Someone instructed their Tesla to drive from San Francisco to Los Angeles, and it was able to do so without any driver intervention. The car even pulled off the highway and went to a charging station when its battery ran low.
I’ve noted before that Waymo and Tesla have two very different strategies for driverless cars. Waymo’s is basically to go into full-fledged driverless mode with the help of maps and remote monitors. Tesla’s is to incrementally improve their self-driving software until it can do everything by itself.
Recently, it looked like the Waymo model (which is also the Cruise model) was winning. Waymo and Cruise are both providing driverless taxi services in San Francisco and Waymo has expanded in Chandler, Arizona services to cover the entire Phoenix urban area. Meanwhile, Tesla’s future was clouded by debate over Musk’s takeover of Twitter, which may have alienated many potential Tesla buyers.
The above video, however, suggests that Tesla was not ignoring the issue. Musk has specifically said that he would like Tesla to sell its cars at cost and make its profits from driverless software. This video suggests this may be possible.
I confess I’ve never been in a Tesla and don’t know how the self-driving software works from an operator view. The speeded-up video shows the driver putting his hand on the controls only once: to put the car in gear so it can leave the mid-way charging station. Does this mean the car automatically recognized that its battery was running low and pulled off the freeway into a charge station to recharge? I can see where that would be possible, but it is also possible that the fast-motion video cut out some interaction between car and driver that led the car to pull off the highway.
Otherwise, the car successfully negotiated streets in both San Francisco and Los Angeles as well as the freeway between. It’s hard not to be impressed by this, especially considering Waymo and Cruise can’t leave the city limits.
This is good news but … as self-driving cars hit the road each and every accident and incident will generate headlines that ignore the relative safety of software driven vehicles.
From Bing Chat …
Car accidents are a major cause of death and injury in the United States. In 2020, there were over 5 million police-reported car accidents, resulting in 38,824 traffic-related deaths and 1,593,390 injuries. The number of accidents decreased by 22% between 2019 and 2020, but the number of fatal accidents per mile driven increased by 21% year over year. In 45% of fatal crashes, the drivers of passenger vehicles were engaged in at least one of the following risky behaviors: speeding, alcohol impairment, or not wearing a seat belt. Over 90 people die in car accidents every day, and around 2 million drivers in car accidents experience permanent injuries every year.
A low bar.
11.3.6 was a massive improvement to full self driving. I use it perhaps 80% of the time I drive, only taking over to enter/exit parking lots, and to turn quickly from unprotected left turns if it is taking too long (it’s too cautious).
By the way, Waymo does not drive on highways, which is a big handicap.
Question – Had the vehicle been in a deadly accident, would the owner be facing manslaughter charges? Execs at Tesla? This was arguably quite reckless.
Still trying to figure out why I should care about this…
Interesting that it frequently exceeds the speed limit on the freeway. Does Tesla software know enough to pull over when flashing red-and-blue lights appear behind it? If a speeding ticket is issued, does Tesla share the fine?