Tesla Chief Elon Musk announced through its social media platform X (formerly known as Twitter) that the automaker has started exploring the “final piece of the Full Self-Driving (FSD) AI puzzle.”
For context, the billionaire has long been proclaiming that Tesla is almost done with developing the actual “Full Self-Driving” technology. As EV-a2z previously reported, CEO Musk even claimed that the automaker would achieve full autonomy this year.
Tesla Chief’s previous comments regarding FSD
Tesla CEO Elon Musk announced in a March conference with famous Morgan Stanley that its next-gen luxury EVs will function “almost entirely in autonomous mode.”
A month later, he indicated that Tesla may finally achieve full autonomy this year following the Q1 2023 Earnings Call.
“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.”
Tesla CEO Elon Musk
The Tesla boss reiterated its FSD forecast during the VivaTech conference in Paris, saying that the team would soon “solve autonomy.”
It is also worth noting that Tesla owners with FSD Beta software remain optimistic about the technology’s progress based on their personal experiences.
Vehicle control as the key element for full autonomy
All that said, CEO Musk continues to work towards fulfilling its long-standing promise. According to his recent post in X, the automaker has finally determined the missing element in achieving full autonomy.
“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.”
Tesla CEO Elon Musk
With this recent announcement, it seems that the automaker aims to significantly leverage neural nets instead of direct coded instructions for vehicle control. As CEO Musk mentioned, these neural nets are now undergoing training. He further noted that the progress limitation is currently “training compute.”
See Also:
- Tesla CEO expects to achieve “full autonomy” in 2023
- Tesla Chief says FSD v12 Alpha build is “mind-blowing”
- Tesla to license self-driving tech to a “major” automaker soon, CEO Musk confirms
- Tesla announces plans to invest $1B in Dojo Supercomputer, aiming for 100 exaflops by 2024
- Tesla boss allusively confirms Dojo supercomputer now operational
CEO Musk also previously disclosed that the company started activating Dojo to perform “useful tasks for a few months” already. For context, Dojo is a crucial part of the company’s strategy in advancing its FSD development. It will aid in processing real-time driving conditions from Tesla’s EVs.
