Washington Post: Welcome to the dawn of the age of robots
Growing up, I believed that very soon we would all have robots like Rosie, from “The Jetsons,” cleaning up after us. For those out there who are too young to remember “The Jetsons,” Rosie is the family’s domestically adroit robot maid. After all, why should anyone waste time doing dishes or folding clothes? I longed for a droid friend like C-3PO, Luke Skywalker’s robot buddy from Star Wars.
But Rosie never came. All I got was a Roomba — an automated vacuum cleaner that goes round and round. And the closest thing to C-3PO on the mass market is Siri, who is cute, but not terribly clever, and, beyond superficial capabilities, conversationally challenged. Even in this exponential era, we seem far away from robots that can talk to us, help us keep our home clean, or perform unstructured tasks.
If you watched the videos of the robots from the DARPA Robotics Challenge, you might believe that we may never see Rosie in real life.
The robots were required to navigate an eight-task course simulating a disaster zone. Tasks included driving alone, walking through rubble, tripping circuit breakers, turning valves, and climbing stairs. Despite the best efforts of the world’s best roboticists, the robots were slow and clumsy. They kept falling over, and moved at the speed of molasses. In fact, no robots can quickly perform tasks we humans consider mundane. Picking a towel up from a laundry basket and neatly folding it is something robots still struggle to do. So it appears that these droids are not yet a threat to human jobs, let alone saviors for countries, such as Japan, that face significant demographic challenges due to their rapidly aging populations.
But after watching the DARPA Challenge and observing the rapid advances of computing, artificial intelligence, and sensor technologies, I see Rosie being very close to reality. These technologies are all advancing at exponential rates. And exponential technologies can be deceptive. Things move very slowly at first, but then disappointment turns into amazement.
That is what I believe will happen with robotics over the next five to 10 years. Amazing progress is being made in the underlying hardware and software. In part, that’s because costs have plunged. The single-axis controller, a core part of most robots’ inner working, has plunged in price from $1,000 to $10. And according to Rob Nail, the chief executive of Singularity University, the price of critical sensors for navigation and obstacle avoidance has fallen from $5,000 to less than $100.
It is notable, too, that three teams — and incidentally, three different designs — completed the DARPA Challenge course. That’s better than in the DARPA Grand Challenge in 2004, in which no self-driving car came close to finishing. Now, only 11 years later, self-driving cars are legal in eight states and a common sight on the streets of the Bay Area.
A clue about how fast this can happen can be found in the world of drones. Less than a decade ago, drones with real capabilities were expensive, rather large, and very hard to fly. Today, consumer drones are starting to match the capabilities of the fancy Predators and other custom-made, expensive unmanned aerial vehicles flown by the U.S. military. Companies such as Shenzhen-based DJI Innovations are selling drones with the same capability as the original military ones, for less than $1,000.
These Chinese firms are, in turn, competing with even cheaper drones, created by amateurs around the world, who share their designs for free in communities online. It’s safe to say that drones are the first technology in history of which the toy industry and hobbyists are beating the military–industrial complex at its own game. These personal drones can do everything that military drones can, aside from blow up stuff.
That drones have achieved autonomy and low cost before robots is not surprising. It takes serious coding to have a semi-autonomous bipedal robot stay on its feet, not to mention detect its surroundings. Moving around in a cluttered, ever shifting environment is precisely the sort of task that humans learn to do from very young ages but robots struggle to replicate, because it is computationally expensive. (Folding clothes and doing dishes are two other examples of these types of tasks.)
For voice recognition, we are already pretty close to C-3PO-like capabilities. Both Apple and Google use artificial intelligence to do a reasonably good job of translating speech to text, even in noisy environments. No bot has passed the Turing Test yet, but they are getting closer and closer. When it happens, your droid will be able to converse with you in complex, human-like interactions.
The computational power necessary to enable these robots to perform these difficult tasks is still lacking. Consider, however, that in about seven or eight years, your iPhone will have the computational ability of a human brain, and you can understand where we are headed.
Robots will be able to walk and talk like human beings.
What are presently halting steps moving up stairs will, in the next DARPA challenge, become sure-footed ascents. The ability to merely open a door will become that of opening a door and holding a bag of groceries and making sure the dog doesn’t get out.
And, yes, Rosie will replace lots of human jobs, and that is reason to worry — and cheer. The good part is that Rosie will do all the things we used to have to do but do them faster and cheaper. Everyone will have more time to enjoy their lives or to focus on their creative passions. That sounds pat, but I sincerely believe that a world full of Rosies could translate into richer opportunities and less stress for us humanoids. And no one will ever have to fold laundry again.