California Suspends GM Cruise Self-Driving Vehicles As 'Not Safe' For Public (reuters.com) 115
An anonymous reader quotes a report from Reuters: California's auto regulator said on Tuesday it has suspended General Motors' Cruise autonomous vehicle deployment and driverless testing permits, saying it had misrepresented information related to the safety of the autonomous technology. Cruise's AVs posed an "an unreasonable risk to public safety," and "are not safe for the public's operation" California's Department of Motor Vehicles (DMV) said in a statement. The suspension, which came after a series of accidents involving Cruise vehicles, is a major setback to GM's self-driving technology unit and to the nascent autonomous vehicle (AV) industry. The suspension was effective immediately.
GM has previously insisted Cruise represents a "giant growth opportunity." In June, GM CEO Mary Barra reiterated a forecast that Cruise could generate $50 billion a year in annual revenue by 2030, a target the company has not backed away from. The automaker reported on Tuesday that it lost $723 million on Cruise during the third quarter. The DMV added that Tuesday's "decision does not impact the company's permit for testing with a safety driver." Cruise issued a statement saying it "will be pausing operations of our driverless AVs in San Francisco." Cruise said the California Department of Motor Vehicles reviewed an incident on Oct. 2, when one of its self-driving vehicles was putting on the brakes but did not avoid striking a pedestrian who had previously been struck by a hit-and-run driver.
Cruise said that in that incident, "a human hit and run driver tragically struck and propelled the pedestrian into the path of the AV" which struck the pedestrian. "When the AV tried to pull over, it continued before coming to a final stop, pulling the pedestrian forward," the company said. "Our teams are currently doing an analysis to identify potential enhancements to the AV's response to this kind of extremely rare event," it added. In August, the regulator said it was investigating "concerning incidents" involving autonomous vehicles operated by Cruise in San Francisco and asked the company to take half its robotaxis off the roads. That month, a Cruise robotaxi was involved in a crash with an emergency vehicle in San Francisco. Barra said the Cruise robotaxis have better safety records than human drivers.
GM has previously insisted Cruise represents a "giant growth opportunity." In June, GM CEO Mary Barra reiterated a forecast that Cruise could generate $50 billion a year in annual revenue by 2030, a target the company has not backed away from. The automaker reported on Tuesday that it lost $723 million on Cruise during the third quarter. The DMV added that Tuesday's "decision does not impact the company's permit for testing with a safety driver." Cruise issued a statement saying it "will be pausing operations of our driverless AVs in San Francisco." Cruise said the California Department of Motor Vehicles reviewed an incident on Oct. 2, when one of its self-driving vehicles was putting on the brakes but did not avoid striking a pedestrian who had previously been struck by a hit-and-run driver.
Cruise said that in that incident, "a human hit and run driver tragically struck and propelled the pedestrian into the path of the AV" which struck the pedestrian. "When the AV tried to pull over, it continued before coming to a final stop, pulling the pedestrian forward," the company said. "Our teams are currently doing an analysis to identify potential enhancements to the AV's response to this kind of extremely rare event," it added. In August, the regulator said it was investigating "concerning incidents" involving autonomous vehicles operated by Cruise in San Francisco and asked the company to take half its robotaxis off the roads. That month, a Cruise robotaxi was involved in a crash with an emergency vehicle in San Francisco. Barra said the Cruise robotaxis have better safety records than human drivers.
Not to worry! (Score:5, Funny)
Each time they experience a situation, new elseif statments will be added to make sure it probably won't happen again.
Re:Not to worry! (Score:5, Insightful)
Re:Not to worry! (Score:4, Insightful)
I don't really understand the desire for people in tech to want to replace taxi drivers, almost obsessively so. When I've had a bad experience in a taxi, I don't recall it ever relating to the driver, but more to cleanliness which I'm not sure how removing the adult supervision from the vehicle helps that?
They don't. They want to replace ALL drivers. But you have to start with the lowest hanging fruit and grow from there as you perfect the tech.
Re: (Score:2)
No; they want to replace public transit.
Save our trains!
Re: (Score:2)
How about you, AC, suspending yourself for not being a nerd, or other than an ignorant idiot?
Replace the drivers (Score:3)
This. But, starting with "low hanging fruit", well,
Taxi driver: $40k/year.
Taxi: $30k. And you might get a couple years out of it.
The driver is the biggest cost driver for taxi services. The cost of the driver is proportionally less for more "mass transit" type options like busses and trains. Trending towards insignificant for the latter if decently full.
There's work to automate trucks - tractor trailers, as well, but there's a lot more complexity when you look at handling a trailer.
and will the robo taxi ask for an tip? (Score:2)
and will the robo taxi ask for an tip?
Re: (Score:2)
Will the tip go to the programmers!?!?!
Re: (Score:2)
Treating this seriously, I can't help but think that the most beneficial party to tip at that point would be the cleaners who ensured that you didn't get into a nasty dirty cab.
And maybe the last rider - for not turning it nasty, or failing to report a nasty taxi.
Re: (Score:2)
I don't really understand the desire for people in tech to want to replace taxi drivers, almost obsessively so. When I've had a bad experience in a taxi, I don't recall it ever relating to the driver, but more to cleanliness which I'm not sure how removing the adult supervision from the vehicle helps that?
They don't. They want to replace ALL drivers.
I don't understand this, a good drive is one of life's great little pleasures.
I highly suspect it's because cars these days are so devoid of life and feeling, also that too many people are remarkably uptight and have no joie da vivre.
However as soon as it's here I guarantee everyone who demanded it will absolutely hate it as the cars will be programmed to obey all traffic laws. No doing 10 over on the motorway, no cutting in, no dodgy overtakes (or undertakes) in fact there will be very little passing
Re: (Score:2)
I don't understand this, a good drive is one of life's great little pleasures.
You're in a minority there. Outside of people who are into driving the vast majority of the populace regard it as a chore and nothing more.
There were likely a lot of people who complained that you had no connection to your transport when horses were replaced by cars, and horses are still there as a hobby for those people, but the rest of the world moved on.
Eventually manual driving will likely be confined to closed tracks or offroad driving using specialized vehicles (since eventually your standard car won
Re: (Score:3, Insightful)
Because they're fearful loners in their dark little hearts.
Pazuzu forbid that they may have to make small talk with a STRANGER. Ewwwwww.
Same idea with them wanting to replace public transit. Just terrible to have to look at strangers or sit across from them.
Re: (Score:2)
That's also stupid: it's quite clear that self driving cars, no matter how good the driving gets, can replace a well designed and well run mass transit system. The roads simply don't have the capacity. To match the capacity, absurdly huge increases in space efficiency (something like 50x) would be needed for self driving cars to match mass transit.
Re: (Score:2)
Just terrible to have to look at strangers or sit across from them.
Especially if they are impaired and obnoxious, or coughing and sneezing. Yeah, I have no problem finding conversation elsewhere.
Re: (Score:2)
Londoners, or Parisiens, or Berliners (not the donut) don't seem to express those sorts of views.
Re: (Score:2)
Re: (Score:2)
You also exhibit kneejerk negativity when confronted with random people on the internet who can see through you.
People on the internet like you my AC friend? Hilarity is indeed a good cure for negativity. Thanks for that.
Re: (Score:2)
Londoners famously avoid eye contact with strangers in transport systems. See e.g. the anthropological study Watching the English by Kate Fox.
Re: (Score:2)
Re: (Score:3)
> When I've had a bad experience in a taxi, I don't recall it ever relating to the driver, but more to cleanliness
Automated car cleaners may have more bang for the buck. Protein detection would be helpful.
But I suspect taxi-bots are a loss leader for R&D into general self-driving cars, which many expect are the future. It just may be, as young people don't want to drive whether it be their own car or renting a taxi; they'd rather fiddle on their phones and tablets.
Re:Not to worry! (Score:4, Insightful)
Re:Not to worry! (Score:4, Insightful)
I'd rather take a train
So, let's say I'm living in suburbia and want a McLardburger. The nearest train station is actually further away than the McDonald's. Actually, I think I'd pass by no less than three McDonald's before I arrived at the train station. Public transit only really works if you're also planning on forcing everyone to move to the city, which is a can of worms in itself.
Now I certainly don't see someone getting a self driving taxi just to grab a crappy fast food burger when paid delivery service is an option, but having my own car means I can just drive to the restaurant and not have to pay anyone to bring the burger to me.
Re: (Score:2)
The problem you have is that your suburbia is designed wrong.
Look at Japan. Their zoning laws allow for a lot of mixed use. Obviously not heavy industry right next to where you live, but small factories, offices, shops, and restaurants can be mixed in with housing. A lot of people live within a few minutes walk of amenities.
Combined with good public transport links, you really don't need a car to live in many parts of Japan, even the more suburban ones.
Maybe you could have some of that if you adjusted your
Re: (Score:2)
Re: Not to worry! (Score:2)
That's depressing.
But city-dwellers are prone to depression, and high density housing is to blame, as described below and studies out of Sweden.
https://www.theguardian.com/su... [theguardian.com]
You have valid points but suburbia is better for the mental health of its inhabitants
Re: (Score:2)
Why is everyone so damn asocial since COVID?
Because large portions of Gen Z and millennials claim to be on the autism spectrum and use it as an excuse not to talk to people when in reality they're whiny cunts who are always looking at their phones.
As for me, I work in IT. I talk to people all day. You know what I don't want to do when I leave work?
Re: (Score:2)
Why is everyone so damn asocial since COVID?
We live in the era of mass media screaming at us "WATCH OUR PROGRAM OR YOU'LL DIE!!!", and a lot of people these days have never known anything else. And it's 99.999999999999999999% bullshit, of course. Along comes COVID, which actually is dangerous, and a lot of people broke.
They've become so addicted to their fear, they literally cannot live without it any more.
Re: Not to worry! (Score:2)
Since COVID? Numbers are up right now.
Re: (Score:2)
Re:Not to worry! (Score:5, Insightful)
I don't really understand the desire for people in tech to want to replace taxi drivers, almost obsessively so.
The idea isn't so much to put taxies out of business as it is to be able to offer cars as a service. Putting taxies out of business is just a side effect of that goal. It's actually pretty common among urban-dwelling tech folks to view private vehicle ownership as inefficient and wasteful, since the majority of the time the vehicle just sits parked. That's why you also see tech companies littering cities with rental scooters and ebikes.
Of course, the flipside to that is it's also one step closer to a "you'll own nothing and be happy" future that many find distasteful, because it represents a step down in the standard of living that previous generations had become accustomed to. Having a car sitting on your driveway didn't just represent a measure of success, it also represented the freedom to go wherever you wanted, whenever you damn well pleased.
Re: (Score:2)
Re: (Score:3)
Clearly, the city you live in isn't in southern California. Nothing is close enough to walk to, riding a bike is an invitation to be run over by some maniac in the Road Rage Capital of the World, and mass transit? Bring a towel, to wipe the homeless guy's urine off the seat (literally) before you sit down.
Re: (Score:2)
It's actually pretty common among urban-dwelling tech folks to view private vehicle ownership as inefficient and wasteful
Especially the poor ones.
Re: (Score:2)
I don't really understand the desire for people in tech to want to replace taxi drivers, almost obsessively so.
The idea isn't so much to put taxies out of business as it is to be able to offer cars as a service.
I have to disagree, there is a huge sentiment, especially in the US, to put taxis out of business. That's why illegal taxi businesses... erm... I mean "ride sharing" businesses like Uber are so popular over there. Sure there are arguments that the taxi systems in various US areas are broken but you wont fix that by removing the human driver, it's not the driver causing the problems and the system (I.E. Medallions) will just adapt to the driverless car.
Also taxis are expensive, people think that driverles
Re: (Score:2)
Having a car sitting on your driveway didn't just represent a measure of success, it also represented the freedom to go wherever you wanted, whenever you damn well pleased.
Kind of. Assuming good alternatives, there's no reason I need to have a car in my driveway to have the freedom to go wherever I want whenever I damn well please. And the car in my driveway also comes with a lot of expense and non-trivial maintenance burden.
Note that I live in a rural area where there currently is no good alternative to a personally-owned vehicle. Uber/Lyft don't operate here, for example, and there is definitely no mass transit. But this is situational, not universal.
Re: (Score:2)
Re: (Score:2)
and self driving taxis will be loaded with puke and other shit with no driver.
Re: (Score:2)
Or loaded with homeless meth heads who refuse to - or can't - leave.
Be a nice way to delay finding a body, too.
Re: (Score:3)
I don't really understand the desire for people in tech to want to replace taxi drivers, almost obsessively so. When I've had a bad experience in a taxi, I don't recall it ever relating to the driver
Are you living on the same planet as the rest of us? About 50% of taxi rides (in any country) make me want to get out, kiss the earth and thank any deity which will listen for granting me another day surviving on this horrendously bad driver infested planet.
Literally today I took two taxi trips, one was doing 85 in a 55 zone weaving through traffic, and on the return the other decided to try and take up two lanes of the highway for as much of the journey possible, and then proceeded to cut across 3 lanes of
Re: (Score:2)
Re: (Score:2)
I certainly find it plausible in Puerto Vallarta, Mexico. 60+ miles an hour down a crowded, one land street, with less than 2 inches of room on either side.
Re: (Score:2)
When I've had a bad experience in a taxi, I don't recall it ever relating to the driver, but more to cleanliness
Who do you think has been slacking off on cleaning?
Re: Not to worry! (Score:2)
Taxi drivers in the USA are convicted of about 14 assaults a year in the US alone. That's not a very big number, but if the taxis are self driven it will be zero.
Also, IME most taxi drivers are assholes and/or idiots, based on how they behave in traffic. In fact I just called one in a week ago or so for cutting me off to pass a truck, then slowing down to 55 instead, endangering everyone involved. No idea what his problem was but now he has another one.
Re: (Score:2)
I don't really understand the desire for people in tech to want to replace taxi drivers, almost obsessively so.
No. People want to replace other drivers. People see how other people drive and think they are insane. For their own safety, they want those people off the road. Of course, they fail to look at it from their own perspective until it is too late. They won't be able to drive either... which is not really a problem except now they have to trust the AI with their own car and life. They will be horrified.
"We couldn't have known that at 12:01 localtime on 7 December 2038, all vehicles using a Ford championed AI w
Hey! What... (Score:2)
ever happened to all the carping about how the Big Automakers were going to ramp up and destroy Tesla?
Now we are hearing that they have failed to master self-driving and are already shutting down electric vehicle assembly lines...
Maybe this is all just too hard for them
Re: (Score:2)
Re: (Score:2)
That is a novel idea, but not something I would "remember", since it ignores that the US public is addicted to rolling about in individual vehicles
One thing about Capitalism that people tend to forget, is that it is Consumer Driven, and failing to meet the consumer's desire is one sure way to fail as a company
That is probably why you only see governments implementing mass transit solutions, since they are not bound to producing a profit
Re: (Score:2)
Re: (Score:2)
Only government implements mass transit because people aren't willing to knowingly pay what it costs. Period. If they were, private companies would be all over it in a heartbeat.
Re: (Score:2)
Re: (Score:2)
The four big private railroads are in a race to the bottom to see how much track they can remove and crews they can terminate and still make record profits. Capitalism promotes inefficiency.
I'd argue that it's actually like democracy - it sucks, but what do you have that's better?
Your railroad example suffers from the problem that the railroad executives got their heads into a very weird non-capitalistic space. They got obsessed about "operating ratio" instead of profit. Operating ratio is basically the inverse of your profit margin. IE they'd turn down any business that didn't earn them 4X as much as it cost them to actually move the cargo.
I've seen signs that they're starting to break out
Re: (Score:2)
hey're losing oodles of money not moving all the profitable to move cargo.
Also, the most successful expansions of passenger rail in recent history in the USA would be the sunrail, which is mostly private.
It's a public private partnership, which is even worse than if it was just privatized infrastructure, because instead of saving for the future when there's a surplus, they're going to burn it off as a dividend and call it profit. But, guess who didn't chip in to build it and who won't be there to maintain it? That's right, Bombardier.
The best private infrastructure success story you have is a railroad that exists primarily to haul money from Florida to a private Canadian company. It's not even good as, s
Re: (Score:2)
ever happened to all the carping about how the Big Automakers were going to ramp up and destroy Tesla?
I don't think anyone expects Tesla to get "destroyed" a la Netflix vs Blockbuster Video. It's far more likely they'll end up as a niche market automaker as companies like Ford, Toyota, Chevrolet, Honda, and Hyundai slowly figure out the market for EVs. You know, basically the same brands that you see all over the roads today.
Re: (Score:2)
You think Tesla is any better? Frankly I'm shocked that they're still allowed to market and sell "Full Self Driving", let along having it enabled on public roads with ordinary drivers.
Tesla Emergency Brake Test — Full Self Driving Beta 11.4.4 [youtube.com]
Watch a Tesla Run A Red Light on FSD Beta! [youtube.com]
Tesla Automatic Emergency Braking - Real World Test [youtube.com]
How to Bypass Tesla Driver Monitoring System [youtube.com]
Speaking as someone who frequently rides a bike to work, Teslas are terrifying.
Re: (Score:2)
ever happened to all the carping about how the Big Automakers were going to ramp up and destroy Tesla?
They realized all they needed to do was sit back with their popcorn and watch Elon Musk do it for them.
Wrong headline (Score:2, Troll)
Re: (Score:2)
Headline should be "California, as always, was dumb enough to approve something clearly idiotic run amuck".
It's written into the state constitution. The more is costs, the more graft and corruption can be embezzled out of it for their union backers. If they ever complete one of these massive public works, then they can't come back and try again in a few years, and where's the graft and corruption in that?
I do not exaggerate.
You don't say? (Score:3, Interesting)
But I thought all the know-it-alls have been absolutely 100% certain that self-driving cars are guaranteed safer than human drivers? Who know that throwing beta-self-driving out on the road could be dangerous?
I'm gonna come back to what I've been harping on since the "we must have self-driving cars to save lives" crowd started barking on the subject. You want a stamp of approval for driverless tech? We need a test-track somewhere that's got nothing but driverless cars on it, with a few stunt folks to take the brunt of difficult situations, some "unexpected" (by the cars) obstacles just randomly popping up here or there, and then compare that all to how well humans do in similar situations. Until these systems prove themselves in a highly watched/sensored/micromanaged environment for a while (more than an hour on the test track), they can't drive on public streets. Why we think the route we're currently taking is the correct one is utterly baffling. I'm all for new tech. I'm not all for new tech that hasn't really been vetted properly being unleashed as driverless technology for public use. Too many gotchyas out there in real driving conditions. Too few really difficult tests for these systems.
Re:You don't say? (Score:5, Insightful)
I thought you were going to go the other way with this. On aggregate, humans are absolutely terrible drivers. I've even had my car get hit while it was parked in my driveway. People suck at driving.
Re: (Score:2)
Re: (Score:3)
The issue is that the manufacturer/operator of the self driving vehicle is liable for all the mistakes it makes. If a haulage company driver kills someone through poor driving, the driver is the one that will get prosecuted and face the consequences. If a self driving truck kills someone, the company that made the self driving tech will be on the hook for it.
So self driving vehicles need to be a lot better than humans, because all the liability is concentrated in one place.
Re:You don't say? (Score:4, Insightful)
But I thought all the know-it-alls have been absolutely 100% certain that self-driving cars are guaranteed safer than human drivers?
Literally no one has said that. Everyone has said that self driving cars are guaranteed to *eventually be* safer than human drivers on account of the fact that when one learns, the entire fleet learns, and combined with the fact that over the many millions of miles driven across fleets of all self driving cars they are already, even in their damn beta state performing exceptionally well.
We need a test-track somewhere that's got nothing but driverless cars on it,
You want to train a driving technology on an artificial track that isn't remotely representative of real world conditions? What are you hoping to achieve? More accidents? More deaths? What next, should we teach all drivers on the road to drive wearing a VR headset while reclined on a lounge chair?
Re: (Score:2)
Re: (Score:2)
You want to train a driving technology on an artificial track that isn't remotely representative of real world conditions?
I don't see a problem having these cars on the streets. But until they are provably safer than humans, they should come with a safety driver.
Re: (Score:2)
You want to train a driving technology on an artificial track that isn't remotely representative of real world conditions?
Ummm, that is EXACTLY how i started training all of the drivers that I have trained. The driver absolutely needs to know the basics before being thrust into chaos.
Re: (Score:2)
You want to train a driving technology on an artificial track that isn't remotely representative of real world conditions?
Ummm, that is EXACTLY how i started training all of the drivers that I have trained. The driver absolutely needs to know the basics before being thrust into chaos.
Sure. I tend to start with a large parking lot. But the parking lot training constitutes less than 1% of the total training. After the parking lot we move to quiet residential streets with very few vehicles, and then to freeways -- where speeds are high but the driving problems are simple. Then we move to move complex city streets with more vehicles, etc.
Self-driving cars achieve parking lot proficiency about 15 years ago. They need training in complex environments.
Re: (Score:3)
Everyone has said that self driving cars are guaranteed to *eventually be* safer than human drivers on account of the fact that when one learns, the entire fleet learns
Sort of.
(Note that I think self-driving cars will be awesome, this isn't a dig at the concept, but your particular point misses a very important nuance, I think.)
I have a semi self-driving car, a Tesla with "Full Self Driving". It's clearly imperfect, but also fairly usable, but neither of those is my point. Whatever degree of capability it has is only from fleet-level learning, and I think that's a problem, and one that needs to be fixed.
Consider a competent human driver, in two situations. In the fi
Re:You don't say? (Score:4, Insightful)
Well, statistically they *are* safer. They just aren't 100% perfect.
Doesn't mean they don't occasionally have accidents, but the rate and severity are both less than human drivers.
I's just that people read about some super-rare but incredibly stupid thing a self-driving car did and assume that means they are not as safe as human drivers. People are bad at statistics.
Re: (Score:2)
Well, statistically they *are* safer.
False. You don't have the data necessary to make that conclusion. These self-driving cars do really weird things.
Re: (Score:2)
They do really weird things, so some of their accidents seem really "bad" to us, like "a human would NEVER do that!".
But then they have less accidents than humans per mile, and they "get hit" by other cars twice as often as they "hit other cars".
Re: (Score:2)
Well, statistically they *are* safer. They just aren't 100% perfect.
Doesn't mean they don't occasionally have accidents, but the rate and severity are both less than human drivers.
I's just that people read about some super-rare but incredibly stupid thing a self-driving car did and assume that means they are not as safe as human drivers. People are bad at statistics.
The problem is, it's not incredibly rare, 1 in a million scenarios that autonomous cars keep failing at. It's normal scenarios that an ordinary driver can negotiate their way around which the cars keep failing at. Things like deciding who gets to go first at stop or worse, forgetting what they're doing and stopping mid intersection. https://edition.cnn.com/2023/08/14/business/driverless-cars-san-francisco-cruise/index.html [cnn.com] https://www.drive.com.au/news/autonomous-car-robotaxi-traffic-jam-in-texas/ [drive.com.au] Also,
Re: (Score:2)
Doesn't the fact that AI cars drive more miles than humans mean that the accident statistics are in fact even better for them than they look to begin with?
Re: You don't say? (Score:2)
Only if you factor in when the milage is clocked up and what conditions.
How many miles does the robocar clock up in the middle of the night to bring the accidents-per-mile figures down compared to human drivers doing most of their milage in heavy traffic?
Re: (Score:2)
That would actually make them appear much safer yet again:
"One of the surprising facts you’ll discover when investigating car accidents is that the most serious crashes occur at night. Despite the roads being less active at night, approximately 50% of serious car accidents and fatalities occur after dark. "
https://www.brunobrunolaw.com/... [brunobrunolaw.com]
Re: (Score:2)
I's just that people read about some super-rare but incredibly stupid thing a self-driving car did and assume that means they are not as safe as human drivers. People are bad at statistics.
Knowing that statistics say that AI is a better driver than myself is not very heartening when I see the AI constantly unable to deal with situations that I could easily and simply deal with. Sure, the AI can apply the brakes faster than I can; however, I drive myself in such a way so that I don't HAVE to press the brakes. I avoid the situation entirely, which an AI can not do.
The numbers says AI is a better driver than myself. The real world says that the AI is not anywhere NEAR close to being as safe as I
Re: (Score:2)
Sure, if you are actually better than 99% of the drivers out there, you are probably safer than an AI car.
That doesn't change the overall analysis of AI car safety versus the average driver.
And of course Dunning-Kruger would like a word.
Re: (Score:2)
Absolutely, they aren;t ready for everywhere yet. Won't be ready for all-season driving anywhere with snow for a long time I think.
But for where they are being used, they seem already highly effective.
I do wonder why everyone is trying so hard to replace taxi drivers though.
Re: (Score:2)
Re: (Score:2)
So your point is what, that AI drivers get an unfair advantage because they are driving better cars?
Re: (Score:2)
In essence, testing imperfect autonomous driving vehicles on public roads is the Trolley Problem (https://en.wikipedia.org/wiki/Trolley_problem), and it boils down to the following trade-off (I'm going to use made-up numbers for illustration):
Testing autonomous vehicles in computer simulations or on privately-owned test tracks maximizes safety during testing, but (it is argued) will result in it taking, say, 20 years to get a fully safe technology. In contrast...
Permitting testing of imperfect autonomous ve
Re: (Score:2)
By developing the fully-safe technology 15 years earlier, you can save hundreds of thousands or millions of lives that would otherwise have been caused by human-driver-error during those 15 years.
This assumes that we'll get the entire universe to stop driving their existing cars the second the technology is ready to go. And thus far, most of us buying used vehicles and driving them for a decade or more before trading? That's not really gonna be the case.
I'm also having a real hard time wrapping my head around this blind-faith that self-driving cars will ever really be safer than humans. Thus far they haven't been, and tossing human lives at the problem seems the most backwards possible way to go abo
it may take an criminal case to fix issue and kids (Score:2)
it may take an criminal case to fix issue and it may take an bus full of kids dieing in really bad crash to get the case into an criminal court. Where they can't hide behind NDA's or forced arbitration.
Move Fast and Break Things (Score:4, Insightful)
https://xkcd.com/1428 [xkcd.com]
You think? (Score:2)
Safe? (Score:4, Insightful)
Still not seeing a statistical analysis if they are safer yet, or not.
Re:Safe? (Score:5, Informative)
https://arstechnica.com/cars/2... [arstechnica.com]
Pretty good article with some decent investigative journalism.
Re: (Score:2)
Re: (Score:2)
Oh, absolutely: they are only safer than humans on certain roads, under certain conditions. They ain't driving in Eastern Canada in the winter for a long while yet!
It is telling that, under the conditions they do drive, they get hit a lot more than they hit. The majority of accidents they *are* involved in are another driver's fault. It's interesting to speculate what they accident record would look like if say San Francisco only allowed AI-driven vehicle on the roads.
Re: (Score:2)
Oh, absolutely: they are only safer than humans on certain roads, under certain conditions.
How do you know? What are the statistics for humans on those certain roads, under those certain conditions?
Re: (Score:2)
Check the ArsTechnica article and associated links posted earlier in this chain of replies. There's been several reports like that, all stating the same basic thing: AI driven cars get in stupid accidents, but they get in way less accidents of all types less often per mile driven than human drivers do under the same circumstances i.e. same areas driven, like say San Francisco.
Re: (Score:2)
"It is telling that, under the conditions they do drive, they get hit a lot more than they hit"
That's not surprising at all, given how often Waymo blocks the road completely.
Re: (Score:3)
The take-away from that article is that only Waymo has a really good system, everyone else is still a decade behind them with an inability to handle many specific and not that uncommon situations.
I think Cruise is just moving faster than their technology warrants, because they know they are behind and are under management pressure to catch up. The others were mostly just jumping on the bandwagon and have largely fallen off it (e.g. Uber). Then you have Elon Musk and his hopelessly over-optimistic delusions
Whaaa?? (Score:2)
But they're so much safer [getcruise.com]!!!
I honestly have no idea if that whitepaper is fairly accurate or completely misleading, but I do find it ironic that they're getting shut down less than a month after its release.
Modest Proposal (Score:2)
High powered self-driving electric bicycles.
A sort of "walk before you run" interval in the march of progress towards the self driving car.
The reduced size and feature set would mean reduced product cost, which in turn would surely mean ALL the people wanting self driving cars have a better shot at getting them, and sooner! I'm sure there's no unpredictable breaking problem that will create significant problems for other road users while the bugs in the code get ironed out. Don't listen to the haters!
Right
Re: (Score:2)
Re: (Score:2)
Isn't saying "aren't ready *yet*" is a way of justifying anything that isn't making progress?
How many people not interested in, say, communism would re-think if an apologist for it would say "We'll it's not got a *perfect* track record but it's improving; it's just not ready yet which is why we should continue to support it!"
Suddenly the argument just sounds like a justification for upholding something causing harm, showing no progress and sweeping reasons to refuse to allow it to continue under the rug (su
Re: (Score:2)
Isn't saying "aren't ready *yet*" is a way of justifying anything that isn't making progress?
Nah, it's still ok as long as:
1) They are willing to pay for it (great!)
2) They are willing to do research in a way that isn't harmful (like having safety drivers)
Better than humans is a low bar (Score:2)
Re: (Score:2)
driverless cars are great because they are usually parked in garages or sides of roads and allows cars with drivers less traffic to deal with.
i vote for more driverless cars, and more public transport for others.
Re: (Score:2)