Tesla’s latest Full Self-Driving Supervised release, FSD v14.2, is not just another incremental software push, it meaningfully changes how much attention the car demands from the person behind the wheel. The update loosens some of the strictest driver monitoring behaviors and, in practice, gives owners more leeway to glance at their phones while the system handles more of the driving workload. At the same time, it layers in new stats, interface tweaks, and neural network upgrades that make the software feel more confident, even as the human is asked to do a little less.
FSD v14.2 arrives as a limited but significant rollout
I see FSD v14.2 as a pivotal test of how far Tesla is willing to push supervised automation while regulators still classify it as driver-assistance. The company has begun a limited rollout of the 2025.38.9.5 build, framed as a major Full Self-Driving Supervised update that reaches a subset of owners first before a broader release. Reporting on the official release notes describes this version as adding a dedicated self-driving statistics feature and other refinements that sit on top of the existing supervised stack, which is still marketed as requiring active human oversight even as it takes over more of the driving task in city streets and on highways, according to the detailed breakdown of the limited rollout.
The 2025.38.9.5 release notes show that Tesla is treating v14.2 as a cohesive package rather than a quiet background patch, with explicit references to FSD Supervised v14.2 and a list of behavioral and interface changes that owners can see as soon as the car reboots. Those notes highlight the new self-driving stats tile, updated visualizations, and a series of driving behavior improvements that build on the company’s end-to-end neural network approach, all bundled under the same software version identifier in the 2025.38.9.5 release notes. Taken together, the documentation positions v14.2 as a step toward a more transparent and less intrusive FSD experience, even though the legal framing still insists the driver is responsible.
New self-driving stats and UI polish change how drivers relate to FSD
One of the most striking shifts in v14.2 is how it surfaces data about the system’s own performance, which subtly reframes the relationship between driver and software. Instead of treating FSD as a black box that either works or disengages, the new build introduces a dedicated stats interface that tracks how often the system is used and how it behaves, giving owners a clearer sense of their own reliance on automation. Coverage of the update notes that this statistics view appears as a new tile in the vehicle’s interface, alongside broader user interface refinements and a neural network upgrade that is described as part of the same package in the detailed overview of the FSD v14.2 stats and UI improvements.
I read this as Tesla trying to normalize continuous FSD use by making its operation feel more like a trackable habit than an experimental feature. The stats view, combined with smoother visualizations and updated prompts, encourages drivers to think of FSD Supervised as a core part of daily driving rather than an occasional assist, while still reminding them that the system is logging their behavior. That framing is reinforced by the way the release notes and early coverage emphasize the neural net upgrade alongside the interface changes, presenting the software as both smarter and more transparent in the same breath, a pairing that is spelled out in the description of the major FSD v14.2 update.
Driver monitoring relaxes, with more tolerance for phone use
The most controversial change in v14.2 is the way it dials back some of the strictest driver monitoring behaviors, especially around phone use. Owners who have installed the update report that the system is less aggressive about issuing nags or disengaging when the driver briefly looks down or interacts with a device, which effectively makes it easier to use a phone while FSD is active. A widely shared summary of the update describes this as easing driver monitoring and explicitly allowing more leniency for cell phone use, characterizing the shift as a deliberate softening of the camera-based attention checks in the FSD v14.2 driver monitoring change.
From a safety perspective, I see a tension between Tesla’s push to make FSD feel less nagging and the reality that the system still requires a fully attentive human. The company has long argued that its in-cabin camera and steering wheel torque checks are enough to keep drivers engaged, but loosening those thresholds risks normalizing more distracted behavior at exactly the moment the software is taking on more complex maneuvers. That tradeoff is at the heart of the current debate, and it is underscored by the way early coverage pairs the monitoring changes with the broader description of v14.2 as a major step forward in capability, a framing that appears in the same discussion of the FSD v14.2 update.
On-road testers report smoother turns and more human-like decisions
Beyond the monitoring tweaks, early testers are focusing on how v14.2 behaves in real traffic, especially in complex scenarios that have historically tripped up FSD. Longtime community evaluators who run repeatable routes are reporting that the new build handles unprotected turns, merges, and lane selection with fewer abrupt corrections and less hesitation, which makes the car feel more predictable to the person in the driver’s seat. A detailed thread comparing runs from Chuck Cook and Dirty Tesla, two well-known FSD reviewers, notes that v14.2 shows improved confidence in challenging maneuvers and fewer disengagements on their usual test routes, based on their observations from Chuck Cook’s and Dirty Tesla’s drives.
In my view, those anecdotal reports matter because they come from drivers who have documented FSD’s quirks across multiple versions, giving them a baseline for comparison. When they describe v14.2 as smoother or more human-like, they are usually reacting to specific behaviors like how the car creeps into intersections, negotiates four-way stops, or chooses gaps in traffic. Video reviews of the new build show the system taking cleaner lines through turns and making fewer last-second lane changes, which aligns with the claim that the underlying neural networks have been upgraded in this release, as seen in one early FSD v14.2 drive test.
Video tests highlight both progress and lingering edge cases
Watching full-length drives of v14.2, I see a system that is more composed but still far from infallible, especially in dense urban environments. Reviewers who record their commutes and test loops are documenting fewer hard braking events and less ping-ponging within the lane, which makes the ride feel less robotic to passengers. At the same time, they still encounter occasional awkward decisions around parked cars, construction zones, and unusual intersections, which require the driver to intervene even though the software is marketed as handling “Full Self-Driving” in supervised form, a balance that is evident in one extended v14.2 urban drive.
Highway behavior appears to benefit from the same neural network improvements, with testers noting smoother merges and more natural following distances that reduce the sense of constant micro-adjustments. Yet the videos also capture moments where the car hesitates longer than a human would when entering fast-moving traffic or navigating complex interchanges, reminding viewers that the driver must be ready to take over at any time. That duality, a more capable system that still demands vigilance, comes through clearly in another recorded FSD v14.2 highway session, where the driver praises the progress but still logs multiple manual interventions.
Balancing comfort, accountability, and the path to wider release
What stands out to me about FSD v14.2 is how it tries to make supervised autonomy feel less burdensome while simultaneously documenting more about how it is used. The relaxed driver monitoring and increased tolerance for brief phone interactions make the system feel less like a nagging co-pilot, which many owners will welcome on long commutes. At the same time, the new self-driving stats interface and expanded release notes signal that Tesla is tracking usage patterns closely, perhaps to inform future policy decisions or to demonstrate to regulators how often the system runs without incident, a dynamic that is reflected in the detailed description of the self-driving stats feature.
As the limited rollout of 2025.38.9.5 expands, I expect the debate over distraction and responsibility to sharpen, because v14.2 makes it easier for drivers to mentally offload more of the driving task while still being legally on the hook. The company’s own documentation and community reports frame this update as a major step forward in capability and comfort, but they also underscore that FSD Supervised is not autonomous in the regulatory sense and that the human must remain ready to intervene. That tension, between a smoother, less intrusive experience and the unchanged requirement for constant oversight, is the core story of v14.2, and it will shape how owners, regulators, and the broader public judge Tesla’s next moves in supervised self-driving.
More from MorningOverview