Stress Testing Tesla FSD V12 is Getting Difficult...

30 Apr 202427:34

TLDRThe video provides a detailed stress test of Tesla's Full Self-Driving (FSD) version 12 in the challenging narrow streets of the Berkeley Hills. The narrator acknowledges FSD's progress from being uncomfortable to use to now managing most situations with occasional odd behavior. Despite some disengagements and cautious moments around pedestrians, the system demonstrates improved autopilot visualizations and better handling of speed bumps and stop signs. The video also highlights FSD's struggle with road closures and construction zones, where it fails to recognize a road closure sign, leading to a disengagement. However, the overall drive is described as the most comfortable and least stressful in the area, showcasing the system's capability to adapt and learn from its surroundings, though it still requires some human intervention.


  • ๐Ÿš— The Tesla FSD V12 has made significant progress in navigating challenging areas like the narrow streets of the Berkeley Hills, though it's not perfect and still exhibits some odd behavior.
  • ๐Ÿ“ˆ FSD has evolved from a system that was not comfortable to use to one that can manage most situations, with occasional disengagements and stress-inducing moments.
  • ๐Ÿ“Š The autopilot visualizations have improved, rendering cars more steadily and accurately, a notable enhancement from previous versions.
  • ๐Ÿšฆ FSD V12 removed the visualizations for cones and replaced them with more accurate occupancy network visualizations, which could้ข„็คบ็€ (indicate) future changes in how FSD represents its surroundings.
  • ๐Ÿ›‘ The system showed good defensive driving behaviors, such as yielding to pedestrians and being cautious around cars with brake lights on, indicative of its ability to anticipate potential hazards.
  • ๐Ÿšง FSD encountered difficulties with road closures and construction zones, highlighting areas that need further improvement.
  • ๐Ÿ’ก The car's handling of speed limits was inconsistent, particularly when it incorrectly adhered to a 5 mph limit in a school parking lot, causing operational challenges.
  • ๐Ÿ›ฃ๏ธ In narrow and complex road scenarios, FSD demonstrated smooth and confident driving, closely mimicking human driving patterns.
  • ๐Ÿšฆ FSD occasionally failed to use turn signals, especially when it was the only vehicle on the road, which could be a quirk stemming from its training on human driving data.
  • ๐Ÿ…ฟ๏ธ The system struggled with parking, particularly avoiding red zones and finding suitable parking spots, indicating that further refinements are needed for these tasks.
  • โฑ๏ธ The drive was one of the most comfortable and least stressful in challenging areas, suggesting that FSD is approaching a level of competence where such areas are less of a challenge.

Q & A

  • What is the main focus of the video?

    -The main focus of the video is to test and showcase the capabilities of Tesla's Full Self-Driving (FSD) version 12 in navigating the challenging and narrow streets of the Berkeley Hills.

  • How has the FSD system evolved over the years according to the video?

    -The FSD system has evolved from a system that the presenter wouldn't feel comfortable using in the Berkeley Hills to one that can manage the majority of the driving with occasional sketchy behavior, and now to a system that is much more comfortable and competent, with undeniable progress over just a few months.

  • What does the presenter think about the news media's portrayal of full self-driving technology?

    -The presenter believes that the news media often paints full self-driving technology as dangerous, focusing on clicks rather than facts and sometimes spreading misinformation.

  • How does Ground News help in navigating media bias?

    -Ground News provides unbiased news coverage, presenting diverse perspectives on global stories. It offers insights into the credibility of news sources, their political biases, and funding, allowing users to see differences in reporting and potentially missed stories.

  • What is the presenter's opinion on the autopilot visualizations in the latest FSD version?

    -The presenter notes a significant improvement in the autopilot visualizations, with cars being rendered more steadily as opposed to previous versions where they were jumpy and unsure of their location.

  • What challenges did the FSD face during the drive?

    -The FSD faced challenges such as navigating narrow roads, handling road closures, and interpreting faded or unclear road signs. It also had issues with speed limit adjustments and occasionally did not use turn signals when expected.

  • How did the FSD system handle speed bumps and stop signs during the test drive?

    -The FSD system handled speed bumps by slowing down significantly and managed to stop at a faded stop sign on the ground. However, it sometimes stopped too early or too late at certain stop signs, indicating there is still room for improvement.

  • What is the presenter's view on the future of autopilot visualizations?

    -The presenter speculates that future versions of autopilot visualizations may be replaced by more detailed and accurate representations, similar to those used in High Fidelity park assist, which could potentially eliminate the need for current car model visualizations.

  • How did the FSD system perform during unprotected left turns and right turns?

    -The FSD system performed well during unprotected left and right turns, showing caution and good decision-making, such as slowing down for oncoming traffic and yielding to pedestrians.

  • What was the presenter's experience with the FSD system's handling of road closures?

    -The FSD system had difficulty with road closures, causing two disengagements during the drive. The presenter hopes for improvements in future updates to better handle such situations.

  • What was the overall conclusion of the drive?

    -The overall conclusion was that the FSD system has improved significantly, making the drive much more comfortable and less stressful than previous versions. Despite a few interventions and disengagements, the system showed competence in handling complex driving scenarios.



๐Ÿš— Self-Driving Progress in Challenging Terrains

The video discusses the advancements in Full Self-Driving (FSD) capabilities, particularly in navigating the narrow streets of the Berkeley Hills. The narrator shares their personal experience with the system's evolution from being uncomfortable to use to now being able to manage most driving situations, albeit with occasional odd behavior. The video also addresses misconceptions spread by local news channels about self-driving technology being dangerous, and recommends using Ground News for unbiased news coverage. The drive starts at a supercharger near Berkeley, showcasing the car's ability to handle speed bumps, recognize stop signs, and yield to pedestrians and other vehicles.


๐Ÿ“Š High Fidelity Visualization and Defensive Driving

The paragraph focuses on the improvements in autopilot visualizations, which now render cars more steadily compared to previous versions. The narrator shares their anticipation that future FSD visualizations might be replaced by more accurate models similar to those used in High Fidelity park assist. The video also highlights the car's defensive driving behaviors, such as cautious navigation around pedestrians, anticipating other vehicles' movements, and smoothly handling narrow roads and turns. However, the narrator had to take manual control due to technical difficulties with a third-person camera, which is not counted as a disengagement.


๐Ÿ›ฃ๏ธ Navigating Narrow Roads and Unprotected Turns

This section of the video script details the car's ability to drive through narrow roads and perform unprotected left-hand turns. The car demonstrates smooth steering, minimal movement, and anticipation of oncoming traffic. It also shows caution around pedestrians at crosswalks and adapts its behavior based on the presence of other vehicles. The car's adherence to road signs, even faded ones, and handling of road obstacles like cones and military trucks are also covered. However, the car's occasional failure to use turn signals when it's the only vehicle on the road is noted.


๐Ÿšฆ Challenges with Speed Limits and Road Closures

The narrator discusses the car's challenges with speed limits, particularly when the system incorrectly adheres to a 5 mph speed limit in a school parking lot, causing the car to move extremely slowly on the road. The video also covers the car's difficulty with road closures and construction zones, leading to disengagements. The car's struggle with rerouting and its tendency to choose less optimal parking spots, such as in red zones, are also mentioned. The narrator expresses hope for improvements in these areas in future software updates.


๐Ÿšง Construction Zone Navigation and Roundabout Behavior

This part of the script addresses the car's interaction with construction zones and road closures. The car's ability to navigate through a roundabout is praised, but its difficulty in consistently handling construction zones is noted. The narrator observes that the car sometimes struggles to find a suitable parking spot and ends up in a red zone. The video also captures an instance where the car does not yield to an oncoming vehicle, which pulls over to let the car pass. The narrator expresses a desire for the car to use its horn in certain situations, like when another car stops and reverses suddenly in the road.


๐ŸŒŸ Smooth Driving and Anticipated Improvements

The final paragraph summarizes the drive, which lasted an hour and 20 minutes, and includes two interventions due to the car's adherence to a 5 mph speed limit and two disengagements related to road closures and dead-end routes. The narrator reflects on the comfort and competence of the FSD system, noting that it has improved significantly over the years. They mention that while the system is not yet perfect, it is getting closer to being able to handle challenging areas like the Berkeley Hills with minimal human intervention. The video concludes with thanks to the viewers and an anticipation for future improvements.



๐Ÿ’กFull Self-Driving (FSD)

Full Self-Driving (FSD) refers to a system in a vehicle that enables it to navigate roads without human intervention. In the video, FSD is tested in challenging environments like the narrow streets of the Berkeley Hills. The system has evolved from being uncomfortable to use to managing most of the driving tasks, showcasing undeniable progress over the years.


Disengagements are instances where the autonomous driving system disengages or stops functioning autonomously, requiring human intervention. In the context of the video, there are a couple of disengagements due to technical difficulties and road closures, highlighting areas where the FSD system still requires improvement.

๐Ÿ’กAutopilot Visualizations

Autopilot Visualizations are the graphical representations displayed by the vehicle's system to show the car's surroundings, including other vehicles, pedestrians, and obstacles. The video mentions an improvement in these visualizations, with the system now rendering them more steadily and accurately, which is crucial for safe navigation.

๐Ÿ’กOccupancy Network

The Occupancy Network is a feature in the FSD system that visualizes objects in the car's environment. It has replaced the older cone visualizations and is more accurate in representing the car's surroundings. The script suggests that this network may become the primary visualization method in future versions of FSD.

๐Ÿ’กDefensive Driving

Defensive driving is a style of driving where the driver anticipates potential hazards and takes actions to avoid them. The video highlights the FSD's ability to perform defensive driving maneuvers, such as slowing down for cars with brake lights on and yielding to pedestrians, which contributes to a safer driving experience.

๐Ÿ’กGround News

Ground News is mentioned as a tool for unbiased news coverage. It provides a comprehensive source for news, presenting diverse perspectives on global stories. In the video, it is used as an example of a reliable source to contrast with local news channels that may spread misinformation about technologies like FSD.

๐Ÿ’กHigh Fidelity Park Assist

High Fidelity Park Assist is a feature that provides detailed visualizations for parking assistance. The video suggests that this feature might eventually replace the current autopilot visualizations due to its detailed and accurate rendering of the environment, which includes plants and other objects.

๐Ÿ’กSpeed Limit

Speed Limit refers to the maximum speed at which a vehicle is permitted to travel on a road. The video discusses an issue where the FSD system incorrectly adheres to a 5 mph speed limit sign, causing the vehicle to travel at an uncomfortably slow pace, which highlights the need for better map data and speed limit recognition.

๐Ÿ’กRed Zone

Red Zone typically refers to an area where parking or stopping is prohibited, often marked by red curbs or signs. In the video, the FSD system pulls over into a red zone, which is not the ideal parking spot, indicating that the system still needs to learn to avoid such restricted areas.

๐Ÿ’กRoad Closure

Road Closure is a situation where a road is blocked off, often due to construction or an accident. The video shows that the FSD system struggles with identifying and responding to road closures, which is a critical area for improvement to ensure the system's reliability and safety.


Interventions are actions taken by a human driver to override or correct the autonomous system's actions. In the video, interventions were necessary when the FSD system was stuck on a low speed limit and when it attempted to enter a road closure, indicating moments where the system's performance was not up to par.


The Tesla FSD V12 is tested on the challenging narrow streets of the Berkeley Hills.

FSD has evolved from uncomfortable usage to managing most situations with occasional stress.

The autopilot visualizations have significantly improved, rendering cars more steadily.

FSD's handling of speed bumps and stop signs has become smoother and more human-like.

The car's cautious behavior around pedestrians and its defensive driving skills are noteworthy.

FSD's ability to navigate through narrow and complex road scenarios has greatly improved.

The system still has some issues with turn signals and occasionally does not use them.

FSD's interaction with traffic, including yielding and right-of-way negotiations, is impressive.

The car's adherence to a 5 mph speed limit sign, despite the faded appearance, shows its keen perception.

FSD struggles with road closures and construction zones, indicating areas for future improvement.

The system's handling of roundabouts and merging traffic is smooth and well-executed.

FSD's difficulty in finding suitable parking spots and avoiding red zones is highlighted.

The car's performance in navigating steep and narrow roads with limited visibility is commendable.

FSD's occasional over-caution at stop signs and its smooth acceleration in certain scenarios are observed.

The system's struggle with road closures is evident, requiring manual intervention in some cases.

FSD's overall progress in challenging areas like the Berkeley Hills is significant, though not flawless.

The test drive concludes with a summary of FSD's capabilities and the areas needing refinement.