3 ms·

Standard should be relative to human

• mjburgess 4y ago
  For reference, the human standard is <1 death in every 100 million miles driven.
  • uoaei 4y ago
    It doesn't seem like you're including classic and obvious externalities. What about causing traffic that prevents vital organs or rescue personnel from reaching those in need in time? Disruption of economic activity (qua output) can be significant in places where traffic is substantial and solvable yet unsolved. In my personal experience (I've spent significant time in the neighborhoods around Nvidia's headquarters) approximately 3 cars fewer get through any given stoplight if an autonomous vehicle is in that lane. I can imagine how in a place like SF this could cause real, measurable negative impacts.
    • mjburgess 4y ago
      I think my comment has been read as being "pro self-driving", whereas I intended it to be anti. Self-driving cars have already killed many more people than 1/100mill -- it's about 1 in 1 mil iirc. My comment was to frame the saftey of driving in the relevant units: amount of driving. When scaled by the right factor, "human driving" is incredibly safe by comparison.
      • krupan 4y ago
        You have to be including Tesla's not actually self driving to get to that statistic for self driving cars.