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

California suspends Cruise's autonomous vehicle deployment

reuters.com

Post reply on HN
Print view View original

991–1000 of 1001 posts

Re: California suspends Cruise's autonomous vehicle deployment

#991

Post by sashank_1509 »

ra7 wrote: post #969
sashank_1509 wrote: post #830

As someone who has had FSD for the past 2 months, I would never trust it with left turn merges into fast lanes, any weird situation (I’ve seen it deal with construction fine but once an excavator was blocking the road and it kept trying to inch forward in the smallest of gaps even though it was clear that the only solution was to wait till it was off the road) and extreme turns (>150 degrees). My drive to work is 20…

> Finally, as someone working in AI, I’m actually very optimistic about their next approach of just learning an end to end model to drive the car. I think this is the only scaleable way to reach human level driving. Any set of rules based + optimization based planner will have those edge cases that ruin the cars capability, in fact it turns out the long tail is really long. Tesla doesn’t have an end-to-end model. The…

Their latest model takes video input and outputs car trajectory, and that will supposedly replace their entire current optimization based stack

Re: California suspends Cruise's autonomous vehicle deployment

#992

Post by ra7 »

sashank_1509 wrote: post #991
ra7 wrote: post #969

Earlier quoted context omitted.

> Finally, as someone working in AI, I’m actually very optimistic about their next approach of just learning an end to end model to drive the car. I think this is the only scaleable way to reach human level driving. Any set of rules based + optimization based planner will have those edge cases that ruin the cars capability, in fact it turns out the long tail is really long. Tesla doesn’t have an end-to-end model. The…

Their latest model takes video input and outputs car trajectory, and that will supposedly replace their entire current optimization based stack

We don't know if that's actually true. Tesla claims "end-to-end AI" in their next iteration, not an "end-to-end model". Wayve [1] is the only one who has an end-to-end model that takes video input and outputs a trajectory, but they are still doing one-off demos.

[1] https://wayve.ai

Re: California suspends Cruise's autonomous vehicle deployment

#994

Post by dchftcs »

Lonestar1440 wrote: post #120

From my perspective as a person who lives in San Francisco and also drives a LOT (10-20k miles per year, and many small drives within the city): Cruise cars do not perform acceptably. They manage to avoid collisions by driving extremely conservatively, but the way they traverse, say, a left turn against traffic is absurd. They slow everyone down, including emergency vehicles and public transit, by performing far belo…

This looks like the moment of truth for Waymo. Waymo had self-driving for years but rolled out slowly as they prioritized safety. Competitors seemingly had caught up pretty quickly, the space got crowded and for a long time people were thinking Waymo moved too slowly.

But now it seems their strategy is paying off, at least in terms of developing an acceptable solution. It still remains whether they can monetize it well, but other people will have to more or less adopt Waymo's model or they won't have a business.

Knowing about software, if the key differential factor is how well you can handle a long tail of edge cases, it'll take years before others can compete with today's Waymo in fully autonomous self-driving cars. If you don't have an architecture and culture to obsess about edge cases, it could even be easier to tear down the software and rebuild.

Re: California suspends Cruise's autonomous vehicle deployment

#995

Post by mayneack »

minwcnt5 wrote: post #804
mayneack wrote: post #773

I just started seeing Cruise cars driving around Santa Monica this week. Are they already paused in SF and we're somehow later or is pausing just a SF thing?

Are you sure those were Cruise and not Waymo?

Yes, although the ones I've seen today all have humans in the driver seat.

Re: California suspends Cruise's autonomous vehicle deployment

#996

Post by tim333 »

bumby wrote: post #949
tim333 wrote: post #703

Earlier quoted context omitted.

Though Waymo et al aren't really competing with unusually good drivers, more with average drivers and they are probably safer there, on the roads they have been trained for. (one study - 100% reduction in the frequency of bodily injury claims and a 76% decrease in property damage - https://www.ktvu.com/news/waymo-says-its-driverless-cars-saf... )

I think this ignores the way society is governed. It's easy for an engineer-mindset to think the bar to be cleared is "better than the average driver". But these systems must operate in a society that is often employing something other than an engineering mindset. I doubt the public and their political representatives (absent regulatory capture) would turn over control to a mere average robot. Do you think, for examp…

I probably should have said better than the average taxi driver as those generally don't include teenagers, 90 year olds and so on. Re the politics say human drivers in an area cause 10 deaths a year vs robot causing 6. Do you think voters are going to say we should have more deaths because humans doing it is better? That said I think people generally expect automated systems to be safe and probably self driving cars will be when they de glitch them. And then self driving cars will compete against other self driving cars more than against humans.

Re: California suspends Cruise's autonomous vehicle deployment

#998

Post by mepiethree »

paxys wrote: post #310

I had an epiphany about self driving recently late at night on the streets of San Francisco. It was a somewhat deserted part of town, and the roads were empty. There was a lone Cruise car chugging along and constantly stopping at red lights. The signals were so badly optimized that the car would be the only one waiting at a completely empty intersection for minutes at a time, only to be stopped again at the next one,…

I agree with your general point but will note that Google has been working on smarter traffic lights (albeit with much less funding than Waymo)

https://sites.research.google/greenlight/

Re: California suspends Cruise's autonomous vehicle deployment

#999

Post by patrick451 »

ethbr1 wrote: post #539
pbarnes_1 wrote: post #521

Earlier quoted context omitted.

I've been in this situation, and it's something that seems to get tweaked every week. They definitely pushed the "aggressive" lever up a little over the last month. I think it's one of those cases where it's sort of obvious it can calculate trajectories and speeds much better than a human, so it's a safe manoeuvre in theory, but it "feels" bad as a passenger. Same thing for where it can see that it can squeeze throug…

One huge advantage sensor-equiped vehicles have is knowing velocity and acceleration of traffic around them. Human drivers are often conservative because they don't know the speed/acceleration of a potentially colliding vehicle. We look at something and estimate (How fast? Speeding up or slowing down?), but we don't know . If you could put an actual number on it, I think you could drive a lot more "aggressively" and…

When a Tesla can accelerate to 60 mph in less than 2 seconds, knowing the instantaneous velocity and acceleration is not very meaningful. You really need to be able to predict what the acceleration profile will look like over the next N seconds of your maneuver. Holding the currently perceived velocity and acceleration constant over the next N seconds is one naive way to do that. But the actual set of possible trajectories of the other actor is much larger, and you need to drive more conservatively to account for that.

Re: California suspends Cruise's autonomous vehicle deployment

#1000

Post by icedistilled »

huytersd wrote: post #409
icedistilled wrote: post #407

Earlier quoted context omitted.

Well if you're objection is that public transportation currently sucks, yes that's the problem. It doesn't have to. That's the point. That might be the first time I've seen privacy as a main objection. Not sure how serious that should be taken considering how much privacy has willingly been given up in the last couple decades in the name of convenience. And also how few cases where it would matter. Like does it matte…

It’s not that it currently sucks. Even if it was the cleanest, safest, most connected system it still sucks compared to a personal vehicle (ideally electric and fully powered by renewable energy). It also means people have to live in dense apartments instead of spread out suburbs and semi rural living. I don’t mean privacy as in personal data, I mean not having to be in close quarters with other humans.

> "It also means people have to live in dense apartments instead of spread out suburbs and semi rural living." > "I mean not having to be in close quarters with other humans."

Suburbs tend to be sterile and devoid of character, but I do agree that they promote living in a bubble away from other humans.

Why that is desired or a widespread model for living needs more explanation. Sure some people like to be hermits, but as a social species it seems odd and also a highly inefficient way to accommodate a larger population. Also peculiar if one likes true "wilderness", because more suburban sprawl means less of that.

And suburbs have got to be one of the key factors in the reported loneliness epidemic. With the reports of increasing prevalence of loneliness, seems odd to want the default way of life to be one that we both agree means living spread out and not close to one another.

And yeah some people do like suburbs. On the other hand, Irvine ca despite being such an upscale and model suburb, is like living in the set of the truman show: artificial feeling and boring. People often describe it as sterile or like living in an office park. The landscaping of suburbs, such as irvine, also tend to be environmental disasters on top of being utterly bland. It's lose lose.

Not to mention, car transport being the paramount design concern makes for truly miserable pedestrian life and running trails. Running next to 6 lanes roads with 55 mph speed limits is just a terrible experience - irvine is crisscrossed with them.

Not sure how one can have a better road system than irvine - but it doesn't end up saving the residents time - things are just more spread out and they have to driver further between them and then fight for parking.

We've tried a car oriented transportation plan for decades - it gets one LA traffic. Parts of the 405 have 12 lanes. If you build it, congestion comes. We've tried it for decades, it's horrible. What is it they say about trying the same thing over and over and expecting another outcome? Oops actually the i5 - 405 merge near irvine has 26!!! lanes amongst all the ramps and truck bypasses. Needs more, right?

Post reply on HN