Tesla’s FSD (Supervised) v12.4.2 is started to roll out to some normal Tesla owners (non-employees) in North America. Tesla FSD having possible rapid improvement is critical for increased attachment rate (the percentages of old and new Tesla owners that adopt FSD worldwide.) Tesla FSD rapid improvement is critical for licensing deals with robotaxi companies and car making OEMs.
Initial feedback from users indicates significant improvements in the software’s performance, including the elimination of frequent lane-changing issues, smoother creeping, and improved vision-based monitoring, particularly at night. Users have reported zero disengagements during their initial drives with the new update, suggesting a notable increase in reliability and safety. This update is part of Tesla’s ongoing efforts to refine its autonomous driving technology, aiming to enhance the driving experience and move closer to achieving full self-driving capabilities.
Elon Musk said he was sorry for the delay. This release had far fewer interventions, but suffered in driving smoothness. Part of the issue was too much training on interventions and not enough on normal driving. It’s like a doctor training too much on patients in the emergency room vs training on preventative care.
Our next-gen AI model after this has a lot of promise: ~5X increase in parameter count, which is very difficult to achieve without upgrading the vehicle inference computer.
BREAKING: Tesla's FSD (Supervised) v12.4.2 is now starting to roll out to some normal Tesla owners (non-employees) in North America! pic.twitter.com/TDxYFyjldR
— Sawyer Merritt (@SawyerMerritt) July 1, 2024
Sorry for the delay. This release had far fewer interventions, but suffered in driving smoothness.
Part of the issue was too much training on interventions and not enough on normal driving.
It’s like a doctor training too much on patients in the emergency room vs training on…
— Elon Musk (@elonmusk) July 1, 2024
Brian Wang is a Futurist Thought Leader and a popular Science blogger with 1 million readers per month. His blog Nextbigfuture.com is ranked #1 Science News Blog. It covers many disruptive technology and trends including Space, Robotics, Artificial Intelligence, Medicine, Anti-aging Biotechnology, and Nanotechnology.
Known for identifying cutting edge technologies, he is currently a Co-Founder of a startup and fundraiser for high potential early-stage companies. He is the Head of Research for Allocations for deep technology investments and an Angel Investor at Space Angels.
A frequent speaker at corporations, he has been a TEDx speaker, a Singularity University speaker and guest at numerous interviews for radio and podcasts. He is open to public speaking and advising engagements.
Ive been harping on this theme for a long time: the fsd computers in Tesla cars do not allow for a sufficiently large model.
HW 3 has 16 GB of RAM and HW 4 has only 8 GB of RAM. That’s it! So the maximum model size is about the sameness size as a small LLM that runs on a cell phone. And ALL the smarts of the model has to be compressed in that size… Recognition of objects, drivable surface, pedestrians + driving behaviour.
Too small.. Finally Elon seems to wake up to this reality. We can only hope that HW 5 has at least an order of magnitude more RAM
FSD articles are getting old. It presently doesn’t work as advertised, and it’ll be a little while longer until the proper computing power to make it work reliably and safely, EVERY TIME, is available.
Yeah, people overhype every single decimal point change.
It’s good, but its not FULL self driving…yet.