scorecard
  1. Home
  2. tech
  3. Tesla hired a top AI expert to lead a critical aspect of Autopilot - here's what we know

Tesla hired a top AI expert to lead a critical aspect of Autopilot - here's what we know

Danielle Muoio   

Tesla hired a top AI expert to lead a critical aspect of Autopilot - here's what we know
Tech3 min read
Tesla autopilot

Tesla

A visual rendering of a Tesla driving with Autopilot activated.

Tesla has completely shaken up its Autopilot team, and its newest addition is Andrej Karpathy, the new director of artificial intelligence and Autopilot vision.

Karpathy will replace David Nistér, Tesla's VP of Autopilot vision, who left the company quietly in March to join NVIDIA, according to his LinkedIn. Karpathy will play a critical role in advancing the company's Autopilot system to the point where a Tesla can drive itself from Los Angeles to New York before the year ends.

Karpathy is considered a leading expert in computer vision. He received a pHd in machine learning and computer vision from Stanford University. During his time at Stanford, Karpathy designed the University's very first course on Deep Learning.

For reference, deep learning has become one of the most important and relevant AI disciplines in the last several years. To put it simply, deep learning is a practice where computers use neural networks, or layers of data, to classify things like images or speech.

Karpathy, for example, has used deep neural nets to teach computers to not only identify a cat is in a picture, but that a cat is orange, spotted, and perched on a skateboard.

That kind of image recognition is key for self-driving cars, which have to identify different obstacles quickly and react accordingly. That makes Karpathy a perfect fit for Autopilot vision.

Karpathy has mostly worked in academia, but he joined Tesla's artificial intelligence group OpenAI last September as a research scientist. Tesla CEO Elon Musk formed OpenAI last year to create safety standards for AI to ensure it doesn't turn evil.

Tesla autopilot

Tesla

Tesla Autopilot.

As a Tesla exec, Karpathy said he will look to apply his work with convolutional nets to Autopilot.

Convolutional nets analyze images by breaking them down pixel-by-pixel. In 2015, Karpathy used convolutional nets to create a computer program that can judge the perfect selfie.

"At least on the short-medium term, the focus will be much more applied than what I've done at OpenAI, and will use techniques more along the lines of ConvNets trained with supervised learning, at scale, and deployed on an embedded system," Karpathy wrote in a Reddit thread.

Karpathy will work report directly to Musk in his new role - but that's nothing new for Tesla's most recently Autopilot hire.

Karpathy has interacted with Musk on a semi-regular basis since joining OpenAI, he said in a 2016 Quora forum. In that forum, Karpathy gave some insight as to what it was like working with Musk:

"I thought he might be super-busy or have this perpetually distracted demeanor, but instead I found him to be very nice, personable, light-hearted and fully invested in the conversation when he talks with you. In fact a little too much so. He asks a question, puts his fingers together, looks you directly in the eyes without moving at all and gives you this complete, undivided attention, it's actually slightly intimidating."

Get the latest Tesla stock price here.

NOW WATCH: We drove a brand-new Tesla Model X from San Francisco to New York - here's what happened

Please enable Javascript to watch this video

Advertisement

Advertisement