Tesla's FSD has successfully completed a long-distance journey across the United States under real traffic conditions for the first time, covering complex scenarios such as highways, urban roads, night driving, and multiple entries and exits to supercharging stations. The entire journey was 2,732 miles and relied 100% on FSD. Meanwhile, the head of NVIDIA’s robotics business noted that Tesla FSD v14 exhibits highly natural driving behavior, making it “difficult to distinguish whether it is driven by a human” in actual experience. Therefore, it may have passed the 'physical Turing test.'
$Tesla (TSLA.US)$ Tesla's Full Self-Driving (FSD) system is crossing a crucial threshold.
Recently, a Model 3 equipped with FSD v14 departed from Los Angeles on the West Coast of the United States, traversed the entire continent, and arrived in South Carolina on the East Coast within two days and twenty hours. The entire journey covered 2,732 miles, relying 100% on FSD, encompassing complex scenarios such as highways, urban roads, nighttime driving, and multiple entries and exits to Supercharger stations, without any human intervention throughout the trip.
This was not an official demonstration or laboratory test but a real-world driving record completed by an ordinary car owner under actual traffic conditions.
For the autonomous driving industry, the significance of this journey goes far beyond 'driving long distances'; it makes one question tangible and specific for the first time: Can FSD completely replace human drivers?
Zero-intervention coast-to-coast drive across the United States
The person who completed this transcontinental journey across the United States was Tesla owner Davis Moss.
According to Moss’s post on social media platform X, the Model 3 he drove was equipped with AI4 hardware and ran FSD v14.2.1.25. Based on data from the FSD database and community trackers, prior to completing this cross-country trip, Moss had driven 10,638.8 miles entirely relying on FSD.
Moss started from Tesla’s restaurant in Los Angeles and eventually reached Myrtle Beach in South Carolina, fully utilizing FSD v14 without any form of human intervention.
This 2,732.4-mile journey covered diverse road environments in the United States, including highways, urban roads, and various traffic conditions. Moss emphasized that FSD v14.2 not only completed the entire drive but also handled all parking operations, including automatic parking at Tesla Supercharger stations, with 'not a single hazardous situation occurring' throughout the trip.
Elon Musk personally retweeted his congratulations immediately. Notably, this route was one that Musk had repeatedly mentioned since the release of Autopilot 2.0 in 2016 but had never been able to fulfill. Back then, he estimated that Tesla could achieve 'coast-to-coast' autonomous driving by 2017. Looking back now, this goal was not unattainable.
The Tesla community has responded enthusiastically to this achievement, as a coast-to-coast drive with zero interventions has long been regarded as a significant indicator of the maturity of autonomous driving technology. Tesla's official North American account confirmed on social media: "The first Tesla to complete a fully autonomous coast-to-coast drive using FSD Supervised, with zero interventions and full FSD operation."
Passing the 'Physical Turing Test'?
Coincidentally, at this juncture, Jim Fan, head of NVIDIA’s robotics division, offered an intriguing assessment: Tesla FSD v14 may have passed the 'Physical Turing Test.' This concept is derived from the classic Turing Test proposed by mathematician Alan Turing in 1950 but shifts the evaluation criteria from text-based dialogue to physical behavior in the real world. If an observer cannot distinguish whether a task was performed by a human or a machine, then the machine has passed the test.
After test-riding FSD v14, Fan remarked that it has become difficult to tell the difference.
During his test ride, Tesla’s driving behavior did not resemble that of a rigid rule-following robot but rather appeared akin to a cautious and experienced human driver—gradually moving forward at intersections, braking smoothly, changing lanes naturally, and responding to subtle cues that are challenging to achieve through manual programming.
Fan believes that passing the Physical Turing Test requires overcoming four major challenges: understanding three-dimensional space, handling objects with precision, mastering real-world contextual knowledge, and bridging the gap between digital commands and physical actions. Driving happens to encompass all four challenges, making it one of the most difficult embodied AI problems to solve.
7 Billion Miles of Data Advantage
The breakthrough performance of FSD v14 stems from Tesla's shift from a rule-based system to an end-to-end neural network. Early autonomous driving systems relied on hard-coded rules, whereas FSD v14 learns driving patterns in a human-like manner through training on tens of billions of miles of real-world driving data.
Tesla claims that its vehicles equipped with FSD have cumulatively driven nearly 7 billion miles, with approximately 2.5 billion miles completed in urban environments. Urban driving is far more complex than highway driving, involving unprotected turns, unpredictable pedestrians and cyclists, traffic lights, and adverse weather conditions—scenarios where autonomous driving systems face their greatest challenges.
A recent case further demonstrated the capabilities of the system: a video posted by a Tesla owner showed FSD operating continuously for 7 hours in a hailstorm with extremely poor visibility and severe road flooding, without any human intervention. This performance made the concept of 'human-like driving' feel increasingly tangible.
Fan likened this transformation to the proliferation of smartphones: initially astonishing, then becoming commonplace, and eventually indispensable. If machines can move and behave in the real world as naturally as humans, it will pave the way for robots that understand intentions rather than merely following instructions.
However, Tesla's system still requires human supervision, and even its most enthusiastic supporters acknowledge that perfection is unattainable. Whether driven by humans or machines, vehicles always carry inherent risks. The system is currently defined as "FSD Supervised," requiring drivers to remain vigilant and ready to take over at any time.
Editor/Rocky