A video played by Elon Musk shows many things people can do in the front seat when not driving, including going to sleep.
Tesla
On the path to developing self-driving cars, Tesla CEO Elon Musk is known for telling the truth at length.
Tesla’s software name, “full self-driving” — or FSD — is itself one miscall. The latest promises of autonomous car AI are making skeptics roll their eyes.
Also: Tesla cuts ‘full self-driving’ subscriptions in half
On Thursday night, at the legendary Warner Bros. Hollywood studio, Musk drove a self-driving car down fake streets “robotaxi“, a product he said would be on the market “before 2027.” The car had neither a steering wheel nor pedals, and Musk sat in the front passenger seat.
“Cybercab“, as Musk has called this vehicle, will enable a new level of autonomy through artificial intelligence (WHO). A video shows people reading books in the front seat, not holding the wheel, and doing a variety of other activities – including reading books, watching movies and even sleeping – as the car continues to drive without assistance their help.
Also: The journey to fully autonomous AI agents and the venture capitalists who fund them
To prepare the market for robotaxi, some of which may be owned by individuals, Musk said, he plans to seed the market next year with an update to FSD, called FSD “unattended” – a step forward, he says, on the present. “Supervised” FSD.
Musk begins the video by entering the Cybercab’s gull-wing door, on the passenger side, as a triumphant, driverless ride.
Tesla
“Before that, you will experience a robot taxi,” Musk said of the Model 3 and Model Y program, as well as Model S and X. Model 3 and Y will achieve completely unattended self-driving mode. , with permission, whenever regulators basically approve it in the United States and then outside the United States — and the Cybertruck as well.”
Initially, Tesla plans to offer unattended FSD in California and Texas, Musk said, if regulators allow it.
Also: Six levels of autonomous work: How AI augments and replaces
Although Musk’s terminology is vague, unmonitored FSD would imply “Level 3” in the US National Highway Traffic Safety Administration’s classification of six different levels of automationcalled “ADAS.”
Level 3 is what NHTSA calls “conditional automation,” in which the car’s software “handles all aspects of driving” while a human remains in the driver’s seat to take over if needed. set. The FSD currently available in Tesla vehicles is considered Level 2, in which a person is responsible for all aspects of driving and is only supported by the ADAS system for discrete tasks such as acceleration.
Musk said the Cybercab will be available sometime “before 2027” and that individuals will eventually be able to buy one for under $30,000.
Tesla
Some Tesla owners have been linked to driving for miles without even touching the steering wheel. However, these designations are important when regulators decide that it can be shipped, as are claims by Tesla and other manufacturers.
Some Wall Street stock analysts following the event criticized Musk for a series of details that went beyond his promise of unmonitored FSD.
“Overall, we found Tesla’s Robotaxi event unimpressive and surprisingly lacking in detail,” wrote Bernstein Research’s Toni Sacconaghi. He noted that Musk was completely silent on whether unmonitored FSD would be backward compatible with existing FSD shipping on Tesla vehicles. The company also made no mention of a “pathway toward regulatory approval.”
Also: Upgrade the productivity of ChatGPT and your new AI colleagues
“Tesla has not provided tangible evidence of progress toward L3 [autonomous driving] or quantify the number of robotaxis planned,” wrote John Coleman, an equity analyst at brokerage firm Jefferies & Co.
Coleman and others also raise a specific technical issue. Musk and his team claim that the robotaxi will be able to navigate roads using only built-in cameras, a simpler setup than the more common method known as “sensor fusion.” In sensor fusion, camera data is combined with LiDAR and radar devices to form a more complete picture of the road.
Using video-based steering alone may be less reliable for automated vehicle navigation than LiDAR and radar support through a hybrid approach.
Also: Where AI avatars serve you 24/7
“There is no precedent for achieving higher levels of autonomy using a vision-only approach (instead of a sensor fusion approach),” wrote analyst Coleman.
However, Musk emphasized the aggregated data collected by Tesla vehicles and fed into the company’s driving simulator. With a million Tesla cars on the road, he explained, it’s like a person living “a million lifetimes,” which he said would make autonomous driving “20 to 30 times safer.” compared to the driver.
