1 Matching Annotations
  1. Last 7 days
    1. Tesla Cybertruck na Polskiej Wsi: Czy FSD nas zabije?
      • Adaptability to Local Conditions: The Cybertruck's Full Self-Driving (FSD) system performed surprisingly well on Polish rural roads, even those covered in snow without visible lane markings [00:01:40]. It uses its cameras to identify the edges of the road and navigate obstacles effectively [00:02:13].
      • Vision-Only Technology: The vehicle relies entirely on eight cameras and neural networks trained on data from other Tesla drivers [00:13:47]. It does not use radar or LiDAR sensors [00:14:22].
      • Urban Driving & Safety: In Warsaw, the car successfully managed roundabouts, traffic lights, and stop signs. It proactively yields to pedestrians and cyclists [00:04:59], [00:11:02]. The car also features a 360-degree awareness that reacts faster than a human driver in emergency situations [00:12:02].
      • Driver Monitoring: The system requires the driver to remain attentive. An internal camera monitors the driver's gaze; if the driver looks at a phone or away from the road, the car issues an immediate warning [00:06:02].
      • Handling Tight Spaces: The truck navigated very narrow underground parking lots without hitting curbs or pillars, even coming to an automatic stop when a collision was imminent [00:09:14]. It also features an autonomous parking function that handles tight spots flawlessly [00:13:20].
      • Current Limitations:
        • Lack of European Maps: Since the Cybertruck is not officially sold in Europe, it lacks native maps, meaning users cannot simply enter a destination for the car to navigate to automatically [00:01:09].
        • Necessary Interventions: The driver occasionally had to take over, particularly during complex left turns or when the car misidentified which exit to take at a roundabout [00:05:51], [00:12:08].
      • Performance in Low Visibility: The FSD system remained functional and reliable during nighttime driving and in heavy fog, often "seeing" the road better than the human eye due to its trained neural networks [00:15:37].