Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

S.F. says incidents by Cruise, Waymo driverless taxis are ‘skyrocketing.’

sfchronicle.com

Post reply on HN
Print view View original

131–140 of 1001 posts

Re: S.F. says incidents by Cruise, Waymo driverless taxis are ‘skyrocketing.’

#131

Post by dekhn »

saalweachter wrote: post #82
cameldrv wrote: post #49

I've just got to say I'm so impressed with both Waymo and Cruise. Last year they did 4.5 million miles put together in California with zero fatalities. This is an incredible achievement.

The US rate is 0.57 fatalities per 100 million passenger miles.

Americans drove 3.2 trillion miles in 2021 and 42,795 people died in traffic fatalities (both values are estimates, and likely do not include every single mile driven or fatality that occurred, depending on how you prefer to count).

So humans drive average of 74 million miles before a person gets killed, and Waymo's driven somewhere above 20M miles but probably fewer than 74M. At this point, I'd expect no self-driving car traffic fatalities, statistically, if the SDC fleet is an average driver.

In some sense Waymo and Cruise are basically waiting around for a fatality (not that they want one) so they start having a denominator that isn't zero.

Re: S.F. says incidents by Cruise, Waymo driverless taxis are ‘skyrocketing.’

#133

Post by szundi »

Eisenstein wrote: post #94
standardUser wrote: post #22

Earlier quoted context omitted.

I know this is a tired argument, but it's not wrong: There is equal or greater cause for concern with cars driven by people, we're just used to those horrors and reluctant to trade them for a different set of horrors, even if they are conclusively less horrible. I lived carless in downtown SF for 10 years, usually with a walking commute, and it was not pleasant. Packed streets full of angry, stressed drivers honking…

Depending on driverless taxis instead of having emotional people who need parking all the time seems to be a win, as long as they are safe and someone is held accountable when they mess up . I am very unwilling to change 'people are flawed but (mostly) held accountable' to 'a corporation owned robot car ran someone over because of a software glitch and it is nobody's fault'. The problems with 'oops we got hacked or l…

Then no one is going to implement it probably. Or is it just jail time when dead are more than with humans?

Re: S.F. says incidents by Cruise, Waymo driverless taxis are ‘skyrocketing.’

#134

Post by jstummbillig »

akira2501 wrote: post #123
parl_match wrote: post #3

I'm wary about using public roads to test these, but I think the way the data is presented is misleading. I'm not sure how it's misleading, but separating "incidents" into categories (safety, traffic, accident, etc) might be a good start. For example, I could start coning cruise cars, and cause these numbers to skyrocket. While that's an inconvenience to other drivers, it's not a safety issue at all. By the way, as a…

> I'm not sure how it's misleading, but separating "incidents" into categories The premise is that these vehicles would be unilaterally "better" than human drivers and that would justify their creation and use. If that's not the case, then perhaps we need to fully reevaluate the value proposition here. > While that's an inconvenience to other drivers, it's not a safety issue at all. It hasn't proved to be a safety is…

> The problem with automation is small problems have a tendency to combine into massive emergent problems once you start scaling your fleet up.

Not in this case, no. Every driver has a small problem. Each ai driver is just another driver, replacing some driver.

Unless the ai driver problem is not small but actually quite big, I don't see how automation economics would worsen the situation. On the contrary, them all being identical and automated makes fixing the small problems much more viable.

Re: S.F. says incidents by Cruise, Waymo driverless taxis are ‘skyrocketing.’

#135

Post by __loam »

HWR_14 wrote: post #40
cwp wrote: post #27

San Francisco won't be satisfied until they've completely driven out all the tech companies and their employees.

Maybe. I'm not sure that's a bad goal for the average SF resident.

The average SF resident probably works for a tech company considering how many natives have been driven out by high housing costs. Tech companies are an enormous source of tax revenue for the city and many public services quite literally depend on that revenue. One of the major issues in the city now is the city government's obstinate refusal to approve more housing development. The idea that tech workers are gentrifying one of the most expensive cities in the world is frankly absurd.

Re: S.F. says incidents by Cruise, Waymo driverless taxis are ‘skyrocketing.’

#136

Post by babl-yc »

dekhn wrote: post #115

"Skyrocket" is a word people use when they don't have the data to support their claim, but want to make it sound as extreme as possible. When I hear "skyrocket" I think "several orders of magnitude in a very short time". But the source of the quote with skyrocket, basically admits that they didn't have any quantitative evidence to support a massive change in incidents. Remember: reality is banal. Things are far bette…

The SF Chronicle has an unfortunate habit of headlines telling you "what to feel" as opposed to "what happened".

Re: S.F. says incidents by Cruise, Waymo driverless taxis are ‘skyrocketing.’

#137

Post by akira2501 »

jstummbillig wrote: post #134
akira2501 wrote: post #123

Earlier quoted context omitted.

> I'm not sure how it's misleading, but separating "incidents" into categories The premise is that these vehicles would be unilaterally "better" than human drivers and that would justify their creation and use. If that's not the case, then perhaps we need to fully reevaluate the value proposition here. > While that's an inconvenience to other drivers, it's not a safety issue at all. It hasn't proved to be a safety is…

> The problem with automation is small problems have a tendency to combine into massive emergent problems once you start scaling your fleet up. Not in this case, no. Every driver has a small problem. Each ai driver is just another driver, replacing some driver. Unless the ai driver problem is not small but actually quite big, I don't see how automation economics would worsen the situation. On the contrary, them all b…

Except Humans are dynamic and the AI is pre-programmed. So Humans can _react_ to emergent conditions whereas AI is completely subsumed by it.

There's a good hint here that when Waymo cars "break down" a remote Human operator takes over. If your view is correct, this could never work.

Re: S.F. says incidents by Cruise, Waymo driverless taxis are ‘skyrocketing.’

#139

Post by LispSporks22 »

dekhn wrote: post #115

"Skyrocket" is a word people use when they don't have the data to support their claim, but want to make it sound as extreme as possible. When I hear "skyrocket" I think "several orders of magnitude in a very short time". But the source of the quote with skyrocket, basically admits that they didn't have any quantitative evidence to support a massive change in incidents. Remember: reality is banal. Things are far bette…

I'm with ya normally, but literally a few paragraphs in they have the chart source: the TA, going from 3 to 91 "incidents" over a year. If you turn it on its side, it kinda looks like rocket blasting off.

Re: S.F. says incidents by Cruise, Waymo driverless taxis are ‘skyrocketing.’

#140

Post by lanstin »

valine wrote: post #68
reso wrote: post #63

Earlier quoted context omitted.

You need AGI for good self-driving because driving requires predicting the actions of other human drivers at an extremely high level. This is second-nature for humans so we barely notice that we are doing it, but it is extraordinarily difficult for non-humans.

No you don’t. I can feed images of crazy driving scenarios into LLaVA and get reasonable responses. That’s a general purpose LLM with $500 worth of fine tuning running locally on my PC. You should look into what can be done with the current state of the art LLMs. Your intuition for what’s possible is out of date. If I can do that with open source LLaMA variants, I can only imagine what’s possible if you have an actua…

You getting six nines of accuracy on that with good latency? Did you watch the “how our large driving model deals with stop signs” from Tesla AI department? Given the multiplicative effect of driving decisions and the weird real world out there, it has be extremely reliable and robust to be a good driver as the miles mount up.
Post reply on HN