Talk for Article "Death in Uber’s self-driving car crash due to software bug"

Talk about this Article

  1. [ This comment is from a user you have muted ] (show)

    We know that there was a death caused by an autonomous car based on this article (and similar ones out there). But, I don’t have the tools to evaluate what that means. Is this saying that autonomous cars are significantly safer than human-driven cars based on number of accidents per mile? Far more dangerous?

    Though there isn’t enough data on fatalities for us to do good statistics on it, there is enough data on non-fatal accidents. And, statistics with an n of 1 would be more valuable than no data.

    Notes: wikipedia says that there were between .52 and 1.89 fatalities per 100 million miles traveled in 2015. Waymo claims on their website to have driven 5 million autonomous miles. The article below claims that uber drives about as many miles per week as Waymo. I don’t think that the total, once other companies are included, gets anywhere near the 100 million a human would take, but it’s also not off by more than one order of magnitude.

    Note 2: This article says really damning things about the safety of ubers autonomous driving program.
    https://www.recode.net/2017/3/16/14938116/uber-travis-kalanick-self-driving-internal-metrics-slow-progress

    1. [ This comment is from a user you have muted ] (show)

      Hi Zeke,

      Off the top of my head, I can only think of two fatalities of self-driving cars. Tesla in 2016 and the recent pedestrian death. So as you eluded, I don’t think there’s enough data to simply divide fatalities by miles driven. No clue where that Wikipedia figure comes from. Could you provide a link?

      The challenge is differentiating between road conditions. Experts I’ve spoken with say that taxi services like Uber will likely need a human driver for decades. The recode article you gave backs this up.

      Humans are far better equipped to deal with unpredictable environments in cities, but they also can space out… especially if they’ve been on the road for a long time.

      Long highway driving is great for robot cars because they don’t get bored or need sleep. So we would need to separate rural and urban driving in order to get reliable data.

      So even if it was found that you were safer jumping into a random autonomous car versus jumping into a random human-driven car, where these cars are driving matters. Does this answer your question?

      1. [ This comment is from a user you have muted ] (show)

        Here’s the link. https://en.wikipedia.org/wiki/Transportation_safety_in_the_United_States There were a lot of numbers in there and I didn’t find any one that seemed right for my purposes, but the .52 to 1.89 option seemed fairly good. To find it on the page, just search for .52.

        And for answering my question, I’d take my comments as more a request than a question. In doing the research needed for my comment, i got the information that I was looking for. But, the comment still seemed valid to me because I think that readers of this article, and comparable article on other sites, could benefit from enough background to fully evaluate the situation.

        I like this version of the article much better. I’d still like to see enough background on risks so readers could put the death into context, but the qualifying information that you included in your comment (and the article) definitely feels valuable to me.

        Thanks.
        z

  2. [ This comment is from a user you have muted ] (show)

    Any more details about how this happened? The cars cannot prevent all types of accidents, but I assume this is the top priority for all innovators of this technology. Tragic event, but I think it would be helpful for the community to understand what went wrong and if it would have been more or less likely with a human driver and anything that can be learned from this

    1. [ This comment is from a user you have muted ] (show)

      Agreed. When it comes to deaths, the details take a bit to come out. I’ll do some digging though.

      I’m interested in seeing whether the fact that it was night time played a role. Night is ideal for testing self-driving cars because there’s less traffic. But if the sensors have trouble detecting people in dark clothing, then that could be a problem.

Subscribe to our newsletter

Be the first to collaborate on our developing articles

WikiTribune Open menu Close Search Like Back Next Open menu Close menu Play video RSS Feed Share on Facebook Share on Twitter Share on Reddit Follow us on Instagram Follow us on Youtube Connect with us on Linkedin Connect with us on Discord Email us