My kneejerk was to say, yeah 1->3 is a tripling. In modern journalistic terms that is 'skyrocketing'.
Then I read TFA. Yep 30->90. 'skyrocketing'.
That's not to say it isn't concerning. The downed wire incident that is the highlight of the story seems quite bad. But human drivers do much, much worse things, with very high frequency. Weekly sideshows, for example. The article is very unfair.
I'm on the streets of SF for hours a day, every day, mostly on bicycle. Often at night, when these AVs are at their peak output. The difference in their behavior is scary noticeable, especially within the last 30 days. I'm very curious about the programming reason behind it.
i have no basis for this guess, but i'd suppose it's more miles being driven, not a significant programming update.
<checking>
yep, TFA linkes to another article that there was a recent significant increase in miles driven.
i guess that doesn't necessarily jibe with your perception of a noticeable difference in singular behavior, but our brains do funny things. when you buy a certain new make of car, suddenly you wonder why so many of them are on the road. (they always were)
But in this case, everyone seems to agree that behavior changed, and data about incident reports matches it. It matches my experience too. It wasn't a subtle change. Cruise cars rapidly went from being timid and predictable to unpredictable.
Then I read TFA. Yep 30->90. 'skyrocketing'.
That's not to say it isn't concerning. The downed wire incident that is the highlight of the story seems quite bad. But human drivers do much, much worse things, with very high frequency. Weekly sideshows, for example. The article is very unfair.