Cruise will reduce robotaxi fleet by 50% in San Francisco
cnbc.com
cnbc.com
Not one accident I saw was cruises fault.
I might not have the full information.
Frankly, this is a good move by the DMV. Meaningful oversight with consequences for messing up the political balance is the best tool we have to ensure industry players properly self-regulate.
Come on man, it's the title of that article ...
>Driverless Cruise car collides with SF fire truck, injuring passenger
Hmm ...
"VIDEO: Driverless Cruise car struck by SF firetruck, injuring passenger, company says"
Like an automobile accident between humans, there's a flurry of fingerpointing and attempts by both parties to shift blame. Like an automobile accident between humans, this is an evolving situation and determining fault is an ongoing process of investigation. That's why it's prudent not to rush to judgment and not take at face value the assertions of one of the involved parties--in this case the fire department--who has an incentive to try to get out in front of the story and shape the narrative to protect its own image (just as Cruise has that same incentive).
I had a good laugh, thanks for that.
Interestingly, the title on HN seems to match the title of the page, they probably changed it here as well amirite? :^)
Also, from that same thread seems like you missed this one as well,
https://www.businessinsider.com/driverless-cruise-car-stuck-...
Now you are well informed :)
https://www.freep.com/story/news/local/michigan/wayne/2021/0...
As for the logic behind half rather than all:
1. Revoking a permit is a drastic action that could cause cruise to leave the DMVs regulatory area entirely and go to another state like Arizona with essentially no oversight.
2. It could open them up to legal action that would consume lots of money for essentially no purpose
3. It gets the incident rate down so the situation is less political while the investigation is ongoing.
Of course, this is absurd. Instead, we slap on some caveats (time/passenger restrictions) and accept the risk with the knowledge the driver will improve with time.
This is exactly the same for self driving vehicles. The only way to make them safer for public roads is to drive them on public roads. So if we slap on some caveats, such as a reduced fleet, we can make them safer while keeping incidents down.
Although, I didn't mean that a zero incidence rate should be the rule, more like, if one knows a system is defective, also in such a dangerous context, it'd be negligent to still ship it under the pretense of a "low incidence rate".
It's easy to imagine that a 50% reduction in fleet size could make this 90% less likely, while they work on ways to prevent that from happening again.
It's not just about the behavior of any individual robotaxi, but it's also about their behavior collectively.
It's also not just about safety, but about being a public nuisance -- if they break down or get confused too often. We can tolerate a small overall amount of that in a city, but not a large overall amount. It's not binary.