Tesla’s Full Self-Driving technology has long promised to transform the driving experience, but a Canadian technology personality’s recent test revealed how quickly confidence can turn into concern. Linus Sebastian, the creator behind the hugely popular Linus Tech Tips YouTube channel, experienced both the impressive capabilities and potentially dangerous limitations of Tesla’s advanced driver-assistance system.
What began as an unexpectedly positive experience took a troubling turn when a Tesla Model 3 behaved unpredictably at an intersection, forcing Sebastian to intervene. The incident, documented in a video released on November 5, 2026, raised important questions about driver supervision, trust in automated systems, and the gap between technology that performs remarkably well and technology that can reliably handle every situation.
A Canadian Technology Star Takes on Tesla’s Biggest Promise
Linus Sebastian has spent years examining consumer technology, challenging manufacturers, and explaining complicated products to millions of viewers. His Canadian-based Linus Tech Tips channel has approximately 17 million subscribers, giving his opinions considerable influence among technology enthusiasts. Tesla, however, has remained a company whose products and decisions he has frequently questioned. That made his decision to test Full Self-Driving particularly interesting. Rather than relying on demonstrations or promotional material, Sebastian wanted to experience how the technology behaved during ordinary driving, where unpredictable traffic conditions could expose limitations that controlled demonstrations might miss.
The experiment involved a 30-day vehicle exchange with colleague Colton, who provided his newer Tesla Model 3 equipped with Tesla’s Hardware 4 computer and camera system. In return, Colton would drive Sebastian’s Porsche. The arrangement gave Sebastian an opportunity to evaluate Tesla’s technology over an extended period rather than during a short dealership test drive. Having previously experimented with comma.ai’s driver-assistance system, he also had another technology to compare against Tesla’s more ambitious approach. The central question was whether Full Self-Driving could genuinely make everyday transportation easier without introducing unacceptable risks.
Early Driving Impressions Leave Even a Tesla Skeptic Impressed
The first portion of Sebastian’s experience was considerably more positive than his previous criticism of Tesla might have suggested. Once Full Self-Driving was activated, the Model 3 began navigating roads, managing steering, and responding to surrounding traffic with surprisingly little assistance. Sebastian openly acknowledged the technology’s capabilities, at one point commenting that he had not touched the steering wheel during the drive. For a technology reviewer accustomed to finding problems with new products, that level of automation represented a meaningful achievement. It demonstrated how far advanced driver-assistance technology has progressed beyond conventional cruise control and basic lane-keeping functions.
Not everything about the vehicle impressed him. Sebastian questioned aspects of Tesla’s minimalist interior, including its touchscreen-based gear selection and the absence of traditional turn-signal stalks. Nevertheless, the driving software challenged some of his expectations. Even when the car made an unusual lane change before returning to its previous position, he did not immediately classify the behaviour as a serious failure. The experience suggested that automation could eventually reduce the effort associated with routine commuting. Yet there was a critical distinction between finding the technology impressive and being able to trust it completely. That distinction became increasingly important as the test continued.
A Troubling Intersection Incident Changes Everything
The most alarming moment arrived approximately 26 minutes into the video, when the Tesla approached an intersection with another vehicle preparing to turn left. According to Sebastian’s account, the Model 3 activated its right-turn signal but then continued travelling straight. The unexpected manoeuvre created a potentially dangerous situation because the other motorist could have interpreted the signal as an indication that the Tesla would turn away from their path. Instead, the vehicle continued toward the intersection, prompting Sebastian to press the brake pedal and disengage Full Self-Driving. The incident occurred without a reported collision, but it exposed the consequences that can arise when a vehicle’s movements do not match the intentions communicated to surrounding drivers.
What made the situation especially concerning was that the Tesla had performed so convincingly during much of the earlier journey. Sebastian had grown accustomed to the software making reasonable decisions without his assistance. Suddenly, he was responsible for recognizing a potentially unsafe manoeuvre and correcting it before the situation escalated. After intervening, he needed a moment to collect himself and discuss what had happened. The experience illustrated an uncomfortable reality of supervised automation: a system can manage numerous complicated driving situations successfully and still make an unexpected decision that requires immediate human judgment.
A Navigation Confusion Raises Questions About What Went Wrong
One important detail complicated the interpretation of the intersection incident. Before the unexpected manoeuvre, Sebastian and his colleague had discovered that the Tesla appeared to be operating without a destination entered into its navigation system. The vehicle was still capable of moving under Full Self-Driving, but it was not following a clearly programmed route. That raised questions about whether the system’s unusual behaviour was connected to the missing destination, an incorrect assumption about the road layout, or another issue entirely. The recording did not establish a definitive technical explanation, making it inappropriate to attribute the incident to a particular software defect.
The uncertainty does not eliminate the safety concern. Even without a navigation destination, other motorists reasonably expect a vehicle’s signals and movements to communicate a consistent intention. An incorrect indicator can influence how surrounding drivers make their own decisions, especially at intersections where several vehicles may be moving simultaneously. Sebastian’s experience also highlighted why understanding an advanced vehicle’s operating conditions matters. Features that appear straightforward can behave differently depending on settings, route information, road markings, and environmental conditions. For drivers, the practical lesson is that automated steering and acceleration do not remove the need to understand what the vehicle is attempting to do.
The Hidden Danger of Becoming Too Comfortable With Automation
Perhaps the most revealing part of Sebastian’s reaction was his explanation of why the intersection incident caught him off guard. During his earlier experience with comma.ai’s driver-assistance technology, he had remained more conscious of its limitations and more prepared to intervene. Tesla’s system, by comparison, had performed so well that he found himself becoming more confident in its decisions. That confidence made the unexpected manoeuvre particularly unsettling. The issue was not simply whether the software could drive effectively, but whether consistently successful automation might encourage drivers to become less prepared for the occasional mistake.
Research supports the broader concern. A 2024 Insurance Institute for Highway Safety study conducted with MIT’s AgeLab found that drivers became more likely to engage in distracting activities when using partial driving automation. One component of the research examined 29 drivers using Volvo’s Pilot Assist over a four-week period, while another followed 14 Tesla Autopilot users. Researchers observed that some participants adapted their behaviour as they became familiar with the systems, including learning how to respond to attention warnings without consistently maintaining their focus. These findings do not establish how Sebastian behaved throughout his test, but they show why driver confidence and attention are important safety considerations as increasingly capable automation becomes commonplace.
Tesla’s Driver-Monitoring System Gets Put to the Test
Sebastian also experimented with Tesla’s driver-monitoring technology, which is designed to discourage motorists from becoming inattentive while Full Self-Driving is operating. During one portion of the recording, he deliberately tested how the system responded when it could not clearly monitor his eyes. The Model 3 issued attention-related warnings, eventually escalating its response and displaying an improper-use notification. The recording also showed the vehicle activating its hazard lights during the test. This demonstrated that Tesla has safeguards intended to detect certain forms of driver inattention, rather than simply allowing the vehicle to continue indefinitely without checking whether someone remains engaged.
Tesla’s Model 3 owner’s manual explains that Full Self-Driving uses the cabin camera to monitor driver attentiveness and can issue progressively stronger warnings if the driver fails to respond. Depending on the circumstances, the system can disable automated assistance for the remainder of a drive. However, attention monitoring cannot guarantee that a driver understands every developing traffic situation. A person might be looking toward the road and still misjudge what the vehicle intends to do next. The distinction matters because safe supervision requires more than satisfying a camera-based monitoring system. Drivers must maintain sufficient awareness to recognize unusual behaviour and respond appropriately, particularly when the technology is controlling steering, braking, and acceleration.
Full Self-Driving Still Requires an Attentive Human Driver
Despite its name, Tesla’s Full Self-Driving (Supervised) is classified as Level 2 driver assistance, not fully autonomous driving. The distinction is important. Under the commonly used SAE automation framework, Level 2 systems can control both steering and vehicle speed, but the human driver remains responsible for monitoring the road and responding to situations the technology cannot safely manage. Tesla’s Canadian documentation explicitly states that Full Self-Driving does not make its vehicles autonomous. It can navigate intersections, change lanes, respond to traffic, and perform other complicated manoeuvres, but those capabilities operate under the assumption that a fully attentive driver remains available to take control.
Tesla’s own Model 3 manual provides particularly relevant guidance for Sebastian’s experience. It warns that drivers should not assume how Full Self-Driving will behave at intersections and specifically identifies complex intersections as situations where intervention may be necessary. Interestingly, the manual also confirms that the system can operate without a programmed destination, choosing what it considers the most probable driving path. That means the missing destination during Sebastian’s test was not necessarily an unsupported use of the technology. Transport Canada similarly emphasizes that partial driving automation does not replace human attention or judgment. These warnings reinforce why the system’s abilities should not be confused with permission to disengage from driving responsibilities.
The Return Journey Shows Why Tesla’s Technology Remains Compelling
Despite the unsettling intersection incident, Sebastian did not abandon his evaluation or conclude that Full Self-Driving was fundamentally incapable. The return journey provided another demonstration of its abilities. The Tesla navigated changing traffic conditions, responded appropriately to a pedestrian, and continued toward the office with little need for manual assistance. Sebastian acknowledged that the experience remained impressive, describing the return drive positively even after the earlier scare. That reaction reflected the complicated reality of modern driver-assistance technology: an isolated but potentially serious mistake can coexist with long stretches of highly capable performance.
The return trip also provided a less dangerous reminder that automation does not always interpret a destination in the way a human would expect. Near the office, the Model 3 reached the wrong parking area and did not park where Sebastian intended. The episode was considerably less alarming than the intersection incident, but it demonstrated another limitation of automated decision-making. The vehicle could successfully perform the mechanical tasks of driving while still producing an inconvenient or unexpected result. For motorists considering Tesla’s technology, that combination is important. Full Self-Driving can reduce the amount of steering and speed management required during a journey, but the driver must remain responsible for judging whether the vehicle’s decisions are appropriate.
U.S. Safety Investigations Add to the Concerns
Sebastian’s experience comes amid broader regulatory scrutiny of Tesla’s automated driving technology. In October 2025, the U.S. National Highway Traffic Safety Administration opened an investigation covering approximately 2.88 million Tesla vehicles equipped with Full Self-Driving. The investigation initially examined 58 reports of potential traffic-safety violations, including 14 reported crashes and 23 injuries. Regulators focused particularly on allegations involving vehicles proceeding through red traffic signals, entering opposing traffic lanes, and performing unexpected manoeuvres at intersections. The inquiry remained open as of November 6, 2026. Importantly, the investigation itself does not establish that a defect exists, and Sebastian’s incident has not been demonstrated to share a technical cause with the reported cases.
The broader debate also involves how manufacturers measure and communicate automated driving safety. Tesla maintains that vehicles operating with Full Self-Driving experience fewer collisions than vehicles driven without it. However, a 2026 Reuters investigation raised concerns about the statistical comparisons supporting some of Tesla’s safety claims, including differences in crash-reporting thresholds and vehicle ages. Independent researchers argued that these factors could exaggerate the technology’s apparent safety advantage. None of this means that automated driving assistance cannot improve road safety. Instead, it highlights the importance of evaluating technology through reliable evidence, transparent reporting, and real-world performance rather than relying solely on successful demonstrations or promotional statistics.
What Sebastian’s Experience Means for Canadian Drivers
For Canadian motorists considering Tesla’s Full Self-Driving technology, Sebastian’s experience provides a useful perspective that is neither an outright endorsement nor a complete rejection. The system demonstrated that it could handle substantial portions of a journey with minimal intervention, potentially reducing some of the physical demands associated with routine driving. At the same time, the intersection incident showed why even capable automation cannot be treated as an independent driver. Transport Canada advises motorists to understand the limitations of their vehicle’s assistance systems, remain attentive, and recognize that performance can vary with road conditions, visibility, and infrastructure. Those considerations are particularly relevant on Canadian roads, where weather and road markings can change significantly throughout the year.
The most important takeaway from Sebastian’s test may be the relationship between trust and responsibility. Technology that performs exceptionally well most of the time can create expectations that become dangerous when an unexpected situation develops. Sebastian’s willingness to acknowledge the Tesla’s strengths while openly discussing its shortcomings provides a more meaningful assessment than either unqualified enthusiasm or blanket criticism. His experience illustrates how advanced driver assistance is progressing toward increasingly sophisticated capabilities without yet reaching the point of complete autonomy. For now, Tesla’s Full Self-Driving may handle much of the driving, but the human behind the wheel remains the final safeguard when something goes wrong.
