We have self driving vehicles already, and amazingly, we know what to do when there's a crash.
Except in the case of a rail system, you know you can blame the owner/operator of it.
But, in a world of EULAs and Terms of Service, I think it far more likely the manufacturers will have said "we bear no responsibility for the quality and suitability of our product to be a self driving car".
And, lawyers being lawyers, will move on to the next person they can blame. And that will be the owners or the person who was allegedly "driving".
I won't pay you for a car that wants to operate itself, but the blame falls to me when your product fucks up. I will simply buy a car which doesn't drive itself.
TFA rightly points out that this is a legal gray area -- which means you really don't want to be the test case.
Except, being on rails provides distinct advantages in terms of things being on auto-pilot.
There's far fewer degrees of freedom in terms of what can happen, because, well, you're on frigging rails.
You need to monitor your speed and your braking, but the turning is enforced by the rails unless you're going way too fast.
So why is non-mass transit any different?
Because cars aren't on rails?
Planes are slightly different, because you can bet that the pilot is still ultimately responsible for the aircraft, and if it crashes due to pilot error, he's going to be the one hung out to dry. (Other than that, we mostly just hope/trust that pilots are professional, qualified, and able to do the job at hand)
I would think the point would be that machines, once properly programmed, can be the worlds safest drivers...statistically. You, as a human, will still be responsible for taking over when the machine doesn't know what to do.
No way that's gonna work.
There's now way you can expect people to be alert and responsive if they have to be on the ball for that small fraction of the time -- they'll have started reading their paper or plenty of other things.
If I'm responsible for the operation of the vehicle, I'll bloody well drive myself and be engaged for the entire time, and don't need your autonomous car.
I'f I'm not responsible for the operation of the vehicle, I want to be in the back seat in one hell of a good safety cage with no pretense whatsoever that I'm in control.
You can't have the vehicle responsible most of the time, and the ostensible operator responsible whenever that stops working suddenly, it defeats the purpose.
Which, to me, is kind of a fairly fundamental problem with self driving cars. It's all or nothing. And if *all* the cars on the road aren't autonomous, then the autonomous ones are mostly a traffic hazard with no clear liability.
The manufacturer will have an EULA which absolves them from guilt.
It won't be the people who sold it, because they'll also have a contract term which says they are absolved from guilt.
So, it will come down to the owner, who will be entirely dependent on the quality of the product, as delivered by two entities who have already said "not us".
So, if you privately buy an autonomous car, and it crashes, you will likely be on the hook for it. If you merely hire them (as in a Taxi), then I'm sure the people who rent them will also absolve themselves from guilt in some strange way -- likely through arms length 3rd parties who do the actual operation.
This won't be so much "buyer beware" as "everyone else on the roadway beware", because you'll have a vehicle driving around that if it crashes, there's a long line of people who have already made sure their asses are covered.
The lawyers for the companies making and selling these will have covered their asses before it ends up in the hands of anybody else.
Yes, in fact, I do. Political will tends to fade pretty quickly on the wrong end of a gun barrel.
If America is going to choose to 'point a gun' at the entire EU, then you can pretty much expect the entire EU to kick the US out of military bases, and generally GTFO of town.
Members of the European Parliament say that it is 'very doubtful that data collection of such magnitude is only guided by the fight against terrorism,' and that there may be other motives such as political and economic espionage.
Of course they do, because part of the mandate is to look out for US commercial interests in general.
The problem is they use the same program to spy for the terrorists, as they do for the economic and political espionage.
Which means, unless the US is willing to carve out JUST the security stuff (which, they won't), every other country more or less has to block this program on the premise that it's just a widespread "spy on everybody, some of them might be security risks, some of it might be political intel, and some can be given to the corporations".
That's kind of the problem from the perspective of the rest of the world -- any form of cooperation with this spying has far broader ramifications than just national security.
Hell, people here routinely defend it, but increasingly you might see other world governments saying they won't allow you to do it any more (in which case, it will be done anyway, just in a more clandestine manner).
Snowden will absolutely be captured if he appears in any of these places and would be a great fool to testify there.
Or, if they're going to have him testify, they have diplomats collect him and bring him in on a plain covered by immunity, move him around in diplomatic cars, and house him in diplomatic residences.
Do you *really* think that it is impossible to basically "fuck you" and bring him there safely if there's the political will?
"No company's terms of service can override the law"
I see you're new here. Welcome to America!
It's not so much that they can't override the law, it's that the courts have determined that they're valid, and therefore lawful (even when unconscionable).
And, it should be noted, they want to find out in controlled conditions with sufficient protective equipment in a facility explicitly configured for this kind of situation.
Of course, if you have no idea of what will happen, how the hell can you know you're got "sufficient" protection and a facility which has been "configured for this kind of situation"?
Because, it seems to me, they don't actually know what constitutes "good enough" in this case, at which point saying you have done enough to mitigate is mostly just guess work, isn't it?
And, no, I don't know a damned thing about nuclear reactions in any meaningful sense of the word, so maybe the answer is fairly obvious to someone who does.
Not so much "Hacking the Concept of Time" as "Hacking camera software to change how it takes pictures"
Let me guess, you either didn't read the article, or didn't understand it:
Magyar mounted the device on a tripod in a busy Shanghai neighborhood and scanned pedestrians as they passed in front of the sensor. He then digitally combined over 100,000 sequential strips into high-resolution photographs.
He's not taking a single exposure. He's taking a very large amount of small slices over a span of time, and stitching them together into a single image.
He hasn't so much taken a 'snapshot in time' like a traditional camera, he's made images out of snapshots which occurred across time.
Which means he's taking objects going by at a pretty good clip, and combining a whole lot of them into something which looks like a single astounding image.
Some of his images have a time lapse quality to them, because they show things which are both in motion and still, over a time sequence:
Eerie distortions of objects in motion and at rest reminded viewers that they were looking at a pictorial representation of time, not space. Speeding buses were compressed into Smart cars. Individuals who paused at a bus stop were elongated like Metroliners. Slower walkers had billowing pants legs, or feet like skis, or Oscar Pistorius-style blades. And because of the peculiar nature of the scanning technology, everyone was moving in the same direction. "The horizontal axis is not about space, it's not about left and right, it's about earlier and later," he says. "If two people are crossing the pixel at the same moment, they will look like they are walking together."
If you read the article, you'll find he's done much much more than "Hacking camera software to change how it takes pictures" -- the resulting images look like a still frame, but are composited from a time lapse, and are MUCH more sophisticated than you seem to realize.
Why do people on Slashdot persist in dismissing things they don't really understand? What he's done is taken what look like still images, but are in fact a cross section in time.
That you think all he's done is to hack camera software means you don't have the barest idea of what it is he's actually done.
Damn all those days drooling over parts and tools at Radio Shack.
Funny, my experience with Radio Shack was more in the category of 'drooling tools that can't find the parts' -- they really went downhill for the last bunch of years, until ultimately becoming the "cheap-ass electronics and toys store".
And yet, the doctor seems to have determined that it had nothing to do with the current stuff and moved on:
There, on MugShots.com, was a younger version of my patient's face, with details about how she had been detained for cocaine possession more than three decades earlier. I looked away from the screen, feeling like I had violated my patient's privacy. I resumed our medical exam, without bringing up the finding on the Internet, and her subsequent hospital course was uneventful.
So, depending on the kinds of tests he was doing, he apparently concluded it was a red herring.
I don't know enough about it to know how long it stays in your system... but I gather from TFS that the doctor decided that the information he had wasn't what he needed and moved on from it, and didn't pursue it.
So, for all I know cocaine is fat soluble and persists for a very long time.
Totally not qualified to speak to the medical stuff.
She told the doctor she had no idea why there was coke in her system. If she's telling the truth (doubt it), ok, fine. But if she lied or is in denial, that's totally on her.
Or, you know, maybe she has been clean for 30 years, doesn't touch the stuff any more, and literally has NO IDEA of why it's in her system.
Reading that I almost got the impression the doctor realized there could be residual cocaine, and that it was likely a false hit.
Next bit of news: Doctors offices use VPN's and Tor to access Google.
Bullcrap. Even if doctors all had the technical sophistication to do this, which I assure you, they don't -- if you can identify the IP address of the VPN (or some of the TOR exit nodes) then you can still determine that 'a' doctor, and possibly 'this group of doctors' is doing searches about people.
When I see this:
When I walked out of the room, a nurse called me over to her computer. There, on MugShots.com, was a younger version of my patient's face, with details about how she had been detained for cocaine possession more than three decades earlier.
I immediately think, "yup, the, the nurses are just googling and finding everything about you, and they're probably doing it with zero anonymity". My impression of the standards of IT and security in the average medical context is that it's barely there (if at all), managed by people who don't know or care, likely woefully out of date and missing security updates, and probably on a network which has been compromised by malware.
Sorry, but the interwebs pretty much guarantee that unless you took some pretty extraordinary measures, determining that a specific doctors office had the mugshots.com up for a patient isn't all that tough, which tells you that patient is associated with that doctor.
I do not believe the average doctor's office has the technical skills, resources, or inclination to be able to do this in a way which would be safe, stay within HIPAA laws, and guarantee you aren't leaking out patient information in the process.
Which means they have no business doing it in the first place, but being doctors, think they know everything and have no idea of the ramifications of this.
... adoctor will fondle and touch and examine your most intimate body parts, yet they shouldn't look at publicly available information? STUPID.
Except when doctors look at this publicly available information, the fact that they looked at it also becomes information which, while not publicly available, is still available to Google and, by extension, the government. Because the search engine knows who did the search (possibly exactly who if you're logged in) and where it came from.
The simple act of the search allows someone to say "this doctor's office looked for this person, and they also looked at this information". You don't think big data can't then determine that "this person has that condition and is being treated by that doctor"?
And then you've violated HIPAA laws and your obligation to patient confidentiality.
Unless you can prove no 3rd party could glean information from you doing that search (and I assure you, the doctors can't), you pretty much have to assume that someone actually could.
Which means the default position here has to be "no, you can't do that". Because it has more potential to cause harm than people realize.
Wouldn't doctors googling their patients essentially violate HIPAA rules?
Because you've now let the fact that you are a doctor treating a specific patient bleed out around the corners, and since Google is keeping track of who you are and what you searched for, they know it too.
Unless you are doing this in such a way that you can guarantee you're not causing patient confidentiality to be breached (which Google sure as hell isn't), I'm of the opinion you've demonstrated a lapse in ethics, and a breech of the law.
And, even if you search in a manner you know was anonymous, if those searches come from something which is identifiable as being the anonymous search of doctors, the content of those searches can still leak information out.
Because when Google see that Dr. Joe Quack has searched for Bob Skippy Smith followed by a quick refresher on the symptoms of herpes.... Google knows (or can infer) that Bob Smith has Herpes.
Doctors are not information theorists, and quite possibly not well educated enough about this technology to be using it in conjunction with their medical practice. Because clearly, if they understood this a little better, they'd realize they've more or less violated their ethics (and possibly the law) by doing this.
Doctors Googling their patients is a terrible idea, and has every possibility of violating the privacy of the patient, as well as the laws meant to protect it.
hahaha. Medicine was largely unchanged during that time. Wasn't until the end of the 19th century before actual science started being applied to medicine, for the most part.
Don't think for one minute the acients people weren't using stuff like science.
Just because the West went through the dark ages and rooted around in the muck for a couple of centuries, there was an awful lot of things people knew before.
There's a reason why Latin is still the language of science. And there's also a reason why several thousand years ago people had some pretty sophisticated cultures.
That the Church made everybody live for a few centuries with little or no advancement doesn't mean it didn't happen before, or elsewhere. But there's plenty of things we are still learning that ancient cultures had that we didn't think they would.
Frankly, so what if someone know where you drove to last year.
Until your wife demands it for divorce proceedings which prove you were at your mistress when you should have been at work.
Or until someone decides that the fact that you were in Little Italy means you might be associated with organized crime.
Or any number of ways in which you don't expect your location to be constantly broadcast to a 3rd party, and be something which comes back to bite you in the ass.
Frankly, I would like the automotive companies to keep it forever but have to get my permission to give it to the government and that the government must serve ME with a warrant to get the data.
You are aware of the Patriot Act, right? The one which says they can walk into a company, demand your data, and it would be illegal for them to tell you about it?
Sorry, but as long as they can use national security laws to get this data, putting it into the hands of companies is no solution.
Sadly, people have known OnStar would have this capability for years.
Because the last thing the Federal government cares about is the privacy of its citizens.
Of course they don't. Because they can demand this information from them and use it themselves.
"Well, we couldn't get a warrant to install a GPS tracker, but since your Escalade had a GPS/OnStar, we'll just ask GM for all of your travel history. Gee, it says here you were in an area which is known to have drug dealers and prostitutes".
Much like the Patriot Act rendered cloud-computing to be a security problem for anybody not in the US but using a US based service, the internet of things will essentially cause all of your information to become the property of a company, and readily accessible to the US government.
I can't possibly put enough layers of tin-foil on to make me feel any better about this stuff. Because we're hurtling towards the dystopian future some of us have been fearing for years.
Only we seem to be voluntarily providing the companies with this stuff in return for shiny baubles.
This is why stuff like OnStar, or the fully connected internet of things is going to be a privacy nightmare.
You can't turn off OnStar and trust they still aren't listening.
When you can't trust that your own property isn't spying on you (which can of course then be subpoena'd by law enforcement), you're pretty much screwed.
It's bad enough everything you do on the internet someone is trying to track -- having your car always telling the company where you are is beyond creepy.
Slashdot Mirror
Except in the case of a rail system, you know you can blame the owner/operator of it.
But, in a world of EULAs and Terms of Service, I think it far more likely the manufacturers will have said "we bear no responsibility for the quality and suitability of our product to be a self driving car".
And, lawyers being lawyers, will move on to the next person they can blame. And that will be the owners or the person who was allegedly "driving".
I won't pay you for a car that wants to operate itself, but the blame falls to me when your product fucks up. I will simply buy a car which doesn't drive itself.
TFA rightly points out that this is a legal gray area -- which means you really don't want to be the test case.
Except, at the heart of this is, has America lived up to their treaty obligations or cared about insults to allies?
If the answer is "no", then finding out the scope of this might be considered something which trumps how America feels about letting him testify.
Except, being on rails provides distinct advantages in terms of things being on auto-pilot.
There's far fewer degrees of freedom in terms of what can happen, because, well, you're on frigging rails.
You need to monitor your speed and your braking, but the turning is enforced by the rails unless you're going way too fast.
Because cars aren't on rails?
Planes are slightly different, because you can bet that the pilot is still ultimately responsible for the aircraft, and if it crashes due to pilot error, he's going to be the one hung out to dry. (Other than that, we mostly just hope/trust that pilots are professional, qualified, and able to do the job at hand)
No way that's gonna work.
There's now way you can expect people to be alert and responsive if they have to be on the ball for that small fraction of the time -- they'll have started reading their paper or plenty of other things.
If I'm responsible for the operation of the vehicle, I'll bloody well drive myself and be engaged for the entire time, and don't need your autonomous car.
I'f I'm not responsible for the operation of the vehicle, I want to be in the back seat in one hell of a good safety cage with no pretense whatsoever that I'm in control.
You can't have the vehicle responsible most of the time, and the ostensible operator responsible whenever that stops working suddenly, it defeats the purpose.
Which, to me, is kind of a fairly fundamental problem with self driving cars. It's all or nothing. And if *all* the cars on the road aren't autonomous, then the autonomous ones are mostly a traffic hazard with no clear liability.
The manufacturer will have an EULA which absolves them from guilt.
It won't be the people who sold it, because they'll also have a contract term which says they are absolved from guilt.
So, it will come down to the owner, who will be entirely dependent on the quality of the product, as delivered by two entities who have already said "not us".
So, if you privately buy an autonomous car, and it crashes, you will likely be on the hook for it. If you merely hire them (as in a Taxi), then I'm sure the people who rent them will also absolve themselves from guilt in some strange way -- likely through arms length 3rd parties who do the actual operation.
This won't be so much "buyer beware" as "everyone else on the roadway beware", because you'll have a vehicle driving around that if it crashes, there's a long line of people who have already made sure their asses are covered.
The lawyers for the companies making and selling these will have covered their asses before it ends up in the hands of anybody else.
If America is going to choose to 'point a gun' at the entire EU, then you can pretty much expect the entire EU to kick the US out of military bases, and generally GTFO of town.
The EU also has their own guns.
Nah, the Americans could tap into it. ;-)
The real tin-foil-hatters would say the US has injected a CG images into the stream to make him say different things. :-P
Of course they do, because part of the mandate is to look out for US commercial interests in general.
The problem is they use the same program to spy for the terrorists, as they do for the economic and political espionage.
Which means, unless the US is willing to carve out JUST the security stuff (which, they won't), every other country more or less has to block this program on the premise that it's just a widespread "spy on everybody, some of them might be security risks, some of it might be political intel, and some can be given to the corporations".
That's kind of the problem from the perspective of the rest of the world -- any form of cooperation with this spying has far broader ramifications than just national security.
Hell, people here routinely defend it, but increasingly you might see other world governments saying they won't allow you to do it any more (in which case, it will be done anyway, just in a more clandestine manner).
Or, if they're going to have him testify, they have diplomats collect him and bring him in on a plain covered by immunity, move him around in diplomatic cars, and house him in diplomatic residences.
Do you *really* think that it is impossible to basically "fuck you" and bring him there safely if there's the political will?
It's not so much that they can't override the law, it's that the courts have determined that they're valid, and therefore lawful (even when unconscionable).
Of course, if you have no idea of what will happen, how the hell can you know you're got "sufficient" protection and a facility which has been "configured for this kind of situation"?
Because, it seems to me, they don't actually know what constitutes "good enough" in this case, at which point saying you have done enough to mitigate is mostly just guess work, isn't it?
And, no, I don't know a damned thing about nuclear reactions in any meaningful sense of the word, so maybe the answer is fairly obvious to someone who does.
Let me guess, you either didn't read the article, or didn't understand it:
He's not taking a single exposure. He's taking a very large amount of small slices over a span of time, and stitching them together into a single image.
He hasn't so much taken a 'snapshot in time' like a traditional camera, he's made images out of snapshots which occurred across time.
Which means he's taking objects going by at a pretty good clip, and combining a whole lot of them into something which looks like a single astounding image.
Some of his images have a time lapse quality to them, because they show things which are both in motion and still, over a time sequence:
If you read the article, you'll find he's done much much more than "Hacking camera software to change how it takes pictures" -- the resulting images look like a still frame, but are composited from a time lapse, and are MUCH more sophisticated than you seem to realize.
Why do people on Slashdot persist in dismissing things they don't really understand? What he's done is taken what look like still images, but are in fact a cross section in time.
That you think all he's done is to hack camera software means you don't have the barest idea of what it is he's actually done.
Funny, my experience with Radio Shack was more in the category of 'drooling tools that can't find the parts' -- they really went downhill for the last bunch of years, until ultimately becoming the "cheap-ass electronics and toys store".
Not just any old tech section, a tech section for non-techies.
And yet, the doctor seems to have determined that it had nothing to do with the current stuff and moved on:
So, depending on the kinds of tests he was doing, he apparently concluded it was a red herring.
I don't know enough about it to know how long it stays in your system ... but I gather from TFS that the doctor decided that the information he had wasn't what he needed and moved on from it, and didn't pursue it.
So, for all I know cocaine is fat soluble and persists for a very long time.
Totally not qualified to speak to the medical stuff.
Or, you know, maybe she has been clean for 30 years, doesn't touch the stuff any more, and literally has NO IDEA of why it's in her system.
Reading that I almost got the impression the doctor realized there could be residual cocaine, and that it was likely a false hit.
Bullcrap. Even if doctors all had the technical sophistication to do this, which I assure you, they don't -- if you can identify the IP address of the VPN (or some of the TOR exit nodes) then you can still determine that 'a' doctor, and possibly 'this group of doctors' is doing searches about people.
When I see this:
I immediately think, "yup, the, the nurses are just googling and finding everything about you, and they're probably doing it with zero anonymity". My impression of the standards of IT and security in the average medical context is that it's barely there (if at all), managed by people who don't know or care, likely woefully out of date and missing security updates, and probably on a network which has been compromised by malware.
Sorry, but the interwebs pretty much guarantee that unless you took some pretty extraordinary measures, determining that a specific doctors office had the mugshots.com up for a patient isn't all that tough, which tells you that patient is associated with that doctor.
I do not believe the average doctor's office has the technical skills, resources, or inclination to be able to do this in a way which would be safe, stay within HIPAA laws, and guarantee you aren't leaking out patient information in the process.
Which means they have no business doing it in the first place, but being doctors, think they know everything and have no idea of the ramifications of this.
Except when doctors look at this publicly available information, the fact that they looked at it also becomes information which, while not publicly available, is still available to Google and, by extension, the government. Because the search engine knows who did the search (possibly exactly who if you're logged in) and where it came from.
The simple act of the search allows someone to say "this doctor's office looked for this person, and they also looked at this information". You don't think big data can't then determine that "this person has that condition and is being treated by that doctor"?
And then you've violated HIPAA laws and your obligation to patient confidentiality.
Unless you can prove no 3rd party could glean information from you doing that search (and I assure you, the doctors can't), you pretty much have to assume that someone actually could.
Which means the default position here has to be "no, you can't do that". Because it has more potential to cause harm than people realize.
Wouldn't doctors googling their patients essentially violate HIPAA rules?
Because you've now let the fact that you are a doctor treating a specific patient bleed out around the corners, and since Google is keeping track of who you are and what you searched for, they know it too.
Unless you are doing this in such a way that you can guarantee you're not causing patient confidentiality to be breached (which Google sure as hell isn't), I'm of the opinion you've demonstrated a lapse in ethics, and a breech of the law.
And, even if you search in a manner you know was anonymous, if those searches come from something which is identifiable as being the anonymous search of doctors, the content of those searches can still leak information out.
Because when Google see that Dr. Joe Quack has searched for Bob Skippy Smith followed by a quick refresher on the symptoms of herpes .... Google knows (or can infer) that Bob Smith has Herpes.
Doctors are not information theorists, and quite possibly not well educated enough about this technology to be using it in conjunction with their medical practice. Because clearly, if they understood this a little better, they'd realize they've more or less violated their ethics (and possibly the law) by doing this.
Doctors Googling their patients is a terrible idea, and has every possibility of violating the privacy of the patient, as well as the laws meant to protect it.
Don't think for one minute the acients people weren't using stuff like science.
Just because the West went through the dark ages and rooted around in the muck for a couple of centuries, there was an awful lot of things people knew before.
There's a reason why Latin is still the language of science. And there's also a reason why several thousand years ago people had some pretty sophisticated cultures.
That the Church made everybody live for a few centuries with little or no advancement doesn't mean it didn't happen before, or elsewhere. But there's plenty of things we are still learning that ancient cultures had that we didn't think they would.
Until your wife demands it for divorce proceedings which prove you were at your mistress when you should have been at work.
Or until someone decides that the fact that you were in Little Italy means you might be associated with organized crime.
Or any number of ways in which you don't expect your location to be constantly broadcast to a 3rd party, and be something which comes back to bite you in the ass.
You are aware of the Patriot Act, right? The one which says they can walk into a company, demand your data, and it would be illegal for them to tell you about it?
Sorry, but as long as they can use national security laws to get this data, putting it into the hands of companies is no solution.
Sadly, people have known OnStar would have this capability for years.
Because it's valuable to them. Because they'd love to have your car recommend a nearby restaurant. Because they can.
Welcome to a world ruled by Terms of Service and End User Licenses, and where corporate greed isn't regulated by privacy laws.
Of course they don't. Because they can demand this information from them and use it themselves.
"Well, we couldn't get a warrant to install a GPS tracker, but since your Escalade had a GPS/OnStar, we'll just ask GM for all of your travel history. Gee, it says here you were in an area which is known to have drug dealers and prostitutes".
Much like the Patriot Act rendered cloud-computing to be a security problem for anybody not in the US but using a US based service, the internet of things will essentially cause all of your information to become the property of a company, and readily accessible to the US government.
I can't possibly put enough layers of tin-foil on to make me feel any better about this stuff. Because we're hurtling towards the dystopian future some of us have been fearing for years.
Only we seem to be voluntarily providing the companies with this stuff in return for shiny baubles.
This is why stuff like OnStar, or the fully connected internet of things is going to be a privacy nightmare.
You can't turn off OnStar and trust they still aren't listening.
When you can't trust that your own property isn't spying on you (which can of course then be subpoena'd by law enforcement), you're pretty much screwed.
It's bad enough everything you do on the internet someone is trying to track -- having your car always telling the company where you are is beyond creepy.