If an autonomous car with all it's sensors can't figure out a tree branch in the road and how to navigate it safely taking into consideration all the cars around it, then autonomous driving has failed. It's that simple. This type of thing is literally a prerequisite for being a safe driver.
You spent all the time arguing and then just agreed with us in this one sentence. Furthermore, this should be obvious to Google by now. So why are their cars still doing it?
Whatever dude, again you're confusing liability in an accident to responsibility on the road. The rear car is at fault in the accident, I said it again. The fact that the rear car in this situation is at fault in the accident doesn't absolve Google from recognizing it as a dangerous situation and fixing it. Google has control of everything their cars do. They can either model them after a 98-year-old grandma with cataracts, or they can model them after a driver that drives defensively and wants to not get in an accident *at all* liability or not. At a certain point their failure to address their issue of cars driving like a 98-year-old grandma is something that they should feel responsible for. I'm not holding them at fault for any accidents, I am simply expecting them to demonstrate that they appreciate the responsibility they have to drive safely, laws or no laws. If they want to give their AI developers all the control, then they must accept all the responsibility for what their cars do.
You are really underestimating the ability for a human driver in car #2 to anticipate a dangerous obstacle for car #1 and thus a fast stop. An automated car with all it's CPU cycles a second should be even better at anticipating a situation and almost NEVER have to stop suddenly.
Safe driving requires BOTH parties to be considerate. Goes Google realize right now this very moment that drivers are irrational? If so then the car should be designed to work with irrational drivers, period, full stop. It doesn't matter if the driver is following too close, and thus caused an accident. In these situations, the human driver in front usually slows down to make the situation safe again, or pull over and let the person pass. An automated car should be even more capable of doing so. Your attitude that a driver should only be concerned about safety in terms of whether they are liable for the accident or not really turns my stomach to tell you the truth.
The onus is on Waymo to drive in a predictable fashion here. The fact that humans sometimes follow too closely is a consideration when driving. I *always* know who is behind me and how close they are so that I know how to handle the situation if I must stop quickly. Is the driver that is following too closely at fault in a specific accident? Sure. But should Waymo identify this as a factor in proper defensive driving and have their cars do it as well? Absolutely. As more time goes on, the responsibility falls more and more on Waymo to recognize this as a dangerous scenario for everyone rather than blaming every driver that gets alarmed when they slam on the brakes for no apparent reason.
This is a problem that requires 99.9% accuracy in order to be used realistically and in a problem space which is almost infinite. Even making an assumption of a 99.99999% common senerio such as an oncoming car being in it's lane instead of on the sidewalk, could cause an accident.
Waymo reported that they drove something like 37,000 miles without a human interaction in November 2017. How does this add up, now knowing that they can't even navigate a normal intersection properly?
Human drivers manage well over 460K miles without an accident. Considering Waymo vehicles can't make a turn without possibly causing an accident what are they up to now? 12 miles?
If they were risk-averse, they would understand that the human drivers behind them have a certain reaction time, and it is stressful for them to have a vehicle in front of them stop suddenly for any reason.
Many good self-installed alternatives to Dropbox on linux. Maybe the problem is most Linux users know how to set up their own and they can make backups too.
Legally, the tailgater is at fault, agreed. But part of the point of autonomous cars is to drive with humans and they are failing at that if they are doing unexpected things.
Never mind driving in bad weather; it is painful enough clearing your normal vehicle of snow in the morning. Who is going to want to climb on top in a snow storm and pick every chunk of ice off the lenses and then use lens cleaner on them?
Oh brother. The difference is one out of a thousand humans may have trouble, *every* Waymo van will have trouble. The fact that this needs to be explained over and over is getting very tiring.
Slashdot Mirror
Whatever. You're just being an apologist for automated driving. The whole point is for them to drive with humans, not the other way around.
If an autonomous car with all it's sensors can't figure out a tree branch in the road and how to navigate it safely taking into consideration all the cars around it, then autonomous driving has failed. It's that simple. This type of thing is literally a prerequisite for being a safe driver.
Yup, also a problem where I am from.
Should Waymo limit spurious hard stops? Yes
You spent all the time arguing and then just agreed with us in this one sentence. Furthermore, this should be obvious to Google by now. So why are their cars still doing it?
Whatever dude, again you're confusing liability in an accident to responsibility on the road. The rear car is at fault in the accident, I said it again. The fact that the rear car in this situation is at fault in the accident doesn't absolve Google from recognizing it as a dangerous situation and fixing it. Google has control of everything their cars do. They can either model them after a 98-year-old grandma with cataracts, or they can model them after a driver that drives defensively and wants to not get in an accident *at all* liability or not. At a certain point their failure to address their issue of cars driving like a 98-year-old grandma is something that they should feel responsible for. I'm not holding them at fault for any accidents, I am simply expecting them to demonstrate that they appreciate the responsibility they have to drive safely, laws or no laws. If they want to give their AI developers all the control, then they must accept all the responsibility for what their cars do.
Does Google realize...*
You are really underestimating the ability for a human driver in car #2 to anticipate a dangerous obstacle for car #1 and thus a fast stop. An automated car with all it's CPU cycles a second should be even better at anticipating a situation and almost NEVER have to stop suddenly.
Safe driving requires BOTH parties to be considerate. Goes Google realize right now this very moment that drivers are irrational? If so then the car should be designed to work with irrational drivers, period, full stop. It doesn't matter if the driver is following too close, and thus caused an accident. In these situations, the human driver in front usually slows down to make the situation safe again, or pull over and let the person pass. An automated car should be even more capable of doing so. Your attitude that a driver should only be concerned about safety in terms of whether they are liable for the accident or not really turns my stomach to tell you the truth.
The onus is on Waymo to drive in a predictable fashion here. The fact that humans sometimes follow too closely is a consideration when driving. I *always* know who is behind me and how close they are so that I know how to handle the situation if I must stop quickly. Is the driver that is following too closely at fault in a specific accident? Sure. But should Waymo identify this as a factor in proper defensive driving and have their cars do it as well? Absolutely. As more time goes on, the responsibility falls more and more on Waymo to recognize this as a dangerous scenario for everyone rather than blaming every driver that gets alarmed when they slam on the brakes for no apparent reason.
This is a problem that requires 99.9% accuracy in order to be used realistically and in a problem space which is almost infinite. Even making an assumption of a 99.99999% common senerio such as an oncoming car being in it's lane instead of on the sidewalk, could cause an accident.
That's a really good point. That's from a still image too; driving requires picking these out from thousands of images a second.
Waymo reported that they drove something like 37,000 miles without a human interaction in November 2017. How does this add up, now knowing that they can't even navigate a normal intersection properly?
Human drivers manage well over 460K miles without an accident. Considering Waymo vehicles can't make a turn without possibly causing an accident what are they up to now? 12 miles?
So you're saying you couldn't set this up yourself? I could.
If they were risk-averse, they would understand that the human drivers behind them have a certain reaction time, and it is stressful for them to have a vehicle in front of them stop suddenly for any reason.
The word 'almost' means she wasn't tailgating, idiot. What it means is that these cars are doing things that are scary for other drivers.
Fix what? If humans don't have a problem with it, it's not broken.
Many good self-installed alternatives to Dropbox on linux. Maybe the problem is most Linux users know how to set up their own and they can make backups too.
Legally, the tailgater is at fault, agreed. But part of the point of autonomous cars is to drive with humans and they are failing at that if they are doing unexpected things.
My taxes are high enough thanks.
Never mind driving in bad weather; it is painful enough clearing your normal vehicle of snow in the morning. Who is going to want to climb on top in a snow storm and pick every chunk of ice off the lenses and then use lens cleaner on them?
The especially sad thing about this is they use it as an excuse to kill people now.
Game problem space is finite. Driving problem space is infinite. You're talking about two ends of the spectrum.
Oh brother. The difference is one out of a thousand humans may have trouble, *every* Waymo van will have trouble. The fact that this needs to be explained over and over is getting very tiring.
I don't think you understood the question. I asked for any other reason why religious fundamentalists continue to support him. No reason was given.