“Jim is developing a specialized AI tool that we think will be the best in the world,” Musk said at a party organized by Tesla.
“Jim” is Jim Keller, a well-known chip engineer who Tesla recruit soldiers last year to lead Tesla’s Autopilot hardware efforts. Keller has held key positions at both AMD and Apple.
The stakes for Tesla are high. The first version of Tesla’s Autopilot is based on chips provided by Mobileye, a leading vendor of autonomous car technology that is now owned by Intel. But the two companies had a bitter falling out last year, and Tesla built a new version of Autopilot without Mobileye’s help. The latest version of Autopilot relies on chips from Nvidia.
But Tesla is apparently looking to reduce its reliance on Nvidia—or any other third-party supplier—for its autonomous car technology. This is consistent with Elon Musk’s broader philosophy—followed by both Tesla and SpaceX—to build as many components in-house as possible.
Tesla needs to move quickly. Musk said April who expects that fully self-driving cars will be available in “about two years.” Waymo was way ahead of its time, with fully self-driving cars already on the roads in the Phoenix area.
Yet Tesla’s Autopilot division has suffered from significant turnover, raising questions about whether the group can deliver on Musk’s ambitious timeline.