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 is already expanding its Rental program aggressively
The program has already launched in a handful of locations, specifically, it has been confined to California for now. However, it does not seem like Tesla has any interest in keeping it restricted to the Golden State.
Tesla is looking to expand its Rental Program aggressively, just weeks after the program was first spotted on its Careers website.
Earlier this month, we reported on Tesla’s intention to launch a crazy new Rental program with cheap daily rates, which would give people in various locations the opportunity to borrow a vehicle in the company’s lineup with some outrageous perks.
Along with the cheap rates that start at about $60 per day, Tesla also provides free Full Self-Driving operation and free Supercharging for the duration of the rental. There are also no limits on mileage or charging, but the terms do not allow the renter to leave the state from which they are renting.
🚨🚨 If you look up details on the Tesla Rental program on Google, you’ll see a bunch of sites saying it’s because of decreasing demand 🤣 pic.twitter.com/WlSQrDJhMg
— TESLARATI (@Teslarati) November 10, 2025
The program has already launched in a handful of locations, specifically, it has been confined to California for now. However, it does not seem like Tesla has any interest in keeping it restricted to the Golden State.
Job postings from Tesla now show it is planning to launch the Rental program in at least three new states: Texas, Tennessee, and Massachusetts.
The jobs specifically are listed as a Rental Readiness Specialist, which lists the following job description:
“The Tesla Rental Program is looking for a Rental Readiness Specialist to work on one of the most progressive vehicle brands in the world. The Rental Readiness Specialist is a key contributor to the Tesla experience by coordinating the receipt of incoming new and used vehicle inventory. This position is responsible for fleet/lot management, movement of vehicles, vehicle readiness, rental invoicing, and customer hand-off. Candidates must have a high level of accountability, and personal satisfaction in doing a great job.”
It also says that those who take the position will have to charge and clean the cars, work with clients on scheduling pickups and drop-offs, and prepare the paperwork necessary to initiate the rental.
The establishment of a Rental program is big for Tesla because it not only gives people the opportunity to experience the vehicles, but it is also a new way to rent a car.
Just as the Tesla purchasing process is more streamlined and more efficient than the traditional car-buying experience, it seems this could be less painful and a new way to borrow a car for a trip instead of using your own.
Elon Musk
Elon Musk’s xAI gains first access to Saudi supercluster with 600k Nvidia GPUs
The facility will deploy roughly 600,000 Nvidia GPUs, making it one of the world’s most notable superclusters.
A Saudi-backed developer is moving forward with one of the world’s largest AI data centers, and Elon Musk’s xAI will be its first customer. The project, unveiled at the U.S.–Saudi Investment Forum in Washington, D.C., is being built by Humain, a company supported by Saudi Arabia’s Public Investment Fund.
The facility will deploy roughly 600,000 Nvidia GPUs, making it one of the world’s most notable superclusters.
xAI secures priority access
Nvidia CEO Jensen Huang stated that the planned data center marks a major leap not just for the region but for the global AI ecosystem as a whole. Huang joked about the sheer capacity of the build, emphasizing how unusual it is for a startup to receive infrastructure of such magnitude. The facility is designed to deliver 500 megawatts of Nvidia GPU power, placing it among the world’s largest AI-focused installations, as noted in a Benzinga report.
“We worked together to get this company started and off the ground and just got an incredible customer with Elon. Could you imagine a startup company, approximately $0 billion in revenues, now going to build a data center for Elon? 500 megawatts is gigantic. This company is off the charts right away,” Huang said.
Global Chipmakers Join Multi-Vendor Buildout To Enhance Compute Diversity
While Nvidia GPUs serve as the backbone of the first phase, Humain is preparing a diversified hardware stack. AMD will supply its Instinct MI450 accelerators, which could draw up to 1 gigawatt of power by 2030 as deployments ramp. Qualcomm will also contribute AI200 and AI250 data center processors, accounting for an additional 200 megawatts of compute capacity. Cisco will support the networking and infrastructure layer, helping knit the multi-chip architecture together.
Apart from confirming that xAI will be the upcoming supercluster’s first customer, Musk also joked about the rapid scaling needed to train increasingly large AI models. He joked that a theoretical expansion one thousand times larger of the upcoming supercluster “would be 8 bazillion, trillion dollars,” highlighting the playful exaggeration he often brings to discussions around extreme compute demand.
Elon Musk
Elon Musk debunks pay package and lip reader claims in double takedown
Musk’s quick debunks highlighted once more that X is an ideal platform for directly countering misinformation.
Elon Musk recently took to X to debunk some misinformation about his 2025 CEO performance award, as well as some comments he made during Donald Trump’s banquet in honor of Saudi Prince Mohammed bin Salman.
Musk’s quick debunks highlighted once more that X is an ideal platform for directly countering misinformation.
Musk’s pay package
Elon Musk’s 2025 CEO performance award was created as a path for him to gain a 25% stake in Tesla. It would also make him a trillionaire, provided that he manages to meet all of the performance award’s aggressive targets. This has not stopped critics from running with the apparent narrative that Musk will be getting the $1 trillion with utmost certainty, however.
This included the More Perfect Union account on X, which noted that “Elon Musk is set to make more than every U.S. elementary school teacher combined, according to the Washington Post.”
Musk responded to the pro-union amount’s post, highlighting that he has not earned any of his $2025 performance award so far. Musk also noted that those who believe he will be getting $1 trillion should invest in TSLA stock, as his compensation is tied to the company’s performance and growth. Investors who hold their TSLA until Musk achieves his full pay package would likely get notable returns.
Lip reader fail
Musk also debunked claims from the Daily Mail, which claimed that he made an “explosive” remark at Trump’s banquet for Saudi Prince Mohammed bin Salman. Citing observations from lip reader Nicola Hickling, the Mail claimed that Musk asked Pfizer CEO Albert Bourla, “What is your opinion, is he a terrorist?” The publication also posted a video of Musk allegedly making the risqué comment on X.
Musk proceeded to correct the publication, stating that the lip reader’s observations were fake. Instead of asking the Pfizer CEO if the Saudi Prince was a terrorist, Musk noted that he was asking the executive about cancer medicine. “False, I was asking about upcoming cancer drugs,” Musk wrote in a response on X.
Musk’s comments resulted in numerous critical responses to the Mail’s video, with some X users joking that the lip reader who analyzed the clip should probably get a visual acuity test, or a better training course on lip reading at least.
Tesla is already expanding its Rental program aggressively
Elon Musk’s xAI gains first access to Saudi supercluster with 600k Nvidia GPUs
