Tesla has now more than 100,000 people in its Full Self-Driving Beta program – is an accelerated rate of improvement coming?
Tesla CEO Elon Musk revealed that the automaker has now more than 100,000 people in its Full Self-Driving (FSD) Beta program.
Does it mean that we are going to start seeing an accelerated rate of improvement?
FSD Beta has been a “two steps forward, one step back” type of program.
Some updates brought significant noticeable improvements while others actually made the product worse. There were also long periods of time without updates when the program appeared to be stalling.
For most of the program (from October 2020 to October 2021), there were only a few thousand people in the FSD Beta and most of them were Tesla employees.
The significant ramp-up in the number of beta testers in the program has given some people hope that the rate of improvement would accelerate as Tesla is able to collect a lot more data from the fleet.
This has yet to prove true with the number of FSD Beta updates being quite stagnant in early 2022.
But in an interview with TED last week, Musk revealed that Tesla has kept adding more people to the FSD Beta program and now has more than 100,000 people onboarded:
“The car currently drives me around Austin most of the time with no interventions – and we have over 100,000 people in our full self-driving beta program.”
The CEO admitted that he has often been wrong about Tesla’s timeline to bring self-driving to market and almost comically added that he believes that “this year is the year”.
However, he has drastically reduced the goal for what happens this year from “1 million robotaxis on the road” to “FSD Beta being safer than the average human driver”.
The CEO is no longer talking about any timeline for the real goal that would lead to true self-driving vehicles without constant driver supervision, which is the case with the current version of FSD Beta.
With now over 100,000 people feeding Tesla with FSD Beta test driver nearly every day, you would think that the automaker would have a lot of data to accelerate the rate of improvement.
For years, Tesla has been boasting about its data advantage when it comes to self-driving thanks to its large and fully-connected customer fleet.
However, the automaker has exaggerated its capacity to use that data in the past. The data needs to be annotated before it can usefully train the neural nets. The suspicion has been that Tesla expanded the program to more people because it was ready to take more data, but that has yet to be proven.
I am repeating myself, but what we (FSD buyers, TSLA shareholders, and just fans of safe driving) really need from Tesla is more data to see some improvements and hopefully for Musk to gain back some credibility when it comes to self-driving timelines.
Tesla should be frequently releasing detailed disengagement reports from its FSD Beta fleet in order for us to see how good this thing really is at driving because the videos don’t cut it.
FTC: We use income earning auto affiliate links. More.
Subscribe to Electrek on YouTube for exclusive videos and subscribe to the podcast.