Fog obstructs view in a way that affects computers even more
Humans rely 100% on vision. SDCs have both cameras and radar. SDCs do better in fog. Ask any Tesla owner that has used Autopilot in foggy conditions.
Also, icy roads, and lateral winds can be quite sudden and affect driving conditions in ways that self-driving cars have little ability to predict, and no experience with to learn, thus far
Tesla Autopilot has already driven several million miles on snowy and icy roads. So far no problems with lateral winds.
Because automated cars will be expensive, only the wealthy will ever be able to afford them.
Autopilot sensors add less than $3k to the cost of a Tesla. With mass production, the cost will likely drop dramatically. You will save more on insurance than you will pay for self-drive capability.
Save us President Trump from outsourcing most of our tech manufacturing to China!
You have it backwards. The Chinese are outsourcing to America. The biggest Chinese AI researcher is Baidu. Baidu's AI research lab is in Sunnyvale, California.
I would do the thing that lands me in the least amount of trouble with the authorities (I do not want to go to jail for a stranger).
Many versions of the "trolley problem" specify that no other person will be aware of your decision, so you shouldn't have to worry about the authorities, unless you blab about what you did.
When I first heard the trolley problem, it seemed obvious to me to throw the switch to kill one guy rather than allowing five to die through inaction. I was surprised to learn that means I am a psychopath. It still makes no sense to me that so many normal people believe that "inaction" somehow absolves them of moral culpability for the five deaths.
So how are you supposed to know if you are increasing or decreasing the risk to you by using a self-driving car?
Test your reaction time. It takes a typical human about 1500 milliseconds from the time they see a hazard, until they start depressing the brake pedal. An SDC takes about 1ms.
Even if you ignore all the other advantages of SDCs, this difference in reaction time alone gives them a huge safety advantage.
they only need to be safer than the average human driver.
Actually, not even that is true. It is likely that the worst human drivers would be the most likely to use SDCs, so even if the SDC was below average, it would still be an improvement.
I tend to daydream while driving, thinking about tech problems or whatever, and to be honest, I am a pretty bad driver. I would love an SDC, so I would no longer have to worry about my train of thought being interrupted by stuff happening on the road.
Yes: better critical thinking skills. Then when someone feeds you a load of baloney like "Back in the golden age of the 1960s everyone had a job for life, a full guaranteed pension, and owned a house" you will ask yourself two questions:
1. Does that make sense? 2. Is it supported by any actual evidence?
The answers are, of course, "no" and "no". Average job tenure has gone UP over the last few decades. Uneducated white men are less likely to have long term employment today, but everyone else is doing better. Most people in the 1960s did not have pensions. Home ownership has not only gone UP, but houses today are significantly bigger.
So what was the world like in the 1960s? Well, people spent a lot of time sitting around reminiscing about the "golden age" of the 1920s.
Here is a flaw: The entire study was done with contrived "vignettes" rather than actual cases. The vignettes were written by human doctors, so just because other human doctors were better than apps at reading between the lines and figuring out the intended diagnosis, does not mean that they would be better at diagnosing actual patients.
I think there is only one clear conclusion from this study: Doctors really don't like these apps.
It was done in real life during WW2. The drones were four engine B24 or B17 bombers packed with high explosives and crashed into high value targets.
Joe Kennedy (JFK's and RFK's older brother) was killed while piloting one of these planes when it prematurely detonated. Joe was considered the high achiever of the family, and his father (also named Joe) intended his son to pursue a political career after the war. It was only after Joe's death than Joe Senior put his effort, influence, and fortune behind John instead.
This study was conducted by medical doctors and published in a journal run by an association of doctors. So it isn't entirely surprising that doctors determined that doctors are really smart.
How can you be so confident that the sensors are right when it isn't even close to autonomous yet?
Autopilot doesn't handle intersections yet, but I fail to see why that would require any additional sensors.
What sensor will differentiate between a floating shopping bag and a flying rock?
Humans do it with their eyes, so an SDC should be able to do it with cameras. An ANN can differentiate a bag from a rock faster and more accurately than a human. And the SDC will certainly have a faster reaction time, usually by 1500 ms or more.
Google/Tesla/Uber have spent many thousands of hours on obstacle avoidance software. This is an area where SDCs are already far better than HDCs.
For many cars, the backup camera and GPS are part of a single package, since they use the same display. So if you skip the GPS, you get no backup camera.
Starting in 2018, backup cameras are mandatory in all new cars, so they will have to either unbundle the GPS, or include it as standard.
It is only partially self driving. But that is because of the software, not the hardware. Once the software is ready, Tesla says that the cars already in the hands of customers will be fully self driving.
So $3000 is enough to make a HDC into a SDC. Once production ramps up, it is likely that the cost will fall dramatically. You will save far more in insurance premiums than the additional cost for the self-driving capability.
Tesla charges an extra $3k for the Autopilot option, which includes all the sensors and actuators needed for automation. That is no where near "double".
Slashdot Mirror
Fog obstructs view in a way that affects computers even more
Humans rely 100% on vision. SDCs have both cameras and radar. SDCs do better in fog. Ask any Tesla owner that has used Autopilot in foggy conditions.
Also, icy roads, and lateral winds can be quite sudden and affect driving conditions in ways that self-driving cars have little ability to predict, and no experience with to learn, thus far
Tesla Autopilot has already driven several million miles on snowy and icy roads. So far no problems with lateral winds.
Because automated cars will be expensive, only the wealthy will ever be able to afford them.
Autopilot sensors add less than $3k to the cost of a Tesla. With mass production, the cost will likely drop dramatically. You will save more on insurance than you will pay for self-drive capability.
... which is why they're looking at papers that have been cited, rather than just published by peer review.
That only works if you eliminate cycles from the citation graph. Otherwise, you cite me, and I will cite you. Win-win with no quality needed.
Save us President Trump from outsourcing most of our tech manufacturing to China!
You have it backwards. The Chinese are outsourcing to America. The biggest Chinese AI researcher is Baidu. Baidu's AI research lab is in Sunnyvale, California.
I would do the thing that lands me in the least amount of trouble with the authorities (I do not want to go to jail for a stranger).
Many versions of the "trolley problem" specify that no other person will be aware of your decision, so you shouldn't have to worry about the authorities, unless you blab about what you did.
When I first heard the trolley problem, it seemed obvious to me to throw the switch to kill one guy rather than allowing five to die through inaction. I was surprised to learn that means I am a psychopath. It still makes no sense to me that so many normal people believe that "inaction" somehow absolves them of moral culpability for the five deaths.
AI would need to break the laws of physics to be comparable.
Last time I checked, computers and human brains adhere to the same laws of physics.
So how are you supposed to know if you are increasing or decreasing the risk to you by using a self-driving car?
Test your reaction time. It takes a typical human about 1500 milliseconds from the time they see a hazard, until they start depressing the brake pedal. An SDC takes about 1ms.
Even if you ignore all the other advantages of SDCs, this difference in reaction time alone gives them a huge safety advantage.
they only need to be safer than the average human driver.
Actually, not even that is true. It is likely that the worst human drivers would be the most likely to use SDCs, so even if the SDC was below average, it would still be an improvement.
I tend to daydream while driving, thinking about tech problems or whatever, and to be honest, I am a pretty bad driver. I would love an SDC, so I would no longer have to worry about my train of thought being interrupted by stuff happening on the road.
Dev machines that spin up VMs are not cheap. Lots of memory and fast drive arrays add up quick! Not to mention EDO motherboards
I run VMs from a thumb drive on my 2012 laptop, and the performance is fine. There is no perceptible UI lag.
This would be the greatest thing ever.
open as "iPad", open as "iPhone 6s", etc.
Yes, that would be really slick, not just for usability, but also for testing.
they were moving so many jobs from high cost countries (i.e. US, Canada and Europe) to low cost countries. Sociopath much?
Why does that make her a sociopath? Is it because subhuman brown people in poor countries don't deserve their jobs as much as white people?
You don't have to completely avoid cannibalism, you just have to avoid eating the brain and spinal cord.
Do you have any constructive suggestions?
Yes: better critical thinking skills. Then when someone feeds you a load of baloney like "Back in the golden age of the 1960s everyone had a job for life, a full guaranteed pension, and owned a house" you will ask yourself two questions:
1. Does that make sense?
2. Is it supported by any actual evidence?
The answers are, of course, "no" and "no". Average job tenure has gone UP over the last few decades. Uneducated white men are less likely to have long term employment today, but everyone else is doing better. Most people in the 1960s did not have pensions. Home ownership has not only gone UP, but houses today are significantly bigger.
So what was the world like in the 1960s? Well, people spent a lot of time sitting around reminiscing about the "golden age" of the 1920s.
Everybody who spends money with them are just as guilty of this as the company, themselves.
Guilty of what? Providing willing workers with jobs? Plenty of companies hire seasonal workers. That doesn't make them evil.
As a CTRL key, what else?
Emacs users configure it as an extra CTRL key.
Vi users configure it as an extra ESC key.
That's called a reputable peer-reviewed journal
... and all the peers are also doctors.
If you can find an actual flaw ...
Here is a flaw: The entire study was done with contrived "vignettes" rather than actual cases. The vignettes were written by human doctors, so just because other human doctors were better than apps at reading between the lines and figuring out the intended diagnosis, does not mean that they would be better at diagnosing actual patients.
I think there is only one clear conclusion from this study: Doctors really don't like these apps.
It was done in real life during WW2. The drones were four engine B24 or B17 bombers packed with high explosives and crashed into high value targets.
Joe Kennedy (JFK's and RFK's older brother) was killed while piloting one of these planes when it prematurely detonated. Joe was considered the high achiever of the family, and his father (also named Joe) intended his son to pursue a political career after the war. It was only after Joe's death than Joe Senior put his effort, influence, and fortune behind John instead.
This study was conducted by medical doctors and published in a journal run by an association of doctors. So it isn't entirely surprising that doctors determined that doctors are really smart.
How can you be so confident that the sensors are right when it isn't even close to autonomous yet?
Autopilot doesn't handle intersections yet, but I fail to see why that would require any additional sensors.
What sensor will differentiate between a floating shopping bag and a flying rock?
Humans do it with their eyes, so an SDC should be able to do it with cameras. An ANN can differentiate a bag from a rock faster and more accurately than a human. And the SDC will certainly have a faster reaction time, usually by 1500 ms or more.
Google/Tesla/Uber have spent many thousands of hours on obstacle avoidance software. This is an area where SDCs are already far better than HDCs.
I wouldn't want any of my financial streams associated with any of this.
Don't worry. Slashdot geezers are not the target market.
For many cars, the backup camera and GPS are part of a single package, since they use the same display. So if you skip the GPS, you get no backup camera.
Starting in 2018, backup cameras are mandatory in all new cars, so they will have to either unbundle the GPS, or include it as standard.
But in reality they are merely fancy decision trees and look ups.
No. Leading AI research and applications are based on deep ANNs, which are neither decision trees nor lookups.
Please how a decision tree or lookup table can play world championship Go.
To my knowledge Google cars have only been driven in near perfect conditions. Ie no blizzards or blowing snow or ice of any kind.
Google has a test center in Michigan. Uber tests their cars in Pittsburgh. Tesla Autopilot has driven millions of miles on snowy and icy roads.
They also claim that it isn't self-driving.
It is only partially self driving. But that is because of the software, not the hardware. Once the software is ready, Tesla says that the cars already in the hands of customers will be fully self driving.
So $3000 is enough to make a HDC into a SDC. Once production ramps up, it is likely that the cost will fall dramatically. You will save far more in insurance premiums than the additional cost for the self-driving capability.
Few will pay double for a car for automation.
Tesla charges an extra $3k for the Autopilot option, which includes all the sensors and actuators needed for automation. That is no where near "double".