The problem with automated cars is that they need to not only prevent the accidents that a human would make, they need to not make their own accidents that only a computer would make. It should be impossible to put the car in a mode that creates an accident, period. Tesla's fault, even if the car owner blatantly activated this mode.
I think you're grasping at straws. What makes you think an autonomous taxi will be any more convenient then a real taxi today? If a person is going to be so eager to use an autonomous taxi then they would already be using a real taxi. There is no reason to think that an autonomous taxi will be less expensive then a real taxi is. If the last 30 years have taught me anything, corporations will always absorb any possible increase in profit brought on by technology. In addition to this, the expense of the live driver will be largely offset by the extra expense and maintenance of the car. As a result, automated taxi's will cost roughly the same. So why aren't people using them now?
"The car can’t start in Autopilot; it requires a set of circumstances (good data, basically) before you can engage the setting. These include clear lane lines, a relatively constant speed, a sense of the cars around you, and a map of the area you’re traveling through—roughly in that order."
You must live in a different world then I do. My bank has the same number of tellers it has always had, because the bank machine does 1/10 of what a teller does. Trains I will give you, but how hard is that, they're on a freaking track. The planes I have been on still have a pilot and a copilot, and I'm not yet asking Siri where I would find a book in the library. So in my lifetime, AI has achieved chess and then GO. I still contest that a computer always had an advantage in GO because of the sheer number of calculations required and low psychological benefit to play, once said computers reached the processing power for such calculations.
When a computer can play poker under a human's rules which makes card counting illegal, and *know* when the player across from it is bluffing; that will be true AI.
Sadly, for this reason it will be just as long before the safety of electronic cars that every keeps touting will carry any kind of statistical relevance at all. Safety only benefits society under a socialist capitalist government with a program to subsidize the things to the masses. If that would ever be affordable, I highly doubt it.
Or your insurance will pay out and it will be up to the insurance vendor to reclaim those millions from Google. I can't see them being too enthusiastic about getting into that scenario.
It doesn't matter how much safer they are. If people can't afford them they won't buy them. Until you can get an almost 100% infallible automated car for in the $20K range, it is not likely to get enough mass adoption to be beneficial on a statistically relevant scale. Adaptive headlights seem like a very good idea to me as well, but it seems all I can afford is a car with an air bag that will fire metal shrapnel towards me in an accident. Oh and I get a $5 seatbelt.
They can push all people to drive autonomous as much as they want if they're going to buy people one. Otherwise it is going to be a bit difficult to get a family barely able to afford to keep their ten year old beater on the road into one.
In winter where I live they don't always plow the roads, almost never down to the pavement in order to save on blade costs. The ruts that remain can turn a car sideways if they are not driven through properly and they rarely match up with the lane. I'm not clear on whether an automated car will actually analyze the ice contour of the road and driving according to that instead of according to where the lane is. Not that the lane matters because the markings are under the ice anyway. The car will have to navigate this while anticipating correctly where the other cars will go as they also deal with the ice.
If the car is truly self driving you would be able to put your five year old in it and have him reach Aunt Sally's safely without your intervention. It won't be a matter of your five year old learning to drive, because they won't have to. It could be a box of fine bone china going to Aunt Sally's.
If they make auto driving cars and they really are safer, then eventually they'll outlaw the ability to drive yourself.
That will require massive assistance from the government. How many people are likely to be able to afford a car with a supercomputer, radar, and a full array of sensors on board?
Personally I think automated cars are going to be terrible for the environment. Why pay $20/hour for parking when you can just have it drive around and around.
Excuse me? I'm not paying a higher premium for damages my automated car causes ever. Insurance should be priced the same way house insurance is, for replacement or repair of the property that may be required due to negligence of others. The car's negligence should not factor into the price of the insurance at all, since I am not controlling it. Today we buy driving insurance, automated cars should only require property insurance.
It depends on who wins.. Automation fanbois and corporate interest to sell the things and recoup the billions being spend in research, or governments and public safety. Personally I can't see insurance companies ever allowing these to be affordable for people to buy. I certainly won't buy one if I am on the hook for possible liability if it fails. When I drive, I can go everywhere at 25 mph if I want and virtually eliminate every possibility of making an accident. I'm sure if I took extreme cautions I would be almost 100% safe. That will never be so for an automated car.
I wouldn't buy a toaster that kills 10,000 people a year by burning down their homes. Why would I buy an automated vehicle that might do the same? I wouldn't be happy with a toaster that has killed 2 people.
But it doesn't even turn on unless conditions are perfect for it, rather than being able to deal with any condition. Therefore I don't consider it self driving.
What do standards have to do with anything? We're talking about one piece of software being compatible with another. A type of document created by application A cannot be used with application B, therefore they are not completely compatible. Standards don't matter to me when I get a spreadsheet with a button I am supposed to press to perform a validation of the data I have entered and the button does nothing.
Most of the documents we have floating around our office were made with Windows XP, or Windows 2003 at the latest.. and they still work fine with newer versions of Office.. So I am not really following you.
I'm not aware of Excel being 100% compatible with ODS files, so I would never open an ODS file with Excel. On the other hand, my place of work has XLS files floating around with macros embedded that have been used for ten years or more. I am talking about basic Excel macros here have have been around forever. Since it would be too much trouble to rewrite those macros in every file that has come to be reused over time, we cannot switch to LibreOffice until those macros work in LibreOffice. I do not call it 100% compatible if there is an XLS file that has been the same since Office 2003 that works in current Excel but that doesn't work in a new version of LIbreOffice.
Because in order for the Nissan to be considered entirely compatible with Ford, it also needs to be compatible with all the Ford extras? Seems perfectly reasonable to me. In my workplace we have hundreds of spreadsheet documents floating around with macros embedded and we need to be able to run those macros just the same, whether we use LibreOffice or not.
Slashdot Mirror
The problem with automated cars is that they need to not only prevent the accidents that a human would make, they need to not make their own accidents that only a computer would make. It should be impossible to put the car in a mode that creates an accident, period. Tesla's fault, even if the car owner blatantly activated this mode.
I think you're grasping at straws. What makes you think an autonomous taxi will be any more convenient then a real taxi today? If a person is going to be so eager to use an autonomous taxi then they would already be using a real taxi. There is no reason to think that an autonomous taxi will be less expensive then a real taxi is. If the last 30 years have taught me anything, corporations will always absorb any possible increase in profit brought on by technology. In addition to this, the expense of the live driver will be largely offset by the extra expense and maintenance of the car. As a result, automated taxi's will cost roughly the same. So why aren't people using them now?
https://www.technologyreview.c...
"The car can’t start in Autopilot; it requires a set of circumstances (good data, basically) before you can engage the setting. These include clear lane lines, a relatively constant speed, a sense of the cars around you, and a map of the area you’re traveling through—roughly in that order."
You must live in a different world then I do. My bank has the same number of tellers it has always had, because the bank machine does 1/10 of what a teller does. Trains I will give you, but how hard is that, they're on a freaking track. The planes I have been on still have a pilot and a copilot, and I'm not yet asking Siri where I would find a book in the library. So in my lifetime, AI has achieved chess and then GO. I still contest that a computer always had an advantage in GO because of the sheer number of calculations required and low psychological benefit to play, once said computers reached the processing power for such calculations.
When a computer can play poker under a human's rules which makes card counting illegal, and *know* when the player across from it is bluffing; that will be true AI.
Sadly, for this reason it will be just as long before the safety of electronic cars that every keeps touting will carry any kind of statistical relevance at all. Safety only benefits society under a socialist capitalist government with a program to subsidize the things to the masses. If that would ever be affordable, I highly doubt it.
Or your insurance will pay out and it will be up to the insurance vendor to reclaim those millions from Google. I can't see them being too enthusiastic about getting into that scenario.
Hey, as long as the glitchy autonomous car is 100 ft in the air instead of on the road with me, they can release them now if they want.
"It's technology, it must be good!"
It doesn't matter how much safer they are. If people can't afford them they won't buy them. Until you can get an almost 100% infallible automated car for in the $20K range, it is not likely to get enough mass adoption to be beneficial on a statistically relevant scale. Adaptive headlights seem like a very good idea to me as well, but it seems all I can afford is a car with an air bag that will fire metal shrapnel towards me in an accident. Oh and I get a $5 seatbelt.
How affordable do you think these autonomous cars will ever be?? I'm thinking they won't really ever be a middle class thing at all.
They can push all people to drive autonomous as much as they want if they're going to buy people one. Otherwise it is going to be a bit difficult to get a family barely able to afford to keep their ten year old beater on the road into one.
In winter where I live they don't always plow the roads, almost never down to the pavement in order to save on blade costs. The ruts that remain can turn a car sideways if they are not driven through properly and they rarely match up with the lane. I'm not clear on whether an automated car will actually analyze the ice contour of the road and driving according to that instead of according to where the lane is. Not that the lane matters because the markings are under the ice anyway. The car will have to navigate this while anticipating correctly where the other cars will go as they also deal with the ice.
If the car is truly self driving you would be able to put your five year old in it and have him reach Aunt Sally's safely without your intervention. It won't be a matter of your five year old learning to drive, because they won't have to. It could be a box of fine bone china going to Aunt Sally's.
If they make auto driving cars and they really are safer, then eventually they'll outlaw the ability to drive yourself.
That will require massive assistance from the government. How many people are likely to be able to afford a car with a supercomputer, radar, and a full array of sensors on board?
Personally I think automated cars are going to be terrible for the environment. Why pay $20/hour for parking when you can just have it drive around and around.
Excuse me? I'm not paying a higher premium for damages my automated car causes ever. Insurance should be priced the same way house insurance is, for replacement or repair of the property that may be required due to negligence of others. The car's negligence should not factor into the price of the insurance at all, since I am not controlling it. Today we buy driving insurance, automated cars should only require property insurance.
Are these sorts of scenarios unlikely?
It depends on who wins.. Automation fanbois and corporate interest to sell the things and recoup the billions being spend in research, or governments and public safety. Personally I can't see insurance companies ever allowing these to be affordable for people to buy. I certainly won't buy one if I am on the hook for possible liability if it fails. When I drive, I can go everywhere at 25 mph if I want and virtually eliminate every possibility of making an accident. I'm sure if I took extreme cautions I would be almost 100% safe. That will never be so for an automated car.
I wouldn't buy a toaster that kills 10,000 people a year by burning down their homes. Why would I buy an automated vehicle that might do the same? I wouldn't be happy with a toaster that has killed 2 people.
Autopilot also handles poor lane markings better than humans
Really? because I've read that it won't turn on without having clear lane markings.
But it doesn't even turn on unless conditions are perfect for it, rather than being able to deal with any condition. Therefore I don't consider it self driving.
What do standards have to do with anything? We're talking about one piece of software being compatible with another. A type of document created by application A cannot be used with application B, therefore they are not completely compatible. Standards don't matter to me when I get a spreadsheet with a button I am supposed to press to perform a validation of the data I have entered and the button does nothing.
Most of the documents we have floating around our office were made with Windows XP, or Windows 2003 at the latest.. and they still work fine with newer versions of Office.. So I am not really following you.
I'm not aware of Excel being 100% compatible with ODS files, so I would never open an ODS file with Excel. On the other hand, my place of work has XLS files floating around with macros embedded that have been used for ten years or more. I am talking about basic Excel macros here have have been around forever. Since it would be too much trouble to rewrite those macros in every file that has come to be reused over time, we cannot switch to LibreOffice until those macros work in LibreOffice. I do not call it 100% compatible if there is an XLS file that has been the same since Office 2003 that works in current Excel but that doesn't work in a new version of LIbreOffice.
Because in order for the Nissan to be considered entirely compatible with Ford, it also needs to be compatible with all the Ford extras? Seems perfectly reasonable to me. In my workplace we have hundreds of spreadsheet documents floating around with macros embedded and we need to be able to run those macros just the same, whether we use LibreOffice or not.
Have they fixed Excel macro support now? Admittedly I am not on a current version.