship this month, it is designed to appeal to companies in the United States and in the Far East, where collaborative robotics as a market segment is just beginning to take off. “Sawyer may be a lifesaver for Rethink,” Tobe says. “It’s conceivable that the new Sawyer can sell in large enough quantities in Asia to help keep Rethink solvent.”
Eric Foellmer rejects the assessment that Rethink is struggling or has anything to worry about from competitors. “We have differentiated technology, strong leadership, and great products,” he says. “In short, we’ve been around since the beginning of the collaborative robot industry and we will continue to pioneer the space for many years to come.” The company has had layoffs in the past but will complete the goal of doubling its full-time workforce from 50 to 100 by the end of the year. And the New York Times recently named Rethink to a list of 50 technology companies likely to be valued at $1 billion or more in years to come.
Foellmer says Sawyer will let the company compete more effectively with Universal, which does most of its business outside the United States. “Since Sawyer will be available throughout North America, Europe, and Asia,” he says, “that will level the playing field significantly in terms of overall volume and share moving forward.”
As for the market size for collaborative robots, it’s estimated at about $100 million worldwide. But an analysis from Barclays predicts it will grow to as much as $3 billion in just three years. “The market size is at least the same as regular robots [estimated at up to $25 billion], maybe even larger,” says Henrik Christensen, a robotics expert at the Georgia Tech College of Computing. Christensen expects the demand for this sort of robot to grow 60 percent to 100 percent a year for the foreseeable future.
BUT WHAT ABOUT our particularly modern nightmare: that plastic-and-steel robots will take the jobs of flesh-and-blood humans?
Last year, the Pew Research Center canvassed some 2,000 technology experts on the subjects of artificial intelligence, robots, and the future of jobs. Close to half predicted “a future in which robots and digital agents have displaced significant numbers of both blue- and white-collar workers.” The result would be as apocalyptic as any science fiction movie, these experts believe, “with many expressing concern that this will lead to vast increases in income inequality, masses of people who are effectively unemployable, and breakdowns in the social order.”
It’s hard to see how a vibrant collaborative robotics sector wouldn’t affect workforce numbers. It will never be cheaper to pay a person to move an object from point A to point B than it will be to purchase Baxter. A monthly recalibration, plus a little extra electricity, is much more cost-effective than paying health insurance and a living wage and offering a lunch break.
The people behind this new wave of collaborative robots, however, argue that humans have nothing to worry about from their new colleagues. “Is there a concern? Sure,” says Universal’s Scott Mabie. “But the reality is that we don’t eliminate people in most, if not all, cases. What we are finding is that the people who have to work with the robot every day are the ones who are finding the new applications.”
So perhaps there is a version of the future where the two kinds of employees work together. The bot performs the drudgery while the human has a chance to take on more ambitious tasks. All the while, the cost of manufacturing comes down so companies can afford to hire more workers. Perhaps they even help to spur the growth of the manufacturing sector in the United States, which fell from 25 percent of all jobs in 1950 to 9 percent in 2013, according to a McKinsey Global Institute report.
As Rethink likes to point out, its factory clients usually don’t lay off any human workers because of Baxter. “We’ve found that it’s really been very complementary to those production workers on the line,” Praxis Packaging’s Scott Hanmer says. “It’s complementing a person. It’s not a labor replacement.”
Some humans have even found themselves growing fond of the big-armed robot with the expressive face. Employees in about a quarter of the workplaces with Baxter take to customizing him. They’ve given him a security badge or dressed him in a company shirt or even hung a lei around his neck. At Praxis, each Baxter has an ID badge. Hanmer, the vice president, refers to the cobots as “our friend Baxter.” Maybe it’s a bit creepy to call a robot a buddy, but it could also be the sign of a better, more sustainable work future. One where Baxter and friends fit right in.
See Rethink Robotic’s latest cobot, Sawyer, at work: