Tesla FSD v15 starts rolling out with Reality View | Magica
Tesla starts FSD v15 rollout with Reality View and an unproven safety claim
Editorial Team
••📖6 min read
Tesla is sending FSD v15 to its first customers, adding a visualization of planned driving and changes to navigation and hazard detection. The software remains supervised, and its claimed 25-fold intervention improvement lacks a disclosed baseline.
Tesla says FSD v15 has begun reaching customers; the initial rollout's size and regional coverage remain undisclosed.
Reality View displays the surroundings and the system's planned path, while navigation changes initially cover trip beginnings and endings.
The update remains supervised. Tesla's claimed 25-fold improvement in critical interventions is not evidence of a 25-fold reduction in crashes.
Tesla began rolling out version 15 of Full Self-Driving (Supervised), its driver-assistance software for Tesla vehicles, on November 21. The electric-car maker is pairing changes to driving behavior with Reality View, a new live visualization intended to show drivers what the system plans to do.
Tesla's announcement claims “25x better handling of safety critical cases,” but the available release material supplies neither an absolute intervention rate nor the testing detail needed to judge that claim. Drivers still have to supervise and be ready to take over.
A clearer view of plans, with limits on the rollout
Reality View generates a live overhead view of the vehicle's surroundings and shows its planned path. Tesla says drivers can choose a full-screen visualization, a map or a split layout. The earlier interface already depicted roads and nearby vehicles, cyclists and pedestrians. Tesla describes the new display as a way to see and control what FSD is planning, according to the rollout report.
Tesla's release notes also call the feature Tesla Augmented Reality. They say it replaces separate model signals with one end-to-end visualization while using less total compute. That is a company claim about the implementation; a more detailed rendering alone does not establish better driving decisions.
Tesla chief executive Elon Musk, whose earlier safety promises accompanied FSD v12 and v14, has targeted a wider December release with the holiday update, distinct from the first customer deliveries, according to an earlier release-timing report. Tesla has not disclosed how many customers are in the first wave or which regions receive it first.
Hardware also divides the audience. Tesla has described v15 as a rewrite with a larger AI model designed for its AI4, also called HW4, driving computer. Older HW3 cars remain on a separate software branch, the rollout coverage notes. The announcement does not specify which vehicles receive it first.
Employee equity at Anthropic’s IPO price is reportedly worth $232 billion, while book rightsholders await a first settlement installment estimated at $2,203.56 per work. Vesting, shared copyrights and founder voting rights complicate the comparison.
Editorial Team
For HW3 owners who bought FSD outright, the gap extends beyond a software download. The same coverage describes Musk’s April acknowledgment that HW3 could not handle unsupervised driving and his proposed alternatives: a discounted trade-in for an AI4 car or a computer upgrade. Tesla has not provided a retrofit date. A separate company target would put some customer cars into its driverless Robotaxi fleet around the second quarter of 2027; that remains distinct from December’s supervised software release.
Navigation and hazards address familiar complaints
Tesla says the update learns navigation preferences from its vehicle fleet at the start and end of trips. Extending that approach across the entire route is reserved for later v15 releases. Parking-location selection uses personalized parking history, and the map can display several destination parking options, according to the reported release notes.
The company also promises better detection of signs, small animals, debris, chains and potholes, plus avoidance of deep standing water. It says speed handling improves for school zones, speed limits, bad weather and the different driver profiles. Slowing for speed cameras is another announced addition.
These are practical tests of the new version because parking, speed control, navigation and potholes were already identified as weaknesses in reporting on the preceding v14 software. Pothole avoidance had already appeared among planned improvements in v14.3 release notes. In the earlier report of his comments, Musk also said on November 21 that v15 should address owner complaints about reluctance to use carpool lanes and last-minute lane changes for turns or merges. Those are promised fixes, rather than measured results.
Bike Rack Mode allows FSD use with a rack or similar hitch accessory, but changes the available maneuvers. The rollout report says it pulls over near the destination without reversing and disables Autopark and Summon.
Other changes affect the driver-system relationship. Tesla says camera-based attention monitoring will work more often with hats and sunglasses, while monitoring becomes less strict in higher-speed driver profiles. Its notes also describe an Autoshift change that uses FSD-based gear predictions. The monitoring relaxation deserves scrutiny alongside the continuing requirement for an attentive driver.
What the safety multiplier leaves unanswered
Tesla attributes the claimed 25-fold improvement in its critical intervention rate to an upgraded vision encoder—the component that interprets camera imagery. According to the reported release notes, it is trained on millions of internet videos.
The promise also has a history. In April, Musk said v15 would exceed human safety even in complex, unsupervised situations, according to the release-day analysis. The continuing supervision requirement is the relevant distinction between those forecasts and the product now reaching customers.
The missing information matters more than the size of the multiplier. The release material does not disclose the starting rate, distance driven, conditions tested or definition of a critical intervention. It therefore cannot establish an absolute reliability level, support a like-for-like comparison with another driving system, or show how crash risk changed. Applying the multiplier to an unrelated v14 driving sample would not resolve those gaps.
The U.S. National Highway Traffic Safety Administration, the federal vehicle-safety agency, opened a preliminary evaluation on October 7, 2025, covering an estimated 2,882,566 vehicles equipped with FSD (Supervised) or FSD (Beta). Its opening record identified red-light violations and travel against the proper direction of traffic among the behaviors under examination. The opening record tabulated 58 incidents, including 14 crashes and 23 reported injuries, with no fatalities listed. Those historical reports are not a v15 performance sample or an exposure-adjusted crash rate: the estimated vehicle population does not establish how far or how often FSD was used.
The agency’s concern extended to whether drivers could supervise effectively. Some reported lane changes into opposing traffic appeared to leave little warning or opportunity to intervene. Its review would examine warnings, drivers’ response time and the system’s ability to detect, display and respond to signals and road markings. The document also says Tesla had taken action to address an intersection in Joppa, Maryland, where multiple incidents had occurred.
NHTSA’s October 2025 opening record tabulates historical FSD reports. These counts do not measure v15 performance or exposure-adjusted crash risk. Source: NHTSA. Source: National Highway Traffic Safety Administration.
The next test is driving evidence
The immediate milestones are the planned wider December release and evidence from the first customers about navigation, hazards and speed control. Older-hardware owners still need a concrete upgrade timetable. Full-route navigation learning remains a promised follow-up rather than a feature delivered across the route at launch.
For the safety claim to become useful evidence, Tesla would need to disclose absolute intervention results, their definition and the exposure and conditions behind them. Independent testing would then need to establish whether gains persist across roads and situations. Reality View adds a practical question: whether displaying planned actions gives drivers enough useful warning to intervene before an unsafe maneuver.
Moore Threads is preparing a December 29 reveal for Lushan, a desktop GPU pitched at gaming and local AI inference. Its claimed memory record remains unquantified, and capacity alone will not establish performance or value.
Amazon is in advanced talks to buy Decart, whose software helps AI models run efficiently across chips. The reported deal could strengthen AWS’s Trainium business, with conditional payments and five-year founder commitments making execution central to its value.
The University of California is expanding ChatGPT and Codex access across 10 campuses, with extra charges for advanced models and an unresolved dispute over faculty authority.
Ray-Ban’s US store cuts the Gen 2 Wayfarer by 25% to $284.25. Buyers still face separate prescription-lens prices, a cheaper Gen 1 option and newer glasses outside Meta’s frame sale.
Ubisoft has made the team behind its Teammates experiment a dedicated subsidiary targeting AI characters with memory and outside game creators. Earlier playtests show both progress and limitations, while commercial plans remain undisclosed amid group cost cuts and refinancing.
Neuralink has shown a trial participant controlling GTA VI through intended movements. Earlier gaming use and company-reported trial results put the demonstration in context, while its reliability across users remains unmeasured in the reports.
SoftBank founder Masayoshi Son reportedly explored financing a $1 trillion acquisition of memory-chip maker SK hynix, but the company says the deal is no longer under active consideration. The target brings surging earnings, existing shareholders and customer commitments that complicate any simple equation between ownership and access to AI-memory supply.
Nvidia reported $115.8 billion in quarterly revenue, but its shares fell 14% Friday amid a wider technology selloff, a leadership transition and a reported accounting investigation. Strong cash generation offers a counterweight to concerns about the durability of AI spending.
The Nasdaq 100 fell 5.6% on November 20 and closed 19.8% below its record after an intraday bear-market breach. Nvidia’s revenue surge and Anthropic’s huge IPO did not halt the rout; infrastructure bills and resilience elsewhere in the market complicate the picture.
UN-backed talks agreed on two more years of work toward an autonomous-weapons instrument, without committing to a treaty. Consensus kept major military powers involved while allowing opponents to weaken the mandate.
OpenAI reports sharply lower reasoning-extraction rates in internal GPT-6.2 tests and backs sanctions against illicit foreign distillation. Its own research and rival Anthropic’s experience show why protecting hidden reasoning leaves capability transfer and customer costs unresolved.
Alberta proposes 600-metre setbacks for large data centres and expanded utility oversight, but exemptions and rules still to be written will determine how much protection communities receive.
President Donald Trump backed Commerce Secretary Howard Lutnick as lawmakers scrutinize alleged delays in restricting a suspected route for AI chips to China. A new congressional inquiry also questions Nvidia’s influence on an existing investigation.
Amazon’s month-old Echo Time is $50 off for Black Friday. Its touchscreen, home hub and optional sleep-aware alarm cost $90 more than the reported Echo Spot price, while Alexa+ access and evidence of better waking remain key buying questions.
Merck will pair its experimental data with Isomorphic Labs’ AI to seek small-molecule cancer drug candidates. The deal follows Isomorphic’s $5 billion financing, but its payments, target and development timetable remain undisclosed.
Thinking Machines Lab released a roughly 20GB open-weights model that reuses its middle layers for reasoning. It brings the Inkling family to local hardware, with competitive coding results but workload-dependent costs and preliminary evidence on oversight.
Apple’s first HomePod mini update since 2020 adds Mac audio and four-speaker surround, but Siri AI remains unavailable, core home-hub features carry over, and reviews disagree on playback reliability.
Treasury Secretary Scott Bessent says an unnamed Anthropic founder will announce a large gift to children's investment accounts within weeks, but the amount, recipients and connection to existing philanthropic pledges remain undisclosed.
Google says air cooling will limit water use at its Visakhapatnam AI hub, but environmental proceedings and residents’ accounts raise broader questions about land, electricity, public incentives and the evidence behind promised benefits.