Tesla CEO Elon Musk talked about the “final piece” of solving the company’s Full Self-Driving suite.
Musk has claimed for years that Tesla is close to releasing a complete version of Full Self-Driving. 2023 is no different. Earlier this year, Musk said on multiple occasions that Tesla will solve FSD this year.
In March, he said at a conference with Morgan Stanley that its next-gen vehicle is expected to operate “almost entirely in autonomous mode.”
Musk says Tesla’s next-gen vehicle will be ‘almost entirely’ autonomous
In April, after the Q1 2023 Earnings Call, he said:
“There’ll be a little bit of two steps forward, one step back between releases for those trying the beta. But the trend is very clearly towards full self-driving, towards full autonomy. And I hesitate to say this, but I think we’ll do it this year. So that’s what it looks like.”
And in June, he said at the VivaTech conference in Paris, France:
“Although I’ve said this before, I think we will solve autonomy soon.”
It is understandable that many would remain skeptical, but the owners who use the Full Self-Driving Beta suite continue to talk positively about their experiences. While there are criticisms, Tesla continues to work to improve the suite.
Others are also interested in adopting the technology as a nod that it is moving in the right direction. Musk said during the most recent Earnings Call that Tesla is in early discussions with a major OEM, and many Tesla fans believe it is Ford, although this is purely speculation.
Musk is still convinced Tesla is close to solving the suite and said in a Tweet last night that “vehicle control is the final piece of the Tesla FSD AI puzzle.”
Vehicle control is the final piece of the Tesla FSD AI puzzle. That will drop >300k lines of C++ control code by ~2 orders of magnitude.
It is training as I write this. Our progress is currently training compute constrained, not engineer constrained.
— Elon Musk (@elonmusk) August 1, 2023
Vehicle control on a level that Tesla is looking for would be the major barrier to solving FSD. It is constantly being trained by the company’s neural network, which helps improve driving behavior. However, although data is coming in at a rapid pace that is helping write “>300k lines of C++ control code by ~2 orders of magnitude,” Musk maintains that solving FSD is not a problem related to manpower.
Instead, he said the progress is training compute-constrained.
Musk hinted that it could Dojo is now online by simply favoriting a Tweet. He also said in June that Dojo “has been online & running useful tasks for a few months.”
Dojo is going to be responsible for continuous growth and progress in the development of FSD, which is supposed to help process real-life driving situations from the company’s cars in an efficient way.
I’d love to hear from you! If you have any comments, concerns, or questions, please email me at joey@teslarati.com. You can also reach me on Twitter @KlenderJoey, or if you have news tips, you can email us at tips@teslarati.com.