• Wilmo@programming.dev
      link
      fedilink
      English
      arrow-up
      18
      ·
      edit-2
      11 hours ago

      A YouTuber already proved that Teslas autopilot would drive straight through a portrait of a road. Since it only uses cameras without other tech like lidar

        • 5too@lemmy.world
          link
          fedilink
          English
          arrow-up
          9
          ·
          7 hours ago

          Human drivers can pull in environmental and other cues though, like “why is there daylight on the road inside a building?”

          • CannonFodder@lemmy.world
            link
            fedilink
            English
            arrow-up
            1
            arrow-down
            2
            ·
            2 hours ago

            Ai can use that info too. And it doesn’t get distracted, or drink beer. Self driving cars have the potential to be much safer than humans. We don’t seem to be there yet, but we will be eventually if we don’t block progress.

            • Couldbealeotard@lemmy.world
              link
              fedilink
              English
              arrow-up
              1
              ·
              24 minutes ago

              When the CEO keeps saying they’ll have full auto pilot in the next few years for over a decade, then complains that the problem is the government stopping then from rolling it out, you start to question if it should ever be rolled out. At least maybe not by a for profit entity who puts profits before safety.

        • Wilmo@programming.dev
          link
          fedilink
          English
          arrow-up
          3
          ·
          6 hours ago

          Sure because humans don’t have radar and lidar. Something Tesla chose not to put on their vehicle.

          • CannonFodder@lemmy.world
            link
            fedilink
            English
            arrow-up
            1
            ·
            2 hours ago

            Yes it is a questionable decision. But I think it wasn’t just being cheap. I think it’s because of liability for errors; when it comes to placing fault on a car in the legal system, a judge / jury is only going to look at video captures. They won’t understand the lidar data. So if the car makes a bad decision based on lidar when there’s a conflict with visual queues, it will be deemed that it made a mistake. So there’s not really any point to trying to work with the lidar data.