Walter Isaacson’s Elon Musk biography is set to be published on Tuesday, and a new preview of the book illustrates details about Tesla’s development of the upcoming Full Self-Driving (FSD) version 12.
In an additional preview of his Musk biography for CNBC, Isaacson discusses the use of AI in the development of Tesla’s FSD v12, in a shift that took place within the last several months. Isaacson talks about Tesla’s recent development of the upcoming FSD v12, which he and Tesla demonstrate has moved away from a “rules-based” approach.
Notably, FSD v12 is expected to use billions of video frames from real-world driving incidents to train its neural network rather than using thousands of lines of code like previous versions. In a conversation with Musk last December, Tesla Autopilot employee Dhaval Shroff had likened the concept to the popular chatbot ChatGPT, instead for use with driving.
“It’s like ChatGPT, but for cars,” Shroff said. “We process an enormous amount of data on how real human drivers acted in a complex driving situation, and then we train a computer’s neural network to mimic that.”
Surprisingly enough, Tesla only shifted toward this “neural network planner” approach recently. By the beginning of this year, however, the neural network had already analyzed 10 million video clips based on the best-case-scenario drivers the system had access to. Musk instructed employees at the company’s Buffalo, New York facility who were in charge of analyzing the footage to train the AI on things “a five-star Uber driver would do.”
-->Moving from a rules-based to a network-path-based AI approach allowed FSD to use human driving data to avoid obstacles, even if breaking some rules was necessary. Shroff helped demonstrate the idea to Musk with a demo featuring trash bins, debris, and upturned traffic cones, which the car handled surprisingly well.
“Here’s what happens when we move from rules-based to network-path-based,” Shroff explained. “The car will never get into a collision if you turn this thing on, even in unstructured environments.”
Musk quickly took to the idea, as can be seen in a recent livestream of Tesla’s FSD v12 software in Palo Alto with Autopilot software director Ashok Elluswamy. He has repeatedly spoken about the upcoming software version’s impressive driving results, despite one small moment in the drive where the car almost ran a red light.
In any case, Musk could argue that the red-light moment is a good case for the need for self-driving software to continually learn. Given that it will constantly be trained from the video data generated by camera footage from real-world drivers, it should theoretically make it safer over time, according to Musk.
-->During development, Musk also reportedly latched onto the fact that it took over a million video clips for the neural network to begin performing well, though he looks forward to what significantly more data will do for FSD.
Still, critics and regulators have expressed concerns about the faults of human drivers training AI-based driving systems, and Tesla has repeatedly been questioned by the National Highway Traffic Safety Administration (NHTSA) about its Autopilot and FSD beta systems.
According to Isaacson, Tesla plans to release FSD v12 as soon as regulators approve it. Meanwhile, an ongoing study by the National Highway Safety Board is looking to determine if self-driving cars should be permitted to imitate human driving actions that blur traffic rules, such as creeping up at stop signs.
Musk said in April that he expects Tesla to reach full autonomy within a year, though he has also been known to share ambitious targets for the software in the past.
You can read Walter Isaacson’s full account of the development of Tesla FSD v12 here, in a CNBC preview of the upcoming Elon Musk biography.
-->What are your thoughts? Let me know at zach@teslarati.com, find me on X at @zacharyvisconti, or send your tips to us at tips@teslarati.com.
News
Tesla Model 3 named New Zealand’s best passenger car of 2025
Tesla flipped the switch on Full Self-Driving (Supervised) in September, turning every Model 3 and Model Y into New Zealand’s most advanced production car overnight.
The refreshed Tesla Model 3 has won the DRIVEN Car Guide AA Insurance NZ Car of the Year 2025 award in the Passenger Car category, beating all traditional and electric rivals.
Judges praised the all-electric sedan’s driving dynamics, value-packed EV tech, and the game-changing addition of Full Self-Driving (Supervised) that went live in New Zealand this September.
Why the Model 3 clinched the crown
DRIVEN admitted they were late to the “Highland” party because the updated sedan arrived in New Zealand as a 2024 model, just before the new Model Y stole the headlines. Yet two things forced a re-evaluation this year.
First, experiencing the new Model Y reminded testers how many big upgrades originated in the Model 3, such as the smoother ride, quieter cabin, ventilated seats, rear touchscreen, and stalk-less minimalist interior. Second, and far more importantly, Tesla flipped the switch on Full Self-Driving (Supervised) in September, turning every Model 3 and Model Y into New Zealand’s most advanced production car overnight.
FSD changes everything for Kiwi buyers
The publication called the entry-level rear-wheel-drive version “good to drive and represents a lot of EV technology for the money,” but highlighted that FSD elevates it into another league. “Make no mistake, despite the ‘Supervised’ bit in the name that requires you to remain ready to take control, it’s autonomous and very capable in some surprisingly tricky scenarios,” the review stated.
-->At NZ$11,400, FSD is far from cheap, but Tesla also offers FSD (Supervised) on a $159 monthly subscription, making the tech accessible without the full upfront investment. That’s a game-changer, as it allows users to access the company’s most advanced system without forking over a huge amount of money.
News
Tesla starts rolling out FSD V14.2.1 to AI4 vehicles including Cybertruck
FSD V14.2.1 was released just about a week after the initial FSD V14.2 update was rolled out.
It appears that the Tesla AI team burned the midnight oil, allowing them to release FSD V14.2.1 on Thanksgiving. The update has been reported by Tesla owners with AI4 vehicles, as well as Cybertruck owners.
For the Tesla AI team, at least, it appears that work really does not stop.
FSD V14.2.1
Initial posts about FSD V14.2.1 were shared by Tesla owners on social media platform X. As per the Tesla owners, V14.2.1 appears to be a point update that’s designed to polish the features and capacities that have been available in FSD V14. A look at the release notes for FSD V14.2.1, however, shows that an extra line has been added.
“Camera visibility can lead to increased attention monitoring sensitivity.”
Whether this could lead to more drivers being alerted to pay attention to the roads more remains to be seen. This would likely become evident as soon as the first batch of videos from Tesla owners who received V14.21 start sharing their first drive impressions of the update. Despite the update being released on Thanksgiving, it would not be surprising if first impressions videos of FSD V14.2.1 are shared today, just the same.
-->Rapid FSD releases
What is rather interesting and impressive is the fact that FSD V14.2.1 was released just about a week after the initial FSD V14.2 update was rolled out. This bodes well for Tesla’s FSD users, especially since CEO Elon Musk has stated in the past that the V14.2 series will be for “widespread use.”
FSD V14 has so far received numerous positive reviews from Tesla owners, with numerous drivers noting that the system now drives better than most human drivers because it is cautious, confident, and considerate at the same time. The only question now, really, is if the V14.2 series does make it to the company’s wide FSD fleet, which is still populated by numerous HW3 vehicles.
News
Waymo rider data hints that Tesla’s Cybercab strategy might be the smartest, after all
These observations all but validate Tesla’s controversial two-seat Cybercab strategy, which has caught a lot of criticism since it was unveiled last year.
Toyota Connected Europe designer Karim Dia Toubajie has highlighted a particular trend that became evident in Waymo’s Q3 2025 occupancy stats. As it turned out, 90% of the trips taken by the driverless taxis carried two or fewer passengers.
These observations all but validate Tesla’s controversial two-seat Cybercab strategy, which has caught a lot of criticism since it was unveiled last year.
Toyota designer observes a trend
Karim Dia Toubajie, Lead Product Designer (Sustainable Mobility) at Toyota Connected Europe, analyzed Waymo’s latest California Public Utilities Commission filings and posted the results on LinkedIn this week.
“90% of robotaxi trips have 2 or less passengers, so why are we using 5-seater vehicles?” Toubajie asked. He continued: “90% of trips have 2 or less people, 75% of trips have 1 or less people.” He accompanied his comments with a graphic showing Waymo’s occupancy rates, which showed 71% of trips having one passenger, 15% of trips having two passengers, 6% of trips having three passengers, 5% of trips having zero passengers, and only 3% of trips having four passengers.
The data excludes operational trips like depot runs or charging, though Toubajie pointed out that most of the time, Waymo’s massive self-driving taxis are really just transporting 1 or 2 people, at times even no passengers at all. “This means that most of the time, the vehicle being used significantly outweighs the needs of the trip,” the Toyota designer wrote in his post.
-->Cybercab suddenly looks perfectly sized
Toubajie gave a nod to Tesla’s approach. “The Tesla Cybercab announced in 2024, is a 2-seater robotaxi with a 50kWh battery but I still believe this is on the larger side of what’s required for most trips,” he wrote.
With Waymo’s own numbers now proving 90% of demand fits two seats or fewer, the wheel-less, lidar-free Cybercab now looks like the smartest play in the room. The Cybercab is designed to be easy to produce, with CEO Elon Musk commenting that its product line would resemble a consumer electronics factory more than an automotive plant. This means that the Cybercab could saturate the roads quickly once it is deployed.
While the Cybercab will likely take the lion’s share of Tesla’s ride-hailing passengers, the Model 3 sedan and Model Y crossover would be perfect for the remaining 9% of riders who require larger vehicles. This should be easy to implement for Tesla, as the Model Y and Model 3 are both mass-market vehicles.
Tesla Model 3 named New Zealand’s best passenger car of 2025
Tesla starts rolling out FSD V14.2.1 to AI4 vehicles including Cybertruck
