Raw LLM Responses

As a person who actually uses FSD Supervised all the time, it is a much better driver than the humans around me. It doesn't drink and drive, text and drive, or engage in road rage. Waymo cars cannot drive on the highway because LIDAR has problems at highway speeds, and multiple sensors can easily create both false positives and false negatives. LIDAR is not as expensive as this professor states. The accident in the tunnel with the fatham braking is the fault of the human driver who rear-ended the Tesla as it was going too fast for the visibility on that curve. Jim Fan, the NVIDIA Director of Robotics, described his experience with Tesla's Full Self-Driving (Supervised) system version 14 as "magical" and an AI that passes the "Physical Turing Test." Passing the "Physical Turing Test": Fan coined the term for a test where a human observer cannot distinguish if a neural network or a human is performing a physical task. He stated it was "perhaps the first time" he experienced an AI that met this benchmark. Indistinguishable from a Human Driver: After a long day, he noted one couldn't tell "if a neural net or a human drove you home". From Surreal to Routine: Fan described the experience as initially "surreal," but it "quickly became a routine part of his daily life," a sign of how humanity gets "rewired and glued to god-like technologies". Indispensability: He compared the technology's integration to that of a smartphone, saying, "Removing it now would actively hurt". Fan, who co-leads NVIDIA's GEAR (Generalist Embodied Agent Research) lab, is a leading researcher in embodied AI and foundation models for robotics. His positive assessment of the FSD v14.2.2 system was a notable endorsement from a high-profile industry expert outside of Tesla.