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 improving Giga Berlin’s free “Giga Train” service for employees
With this initiative, Tesla aims to boost the number of Gigafactory Berlin employees commuting by rail while keeping the shuttle free for all riders.
Tesla will expand its factory shuttle service in Germany beginning January 4, adding direct rail trips from Berlin Ostbahnhof to Giga Berlin-Brandenburg in Grünheide.
With this initiative, Tesla aims to boost the number of Gigafactory Berlin employees commuting by rail while keeping the shuttle free for all riders.
New shuttle route
As noted in a report from rbb24, the updated service, which will start January 4, will run between the Berlin Ostbahnhof East Station and the Erkner Station at the Gigafactory Berlin complex. Tesla stated that the timetable mirrors shift changes for the facility’s employees, and similar to before, the service will be completely free. The train will offer six direct trips per day as well.
“The service includes six daily trips, which also cover our shift times. The trains will run between Berlin Ostbahnhof (with a stop at Ostkreuz) and Erkner station to the Gigafactory,” Tesla Germany stated.
Even with construction continuing at Fangschleuse and Köpenick stations, the company said the route has been optimized to maintain a predictable 35-minute travel time. The update follows earlier phases of Tesla’s “Giga Train” program, which initially connected Erkner to the factory grounds before expanding to Berlin-Lichtenberg.
-->Tesla pushes for majority rail commuting
Tesla began production at Grünheide in March 2022, and the factory’s workforce has since grown to around 11,500 employees, with an estimated 60% commuting from Berlin. The facility produces the Model Y, Tesla’s best-selling vehicle, for both Germany and other territories.
The company has repeatedly emphasized its goal of having more than half its staff use public transportation rather than cars, positioning the shuttle as a key part of that initiative. In keeping with the factory’s sustainability focus, Tesla continues to allow even non-employees to ride the shuttle free of charge, making it a broader mobility option for the area.
News
Tesla Model 3 and Model Y dominate China’s real-world efficiency tests
The Tesla Model 3 posted 20.8 kWh/100 km while the Model Y followed closely at 21.8 kWh/100 km.
Tesla’s Model 3 and Model Y once again led the field in a new real-world energy-consumption test conducted by China’s Autohome, outperforming numerous rival electric vehicles in controlled conditions.
The results, which placed both Teslas in the top two spots, prompted Xiaomi CEO Lei Jun to acknowledge Tesla’s efficiency advantage while noting that his company’s vehicles will continue refining its own models to close the gap.
Tesla secures top efficiency results
Autohome’s evaluation placed all vehicles under identical conditions, such as a full 375-kg load, cabin temperature fixed at 24°C on automatic climate control, and a steady cruising speed of 120 km/h. In this environment, the Tesla Model 3 posted 20.8 kWh/100 km while the Model Y followed closely at 21.8 kWh/100 km, as noted in a Sina News report.
These figures positioned Tesla’s vehicles firmly at the top of the ranking and highlighted their continued leadership in long-range efficiency. The test also highlighted how drivetrain optimization, software management, and aerodynamic profiles remain key differentiators in high-speed, cold-weather scenarios where many electric cars struggle to maintain low consumption.
Xiaomi’s Lei Jun pledges to continue learning from Tesla
Following the results, Xiaomi CEO Lei Jun noted that the Xiaomi SU7 actually performed well overall but naturally consumed more energy due to its larger C-segment footprint and higher specification. He reiterated that factors such as size and weight contributed to the difference in real-world consumption compared to Tesla. Still, the executive noted that Xiaomi will continue to learn from the veteran EV maker.
“The Xiaomi SU7’s energy consumption performance is also very good; you can take a closer look. The fact that its test results are weaker than Tesla’s is partly due to objective reasons: the Xiaomi SU7 is a C-segment car, larger and with higher specifications, making it heavier and naturally increasing energy consumption. Of course, we will continue to learn from Tesla and further optimize its energy consumption performance!” Lei Jun wrote in a post on Weibo.
-->Lei Jun has repeatedly described Tesla as the global benchmark for EV efficiency, previously stating that Xiaomi may require three to five years to match its leadership. He has also been very supportive of FSD, even testing the system in the United States.
Elon Musk
Elon Musk reveals what will make Optimus’ ridiculous production targets feasible
Musk recent post suggests that Tesla has a plan to attain Optimus’ production goals.
Elon Musk subtly teased Tesla’s strategy to achieve Optimus’ insane production volume targets. The CEO has shared his predictions about Optimus’ volume, and they are so ambitious that one would mistake them for science fiction.
Musk’s recent post on X, however, suggests that Tesla has a plan to attain Optimus’ production goals.
The highest volume product
Elon Musk has been pretty clear about the idea of Optimus being Tesla’s highest-volume product. During the Tesla 2025 Annual Shareholder Meeting, Musk stated that the humanoid robot will see “the fastest production ramp of any product of any large complex manufactured product ever,” starting with a one-million-per-year line at the Fremont Factory.
Following this, Musk stated that Giga Texas will receive a 10 million-per-year unit Optimus line. But even at this level, the Optimus ramp is just beginning, as the production of the humanoid robot will only accelerate from there. At some point, the CEO stated that a Mars location could even have a 100 million-unit-per-year production line, resulting in up to a billion Optimus robots being produced per year.
Self-replication is key
During the weekend, Musk posted a short message that hinted at Tesla’s Optimus strategy. “Optimus will be the Von Neumann probe,” the CEO wrote in his post. This short comment suggests that Tesla will not be relying on traditional production systems to make Optimus. The company probably won’t even hire humans to produce the humanoid robot at one point. Instead, Optimus robots could simply produce other Optimus robots, allowing them to self-replicate.
-->The Von Neumann is a hypothetical self-replicating spacecraft proposed by the mathematician and physicist John von Neumann in the 1940s–1950s. The hypothetical machine in the concept would be able to travel to a new star system or location, land, mine, and extract raw materials from planets, asteroids, and moons as needed, use those materials to manufacture copies of itself, and launch the new copies toward other star systems.
If Optimus could pull off this ambitious target, the humanoid robot would indeed be the highest volume product ever created. It could, as Musk predicted, really change the world.
Tesla is improving Giga Berlin’s free “Giga Train” service for employees
Tesla Model 3 and Model Y dominate China’s real-world efficiency tests
