The transparency site PlainSite recently published some letters that Tesla wrote to the California Department of Motor Vehicles, late in 2020. The letters leave Elon Musk’s optimistic timeline for developing fully driverless technology in doubt.
Elon Musk has been predicting for years that technology without a driver is around the corner. At an event in April 2019, Musk predicted that by the end of 2020, Teslas would be able to operate completely driverless – what is known in industry jargon as ‘level 5’.
“There are three steps to self-management,” Musk told Tesla investors. “There’s a full function. Then there’s a full function to the extent that we think the person in the car does not have to pay attention. where is.”
Tesla, of course, missed Musk’s deadline for 2020. But you can be forgiven for thinking that Tesla is now too late to carry out the strategy it described two years ago. In October, Tesla introduced its ‘complete self-driving beta’ software to several dozen Tesla owners. A few days ago, Musk planned to expand the program to more customers.
Since the product is called ‘full self-management’, it may seem like the first step in Musk’s three-step progress. After a few more months of testing, it might be reliable enough to work without human supervision. This could enable Musk to achieve its latest optimistic timeline for Autopilot: in an interview in December 2020, Musk said he was “very confident” that Tesla vehicles would reach level 5 by the end of 2021.
But a letter Tesla sent to regulators in California in the same month had a different tone. Despite the name “full self-management”, Tesla conceded that it does not consider the current beta software to be suitable for full driverless operation. The company said it would not begin testing a true autonomous feature before an unspecified point in the future.
“We do not expect significant improvements”
In several letters last November and December, California DMV officials asked Tesla for details on the FSD beta program. Tesla requires drivers who use the beta software to actively monitor so that they can intervene quickly if necessary. The DMV wanted to know if Tesla planned to relax the requirements for human supervision once the software was made available to the general public.
In its first response, sent in November, Tesla emphasized that the beta software has limited functionality. Tesla has told state regulators that the software ‘is unable to’ recognize or respond to static objects and road debris, emergency vehicles, construction zones, large uncontrolled intersections with multiple incoming roads, occlusions, adverse weather, complicated or conflicting vehicles in management. not. path and unmarked roads. ‘
In a follow-up in December, Tesla added that “we expect the functionality to remain largely unchanged in the future, fully released to the customer fleet.” Tesla added that “we do not expect significant improvements” which would shift the responsibility for the entire dynamic management task to the system. “The system” will continue to be an advanced SAE Level 2 Driver Assistance feature. “
SAE Level 2 is an operational jargon for driver assistance systems that perform functions such as lane keeping and adaptive cruise control. By definition, level 2 systems constantly require human supervision. Fully driverless systems – such as the taxi service that Waymo operates in the Phoenix area – are considered level 4 systems.
In its letter to California officials, Tesla added that ‘Tesla’s development of true autonomous features will follow our iterative process (development, validation, early release, etc.) and such features will not be disclosed to the general public until we fully confirm are not they. “
Critics have spoken out about the revelation. “Here it is, directly from Tesla,” prominent Tesla skeptic Ed Niedermeyer tweeted. “‘Full self-government’ is not and will never be self-government.”
This may not be entirely fair to Tesla – the company apparently intends to eventually develop more advanced software. But at least Tesla’s public communication about the complete self-management package can easily give customers the wrong impression about the future capabilities of the software.
Full autonomy is always around the corner
Since 2016, Tesla has given customers every reason to expect the software for “full self-management” to be good, self-management.
Early promotional material for the FSD package described how driver got out of the vehicle and found a parking space on his own. Tesla has repeatedly talked about the FSD package that enables a Tesla vehicle to function as an autonomous taxi – an application that requires the car to drive itself without anyone being behind the wheel. In 2016, Musk predict that a Tesla owner in Los Angeles could call their vehicle from New York City within two years.
If Tesla is really going to achieve a driverless operation in 2021, it does not leave much time to develop, test and validate complex, safety-critical software. So it’s natural for customers to assume that the Tesla software called “Full Self Driving beta” is in fact a beta version of Tesla’s long-awaited self-driving software. But in its communication with officials in California, Tesla makes it clear that this is not true.
Of course, Elon Musk has a long history of over-optimistic timelines for his products. It’s not really news that Tesla could not meet an optimistic deadline by its CEO.
But there is a deeper philosophical issue that can go beyond a few deadlines.
The long road to full autonomy
Waymo
Tesla’s overall Autopilot strategy is to start with a driver assistance system and gradually develop it into a complete driverless system. A number of other companies in the industry – led by Google’s Waymo – believe this is a mistake. They think that the requirements of the two products are so different that it makes more sense to create a driverless taxi, shuttle service or delivery service from scratch.
In particular, companies like Waymo argue that it is too difficult to get regular customers to pay attention to a vehicle that is almost driverless, but not completely. If a car drives 1000 kilometers perfectly and then makes a big mistake, the risk is great that the driver will not pay enough attention to prevent an accident. Waymo initially considered creating a driver assistance system such as Autopilot and licensing it to automakers, but the company eventually decided it would be too risky.
Musk has always picked up this critique. As we have seen, he believes that improvements to Autopilot’s driver assistance features will transform it into a system that can drive completely driverless.
But in its remarks to the DMV, Tesla seems to endorse the opposite view: that adding “true autonomous features” to Autopilot should do more than just improve the performance of its existing software. Tesla has acknowledged that it needs more sophisticated systems for handling ‘static objects, debris, emergency vehicles, construction zones’.
And that makes it a little hard to believe that Musk boasts that Tesla will achieve level 5 autonomy by the end of 2021. It is especially noteworthy that Google’s prototype self-driving vehicles have been able to navigate most road conditions, such as the Tesla FSD beta software of today, since about 2015. Yet the company still needed five years to refine the technology enough to enable driverless operation. to make.
And it was within a limited geographical area and with the help of powerful lidar sensors. Tesla strives to achieve the same performance for every street nationwide – using only cameras and radar.
Maybe Tesla will move faster than Waymo and it will not take another five years to achieve a completely driverless operation. But customers considering paying $ 10,000 for Tesla’s complete self-driving software package should take Musk’s optimistic timeline with a pinch of salt.