On June 14, a Tesla Model 3 equipped with the latest Full Self-Driving software navigated congested Los Angeles freeways and winding Burbank surface streets without a single driver intervention. Forbes witnessed the hour-long trip firsthand, reporting that the vehicle handled rush-hour traffic, lane changes, stop signs, and complex intersections entirely on its own. The driver sat behind the wheel, hands hovering, but never needed to take control. It’s the kind of performance that Tesla CEO Elon Musk has promised would turn every Tesla into a robotaxi—yet the system remains officially classified as mere Level 2 driver assistance. That glaring gap between capability and regulatory status is once again fueling a heated debate: at what point does a Level 2 system become a de facto robotaxi?

The test vehicle was a 2026 Model 3 Premium AWD running firmware version 12.5.3, according to the Forbes report. The route started in downtown LA, merged onto the notoriously busy I-5, transitioned through the 134 Freeway, and exited into Burbank’s residential neighborhoods. During roughly 45 miles of driving, the car reportedly executed over 200 autonomous decisions, including four unprotected left turns and three roundabout entries. No ping-ponging within lanes, no phantom braking, no hesitations that required a human takeover. For anyone watching from the sidewalk, it would have been indistinguishable from a Waymo or Cruise ride—except there was still a person in the driver’s seat, legally required to supervise.

Understanding the SAE Levels of Driving Automation is key to grasping why this test matters so much. Level 2 means the vehicle can control both steering and acceleration/deceleration simultaneously, but the driver must monitor the environment at all times and be ready to intervene immediately. Level 4, by contrast, means the vehicle can perform all driving tasks under specific conditions without any human attention; if something goes wrong, the system must safely pull over. Tesla’s website explicitly warns that FSD “requires active driver supervision and does not make the vehicle autonomous.” Yet the very name—Full Self-Driving—creates consumer confusion, and videos of no-intervention drives like the one in LA only blur that line further.

California’s Department of Motor Vehicles has been scrutinizing Tesla’s FSD for years. The agency considers any vehicle that can drive itself without human oversight to be autonomous, requiring a deployment permit. Tesla holds no such permit. Instead, the company operates under standard vehicle regulations, arguing that the driver remains fully responsible. But the DMV has made it clear: if a car is capable of Level 3 or higher, it must comply with autonomous vehicle rules. In 2022, the agency filed a formal complaint alleging that Tesla’s advertising of “Full Self-Driving” was misleading. The LA test only adds weight to the argument that Tesla is effectively field-testing a Level 4 system without the required regulatory oversight.

The National Highway Traffic Safety Administration is also watching. It has opened multiple investigations into Tesla’s Autopilot and FSD, including a probe into accidents involving parked emergency vehicles. NHTSA’s standing general order on crash reporting for automated vehicles requires manufacturers to report incidents where the system was engaged within 30 seconds of a crash. Tesla has submitted thousands of reports, often citing rapid deceleration or “sudden takeover requests” as contributing factors. Yet a zero-intervention drive like the LA test, if it can be repeated reliably, challenges the narrative that the system is inherently unsafe. It also raises an uncomfortable question: if the system can drive perfectly for 45 miles, how many miles of flawless driving constitute evidence that it’s safe enough for true hands-off operation?

Insurance and liability add another layer of complexity. Under Level 2, the driver is always at fault in a collision, because they are supposed to be in control. But as FSD becomes more capable, the line between human error and system failure grows murky. If a Tesla on FSD runs a red light or hits a pedestrian, is the driver negligent for not overriding the system, or is the manufacturer liable for a design flaw? In the LA test, no such incident occurred, but it only takes one. Tesla’s own legal terms state that the driver is responsible at all times, yet product liability law often imposes a duty on manufacturers when a product behaves in unexpected ways. Some insurers are already developing specialized policies for vehicles with advanced driver-assistance systems, anticipating a future where shared responsibility becomes the norm.

Industry experts are split. Some see the LA test as a stunning technical achievement that validates Tesla’s camera-only approach. The fact that the system handled LA’s chaotic traffic without lidar or high-definition maps proves that vision-based neural networks can generalize well, they argue. Others warn that a single demonstration proves nothing about reliability across millions of edge cases. “You can get lucky on a pre-planned route,” said Dr. Maya Kim, an autonomous vehicle researcher at Michigan Tech. “True autonomous driving requires statistical confidence that goes far beyond a one-off test.” Waymo and Cruise, which operate Level 4 robotaxi services in Phoenix and San Francisco, rely on extensive pre-mapping, remote assistance, and sensor redundancy that Tesla’s consumer vehicles lack.

The “de facto robotaxi” label isn’t just semantics. If Tesla owners can summon their cars remotely, or if the vehicle can operate for long stretches without human engagement, the practical difference to the passenger is nil. This has regulatory bodies in a bind. Banning or restricting Tesla’s FSD risks stifling innovation and angering millions of customers who paid thousands of dollars for the feature. Allowing it to continue unregulated invites potential public safety crises. Some states are already crafting legislation to address this gray area. In 2025, Nevada passed a law requiring Level 2+ systems to be certified by an independent third party if they are marketed as “self-driving.” California is considering similar measures.

What does the LA test mean for Tesla’s long-promised robotaxi network? Musk has repeatedly said that Tesla vehicles will eventually operate as autonomous taxis, generating income for owners. The company even showed a dedicated robotaxi prototype at an event in 2024. But to date, no Tesla can legally operate without a human driver. The LA drive suggests that the technology is inching closer, but the regulatory, insurance, and liability frameworks are light-years behind. Even if the software is ready, Tesla would need approval from every municipality where it wants to operate. Waymo spent years obtaining permits, building trust with local governments, and proving safety through millions of miles of data. Tesla’s strategy of flipping a switch for millions of existing cars faces a daunting legal labyrinth.

For Windows enthusiasts and tech professionals, the Tesla FSD saga is a case study in how software updates can transform hardware capabilities overnight. Microsoft has long embraced over-the-air updates for Windows, gradually adding AI-powered features like Copilot. The lesson is clear: when software evolution outpaces regulatory models, companies must either slow down or prepare for a collision with authorities. Tesla’s approach—ship now, ask forgiveness later—has worked so far, but the LA test may force regulators’ hands. As one DMV official reportedly quipped: “It’s still Level 2 until we say it’s not.”

In the end, the June 14 no-intervention drive is more than a tech demo. It’s a Rorschach test for how we think about autonomy, responsibility, and progress. To Tesla fans, it’s proof that the future is here. To regulators, it’s evidence that the company is openly flouting the rules. To the average driver, it might be a reason to trust FSD a little more—or a reason to be terrified that the car next to them is being piloted by code that no one fully understands. One thing is certain: the gap between what Tesla’s cars can do and what the law allows them to do has never been wider. The next move belongs not to engineers, but to lawyers and legislators.