If self driving cars aren't on the road yet how does one conduct a study? Sounds dubious at best.
simulations!
Self driving cars behave in a more or less predictable manner, after all they are basically computer programs. Since safety and comfort will be their biggest concern it is highly unlikely that it will be more aggressive than most humans. It will accelerate more slowly and reduce speed earlier... that increases time of response
Rounded corners. An iPad mini for navigation. A horn that plays the Mac startup chime. 100 miles to the gallon. Comes in red or yellow. Can be parked in disabled spaces. An Apple logo on the bonnet conceals a gun for firing bullets at the tyres of Android-using vehicles.
There are more safety issues with self driving cars as well. All you need is one accident caused by a bug or fault in the self driving car and whichever company owns that car will go bankrupt. Tell me which insurance company will insure a car that has no driver? How are you going to sue a car for an accident? If people aren't driving the car then auto insurance companies won't have anything to insure because the people inside the car can't be at fault because they weren't driving. Optical recognition of streets/roads/highways and other vehicles would only work if all the roads were changed so nobody could change lanes or speed. Let's spend more time on developing something that would actually help society in the near future instead of trying to copy iRobot.
You don't have to drive something in order to insure it.
Self driving cars behave in a more or less predictable manner, after all they are basically computer programs. Since safety and comfort will be their biggest concern it is highly unlikely that it will be more aggressive than most humans. It will accelerate more slowly and reduce speed earlier... that increases time of response
And if that did eventuate, the algorithms would simply be adjusted.
I wouldn't leave Tesla either for some "experimental" Apple project that may end up just being another "hobby". If Apple is failing to recruit Tesla employees it's because those people don't want to leave a good thing for something that may fail.
Perhaps one day a full CarPlay system will include a 360-degree LIDAR scanner that continually compares what it sees to what's in iCloud and submits updates in real time, keeping Apple's "ground truth" data always up-to-date, maintaining a comprehensive real-time traffic database for routing of all CarPlay subscribers, and performing collision avoidance from all angles.
I wouldn't leave Tesla either for some "experimental" Apple project that may end up just being another "hobby". If Apple is failing to recruit Tesla employees it's because those people don't want to leave a good thing for something that may fail.
If Tesla employees can't be recruited it's more likely they feel Tesla options are worth more--more likely to skyrocket (no pun intended).
This is true. It is also true that regulation will lag behind technology by about 20 years in this case due to misinformed legislators and members of the public and media beating the fear drum, so it won't actually be legal until we're old.
EDIT
"due to the deeply ingrained monetary relationships between legislators and powerful corporate entities in whose interest it is to not change these laws"
Yes, there is a lot of upside to taking driving out of the hands of humans, but there's a lot of work to be done yet.
Car ads these days show grinning technicians as they watch their car screech to a halt after showing some collision avoidance heads-up display warning.
I'd look in the mirror first before slamming on the brakes or deciding to perhaps swerve. I can't think of any AI programs that have that sensitivity to events yet.
Quote:
Originally Posted by SolipsismY
I'll even say that's an understatement as I assume there are many core technologies needed that haven't yet been invented.
If we don't have this in 50 years ill be shocked. The only think I consider as an unknown at this point is will it being city or highway, first?
Oh we'll have it in much less than 50 years - down-town New York and in rural America - despite appropriate core technologies being uninvented and the appropriate ethical constraints on AI systems also being unimplemented.
Too much money & too many geeks getting off to stop it. Just relax and enjoy ;-(
This is true. It is also true that regulation will lag behind technology by about 20 years in this case due to misinformed legislators and members of the public and media beating the fear drum, so it won't actually be legal until we're old.
The safety benefits are largely over blown, much like antilock brakes. The level of AI required to actually improve safety on the highway isn't going to be realized anytime soon.
What I'd love to see is a hybrid car or preferrably in my case pickup that doesn't suck. A hybrid in this concept would have the internal combustion engine driving a generator only, running only when directed by the driver. I really like the idea of electric cars as I've worked large scale automation for years and have plenty of experience with motors. The thing that bothers me is the short range offered by current battery tech. Well that and the whole industry can't seem to realize anything new in functional design.
Yes, there is a lot of upside to taking driving out of the hands of humans, but there's a lot of work to be done yet.
Car ads these days show grinning technicians as they watch their car screech to a halt after showing some collision avoidance heads-up display warning.
I'd look in the mirror first before slamming on the brakes or deciding to perhaps swerve. I can't think of any AI programs that have that sensitivity to events yet.
I predict the Traffic Wave could be reduced significantly and eliminated with self-driving cars.
Oh we'll have it in much less than 50 years - down-town New York and in rural America - despite appropriate core technologies being uninvented and the appropriate ethical constraints on AI systems also being unimplemented.
Too much money & too many geeks getting off to stop it. Just relax and enjoy ;-(
Half a century is my prediction for commonplace, but I could see that happening much sooner, and that also means it needs to be available for the public long before that.
There are studies that show that self driving cars might actually be worse for congestion. Human drive more aggressively with higher acceleration curves which reduces latency/lag specially on crossroads.
Hogwash.
A computer can safely accelerate more rapidly than a human, which is precisely why they have traction control in modern cars.
And in the case of humans, it is precisely the aggression that causes the bottlenecks, as they hit the breaks causing the shock wave up the line.
In every other way, lag is way, way better with a computer. Which is why they can ride the other guy's bumper 2 cm back.
Yes, there is a lot of upside to taking driving out of the hands of humans, but there's a lot of work to be done yet.
A massive amount of work, I can't even imagine an AI capable of navigating the roads I'm driving today. You have low visibility, roads literally covered with snow and at times white outs. So how would current AI handle this compared to the stage events we have seen so far.
Car ads these days show grinning technicians as they watch their car screech to a halt after showing some collision avoidance heads-up display warning.
I'd look in the mirror first before slamming on the brakes or deciding to perhaps swerve. I can't think of any AI programs that have that sensitivity to events yet.
This will be a huge problem, humans would still on average take the route that does the least damage to other humans.
Oh we'll have it in much less than 50 years - down-town New York and in rural America - despite appropriate core technologies being uninvented and the appropriate ethical constraints on AI systems also being unimplemented.
AI tech will have to move far more rapidly than in the past. Current self driving systems are not even close to implementing AI, the are more or less elaborate terrain following robots.
Too much money & too many geeks getting off to stop it. Just relax and enjoy ;-(
A few dead people will slow it down.
I know that there are interests in the government pushing for this technology with the sad belief that it would increase safety on the road. I don't see this working out well at all. Even though there are examples on the road today, driving is not the same thing as a line following robot. An average driver gets new experiences thrown at them constantly requiring them to adopt instantly. No AI is capable of this at the moment.
Comments
If self driving cars aren't on the road yet how does one conduct a study? Sounds dubious at best.
simulations!
Self driving cars behave in a more or less predictable manner, after all they are basically computer programs. Since safety and comfort will be their biggest concern it is highly unlikely that it will be more aggressive than most humans. It will accelerate more slowly and reduce speed earlier... that increases time of response
My wish list:
Rounded corners.
An iPad mini for navigation.
A horn that plays the Mac startup chime.
100 miles to the gallon.
Comes in red or yellow.
Can be parked in disabled spaces.
An Apple logo on the bonnet conceals a gun for firing bullets at the tyres of Android-using vehicles.
There are more safety issues with self driving cars as well. All you need is one accident caused by a bug or fault in the self driving car and whichever company owns that car will go bankrupt. Tell me which insurance company will insure a car that has no driver? How are you going to sue a car for an accident? If people aren't driving the car then auto insurance companies won't have anything to insure because the people inside the car can't be at fault because they weren't driving. Optical recognition of streets/roads/highways and other vehicles would only work if all the roads were changed so nobody could change lanes or speed. Let's spend more time on developing something that would actually help society in the near future instead of trying to copy iRobot.
You don't have to drive something in order to insure it.
I'll even say that's an understatement as I assume there are many core technologies needed that haven't yet been invented.
If we don't have this in 50 years ill be shocked. The only think I consider as an unknown at this point is will it being city or highway, first?
Please - no "self driving cars." Self driving cars aren't going to happen.
And I'm definitely not going to own a self-driving car made by a data-mining advertising company.
simulations!
Self driving cars behave in a more or less predictable manner, after all they are basically computer programs. Since safety and comfort will be their biggest concern it is highly unlikely that it will be more aggressive than most humans. It will accelerate more slowly and reduce speed earlier... that increases time of response
And if that did eventuate, the algorithms would simply be adjusted.
I wouldn't leave Tesla either for some "experimental" Apple project that may end up just being another "hobby". If Apple is failing to recruit Tesla employees it's because those people don't want to leave a good thing for something that may fail.
Perhaps one day a full CarPlay system will include a 360-degree LIDAR scanner that continually compares what it sees to what's in iCloud
and submits updates in real time, keeping Apple's "ground truth" data always up-to-date, maintaining a comprehensive real-time traffic database for routing of all CarPlay subscribers, and performing collision avoidance from all angles.
I wouldn't leave Tesla either for some "experimental" Apple project that may end up just being another "hobby". If Apple is failing to recruit Tesla employees it's because those people don't want to leave a good thing for something that may fail.
If Tesla employees can't be recruited it's more likely they feel Tesla options are worth more--more likely to skyrocket (no pun intended).
EDIT
"due to the deeply ingrained monetary relationships between legislators and powerful corporate entities in whose interest it is to not change these laws"
Yes, there is a lot of upside to taking driving out of the hands of humans, but there's a lot of work to be done yet.
Car ads these days show grinning technicians as they watch their car screech to a halt after showing some collision avoidance heads-up display warning.
I'd look in the mirror first before slamming on the brakes or deciding to perhaps swerve. I can't think of any AI programs that have that sensitivity to events yet.
I'll even say that's an understatement as I assume there are many core technologies needed that haven't yet been invented.
If we don't have this in 50 years ill be shocked. The only think I consider as an unknown at this point is will it being city or highway, first?
Oh we'll have it in much less than 50 years - down-town New York and in rural America - despite appropriate core technologies being uninvented and the appropriate ethical constraints on AI systems also being unimplemented.
Too much money & too many geeks getting off to stop it. Just relax and enjoy ;-(
Siri is involved.
They're doing a June 2007. But this time it's not a phone.
The ...
AppleCart
or
carplay installers to push the envelope
i'd love to have both
i don't think the auto makers and after market services are carplay enthusiasts
sure some after market companies do carplay but for 600 is just too expensive
battery life for iOS is critical.....fuelcell please
apple has experience with fuel cells at its data centers
> Tell me which insurance company will insure a car that has no driver?
The one that has to insure the car that *has* a driver.
If Bob Smith gets into the car drunk and crashes it, the insurance company isn't going to get squat suing him.
If Apple has a bug in their software and crashes a car, the insurance company has a potential $100 billion payout.
I suspect they'll be lining up around the block. Like iPhone release day.
The safety benefits are largely over blown, much like antilock brakes. The level of AI required to actually improve safety on the highway isn't going to be realized anytime soon.
What I'd love to see is a hybrid car or preferrably in my case pickup that doesn't suck. A hybrid in this concept would have the internal combustion engine driving a generator only, running only when directed by the driver. I really like the idea of electric cars as I've worked large scale automation for years and have plenty of experience with motors. The thing that bothers me is the short range offered by current battery tech. Well that and the whole industry can't seem to realize anything new in functional design.
And if that did eventuate, the algorithms would simply be adjusted.
really? Bigger accelerations?
What if I tell that that is inversely proportional to safety and comfort? You can’t have everything.
I predict the Traffic Wave could be reduced significantly and eliminated with self-driving cars.
Half a century is my prediction for commonplace, but I could see that happening much sooner, and that also means it needs to be available for the public long before that.
"Hey Siri, go to the store."
"OK <name>, opening your door."
Seriously though, and off topic, Amazon Echo is putting Siri to shame in her ability to understand a command and execute it quickly.
There are studies that show that self driving cars might actually be worse for congestion. Human drive more aggressively with higher acceleration curves which reduces latency/lag specially on crossroads.
Hogwash.
A computer can safely accelerate more rapidly than a human, which is precisely why they have traction control in modern cars.
And in the case of humans, it is precisely the aggression that causes the bottlenecks, as they hit the breaks causing the shock wave up the line.
In every other way, lag is way, way better with a computer. Which is why they can ride the other guy's bumper 2 cm back.
A few dead people will slow it down.
I know that there are interests in the government pushing for this technology with the sad belief that it would increase safety on the road. I don't see this working out well at all. Even though there are examples on the road today, driving is not the same thing as a line following robot. An average driver gets new experiences thrown at them constantly requiring them to adopt instantly. No AI is capable of this at the moment.