Photo: Vimeo screenshot
Somehow, the talk about Autopilot has dialed down a little over the past few weeks, and that is despite the fact that Tesla claims all the cars it will produce from now on will have the necessary hardware for Level 5 autonomy installed.
At the center of that hardware is NVIDIA's Drive PX2 supercomputer, the processing unit that replaces the one provided by Mobileye in the older cars. It's too early to credit Tesla with anything to do with NVIDIA's excellent financial results in the last quarter, but the Santa Clara-based company registered a $2 billion revenue on a $1.7 billion forecast. Its earnings followed a similar path, Electrek reports.
NVIDIA CEO Jen-Hsun Huang followed up the excellent news with a conference call in which the discussion quickly veered toward his company's involvement in the development of autonomous technology. According to Huang, building self-driving cars is not a "detection problem," but an "AI problem."
Realizing that sounds like he's putting a nail in his own coffin, he went on by saying it will all be solved in 2017, presumably when Tesla's Autopilot 2 will begin working at full capacity. And if you think about it, that makes perfect sense. A human driver only has his two eyes and ears to probe the environment with, whereas Teslas have no less than eight cameras offering a 360-degree view around the vehicle. Surely, that's better than our narrow field of view. The difference lies in what happens once this information reaches the central processing unit, and that's where Huang thinks lies the key.
The price of NVIDIA's computer remains a mystery, but during the call Mr. Huang did mention a "few thousand dollars." However, looking at the price Tesla is asking for its Enhanced Autopilot ($5,000) and the Full Self-Driving Capability ($3,000), you can get a rough idea. Still, considering the hardware goes into every vehicle, whether the owner buys the two packages or not, Tesla must have a pretty high profit margin, even though Musk did say he expects everybody to unlock those features at one point. In the meantime, even if dormant, the hardware will still collect information for Tesla, so even if not activated, it's still valuable.
Predictably, Mr. Huang is happy about Tesla's decision to release this technology so early: “And I think what Tesla has done by launching and having on the road in the very near-future here, a full autonomous driving capability using AI, that has sent a shock wave through the automotive industry. It’s basically five years ahead. Anybody who’s talking about 2021 and that’s just a non-starter anymore. And I think that that’s probably the most significant bit in the automotive industry. I just don’t – anybody who is talking about autonomous capabilities in 2020 and 2021 is at the moment re-evaluating in a very significant way.”
NVIDIA CEO Jen-Hsun Huang followed up the excellent news with a conference call in which the discussion quickly veered toward his company's involvement in the development of autonomous technology. According to Huang, building self-driving cars is not a "detection problem," but an "AI problem."
Realizing that sounds like he's putting a nail in his own coffin, he went on by saying it will all be solved in 2017, presumably when Tesla's Autopilot 2 will begin working at full capacity. And if you think about it, that makes perfect sense. A human driver only has his two eyes and ears to probe the environment with, whereas Teslas have no less than eight cameras offering a 360-degree view around the vehicle. Surely, that's better than our narrow field of view. The difference lies in what happens once this information reaches the central processing unit, and that's where Huang thinks lies the key.
The price of NVIDIA's computer remains a mystery, but during the call Mr. Huang did mention a "few thousand dollars." However, looking at the price Tesla is asking for its Enhanced Autopilot ($5,000) and the Full Self-Driving Capability ($3,000), you can get a rough idea. Still, considering the hardware goes into every vehicle, whether the owner buys the two packages or not, Tesla must have a pretty high profit margin, even though Musk did say he expects everybody to unlock those features at one point. In the meantime, even if dormant, the hardware will still collect information for Tesla, so even if not activated, it's still valuable.
Predictably, Mr. Huang is happy about Tesla's decision to release this technology so early: “And I think what Tesla has done by launching and having on the road in the very near-future here, a full autonomous driving capability using AI, that has sent a shock wave through the automotive industry. It’s basically five years ahead. Anybody who’s talking about 2021 and that’s just a non-starter anymore. And I think that that’s probably the most significant bit in the automotive industry. I just don’t – anybody who is talking about autonomous capabilities in 2020 and 2021 is at the moment re-evaluating in a very significant way.”
Google News