This is a completely irrelevant statistic at this point. You're comparing accidents per miles driven of regular vs. experimental self-driving cars...the two "sample sizes" so to speak are so vastly different that no valid comparison is possible.
You are quite mistaken. The comparison is fair in deciding whether AVs are safe enough for real world testing in limited environments, not whether they are ready to be left unsupervised in all conditions.
It's just cheaper to put real cars on real roads and endanger real people.
That's all that matters, really. If it would have taken $100 million in extra testing to save this one life, it would not have been worth it. When you also consider that such rigorous testing would delay the deployment of safe AVs (even if only by days), it makes real world testing even more imperative.
Yes, people dying is part of the calculus in testing new vehicles just as it is in testing new drugs. Taking no risk results in far more deaths and only a whiny name calling child thinks otherwise.
"Failure" is defined as the backup driver having to take over. That says nothing about the overall safety of the car (and driver) on the road, just that the cars are not ready to be left on their own.
Google's relative caution may have less to do with caring about public safety and more about fearing public and regulatory backlash.
Except nobody is making the point that AVs are that good driver all the time right now. Just that they will be someday.
However, your standard that they have to be that good before being a useful technology (your "unacceptable compromise") is also absurd. To increase road safety they only need to be better than a decent fraction of human drivers (certainly less than average).
There are videos on Youtube showing this road to be much better lit than the Uber video lets on. This one even has pedestrians on the side of the road at almost the same site of the accident, and the driver has no problems seeing them in advance.
Videos that are just as misinterpreted (by you, for example) as the Uber video. The pedestrians in this video are on the same side of the road as the lights, in the Uber video the ped comes from the other side and is in the darkest spot between lights. This driver is also specifically looking for peds, not your typical driver. There is no fair comparison.
The other part of this is, you could take a hundred human drivers and put them in the same conditions, and you may find one drunk or inattentive enough to kill the same pedestrian. But most would have seen her in time to at least slow down (for a non-fatal accident) or avoid her entirely.
A million dead pedestrians (and counting) from human drivers would disagree with you. The video you linked shows several drivers going 10 to 20 miles over the speed limit, driving recklessly past other vehicles and no doubt changing lanes, listening to the radio or on the phone etc. You may be the supreme driver you think, but even you would not handle every pedestrian situation perfectly. Most drivers are not so supremely focused (again ask the million dead).
But you could take a hundred Uber vehicles with the same condition and most would have killed the same pedestrian, because they would share the same design flaw.
In the same exact spot under the same conditions most Uber cars would have failed, but most safety drivers would not have if the conditions were as ideal as you think. Not to mention that the Uber AV might have done far better than the average human in conditions that were slightly different. A ped coming from the right in a better lit area might get detected by Uber 99.9% vs. only 95% for humans. The same ped coming in the same light would have been flattened by the majority of those SUVs going 60 mph in the left lane, that is the fair comparison to the Uber. Not you going 38 mph in the right lane and totally focused on this one moment.
The big difference between AVs and humans is that the next version of Uber's car will handle this situation much better (as will all other AVs too, I imagine). Whereas humans don't get any better, they make the same mistakes year after year.
Many people might avoid this accident some of the time, but few would avoid it all the time as you suggest. You're living in a fantasy world if you would blindly jaywalk in this situation and assume you'd be fine. That's what you're saying.
Good lord, why? Maybe it is an engineering failure, but why wouldn't you hold humans to the same standard?
Nuclear plants "that blow up occasionally" are still better than coal plants that kill people predictably. That is, if you care care about human life and aren't just anti technology.
The only difference is that you are holding Uber (and other self driving car pioneers) to an unreasonable standard.
Some level of professional driver could have avoided this accident. Some level of self-driving car could have avoided this accident. But could the minimal level of driver we give a license to have avoided this accident? Certainly not. That is the ridiculous double standard you are trying to advocate for.
This car was not ready for limited testing under these conditions WITH a safety driver.
Again, by what standard? Thousands of people have died in auto accidents since this one from human drivers. Should we suspend all driving?
The trouble is someone who is Not driving is going to automatically become fatigued watching and get distracted.
So what? The car doesn't get tired or exhausted. As long as the overall system is roughly as safe as the average driver it deserves to be on the road. All drivers get bored and distracted by such driving, autocars have a much better chance of avoiding freak accidents such as this.
They'd be better off with a remote datacenter of staff watching the car's cameras 24x7 in 10-minute shifts.
That's a ridiculous conclusion based on no evidence. If other self-driving car companies want to develop such a system and show that it's safer (or cheaper) then by all means go for it. To state it as an obvious solution is again, ridiculous.
Yes, there are thousands of logged accounts of the safety driver taking control of the vehicle.
Thousands where the driver saved a life by taking control at the last second? Of course not, that is an absurd standard.
Just the other day a student driver crashed into the DMV building while taking her driving test. Does that prove that the current system of human driver testing is a complete failure?
Didn't work? It has in fact worked in thousands of other cases. It isn't foolproof but maybe it's good enough. Not that you could even clearly say this is an example of the system failing.
This is a serious (as in catastrophic level) failure on Uber's part. It doesn't _matter_ that the pedestrian was crossing illegally. The car needs to detect and avoid obstacles on the road.
Complete the thought, please. Some obstacles, most obstacles, most hazardous obstacles, all obstacles? All obstacles in every condition? As well as a human, better than a human, perfectly?
Firstly: In most most countries what she did was perfectly legal and Uber wants to deploy this software outside the USA.
Hell, in some countries that Uber would have been driving on the wrong side of the road. What the hell were they thinking?
This demonstration of incompetence has set back licensing of self-driving vehicles by a decade (thanks to public pushback) or more and will ensure higher levels of care in certification and more rigorous testing procedures.
That the public overreact should not be blamed on Uber. It certainly affects them and their bottom line, which they will have to reckon with, but you can't blame them for others being luddites. Blame the luddites.
Secondly: If that had been a cow or other large animal, it wouldn't be a dead pedestrian, it would be a set of dead car occupants. There's a reason that the "Moose Manouveure" is a required part of the finnish driving test.
And when Uber is ready to test in Finland they can add software to handle that.
Thirdly: Hazard perception is an important part of most countries' driving tests. The car utterly failed - but then again most american drivers utterly fail, which is why your USA license is generally not transferrable to another country without a full driving test.
So again, why the problem if the car is generally safer than human drivers on the road it is operating on?
Fourthly: Even if being on the road _is_ illegal, this kind of thing happens and humans take account of/react to it. If there's a crash or other hazard on a freeway I want my robocar to stop, not plow on and end up being part of the mayhem. Uber have demonstrated a 100% fail at hazard handling. California was right to order them off the road and Arizona should follow suit.
So basically you're saying that these cars are not ready to be deployed around the world with no safety driver. Good thing that nobody is arguing for that then, isn't it?
For the same reason we can automatically assign some fault for an accident if we detect that a human driver has high blood alcohol content or was texting, on that basis alone. If the driver was operating below the maximum possible level of ability for that specific driver, then there is responsibility to be assigned to the driver.
First, that is not the standard. Drivers do not need to be in peak condition for every drive, that is absurd. People drive all the time while a little tired, after a fight with their ex, with children in the car, while listening to the radio, and yes even after having a glass of wine. Plus people, ALL people make reasonable mistakes. They are NOT held responsible for every accident, only those that likely would have been avoided by another reasonable driver.
Second, if this were true it would lead to perverse and damaging outcomes. The geriatric mother would drive home despite having poor night vision and terrible reflexes because the son had a single glass of wine (or even had a long day at work).
Because.... we say every vehicle's driver has An absolute duty to prevent their vehicle from colliding with pedestrians, And they have some fault for an incident unless it is clearly impossible that they would have prevented it.
"impossible?" Again, this is a fantasy of your own making. Drivers have the obligation to avoid accidents to the best of their ability at the time and are not found responsible unless they were doing something illegal. There would be thousands more people in jail for manslaughter if your standard were applied, just about every pedestrian death is avoidable if the driver did something different. A human driver in this accident would not be in jail if it were simply pointed out that some drivers (or even this driver) could have possibly seen that tiny bit of light.
That is: "Assuming the driver makes no errors -- this accident could still not possibly be prevented; would be an adequate defense", BUT for this case that doesn't hold..... The Self-Driving car COULD have prevented the accident and FAILED to prevent the accident, and it was ultimately caused by Unsafe driving and some defect in the Uber system that they have yet to determine.
Aaaand this is why your argument is totally vacuous. If we use your standard self-driving cars will never be deployed because EVERY accident could be avoided if they had more sensors if they had better algorithms if they launched bubble wrap safety balls from the grill to cocoon a pedestrian that darts in front of them. Your extreme position will cost millions of lives and countless injuries and damage.
Going back to perverse incentives. Which do we want on the road, a human with a 97% safety record or an autocar with a 99% safety record? Your answer is the human and your reasoning is simply ridiculous FUD.
Sure, and Backpage 100% certainly is used for murder. Why don't they just prevent all murders while they're at it?
As an expert I'm sure you can tell us all the things Backpage can easily do to prevent illegal trafficking without damaging the rest of their business. I'll wait.
Ouch. Hal is clearly an idiot with an agenda. His name is an anagram for "Harlot Rep", so I think he has some unresolved sexual issues.
It is clear that Backpage knowingly facilitated prostitution by seeking out advertisers and helping them with their ads. It's likely that they turned a blind eye to underage prostitution, but it's unclear if they didn't want to do the job of policing for legal or moral reasons or if they just wanted the extra revenue.
BTW, while the age of consent is lower in much of the world, in most of those places the age of consent for prostitution is significantly higher, usually 18, I believe. This is a reasonable compromise between the reality of sexual relations and protection of adolescents. This has nothing to do with the issue here, just wanted to clarify.
More lying. Jody Bruchon has refuted your distortions and outright lies several times, yet you keep on going. None of the links you cite support your claims.
You are positively Trump-like in your ability to shrug off the truth.
Except the number of things that computers are good at is growing all the time. The things that humans are good at is not, in fact it might be shrinking.
So lying for a good cause is okay? You might be right, most people aren't really paying enough attention to spot your lies and those who call you on your BS are quickly labeled as pedophiles. Nice angle.
I'll give you credit for attempting to prove things I don't believe using assumptions and "suspicions" that aren't even close to being true. That takes chutzpah.
Bullshit. I don't disparage those who work hard in whatever job they choose (or are stuck with). I also don't envy or admire them for it.
For whatever reason, you fell for the religious and political lie that hard manual labor is a virtue. Now you are passing that lie on to others, "good for you guys, keep up the good work, I really respect you (while I earn 5 times your pay with half the effort)."
The fact is that some jobs ARE better than others and ARE more desired than others and ARE paid more than others which (usually) leads to a higher quality of life for the worker and his/her family. Sending kids off to a work camp rather than science camp is NOT a good thing for those individuals even if it is a necessary byproduct of our flawed economy.
We've also established that you're an ignorant, small minded bigot who thinks that manual labor is the only challenging work. News flash: there are just as many lazy millennials doing shitty manual labor jobs as there are in other careers. The only difference is that you get paid less and you smell worse at the end of the day.
So sure, send your kids to pick tobacco instead of a science program. In 20 years my kids will be glad to have someone willing to do their gardening and clean their toilets for pennies.
Slashdot Mirror
So it wouldn't matter if such a system saved thousands of lives if it has even a single failure that kills one?
This is a completely irrelevant statistic at this point. You're comparing accidents per miles driven of regular vs. experimental self-driving cars...the two "sample sizes" so to speak are so vastly different that no valid comparison is possible.
You are quite mistaken. The comparison is fair in deciding whether AVs are safe enough for real world testing in limited environments, not whether they are ready to be left unsupervised in all conditions.
It's just cheaper to put real cars on real roads and endanger real people.
That's all that matters, really. If it would have taken $100 million in extra testing to save this one life, it would not have been worth it. When you also consider that such rigorous testing would delay the deployment of safe AVs (even if only by days), it makes real world testing even more imperative.
Yes, people dying is part of the calculus in testing new vehicles just as it is in testing new drugs. Taking no risk results in far more deaths and only a whiny name calling child thinks otherwise.
"Failure" is defined as the backup driver having to take over. That says nothing about the overall safety of the car (and driver) on the road, just that the cars are not ready to be left on their own.
Google's relative caution may have less to do with caring about public safety and more about fearing public and regulatory backlash.
Yeah, that's not the law or the reality. But if it makes you feel safer in a scary world that is beyond your control, go on believing it.
Except nobody is making the point that AVs are that good driver all the time right now. Just that they will be someday.
However, your standard that they have to be that good before being a useful technology (your "unacceptable compromise") is also absurd. To increase road safety they only need to be better than a decent fraction of human drivers (certainly less than average).
There are videos on Youtube showing this road to be much better lit than the Uber video lets on. This one even has pedestrians on the side of the road at almost the same site of the accident, and the driver has no problems seeing them in advance.
Videos that are just as misinterpreted (by you, for example) as the Uber video. The pedestrians in this video are on the same side of the road as the lights, in the Uber video the ped comes from the other side and is in the darkest spot between lights. This driver is also specifically looking for peds, not your typical driver. There is no fair comparison.
The other part of this is, you could take a hundred human drivers and put them in the same conditions, and you may find one drunk or inattentive enough to kill the same pedestrian. But most would have seen her in time to at least slow down (for a non-fatal accident) or avoid her entirely.
A million dead pedestrians (and counting) from human drivers would disagree with you. The video you linked shows several drivers going 10 to 20 miles over the speed limit, driving recklessly past other vehicles and no doubt changing lanes, listening to the radio or on the phone etc. You may be the supreme driver you think, but even you would not handle every pedestrian situation perfectly. Most drivers are not so supremely focused (again ask the million dead).
But you could take a hundred Uber vehicles with the same condition and most would have killed the same pedestrian, because they would share the same design flaw.
In the same exact spot under the same conditions most Uber cars would have failed, but most safety drivers would not have if the conditions were as ideal as you think. Not to mention that the Uber AV might have done far better than the average human in conditions that were slightly different. A ped coming from the right in a better lit area might get detected by Uber 99.9% vs. only 95% for humans. The same ped coming in the same light would have been flattened by the majority of those SUVs going 60 mph in the left lane, that is the fair comparison to the Uber. Not you going 38 mph in the right lane and totally focused on this one moment.
The big difference between AVs and humans is that the next version of Uber's car will handle this situation much better (as will all other AVs too, I imagine). Whereas humans don't get any better, they make the same mistakes year after year.
Still you who's trolling actually.
Many people might avoid this accident some of the time, but few would avoid it all the time as you suggest. You're living in a fantasy world if you would blindly jaywalk in this situation and assume you'd be fine. That's what you're saying.
Good lord, why? Maybe it is an engineering failure, but why wouldn't you hold humans to the same standard?
Nuclear plants "that blow up occasionally" are still better than coal plants that kill people predictably. That is, if you care care about human life and aren't just anti technology.
The only difference is that you are holding Uber (and other self driving car pioneers) to an unreasonable standard.
Some level of professional driver could have avoided this accident. Some level of self-driving car could have avoided this accident. But could the minimal level of driver we give a license to have avoided this accident? Certainly not. That is the ridiculous double standard you are trying to advocate for.
This car was not ready for limited testing under these conditions WITH a safety driver.
Again, by what standard? Thousands of people have died in auto accidents since this one from human drivers. Should we suspend all driving?
The trouble is someone who is Not driving is going to automatically become fatigued watching and get distracted.
So what? The car doesn't get tired or exhausted. As long as the overall system is roughly as safe as the average driver it deserves to be on the road. All drivers get bored and distracted by such driving, autocars have a much better chance of avoiding freak accidents such as this.
They'd be better off with a remote datacenter of staff watching the car's cameras 24x7 in 10-minute shifts.
That's a ridiculous conclusion based on no evidence. If other self-driving car companies want to develop such a system and show that it's safer (or cheaper) then by all means go for it. To state it as an obvious solution is again, ridiculous.
Yes, there are thousands of logged accounts of the safety driver taking control of the vehicle.
Thousands where the driver saved a life by taking control at the last second? Of course not, that is an absurd standard.
Just the other day a student driver crashed into the DMV building while taking her driving test. Does that prove that the current system of human driver testing is a complete failure?
Didn't work? It has in fact worked in thousands of other cases. It isn't foolproof but maybe it's good enough. Not that you could even clearly say this is an example of the system failing.
This is a serious (as in catastrophic level) failure on Uber's part. It doesn't _matter_ that the pedestrian was crossing illegally. The car needs to detect and avoid obstacles on the road.
Complete the thought, please. Some obstacles, most obstacles, most hazardous obstacles, all obstacles? All obstacles in every condition? As well as a human, better than a human, perfectly?
Firstly: In most most countries what she did was perfectly legal and Uber wants to deploy this software outside the USA.
Hell, in some countries that Uber would have been driving on the wrong side of the road. What the hell were they thinking?
This demonstration of incompetence has set back licensing of self-driving vehicles by a decade (thanks to public pushback) or more and will ensure higher levels of care in certification and more rigorous testing procedures.
That the public overreact should not be blamed on Uber. It certainly affects them and their bottom line, which they will have to reckon with, but you can't blame them for others being luddites. Blame the luddites.
Secondly: If that had been a cow or other large animal, it wouldn't be a dead pedestrian, it would be a set of dead car occupants. There's a reason that the "Moose Manouveure" is a required part of the finnish driving test.
And when Uber is ready to test in Finland they can add software to handle that.
Thirdly: Hazard perception is an important part of most countries' driving tests. The car utterly failed - but then again most american drivers utterly fail, which is why your USA license is generally not transferrable to another country without a full driving test.
So again, why the problem if the car is generally safer than human drivers on the road it is operating on?
Fourthly: Even if being on the road _is_ illegal, this kind of thing happens and humans take account of/react to it. If there's a crash or other hazard on a freeway I want my robocar to stop, not plow on and end up being part of the mayhem. Uber have demonstrated a 100% fail at hazard handling. California was right to order them off the road and Arizona should follow suit.
So basically you're saying that these cars are not ready to be deployed around the world with no safety driver. Good thing that nobody is arguing for that then, isn't it?
For the same reason we can automatically assign some fault for an accident if we detect that a human driver has high blood alcohol content or was texting, on that basis alone. If the driver was operating below the maximum possible level of ability for that specific driver, then there is responsibility to be assigned to the driver.
First, that is not the standard. Drivers do not need to be in peak condition for every drive, that is absurd. People drive all the time while a little tired, after a fight with their ex, with children in the car, while listening to the radio, and yes even after having a glass of wine. Plus people, ALL people make reasonable mistakes. They are NOT held responsible for every accident, only those that likely would have been avoided by another reasonable driver.
Second, if this were true it would lead to perverse and damaging outcomes. The geriatric mother would drive home despite having poor night vision and terrible reflexes because the son had a single glass of wine (or even had a long day at work).
Because.... we say every vehicle's driver has An absolute duty to prevent their vehicle from colliding with pedestrians, And they have some fault for an incident unless it is clearly impossible that they would have prevented it.
"impossible?" Again, this is a fantasy of your own making. Drivers have the obligation to avoid accidents to the best of their ability at the time and are not found responsible unless they were doing something illegal. There would be thousands more people in jail for manslaughter if your standard were applied, just about every pedestrian death is avoidable if the driver did something different. A human driver in this accident would not be in jail if it were simply pointed out that some drivers (or even this driver) could have possibly seen that tiny bit of light.
That is: "Assuming the driver makes no errors -- this accident could still not possibly be prevented; would be an adequate defense", BUT for this case that doesn't hold..... The Self-Driving car COULD have prevented the accident and FAILED to prevent the accident, and it was ultimately caused by Unsafe driving and some defect in the Uber system that they have yet to determine.
Aaaand this is why your argument is totally vacuous. If we use your standard self-driving cars will never be deployed because EVERY accident could be avoided if they had more sensors if they had better algorithms if they launched bubble wrap safety balls from the grill to cocoon a pedestrian that darts in front of them. Your extreme position will cost millions of lives and countless injuries and damage.
Going back to perverse incentives. Which do we want on the road, a human with a 97% safety record or an autocar with a 99% safety record? Your answer is the human and your reasoning is simply ridiculous FUD.
Why on earth should they be held to a higher standard? That's ridiculous.
Sure, and Backpage 100% certainly is used for murder. Why don't they just prevent all murders while they're at it?
As an expert I'm sure you can tell us all the things Backpage can easily do to prevent illegal trafficking without damaging the rest of their business. I'll wait.
Ouch. Hal is clearly an idiot with an agenda. His name is an anagram for "Harlot Rep", so I think he has some unresolved sexual issues.
It is clear that Backpage knowingly facilitated prostitution by seeking out advertisers and helping them with their ads. It's likely that they turned a blind eye to underage prostitution, but it's unclear if they didn't want to do the job of policing for legal or moral reasons or if they just wanted the extra revenue.
BTW, while the age of consent is lower in much of the world, in most of those places the age of consent for prostitution is significantly higher, usually 18, I believe. This is a reasonable compromise between the reality of sexual relations and protection of adolescents. This has nothing to do with the issue here, just wanted to clarify.
More lying. Jody Bruchon has refuted your distortions and outright lies several times, yet you keep on going. None of the links you cite support your claims.
You are positively Trump-like in your ability to shrug off the truth.
Except the number of things that computers are good at is growing all the time. The things that humans are good at is not, in fact it might be shrinking.
So lying for a good cause is okay? You might be right, most people aren't really paying enough attention to spot your lies and those who call you on your BS are quickly labeled as pedophiles. Nice angle.
I'll give you credit for attempting to prove things I don't believe using assumptions and "suspicions" that aren't even close to being true. That takes chutzpah.
Bullshit. I don't disparage those who work hard in whatever job they choose (or are stuck with). I also don't envy or admire them for it.
For whatever reason, you fell for the religious and political lie that hard manual labor is a virtue. Now you are passing that lie on to others, "good for you guys, keep up the good work, I really respect you (while I earn 5 times your pay with half the effort)."
The fact is that some jobs ARE better than others and ARE more desired than others and ARE paid more than others which (usually) leads to a higher quality of life for the worker and his/her family. Sending kids off to a work camp rather than science camp is NOT a good thing for those individuals even if it is a necessary byproduct of our flawed economy.
We've also established that you're an ignorant, small minded bigot who thinks that manual labor is the only challenging work. News flash: there are just as many lazy millennials doing shitty manual labor jobs as there are in other careers. The only difference is that you get paid less and you smell worse at the end of the day.
So sure, send your kids to pick tobacco instead of a science program. In 20 years my kids will be glad to have someone willing to do their gardening and clean their toilets for pennies.