← Back to context

Comment by greesil

6 hours ago

I have seen a Waymo do a very stupid thing where it darted across a busy street, and it left very little margin of error for the oncoming traffic, which happened to be a loaded dump truck that could not have stopped. The dump truck driver was clearly surprised. It was a move that I never would have made as a driver. Did they dial the aggression up? I'm sure they're safer than humans in aggregate as there are some dumb humans out there but it's not infallible.

Waymo continues to improve every year, but dumb drivers never will.

  • It is probably possible to get drivers to improve if the incentives were there or if they had no choice due to external factors. I bet it would be cheaper than money spent on self driving tech too.

    Or public transit on a track.

    • Drivers can improve, but they won't. They will talk about the abstract just fine, but always in context of how "the other guy" is so bad, they resist any suggestion that they might not be good either. As soon as your point out something that nearly everyone is doing wrong (as backed up by statistics and traffic safety engineers who study this) and suddenly they will shut you down. As the other reply said: drivers vote and so any change that would affect all of them is impossible.

      I'd love to see better public transit, but transit is so bad for most of us that it would take a massive investment before there is any return, and half measures won't work. You have to go all in on transit before you can see any significant change - if you invest in the wrong network you won't know until a massive amount as been invested and there is no return (leaving open the question of if a different investment would have worked).

  • Weird, because per capita deaths leveled off in the 1930's and declined from that plateau in '70's to lows in the 2010's.

    Did we get less dumb drivers starting in the '70's?

    • Deaths and accidents are different measurements. Cars are much safer in an accident than they were in the early/middle 20th century.

      Per-mile-driven deaths started climbing again around 2012 in the U.S., I'd wager due to the trend towards larger vehicles causing more collateral damage.

That reminds me of the Feb 14, 2016 collision in Mountain View [1] (sorry for pdf, but it has the best images of articles I saw) between a Google self-driving car and a VTA articulated bus. TLDR, the software and the safety driver thought the bus would move out of the way because it was a big vehicle and a professional driver. From the report:

> Google said it has tweaked its software to "more deeply understand that buses and other large vehicles are less likely to yield to us than other types of vehicles."

Maybe that got lost.

[1] https://phys.org/news/2016-03-apnewsbreak-video-google-self-...