A recent video posted by popular YouTuber Mark Rober comparing Tesla's camera-only autopilot system to competing LiDAR technology has ignited lively discussions around autonomous vehicles. The video, released during the week of March 17, 2025, aimed to examine the effectiveness of Tesla's Full Self-Driving (FSD) capabilities against those equipped with LiDAR.
Taking inspiration from the Road Runner cartoon series, Rober’s experiment tested how well each vehicle could react to seeing a child unexpectedly dash into the road across six scenarios, including fog, rain, bright lights, and a whimsical test featuring a wall painted to resemble a road.
In general, the results indicated that the LiDAR-equipped vehicle outperformed the Tesla in most tests, especially in adverse weather conditions such as fog and rain, where visibility was severely limited. However, both cars displayed impressive capabilities in scenarios involving a child sprinting into the street or facing bright headlights.
Tesla, opting for a vision-based approach, raised eyebrows within the industry, especially after CEO Elon Musk insisted on removing radar sensors to exclusively rely on visual data through cameras. This decision has resulted in a vocal divide: some experts advocate for a multi-sensor approach, while others endorse Musk's vision-only strategy.
The video has gained considerable attention, attracting around 17.5 million views in its first week after release, reflecting not just the interest in Tesla but also the escalating stakes within the burgeoning EV market. Following the video's debut, stock prices for Luminar, the company behind the LiDAR test vehicle's technology, surged from $5.05 to $8.35, a remarkable 65% increase.
Critics of Rober's test have raised serious concerns about the reliability and validity of the results. Many have pointed out that the test should have utilized Tesla's Autopilot correctly, as Rober engaged the system merely moments before hitting the wall, an action that can compromise safety and accuracy in results.
Rober did not disclose which version of Tesla's FSD software was utilized for his test. This omission is vital, as it influences how the vehicle engages with obstacles. A lack of software transparency regarding the model year of the Tesla—whether it features the legacy Hardware 3 (HW3) or the newer HW4—compounded doubts about the test's rigor.
In response to the criticism, Rober later posted raw footage on X (formerly known as Twitter), aiming to defend the authenticity of his experiment. However, many in the Tesla community believe that improvements are necessary to mitigate flaws evident in contemporary Autopilot functionalities.
To further assess the accuracy of Rober's findings, another Tesla YouTuber, Kyle Paul, attempted to recreate the painted wall experiment. His tests confirmed that the FSD v12.5.4.2 model did not recognize the obstacle until it was alarmingly close, necessitating human intervention to prevent a collision.
Notably, the Cybertruck equipped with FSD v13.2.8 software passed the painted wall test without human assistance, indicating that Tesla's ongoing advancements in AI-driven technology could be improving performance, even if earlier iterations underperformed.
As excitement and skepticism grow regarding the trajectory of autonomous driving, the broader question remains: Is it still possible for Tesla to catch up with other technologies that utilize LiDAR's capabilities, especially as they develop? Some experts emphasize that having more sensors translates to enhanced safety and functional redundancy.
Yet, proponents of Tesla's vision-based approach argue that the extensive data collection by the company over the years fortifies its basis for relying solely on cameras. They assert that with millions of Teslas collecting driving data, the vehicle can continuously learn and adapt to diverse driving scenarios.
This ongoing discourse around the effectiveness of LiDAR versus camera-only systems in self-driving technology is critical, especially as manufacturers strive for feats in autonomous driving safety and reliability. As for the current innovation race, it looks as if Tesla and its rivals will continue to embark on different paths, each convinced of their methods to reach a fully autonomous future.
