In response to a Tweet from ARK Invest’s Tasha Keeney, Elon Musk confirmed plans to release a new limited beta version of Tesla’s Full Self-Driving (FSD) software. He also acknowledged progress being made by other creators of self-driving and driver assist software such as Waymo, which he referred to as an “impressive but highly specialized solution.”
Waymo is impressive, but a highly specialized solution. The Tesla approach is a general solution. The latest build is capable of zero intervention drives. Will release limited beta in a few weeks.
— Elon Musk (@elonmusk) October 8, 2020
Musk indicated that the new FSD version will be capable of managing drives without any input from a driver. If successful, the FSD will be capable of managing fleets of cars for applications like its planned Robotaxi.
The Robotaxi is intended to be Tesla’s in-house ride hailing service that could theoretically compete with existing services like Uber, Lyft, and the highly localized Via. Owners of Tesla models with the FSD suite could make as much as $30,000 per year by leasing their vehicles to the Robotaxi network. Musk has indicated that it could be ready for a limited rollout as early as 2021.
Recent progress on Robotaxi-related features includes the release of “Car Access” controls that owners can use to determine who can access the vehicle. Car Access includes the ability to locate a Tesla vehicle, unlock it, and start it remotely. It can also grant the ability to control the vehicle directly from the app.
Testing out Tesla’s new “Car Access” feature with my colleague in California. This is really a big step for Robotaxi! What’s even more impressive is I controlled the Model Y in Silicon Valley from my home in South Central Pennsylvania! @Teslarati @tesla @elonmusk pic.twitter.com/6OJfprKqbN
— Joey Klender (@KlenderJoey) June 9, 2020
The FSD is likely to require fine-tuning through intense beta-testing and collection of data for Tesla’s Neural Network, an AI application designed to learn from data sent to it by Tesla vehicles on the road. The Neural Network will be especially useful for enabling vehicles to react to unsafe conditions such as ice on the road or an incoming blizzard in real-time.
If the FSD becomes representative of a new industry standard of fully autonomous self-driving cars, it is likely to improve safety on the road by taking human error out of the equation. With applications and infrastructure the Neural Network, cars can effectively update one another about conditions on the road.