Elon Musk confirmed last week that Tesla is working to create custom AI hardware for autonomous driving vehicles, reports CNBC. Tesla currently uses Nvidia's graphic cards in Autopilot hardware.
Jim Keller, VP of hardware for Tesla since 2016, was with Musk at the announcement. Keller's hire, reports of a working relationship with AMD and reports of more than 50 Tesla employees working on AI chips under the chip maker's direction sparked months of speculation Tesla was working on this project, according to a CNBC report.
- The move to rethink how Tesla integrates AI into self-driving systems has been ongoing. Musk hired deep learning expert Andrej Karpathy as the Director of Artificial Intelligence in June, reports MIT Technology Review. Karpathy previously worked at OpenAI, a non-profit AI research company Musk co-founded.
Elon Musk, hat salesman extraordinaire, is at it again — despite a renowned wariness of AI.
With vast resources for R&D, the automobile industry is pulling AI experts away from traditional tech companies in pursuit of autonomous vehicles. The competition is subsequently ramping up salaries, making AI development even harder for SMBs and companies with limited resources.
AI is already being deployed in automobile test cases across the country. Law enforcement has teamed up with tech companies to use AI to identify passing cars, pedestrians and objects through car and body cameras. Yet AI is still not at the point where humans can fall asleep in the backseat as their cars take them to work.
But tools to train these AI systems are getting better. IBM recently released a data set identifying verbs through one million three-second videos. Democratized AI tools are affording AI applications to a broader customer base.
But knowing Elon, this is definitely worth keeping an eye on.
It won't even need to ask you most of the time— Elon Musk (@elonmusk) October 21, 2017