Just consider how fast the brains of automobiles are evolving.
Last year, Nvidia Corp., the Silicon Valley graphics and processor company founded by Taiwan-born Jen-Hsun Huang, developed the Jetson TX1 "mobile supercomputer." Roughly the size of a credit card, the Jetson promised to give autopiloted vehicles, like drones, the ability to recognize faces and remember what they have seen along the road.
But then, Huang's company leapfrogged its own achievement. Nvidia -- a company created in 1993 to put cool graphics into video games -- this year unveiled a vastly more powerful "deep learning" automotive device called the Drive PX, leaving last year's Jetson cards to retail to hobbyists on Amazon.com for only about $197.
Drive PX will have the brains to steer autonomous cars through situations in which they need to read street signs and recognize the difference between an ambulance and a FedEx delivery truck or a police car and a taxi. It will know that a ball in the street probably means that a child soon will follow or that a partially opened door on a parked vehicle likely will yield someone stepping out of it, says Danny Shapiro, Nvidia's senior director of automotive.
Nvidia is working with automakers, Tier 1 suppliers and universities to apply its processing products into upcoming vehicles and component systems.
"We believe that features like these will be mandated in the coming years because they save lives," Shapiro says. "I see a time when this is standard technology."