James Kuffner, the head of Google's robotics division and one of the original team of ten who started its self-driving car work, has left the company for a job at Toyota's $1 billion research institute in Silicon Valley.
His departure will come as a blow to the search and advertising giant, which has been plowing forward with a number of robotics projects including the self-driving car, which it hopes to offer for public use some time next year.
“It’s becoming clear that in the next phase of machine learning, access to lots of data to find and fix corner cases and to make a robust system is going to be very important, and I think Toyota is very well positioned to do that with its resources and its data,” Kuffner said in an interview at the CES expo in Las Vegas on Tuesday.
It was his second day on the job.
Kuffner joins Gill Pratt, former head of DARPA's Robotics Challenge, and a number of other prominent robotics and artificial intelligence researchers at the Toyota Research Institute (TRI).
Toyota's billion-dollar investment in the center was only announced in November, but the institute has already opened for business in two locations: one at the Stanford Research Park in Palo Alto and one in Kendall Square in Cambridge. They were chosen for their proximity to Stanford University and the Massachusetts Institute of Technology.
TRI's mission is to take fundamental robotics research into products that can benefit all of society. One of the loftier visions is the development of cars that are incapable of crashing due to their complex AI systems, but the institute will also look at home-help robotics for the elderly and other projects.
Pratt, who will lead the lab, said one of its chief goals is the development of systems that can handle and react to occasional complex and unexpected situations on the road.
"In truth, we are a long way from the finish line of fully autonomous cars," he said. "Most of what has been accomplished collectively by all of us working in this field has been easy. We need to solve driving when it’s hard. It’s that hard part that TRI intends to address."
Kuffner said Toyota's massive investment in AI and robotics research persuaded him that the company was serious about its work and the difference it could make to society.
So how long until some of this starts paying off?
To be sure, the goal of a completely self-driving car that handles any situation and cannot crash is some distance away, but Kuffner said a lot will be possible in the next few years.
"We're actually closer than people think to having self-driving cars on the road," he said. "It is an evolution. There is a continuous spectrum between full manual control and full autonomous control, and there's going to be phased deployments."
He cited some of the current technologies making their way into cars, such as lane assist and adaptive cruise control.
"These safety features are creeping into lots of cars you can buy today, and the pace is increasing, so I think people will be happily surprised in the next five years at how our vehicles have changed."