Elon Musk looks guilty of distracted driving

A pedestrian looks inside a closed Tesla Inc. store in Palm Desert California U.S. on Thursday

A pedestrian looks inside a closed Tesla Inc. store in Palm Desert California U.S. on Thursday

October 19 - Musk says he expects that by the end of 2017 a Tesla would be able to drive in full autonomous mode from Los Angeles to NY "without the need for a single touch" on the wheel.

Musk says that all Tesla cars now being produced have the ability to be autonomous through the Navigate feature of Autopilot, it would only take a software update to enable it.

Today's presentation started with Bannon giving a tour of the AI hardware and who was followed by Andrej Karpathy, Tesla's director of AI, presenting a rundown on the tightly integrated software that runs the chip.

Musk took a swipe at competitors relying on Lidar, light-based sensors that are a key element in most other self-driving systems. Tesla, on the other hand, has already committed to a sensor hardware design for its autonomy system, and a particular software approach.

The company also sells a "full self-driving option" for an additional $5,000, explained on Tesla's website as "automatic driving from highway on-ramp to off-ramp", automatic lane changes, the ability to autopark and to summon a parked vehicle.

Musk has been predicting fully self-driving cars are "only a couple of years away" for some time now. While the company is best known for graphics cards that make games look pretty, in recent years, it's focused much of its attention on self-driving technologies and other enterprise sectors. If there are any lingering doubts for non-tech folks, it's reassuring to know that Nvidia itself has acknowledged that the new Tesla computer has "raised the bar for self-driving computers". Musk says full self-driving features will be rolled out in future via over-the-air updates when available.

Tesla has claimed it will have a million self-driving taxis on the road in 2020.

The company has also begun to stock its own fleet of vehicles for the service to deploy in areas where there simply aren't enough owners to service anticipated demand for the platform. "It's like having a whole bunch of appendixes".

The so-called "neural network" uses human annotators to identify cars, pedestrians, and objects in the road from data fed into the computers from other drivers.

"Five or six years from now I think we'll be able to achieve true autonomous driving where you could literally get in the auto, go to sleep and wake up at your destination".

Now, of course, Tesla will have chosen a route that its vehicle can comfortably tackle, and they're not the first to do it, though the car's adeptness in dealing with on ramps, junctions and traffic (the latter which can't entirely be planned for) is remarkable.

Leading up to that point, steering wheels will remain in Tesla's robotaxis, but users will also have the ability to intervene while driving if they want to.

Latest News