2 days and 20 hours, zero takeover crossing the United States—has Tesla FSD passed the "physical Turing test"?

Tesla's Full Self-Driving system (FSD) is crossing a critical threshold.

Recently, a Model 3 equipped with FSD v14 set off from Los Angeles on the US West Coast, crossed the entire continent, and arrived at Myrtle Beach, South Carolina on the East Coast in 2 days and 20 hours. The total distance was 2,732 miles, relying 100% on FSD, covering highways, city roads, nighttime driving, and multiple entries and exits from Supercharger stations—handling complex scenarios with no human intervention throughout the trip.

This wasn't an official demonstration, nor a lab test, but a real road record completed by a regular owner in genuine traffic conditions.

For the autonomous driving industry, the significance of this journey goes far beyond just “going the distance”: for the first time, it makes a certain question tangible and realistic: Has FSD now become fully capable of replacing human drivers?

Zero Interventions Across the US

The one who completed this cross-country journey was Tesla owner Davis Moss.

According to Moss's posts on social media platform X, his Model 3 is equipped with AI4 hardware running FSD v14.2.1.25. According to FSD databases and community trackers, before this coast-to-coast drive, Moss had already driven 10,638.8 miles using FSD, relying on it 100% of the time.

Moss set off from the Tesla restaurant in Los Angeles, ultimately arriving at Myrtle Beach, South Carolina, using FSD v14 for the entire trip, with no manual interventions whatsoever.

This 2,732.4-mile journey covered America's diverse road environments, including highways, city streets, and various traffic conditions. Moss emphasized that not only did FSD v14.2 complete the entire drive, it also handled all parking operations, including automatic parking at Tesla Supercharger stations. There was “not even a single dangerous moment” throughout.

Elon Musk himself was quick to share his congratulations. Notably, this particular route is the same one Musk has mentioned repeatedly since the release of Autopilot 2.0 in 2016—a goal that was spoken of but never delivered. Back then, he predicted Tesla could achieve “coast to coast” autonomous driving by 2017. In hindsight, this goal was not impossible.

The Tesla community responded enthusiastically to this achievement. Crossing the continent with zero interventions has always been viewed as a crucial milestone for the maturity of autonomous driving technology. Tesla’s North America official account confirmed on social media: “The first Tesla to drive autonomously coast-to-coast using FSD Supervised, zero interventions, full FSD.”

Passing the “Physical Turing Test”?

Coincidentally, at this very moment, NVIDIA’s Head of Robotics, Jim Fan, offered an intriguing assessment: Tesla FSD v14 may have already passed the “Physical Turing Test.” This concept stems from mathematician Alan Turing’s classic test from 1950, shifting the standard from textual interaction to physical behavior in the real world. If an observer cannot distinguish whether a task has been done by human or machine, the machine passes the test.

After riding in FSD v14, Fan’s impression is: It’s already hard to tell.

In his ride experience, Tesla’s driving behavior didn’t feel like a rigid, rule-following robot, but rather resembled a cautious, experienced human driver—proceeding slowly at intersections, braking smoothly, changing lanes naturally, and reacting to subtle cues that are nearly impossible to replicate with manual programming.

Fan believes that passing the Physical Turing Test requires solving four challenges: understanding three-dimensional space, handling objects with precision, mastering real-world contextual knowledge, and bridging the gap between digital instructions and physical actions. Driving happens to incorporate all four, making it one of embodied AI’s hardest problems to conquer.

7 Billion Mile Data Advantage

The breakthrough with FSD v14 comes from Tesla’s transition from rule-based systems to end-to-end neural networks. Early autonomous systems relied on hard-coded rules, but FSD v14 learns driving behavior in a human-like way through training with billions of miles of real-world driving data.

Tesla reports that its vehicles equipped with FSD have now driven nearly 7 billion miles in total, about 2.5 billion of which were in city environments. City driving is far more complex than highways, involving unprotected turns, unpredictable pedestrians, cyclists, traffic lights, and severe weather—these are the scenarios most challenging for autonomous driving systems.

A recent case further demonstrates the system’s capability: a Tesla owner posted a video showing FSD driving continuously for 7 hours in an intense hailstorm with severely limited visibility and waterlogged roads, with no human intervention at all. This brings the concept of “human-like driving” closer to reality.

Fan compares this evolution to the popularization of smartphones: at first astounding, then commonplace, ultimately indispensable. If machines can move and act in the real world as naturally as humans, this opens the way for robots that understand intent, not just follow instructions.

However, Tesla’s system still requires human supervision—even the most enthusiastic supporters concede that perfection is impossible. Whether driven by humans or machines, cars always face risks. The system is currently defined as “FSD Supervised,” requiring drivers to stay alert and ready to intervene at any time.

Risk disclosure and disclaimerThe market has risks, and investments need caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of any particular individual. Users should consider whether any opinions, viewpoints, or conclusions in this article are suitable for their circumstances. Investing based on this article is at your own risk.