When your company absolutely, positively needs a chip that kicks a lot of butt, you bring in Jim Keller.
The legendary CPU architect earned his chops helping with the design of AMD’s K7 processors before moving onto the K8 core, a.k.a. the first 64-bit consumer desktop chips. After that, Keller eventually wound up at Apple, where he spearheaded the creation of Apple’s A-series processor, which made their debut in the iPad and iPhone 4. Then he returned to AMD as the mastermind behind the company’s forthcoming Zen processor architecture—a design that promises to battle Intel for high-end computing supremacy after years of AMD’s chips being relegated mostly to budget and mid-range PC designs.
Then, last September, Keller abruptly left AMD (though the timing indicates that it was after the Zen architecture work was done). The geek world immediately burst into speculation. Where would he land? Samsung? Intel? ARM?
Late Thursday, we learned the answer: Tesla, where he’ll be heading up the company’s autopilot endeavors, according to a statement sent to Electrek:
“Jim Keller is joining Tesla as Vice President of Autopilot Hardware Engineering. Jim will bring together the best internal and external hardware technologies to develop the safest, most advanced autopilot systems in the world.”
Some of the biggest names in the technology world are devoting serious time and resources to infusing automobiles with smarts: Google’s self-driving cars can frequently be seen cruising California streets, Nvidia’s doubling down hard on automotive processors, Uber’s testing autonomous vehicles with top university researchers, and even Apple’s been strongly rumored as having an interest in the field.
But creating self-driving cars isn’t easy. There are a ton of variables an autonomous needs to keep track of out on the wild streets, with data zooming in from numerous sensors of various types, and it’s literally a matter of life and death that the system doesn’t fail—both for you as the driver, and for children chasing balls into streets. Currently, Tesla’s Autopilot mode only really works on highways, Jalopnik reports, as it “needs to be able to see clear lines marking out its lane before it can work.” Think of it like an advanced form of cruise control.
Why this matters: We’re still a long way from truly autonomous vehicles being a regular sight on city streets, and all the top names in tech are racing to be the first to make that breakthrough. By bringing in “the fixer”—Keller’s designed many of the most revolutionary processors of the past decade-plus—Tesla’s putting its money where its mouth is. We’ll know in the next few years whether Keller’s still got the magic touch.
This story, "Guru behind Apple's A-series, AMD's Zen chips takes over Tesla's Autopilot hardware" was originally published by PCWorld.