On Saturday, April 23, hundreds of people filed into MIT’s Stata Center, home to the Computer Science and Artificial Intelligence Laboratory (CSAIL). They’d come to the sixth TEDxMIT event to hear some two dozen speakers—MIT students, faculty, alumni, and corporate CEOs—present their views on the “superpowers” that might soon enhance human performance and capabilities.
After some rousing choruses turned in by the Logarhythms, MIT’s undergraduate acapella group, CSAIL director Daniela Rus took the stage. Rus—who organized the event with John Werner (an MIT fellow) and the TEDxMIT student group—began by talking about whales, a subject that was probably not anticipated by most audience members. In 2009, Rus struck up a collaboration with the biologist Roger Payne, one of the world’s preeminent whale experts. Payne had established a laboratory on Argentina’s Valdes Peninsula, where he was studying the lifespan of whales and carrying out the first long-term census. In that same year, Rus encouraged two former students to develop a drone that could help with whale surveillance—a tool that almost instantly “expanded [Payne’s] reach beyond anything he imagined.”
That kind of expansion in range and scope was one of many examples of technology-abetted capabilities that were presented at the TEDxMIT symposium. And speaking of range and scope, Rus noted, drone technology has, for the past year, been providing us with unprecedented views of the Martian surface—thanks to the ingenuity of NASA’s Ingenuity drone/helicopter.
Visually-impaired people, aided by sophisticated sensors embedded into wearable devices, may soon be able to see through walls and peer around corners, Rus added, thereby gaining a much “richer experience of the world than they can get through a white walking stick.” The future is indeed “extraordinary,” she said, and marvels of the sort she alluded to were unveiled in subsequent talks.
In a conversation with the writer Patrick Kane, Hugh Herr—who co-leads MIT’s Yang Center for Bionics—discussed the future of prosthetic devices. Herr—a double amputee who lost both legs during a mountain climbing accident when was 17—predicted that “in the future, prosthetic limbs will no longer be separate devices” but will instead act like biological limbs that are “intimately integrated into our tissues.” He and his colleagues are devising a new approach to prosthesis in which a titanium shaft penetrates and attaches to the residual bone. Wires running through the shaft can communicate directly with the body’s nervous system, providing sensory and mobility functionality much closer to that of a natural limb, while improving the lives of millions who require protheses.
Herr’s lab is also pursuing “human augmentation.” Dephy, a company Herr cofounded in 2016, has produced a bionic shoe that can enable people to run faster and jump higher. “In 20 years,” he mused, “we’re going to look back and say, ‘Remember that time when shoes didn’t do anything?’”
There was also a time, not so long ago, when robots couldn’t do much either. But Marc Raibert, a former MIT professor and co-founder of Boston Dynamics, has spent years trying to improve their abilities and usefulness. He brought a prop with him—a scene-stealing, four-legged robot called Spot—to illustrate what modern robots are capable of. Spot—about four feet long and 70 pounds (with battery)—nimbly pranced up and down the steps of the Kirsch Auditorium. This entertaining talk continued with a video clip, which showed Spot dancing alongside Mick Jagger to the Rolling Stones hit, “Start Me Up.”
The work with dancing robots is “purely for fun,” Raibert said, but Spot robots are already carrying out more practical tasks. Roughly 1,000 of them are “out in the world,” checking for gas leaks at a British Petroleum refinery and inspecting high-radiation zones at the Chernobyl nuclear power plant, among other chores. Another Spot is currently patrolling an archaeological site in Pompeii, Italy, to ward off souvenir hunters. The robots have been deployed in various risky environments like high-voltage electric facilities that have areas where humans cannot go. “If you fry a robot, who cares?” Raibert asked (although many would not consider the $75,000 price tag trivial).
MIT Electrical Engineering professor Tomás Palacios did not have a dancing robot to show off, but he did regale the crowd with tales of new materials that can confer abilities far beyond what science fiction authors ever envisioned. The first material he extolled was graphene, consisting of a single layer of carbon atoms arranged in a honeycomb lattice. “It’s actually the strongest material we have,” he said, “100 times stronger than steel and much, much lighter. And it will very soon change the way we interact with the world.” His MIT group is building a new generation of sensors—consisting of a layer of graphene placed on a glass substrate—which will have the capacity to smell and taste, “two senses that machines have not been able to conquer yet,” Palacios said.
Another material he touted is ordinary air, “but air with a twist. What if we take computers and make them so tiny, they can float in air and start measuring everything around us?” It sounds crazy, Palacios admitted, but that is actually being pursued in his MIT lab. And to bring home the point, he held up a small vial containing roughly 100,000 miniscule computer chips. “We are not far from the day when these tiny computer chips will be everywhere, interacting with the world around us, or inside us, giving us new superpowers we cannot even imagine today.” It is hard to grasp, he said, until one remembers that today’s computers, and the amazing software running on them, are all based on a single material—silicon.
MIT graduate Daniel Kokotov—a cofounder of Rev.com—spoke about the subject of speech itself. Kokotov called the ability to understand the human voice “the original superpower.” He charted progress in the field of speech recognition, from work at Bell Labs in the early 1950s to present technology, which—thanks to the advent of deep neural nets—is “almost good enough to being actually good enough.” A steady stream of advances, he added, is getting us closer to the “’speakularity,’ defined as the moment when speech recognition becomes free, fast, and decent.” Achieving that level of proficiency in speech recognition and transcription technology could be life-changing for many people, including those who are hearing impaired, “who will suddenly gain access to a whole new set of information.”
Breakthroughs in artificial intelligence, which have yielded new capabilities, are also contributing to a new kind of problem—a surfeit of information that will require constant filtering to separate the signal from the noise. In contemplating the future of AI, Kokotov drew a comparison with the game “kiss, marry, or kill.” Initially, one might be so excited by a new development as to be tempted to select the “kiss” option. But upon reflection, one might remember that things don’t always turn out as well as expected, so maybe we ought to kill it. That’s not an appealing, or realistic, option either, according to Kokotov. “AI is here. It’s too useful and too important to kill. That leaves us with our third choice, and it’s actually the best choice of all—to marry. Like any good marriage, it will require a lot of work. But it’s work we have to do.” And based on the breadth of ideas presented throughout the day, there is no shortage of work to be done, and no limit to humanity’s quest for superpowers.