FSD Performance & Advancements - Tesla's FSD is rapidly improving, potentially surpassing human driving capabilities in many scenarios [1][2][3] - Emergency response feature now pulls the car over and activates hazard lights, improving safety for drivers with medical conditions and enabling potential robo-taxi functionality [5][6] - Driver monitoring has been relaxed, allowing for more comfortable use of the center screen and phone, indicating increased confidence in FSD [6][7] - Camera visibility issues related to sun glare and take-over requests have been significantly reduced [8][9] FSD Regressions & Issues in Version 142% - Parking performance has regressed, with the system struggling to park correctly and exhibiting illogical parking choices [16][17][18][19] - Navigation accuracy has declined, with the car randomly abandoning its route, potentially due to the integration of vision-based neural networks for real-time handling of blocked roads [21][22][23][24] - Increased passiveness in certain situations, such as yielding to cyclists and pedestrians even when not necessary, leading to potential safety risks and driver frustration [25][26][27][28] Speed Control & Mapping Challenges - Speed control remains a significant issue, with the car often driving too fast, especially when speed limits suddenly drop [30][31][32] - Tesla's solution of capping speed at 5 mph over the limit doesn't address the core problem, as it's too fast in low-speed zones and too slow when speed limits are broken [34][35][36][37] - Inaccurate maps and navigation continue to be a problem, with the car making illegal turns or taking inefficient routes, despite repeated reporting [41][42][43][44][45] Logic & Camera Issues - FSD exhibits a lack of logic in certain situations, such as choosing inappropriate parking spots or failing to follow simple directional signs [46][47][48][49] - Camera visibility issues, specifically off-gassing, affect FSD performance by causing sun glare and take-over requests [51][52][53]

Robo Taxi Experience & Performance - Robo Taxi demonstrates impressive on-the-fly pickup and drop-off adjustments, considering traffic flow and finding safe spots [5][6][8] - The system exhibits an understanding of human driver intentions, reacting smoothly to unexpected situations [19] - Robo Taxi's performance in heavy rain is notable, showing confidence and smoothness even in challenging conditions [55] - The system can reroute effectively when necessary, finding alternative paths without hesitation [51][52] Technology & System Analysis - The analysis suggests Robo Taxi might be an improved version of full self-driving version 13, potentially not utilizing the front bumper camera [38] - Hardware 4 is considered sufficient for unsupervised and driverless operation, with the absence of a front bumper camera not being a critical issue [43][45] - The system exhibits some randomness in decision-making, likely a byproduct of neural networks [67] - Safety monitors are present due to the system being a supervised one and to address potential issues [26][27] Safety & Human Interaction - The most significant danger comes from other human drivers who are inattentive or make mistakes [84] - Robo Taxi handles mistakes well and recovers effectively [64] - The system sometimes struggles to anticipate and avoid bad drivers [85][86]

Incident Overview - The incident is classified as a Full Self-Driving (FSD) crash because it occurred within 5 seconds of an autopilot or FSD disengagement, which will be included in future safety reports [7] - Public perception has already framed the incident as an FSD failure, regardless of what the data reveals [5] Data Analysis - Vehicle logs indicate leftward torque was applied to the steering wheel before FSD disengaged [15] - The steering wheel position remained straight initially, suggesting FSD was not responsible for the initial leftward torque [16][17] - FSD disengaged at the same time the steering wheel started to turn left [17] Possible Explanations - The driver may have unintentionally applied the leftward torque, possibly due to inexperience with the system or accidental contact [22][23] - The incident could be similar to cases of unintentional acceleration, where drivers mistake the accelerator for the brake [25] - Human error and perception can be unreliable, with drivers sometimes creating false narratives after accidents [24]

FSD Release & Technology - Tesla AI team is hinting at a June unsupervised FSD release, with a good track record around releases [2][3] - Tesla has been testing with 300 test operators in Austin to collect necessary safety data [4] - Robo taxi features are present but not yet available in the latest Tesla iOS app, including pickup, drop off, opening and closing doors, emergency stops [5][6] - The refreshed Model Y might be the initial vehicle for unsupervised FSD due to its additional front bumper camera [7][8] - FSD may attempt to drive anywhere within its geoence, unlike competitors that avoid certain roads [25] Market Perception & Challenges - Even with a successful launch, a narrative shift might not happen overnight due to psychic numbing and media coverage [13][14] - Perfection is impossible as long as driverless cars share roads with humans; a significant reduction in incidents is the realistic goal [17] - Tesla may not receive the same media leeway as other self-driving companies like Waymo [20] - Legacy news media is likely to paint a terrible picture of full self-driving, even if data shows it is saving lives [29] Future Outlook - AI drivers are improving while human drivers are becoming worse, making it a matter of time until AI's superiority is clear [30] - AI is expected to change the world in unimaginable ways [31]

Video Analysis & Methodology Concerns - The video features a Tesla but doesn't consistently use its autopilot or full self-driving (FSD) features, raising questions about the validity of the tests [1] - There are instances where autopilot appears to be disengaged, such as driving over double yellow lines or during the finale shot [1] - The video's methodology is questioned, particularly the decision to test autopilot instead of FSD, as they are different technologies with varying capabilities [1] - The presenter admits the video is about LiDAR and its uses, which contrasts with the framing of a scientific comparison between technologies [2] Luminar Relationship & Potential Bias - The presenter claims no financial relationship with Luminar, but acknowledges a friendship with Luminar's CEO, who donated $4 million to the presenter's project [1][2] - Luminar promoted the video on their homepage, raising questions about potential bias [2] - The video highlights positive results for the Luminar car, while downplaying potential issues with LiDAR technology, such as detecting water spray as a solid barrier [2] Autopilot Disengagement & Steering Wheel Force - The presenter suggests ultrasonic sensors caused autopilot to disengage before impact, but an alternate theory suggests force on the steering wheel led to disengagement [2] - Evidence is presented suggesting the presenter was gripping the steering wheel tightly, potentially causing the disengagement [2][3] Multiple Takes & Conflicting Explanations - The presenter admits to multiple takes of the grand finale scene, with inconsistencies in explanations regarding the car's ability to detect the wall [3] - The presenter initially suggests ultrasonic sensors detected the wall, but later states the car couldn't detect the wall due to its algorithms [3] External Influences & Irrational Behavior - The analysis suggests that negative sentiment towards a certain individual (Elon Musk) may have influenced the video's approach and conclusions [4] - Examples are given of other individuals who seem to exhibit bias against Tesla's technology due to their views on its CEO [4]

these latest versions of FSD coming out recently don't seem like anything major at least on paper the release notes are exactly the same as before and there's no mention of any new improvements or features but there is something about these newest versions that feels very different to me so I'm going to do my best to explain why that is and also kind of where I think we're at in the grand scheme of full self-driving so first off the car seems to be parking when it gets to destinations more than ever and I'm ...

man do we have a lot to talk about on the latest earnings call Tesla not only announced that they'll be launching unsupervised full self-driving in Austin Texas June of this year which by the way is ahead of schedule that they laid out at the Wii robot event but have also already started rolling out a version of it in their California Factory that drives the cars from the production lines to the outbound Logistics Lots completely autonomously and this marks a very historical moment for FSD as this is the fi ...