Speaking at an earnings call earlier this week, Tesla CEO Elon Musk decided to “let the cat out of the bag” by announcing that the company has been building its own artificial intelligence chips which will power the self-driving cars of the future.
According to a leading tech portal, Pete Bannon, director of the Hardware 3 project, said, “We had the benefit of knowing what our neural networks look like, and what they’ll look like in the future…”
The company has been relying on Nvidia’s Drive platform. But now, Musk reportedly said that Tesla’s specialised chip will be able to crunch out 2,000 frames per second, as opposed to Nvidia’s hardware handling about 200 frames per second.
“The key is to be able to run the neural network at a fundamental, bare metal level. You have to do these calculations in the circuit itself, not in some sort of emulation mode, which is how a GPU or CPU would operate. You want to do a massive amount of [calculations] with the memory right there,” added Musk.
2/ This is true circa 2016, but as of 2018, Nvidia Volta chips have tensor units that's very much custom to deep learning and dozens of companies are doing custom DL processors.
— James Wang (@jwangARK) August 1, 2018
4/ Apple's custom SoC effort is a key reason why the iPhone leads in performance and go-to-market. Tesla's AI chip effort will do the same—they will be able to deliver performance & features *ahead* of other auto makers that must wait for the next chip from Nvidia/Intel.
— James Wang (@jwangARK) August 1, 2018
Three years ago, Tesla CEO Elon Musk predicted that autonomous driving would be ready in two years. He had clarified that it was dependent on regulatory approval and software validation. Recently, Musk disclosed on Twitter that the company plans to roll out autopilot software, Version 9, which will enable full self-driving features. He however added that Tesla will activate a subset of self-driving car featres, but did not promise that their vehicles would be fully autonomous.
That issue is better in latest Autopilot software rolling out now & fully fixed in August update as part of our long-awaited Tesla Version 9. To date, Autopilot resources have rightly focused entirely on safety. With V9, we will begin to enable full self-driving features.
— Elon Musk (@elonmusk) June 10, 2018
In 2017, Tesla’s autopilot-enabled Model X crashed into a barrier in Mountain View, California. Reports had suggested that Tesla’s autopilot itself was at fault. Another report said that in 2016, Tesla’s autopilot-enabled Model S was responsible for a man’s life as the AI mistook a white lorry for a clear sky. The driver could not correct it in time, and it resulted in an accident.
The post Tesla To Build Its Own AI Chips For Self-Driving Cars, Announces Elon Musk appeared first on Analytics India Magazine.