Do you know that for a fact, did you see that stated somewhere? It *looks* like the car didn't have lidar, so if it did, then some serious forensic investigating needs to happen to figure out what failed and why.
WTF happened to LIDAR and sub millisecond braking reactions?
That's exactly the point here. With lidar, and with the reaction times that we know for a fact these cars have, this should never have been an issue. This should be a video that Uber uses to show why its cars are valuable on the roads, because they can avoid a situation that humans wouldn't be aware of until too late. This should be a promo video for Uber, but instead it's going to be a case study in computer science classes about what not to do. This is exactly the situation where autonomous vehicles should show their usefulness and they completely blew it.
That brings up some legal issues also. Sure, maybe a human wouldn't have been able to respond, and maybe she would have died with 50% of the human drivers out there. But the stated capabilities of this car should have allowed it to avoid her, so if that fails then why wouldn't Uber be at fault? It's kind of like drunk driving. A drunk driver might get into a minor accident that any sober driver would have been able to avoid, but the driver is fucked because they were drunk. If the Uber car was technically capable of avoiding this, but didn't, maybe Uber should be liable regardless of what a "normal" human would have done in that situation. The Uber car was drunk.
It should have slowed down and turned to the left if it knew that lane was clear. It should have done that before she even entered the headlights. Excuse the technical term, but something with that car was seriously fucked up. Hopefully there's enough data to identify what that is and make sure it doesn't happen again. In the meantime, vehicles with that software or hardware shouldn't be self-driving until we know what happened and we know that it has been fixed and won't happen again.
While all of that is true, when we think about the advantages of autonomous cars, this is exactly the kind of situation we think they should be able to avoid. Most human drivers might well have hit her, but a vehicle that is covered in sensors and cameras should specifically be able to avoid exactly this kind of situation. Something obviously went very wrong if the car didn't even react while she was in the headlights.
Especially considering the video is way darker than it would have been in real life.
That really is a very dark stretch, there aren't a ton of lights there, and her choice of clothing wouldn't help a human driver see her. It's just coming out from under a bridge overpass. But lidar should make all of that a moot point. She was even moving at nearly 90 degrees to the car. I don't know if it got confused by the bike spokes, or plastic bags, or if a wire fell out right then, but something obviously went wrong. Hopefully they have enough data to know what that is. I'm a little surprised that the police chief said it didn't look like Uber was at fault so quickly, although I don't know what the police would do if that was a person with a dashcam. If that's a human driver I'd say it's almost 50/50 whether she gets hit or the car swerves across the road at the last second once her shoes became visible in the headlights. I don't know how long she was in the headlights for, but even without lidar it should have recognized an object and slowed and turned. It has the reaction time to avoid that accident, but it didn't react at all. I wonder if the "safety driver" did anything to interfere, maybe she was turning it to manual control in between checking her phone for some reason.
Me neither, that's exactly the situation where anyone would expect an autonomous car to perform well. Anything running that software or with that hardware should be off the road until they figure out what went wrong and certify a fix. I doubt I would have been able to avoid that because she's only visible when the headlights hit her shoes (although, for some reason people think Twitter is a video distribution service so the only video I saw had a lot of artifacts, maybe she was more visible than it seemed), but something with as many sensors as those cars should be expected to avoid that situation. The fact that the car took no action while she was in the headlights is worrying also. We know for a fact that the cars (should)have a very low reaction time, so the fact that it didn't react at all needs to be analyzed. It could just be a simple hardware failure right at that time but, again, this is exactly the situation where anyone would expect an autonomous car to perform well.
This is the kind of thing you'd expect from a car running Windows 98, not something bristling with sensors and cameras.
Oh, OK. So, if I call up Tesla, or whoever is developing those, and say "listen, I hear that your goal is to have these half-assed things on the road without anyone inside them to stop them from killing however many people" they'll say "yeah, that's exactly what we're going for." OK, well you've got me convinced.
Think how many people a driverless truck can kill when it fucks up.
Would you say it's more or less than an angry guy with a cache of weapons? Or, for that matter, an angry guy with a several thousand pound truck?
Didn't you pay attention in Trolling 101?
I must have been absent that day.
We have all the evidence we need -- IN THE MORGUE. If that's not enough for you then maybe there's something wrong with YOU.
Ignore everything, everyone quick look at this emotion! Now! Now is the time to legislate! Just like when 9/11 happened, did we wait until rationality took over? Hell no, we didn't! We immediately outlawed privacy! That's the way forward, let's run on pure emotion, baby! One person DIED! A PERSON! Do you know what that MEEEANS! One death is TOO MANY! Human lives are TOO VALUABLE! Unless we're talking about gun ownership, or the opioid crisis, or sending our soldiers to other countries, or diabetes, or capital punishment, or world hunger, or suicide because of a lack of mental health care, or whatever else. No, this one thing, regardless of the fact that this could completely revolutionize transportation all over the world, for some reason the talented and brilliant Rick Schumann has made the decision for us that our tolerance for death in this case is exactly 0. Since we've already exceeded that limit, we should shut down progress on this front forever and ever.
That's all you SDC/driverless car fanboys talk about is how "self driving cars will SAVE LIVES!
That's part of the reason, yeah. And we can get there, too. I don't know why you think we aren't capable of ever doing this one thing, but we can, and I'm glad people who aren't ruled by emotion are actively working on the problem.
Or do you just not give a fuck that someone is dead?
OK, you understand that appeal to emotion is one of the logical fallacies, and your entire argument is based on that right? When your entire argument is based on a logical fallacy, what do you think that says about the strength of your argument?
If you're NOT having any emotional response to a human being dying for no damned good reason then maybe you're a sociopath.
You know, I would agree with that statement. But that statement and what's going on here are not related to each other.
They can try to 'perfect' this 'technology' all they want
Oh gee Rick, do the people actively working towards progress really have your permission to keep working? Wow, that's super.
OFF OF PUBLIC ROADS, AWAY FROM PEOPLE WHO COULD BE KILLED.
Rick, you understand that at some point they're going to need to be on public roads, right? I mean, you understand that's the end goal, right?
SDCs/Driverless Cars have been RUSHED TO MARKET and are not ready to be used on public roads!
As far as general use is concerned? Sure, I agree with you, there's still a lot to learn. That's why we're continuing to work on this problem instead of claiming that the problem has been solved and everyone can transition immediately.
Get them off my roads NOW before more people get killed!
Rick, the problem is more widespread than any of us are willing to admit. You know what the highest cause of death is other than disease? Accidents, Rick. Specifically, car accidents. You know what that means, Rick? It means that we need to BAN ALL VEHICLES FROM ALL ROADS BECAUSE ONE HUMAN DEATH IS TOO MANY!
This is becoming a broken record, but maybe I enjoy pain or something, so let's try this again - cite your evidence from that claim. How do you know she was walking across the road? How do you know she wasn't running, or riding the bike, or tripped and fell off the sidewalk?
It is not clear to me how a car could approach and not once see any evidence that there was a human there that may walk onto the road.
It's not clear to me either, that's why I'd love to see the video and other data so that we can actually figure this stuff out instead of guessing.
If a human saw a bike tire moving from behind a tree, they would slow down because they would come to the conclusion that there is a person there.
That's pretty idealistic. You know very well that a lot of people would not slow down even if they noticed the person.
Why did the self driving car not make the same inference and slow down?
Again, a question in need of evidence to answer, so excuse me if I don't just jump to a conclusion.
they want to let more of these half-assed pieces of crap on the roads, without anyone inside it to have any chance to stop it from killing however many people
Who wants that situation? Specifically, who are you referring to? Do you have any evidence that the situation you're describing is the goal they are actually trying to achieve? No, of course you don't. Your entire replies are nothing more than appeals to emotion. That's part of the problem.
Ban the technology, stop this madness.
There's that good old human emotional knee-jerk reaction. Yeah, let's not try to improve anything at all, let's just ban everything and think of the children while we do it. No need to try to progress, let's just ban everything and cower in fear.
Even the programmers who code these pseudo-intelligences don't know what the hell is going on under the hood
Again, do you have any reason to believe that which you would care to share, or is that just what your gut is telling you?
how do you expect me to provide you with your 'evidence'?
I know you can't, that's what I'm pointing out. If you want to continue making unfounded claims, that's on you. I just want to point out that your claims are unfounded.
The damned thing is not a mind, it is just a machine, it doesn't think, therefore how can it 'know' there's a significance to a living being compared to an inanimate object?
It's pretty trivial to add object recognition to be able to identify something as a person or animal, and if they want to assign various weights to various objects based on whether they're living, or a person, or whatever, and identify which is "more important", they can do that too. That's basic game theory. This is very similar to what humans do.
So you're saying if you're driving and you find yourself in the unenviable situation of having to decide in a split second whether to run over a kid or run into a telephone poll, it's just a coin-flip to you?
No, that's not what I'm saying.
Given the theoretical choice I put to you earlier in this paragraph, it WOULD flip a coin instead of automatically crashing into a telephone pole to avoid hitting a human being.
Again, do you have any evidence that the system puts the same decision weight on impacting a person versus a telephone pole? No, you don't, so excuse me if I don't automatically assume that what you're saying is in any way accurate. Excuse me if I know that this is not a computer in a vacuum, that it was designed and programmed by actual humans.
What if that car is carrying a load of children, and impacting the pole results in a high probability that many of them would be seriously injured or die? What if impacting one person really is the lesser evil?
it needs to make human-level judgements, not just follow some cold, dead algorithm.
You know, human judgment leads to a lot of terrible things happening in the world. Maybe removing emotion from the equation isn't the evil you're making it out to be. Maybe we need more rationality and less emotion when making consequential decisions.
A woman entered a roadway with her bike and was hit by an automated car which did not slow down so we can conclude the car did not see her at all.
Well, there's one mistake. No, we can't draw that conclusion from that evidence. It would be more accurate to say that the vehicle did not have time to identify her and react (more likely), or for some other reason chose not to react (less likely). That's all we can say.
I find it hard to believe that there was not some indication that there was a person there that a human would have noticed.
I don't see what your personal beliefs have to do with the facts of the case, which may or may not be known to us. For example, what specifically was in the field of view of the vehicle is not known to us, so I don't know what your personal beliefs have to do with anything.
How did this car not see any sign of a human that would walk onto the road?
Is this just a rhetorical question? Because none of this can even be addressed without seeing the video and other data from the car. Are you "just asking questions" again? No need to wait for any evidence, right? Your gut is a'feelin', so better jump in and shit out your opinion for everyone to smell.
Thought exercise for you: is it possible to develop an autonomous vehicle that will always, and I mean with 100% certainty, avoid an accident with a suicidal person trying to get the vehicle to impact themselves? If that is not possible, then isn't it also possible for a person to unintentionally enter the street at that exact same time from that exact same location?
She, actually, the Tempe police chief is a woman. But from what I can tell the car could have only stopped in 1 location, and if it's going north on Mill the speed limit should be 45. The assumptions I am making are that the speed limit doesn't change after crossing the Salt River/202, and that the speed limit shown on Google Maps is current. Maybe it changes without a posted sign, or maybe the imagery is not accurate.
But the speed limit isn't really a major factor. It's a minor detail to how the car struck the person.
I don't think it actually was, from what I can tell the speed limit where it was is 45. I don't know why the cop said 35. This comment and my follow-up should explain why I think that. I don't think she was coming off the center median either, I think she was coming off the right sidewalk based on where the vehicle actually stopped. The speed limit on that street shows 45 at the closest sign on the other side of the 202 though.
The article says center median, but it also says 35mph speed limit. Based on the actual physical location of the vehicle in the picture (I know, fucking evidence, it's fucking crazy!), it looks like - in reality (reality is the set of things that are true, not the set of things you feel in your gut should be true) - that the car was in the rightmost lane of a street with a speed limit of 45 getting ready to turn and the woman came off the sidewalk on the right side. But this is just what I see, based on reality, based on the actual evidence that I'm looking at, and the fact that I have actually been there plenty of times.
Everyone knows that... sorry, I forgot who I'm talking to. Most rational adults are aware that eyewitness testimony is notoriously unreliable, so I don't know who said the phrase "center median" to a reporter ("she was going to the center median", maybe?), or if anyone did at all, and why the cop said the speed limit was 35 when the nearest speed limit sign on the other side of the 202 shows 45. But what I see from the actual picture of the actual stopped vehicle next to the actual bike that it hit, this occurred in the right hand turning lane on a street with a speed limit of 45.
I'll bet that if I go down there some night, there will be a group of homeless people hanging out on that bench/table in the grassy area near the tree that obscures the road, and I'll also bet that they can tell me that the woman took off on her bike from that table, went into the road, and got hit immediately. I'm not going to actually assert that all of that is true, because I don't have evidence of that, but from the evidence I do have it looks highly likely. What I do have evidence of is where the car and the bike came to a stop. You, I don't know what hopes and dreams are pushing all of your assumptions that, for some reason, after well over a decade of development, the sensors on an automated car can't detect a person. I don't know why you're assuming that the sensors on a car designed specifically to see people in any condition from a distance are lower quality than the automatic door sensors on a Wal-Mart, but for some reason that's the hill you've chosen to die on. It's a stupid argument, this entire discussion is completely stupid. I'm trying to force evidence down the throat of someone who doesn't give a shit about evidence. Listen, if you want to go on your stupidity hunger strike and refuse to eat anything based on reality or evidence, then I'll let you do that. There's no reason to fight stupidity, it's like playing chess with a pigeon. It's just going to shit all over the board and strut around like it won.
It's hard to win an argument with a smart person. It's damn near impossible to win an argument with a stupid person.
The driver is in a car that is driving itself, she was probably not attentive as a result.
OK, once again, do you have any evidence for anything you're saying, or do you just feel things in your gut and decide to spew them out? Seriously, any evidence at all?
How big is this freaking tree?
Fuck, if only I had linked to a picture of it. If only I had done that, then you could not ask stupid questions with obvious answers. If only...
Why was the car driving faster than it takes a person with a shopping cart loaded with bags to walk from one side of a tree to the other?
Shopping cart? What shopping cart? Are you inventing things again? Did you just completely ignore everything I wrote? Is it necessary for the woman to traverse the entire width of the tree for some reason, have you added that as a requirement? You're having a really hard time keeping this discussion on the rails, you know that?
Ok so that changes the number to 0.000024 accidents per year per licensed driver.
That means absolutely nothing, it's literally a meaningless statistic.
Is automation looking better now?
Yes it is, every day. Every day the systems improve, and eventually we will get the point when private ownership of vehicles is completely unnecessary for the vast majority of the population. There's no reason to stop that progress because we broke an egg. That's a reason to improve the process.
We might be living near the historical peak of private vehicle ownership. The future is in fact automation and shared resources, there's going to be a day when a garage is just another room where the whole wall opens.
Sorry, but we aren't going back, there's no reason to. We aren't where we want to be yet, but that doesn't mean we stop trying to get there.
The thing is so far nobody has made any credible claim anyone broke any laws as related to this! Not the Obama campaign, not the Trump campaign, not even Cambridge Research.
Well, maybe we can take the opportunity to pass some legislation which would make this kind of thing illegal.
Or, we can sit here and yell about political parties. Whatever you prefer.
What's the point of this message? Should we allow everyone to do it, since Obama did it, is that your point? Should we get in a political partisan pissing contest and argue amongst ourselves while nothing gets done about the actual issue? Is that what you're suggesting? If Obama did anything to break the law, great, let's go after him for that. Otherwise, let's focus on individual data privacy while it's actually being talked about and on everyone's minds instead of some stupid political bullshit. Maybe we can take the opportunity to pass some legislation protecting individual data, or are you trying to derail that and get everyone to start slinging mud and shit at each other.
They allowed a researcher access for research purposes and that guy violated the terms of his agreement and sold the data.
From what I hear, that's not true. It sounds like one of the apps that CA developed, some survey app, was used by 270,000 people. Per the terms (and capabilities) of Facebook's Social Graph API, CA then had access to all of the data for every "friend" of those 270,000 people, which got them to the 50 million number. And those people never agreed to the terms of the app or necessarily had anything to do with CA, they're just Facebook's product. It sounds like CA were doing exactly what Facebook allowed at the time, and the damage control is Facebook trying to convince everything that they don't do that now.
Slashdot Mirror
The car obviously had no lidar.
Do you know that for a fact, did you see that stated somewhere? It *looks* like the car didn't have lidar, so if it did, then some serious forensic investigating needs to happen to figure out what failed and why.
WTF happened to LIDAR and sub millisecond braking reactions?
That's exactly the point here. With lidar, and with the reaction times that we know for a fact these cars have, this should never have been an issue. This should be a video that Uber uses to show why its cars are valuable on the roads, because they can avoid a situation that humans wouldn't be aware of until too late. This should be a promo video for Uber, but instead it's going to be a case study in computer science classes about what not to do. This is exactly the situation where autonomous vehicles should show their usefulness and they completely blew it.
That brings up some legal issues also. Sure, maybe a human wouldn't have been able to respond, and maybe she would have died with 50% of the human drivers out there. But the stated capabilities of this car should have allowed it to avoid her, so if that fails then why wouldn't Uber be at fault? It's kind of like drunk driving. A drunk driver might get into a minor accident that any sober driver would have been able to avoid, but the driver is fucked because they were drunk. If the Uber car was technically capable of avoiding this, but didn't, maybe Uber should be liable regardless of what a "normal" human would have done in that situation. The Uber car was drunk.
Haha, was that a serious reply? I can't tell any more.
Yeah, good job on her "saving her night vision." That worked out really well. You know, maybe the Uber car was "saving its brake pads."
The video appears to be deceiving. It is almost like it was purposely dimmed before being released.
That really is a dark stretch of road, but the fact that the thing has lidar should put light out of the question completely.
It should have slowed down and turned to the left if it knew that lane was clear. It should have done that before she even entered the headlights. Excuse the technical term, but something with that car was seriously fucked up. Hopefully there's enough data to identify what that is and make sure it doesn't happen again. In the meantime, vehicles with that software or hardware shouldn't be self-driving until we know what happened and we know that it has been fixed and won't happen again.
While all of that is true, when we think about the advantages of autonomous cars, this is exactly the kind of situation we think they should be able to avoid. Most human drivers might well have hit her, but a vehicle that is covered in sensors and cameras should specifically be able to avoid exactly this kind of situation. Something obviously went very wrong if the car didn't even react while she was in the headlights.
Especially considering the video is way darker than it would have been in real life.
That really is a very dark stretch, there aren't a ton of lights there, and her choice of clothing wouldn't help a human driver see her. It's just coming out from under a bridge overpass. But lidar should make all of that a moot point. She was even moving at nearly 90 degrees to the car. I don't know if it got confused by the bike spokes, or plastic bags, or if a wire fell out right then, but something obviously went wrong. Hopefully they have enough data to know what that is. I'm a little surprised that the police chief said it didn't look like Uber was at fault so quickly, although I don't know what the police would do if that was a person with a dashcam. If that's a human driver I'd say it's almost 50/50 whether she gets hit or the car swerves across the road at the last second once her shoes became visible in the headlights. I don't know how long she was in the headlights for, but even without lidar it should have recognized an object and slowed and turned. It has the reaction time to avoid that accident, but it didn't react at all. I wonder if the "safety driver" did anything to interfere, maybe she was turning it to manual control in between checking her phone for some reason.
Me neither, that's exactly the situation where anyone would expect an autonomous car to perform well. Anything running that software or with that hardware should be off the road until they figure out what went wrong and certify a fix. I doubt I would have been able to avoid that because she's only visible when the headlights hit her shoes (although, for some reason people think Twitter is a video distribution service so the only video I saw had a lot of artifacts, maybe she was more visible than it seemed), but something with as many sensors as those cars should be expected to avoid that situation. The fact that the car took no action while she was in the headlights is worrying also. We know for a fact that the cars (should)have a very low reaction time, so the fact that it didn't react at all needs to be analyzed. It could just be a simple hardware failure right at that time but, again, this is exactly the situation where anyone would expect an autonomous car to perform well.
This is the kind of thing you'd expect from a car running Windows 98, not something bristling with sensors and cameras.
Well, in that case let's skip the yelling about parties and just act like it never happened. Let's just jump straight to the end.
Go read about driverless trucks on the highways.
Oh, OK. So, if I call up Tesla, or whoever is developing those, and say "listen, I hear that your goal is to have these half-assed things on the road without anyone inside them to stop them from killing however many people" they'll say "yeah, that's exactly what we're going for." OK, well you've got me convinced.
Think how many people a driverless truck can kill when it fucks up.
Would you say it's more or less than an angry guy with a cache of weapons? Or, for that matter, an angry guy with a several thousand pound truck?
Didn't you pay attention in Trolling 101?
I must have been absent that day.
We have all the evidence we need -- IN THE MORGUE. If that's not enough for you then maybe there's something wrong with YOU.
Ignore everything, everyone quick look at this emotion! Now! Now is the time to legislate! Just like when 9/11 happened, did we wait until rationality took over? Hell no, we didn't! We immediately outlawed privacy! That's the way forward, let's run on pure emotion, baby! One person DIED! A PERSON! Do you know what that MEEEANS! One death is TOO MANY! Human lives are TOO VALUABLE! Unless we're talking about gun ownership, or the opioid crisis, or sending our soldiers to other countries, or diabetes, or capital punishment, or world hunger, or suicide because of a lack of mental health care, or whatever else. No, this one thing, regardless of the fact that this could completely revolutionize transportation all over the world, for some reason the talented and brilliant Rick Schumann has made the decision for us that our tolerance for death in this case is exactly 0. Since we've already exceeded that limit, we should shut down progress on this front forever and ever.
That's all you SDC/driverless car fanboys talk about is how "self driving cars will SAVE LIVES!
That's part of the reason, yeah. And we can get there, too. I don't know why you think we aren't capable of ever doing this one thing, but we can, and I'm glad people who aren't ruled by emotion are actively working on the problem.
Or do you just not give a fuck that someone is dead?
OK, you understand that appeal to emotion is one of the logical fallacies, and your entire argument is based on that right? When your entire argument is based on a logical fallacy, what do you think that says about the strength of your argument?
If you're NOT having any emotional response to a human being dying for no damned good reason then maybe you're a sociopath.
You know, I would agree with that statement. But that statement and what's going on here are not related to each other.
They can try to 'perfect' this 'technology' all they want
Oh gee Rick, do the people actively working towards progress really have your permission to keep working? Wow, that's super.
OFF OF PUBLIC ROADS, AWAY FROM PEOPLE WHO COULD BE KILLED.
Rick, you understand that at some point they're going to need to be on public roads, right? I mean, you understand that's the end goal, right?
SDCs/Driverless Cars have been RUSHED TO MARKET and are not ready to be used on public roads!
As far as general use is concerned? Sure, I agree with you, there's still a lot to learn. That's why we're continuing to work on this problem instead of claiming that the problem has been solved and everyone can transition immediately.
Get them off my roads NOW before more people get killed!
Rick, the problem is more widespread than any of us are willing to admit. You know what the highest cause of death is other than disease? Accidents, Rick. Specifically, car accidents. You know what that means, Rick? It means that we need to BAN ALL VEHICLES FROM ALL ROADS BECAUSE ONE HUMAN DEATH IS TOO MANY!
In other news, Hitler said he "made mistakes" in protecting Jews.
That Godwin was too easy.
The woman was walking a bike across the road.
This is becoming a broken record, but maybe I enjoy pain or something, so let's try this again - cite your evidence from that claim. How do you know she was walking across the road? How do you know she wasn't running, or riding the bike, or tripped and fell off the sidewalk?
It is not clear to me how a car could approach and not once see any evidence that there was a human there that may walk onto the road.
It's not clear to me either, that's why I'd love to see the video and other data so that we can actually figure this stuff out instead of guessing.
If a human saw a bike tire moving from behind a tree, they would slow down because they would come to the conclusion that there is a person there.
That's pretty idealistic. You know very well that a lot of people would not slow down even if they noticed the person.
Why did the self driving car not make the same inference and slow down?
Again, a question in need of evidence to answer, so excuse me if I don't just jump to a conclusion.
they want to let more of these half-assed pieces of crap on the roads, without anyone inside it to have any chance to stop it from killing however many people
Who wants that situation? Specifically, who are you referring to? Do you have any evidence that the situation you're describing is the goal they are actually trying to achieve? No, of course you don't. Your entire replies are nothing more than appeals to emotion. That's part of the problem.
Ban the technology, stop this madness.
There's that good old human emotional knee-jerk reaction. Yeah, let's not try to improve anything at all, let's just ban everything and think of the children while we do it. No need to try to progress, let's just ban everything and cower in fear.
Even the programmers who code these pseudo-intelligences don't know what the hell is going on under the hood
Again, do you have any reason to believe that which you would care to share, or is that just what your gut is telling you?
how do you expect me to provide you with your 'evidence'?
I know you can't, that's what I'm pointing out. If you want to continue making unfounded claims, that's on you. I just want to point out that your claims are unfounded.
The damned thing is not a mind, it is just a machine, it doesn't think, therefore how can it 'know' there's a significance to a living being compared to an inanimate object?
It's pretty trivial to add object recognition to be able to identify something as a person or animal, and if they want to assign various weights to various objects based on whether they're living, or a person, or whatever, and identify which is "more important", they can do that too. That's basic game theory. This is very similar to what humans do.
So you're saying if you're driving and you find yourself in the unenviable situation of having to decide in a split second whether to run over a kid or run into a telephone poll, it's just a coin-flip to you?
No, that's not what I'm saying.
Given the theoretical choice I put to you earlier in this paragraph, it WOULD flip a coin instead of automatically crashing into a telephone pole to avoid hitting a human being.
Again, do you have any evidence that the system puts the same decision weight on impacting a person versus a telephone pole? No, you don't, so excuse me if I don't automatically assume that what you're saying is in any way accurate. Excuse me if I know that this is not a computer in a vacuum, that it was designed and programmed by actual humans.
What if that car is carrying a load of children, and impacting the pole results in a high probability that many of them would be seriously injured or die? What if impacting one person really is the lesser evil?
it needs to make human-level judgements, not just follow some cold, dead algorithm.
You know, human judgment leads to a lot of terrible things happening in the world. Maybe removing emotion from the equation isn't the evil you're making it out to be. Maybe we need more rationality and less emotion when making consequential decisions.
A woman entered a roadway with her bike and was hit by an automated car which did not slow down so we can conclude the car did not see her at all.
Well, there's one mistake. No, we can't draw that conclusion from that evidence. It would be more accurate to say that the vehicle did not have time to identify her and react (more likely), or for some other reason chose not to react (less likely). That's all we can say.
I find it hard to believe that there was not some indication that there was a person there that a human would have noticed.
I don't see what your personal beliefs have to do with the facts of the case, which may or may not be known to us. For example, what specifically was in the field of view of the vehicle is not known to us, so I don't know what your personal beliefs have to do with anything.
How did this car not see any sign of a human that would walk onto the road?
Is this just a rhetorical question? Because none of this can even be addressed without seeing the video and other data from the car. Are you "just asking questions" again? No need to wait for any evidence, right? Your gut is a'feelin', so better jump in and shit out your opinion for everyone to smell.
And instead just make wild assumptions with zero evidence, like you've been doing? I'm not sure that's any better.
Thought exercise for you: is it possible to develop an autonomous vehicle that will always, and I mean with 100% certainty, avoid an accident with a suicidal person trying to get the vehicle to impact themselves? If that is not possible, then isn't it also possible for a person to unintentionally enter the street at that exact same time from that exact same location?
Haha, "thought exercise", who am I kidding?
She, actually, the Tempe police chief is a woman. But from what I can tell the car could have only stopped in 1 location, and if it's going north on Mill the speed limit should be 45. The assumptions I am making are that the speed limit doesn't change after crossing the Salt River/202, and that the speed limit shown on Google Maps is current. Maybe it changes without a posted sign, or maybe the imagery is not accurate.
But the speed limit isn't really a major factor. It's a minor detail to how the car struck the person.
I don't think it actually was, from what I can tell the speed limit where it was is 45. I don't know why the cop said 35. This comment and my follow-up should explain why I think that. I don't think she was coming off the center median either, I think she was coming off the right sidewalk based on where the vehicle actually stopped. The speed limit on that street shows 45 at the closest sign on the other side of the 202 though.
The article says center median, but it also says 35mph speed limit. Based on the actual physical location of the vehicle in the picture (I know, fucking evidence, it's fucking crazy!), it looks like - in reality (reality is the set of things that are true, not the set of things you feel in your gut should be true) - that the car was in the rightmost lane of a street with a speed limit of 45 getting ready to turn and the woman came off the sidewalk on the right side. But this is just what I see, based on reality, based on the actual evidence that I'm looking at, and the fact that I have actually been there plenty of times.
Everyone knows that... sorry, I forgot who I'm talking to. Most rational adults are aware that eyewitness testimony is notoriously unreliable, so I don't know who said the phrase "center median" to a reporter ("she was going to the center median", maybe?), or if anyone did at all, and why the cop said the speed limit was 35 when the nearest speed limit sign on the other side of the 202 shows 45. But what I see from the actual picture of the actual stopped vehicle next to the actual bike that it hit, this occurred in the right hand turning lane on a street with a speed limit of 45.
I'll bet that if I go down there some night, there will be a group of homeless people hanging out on that bench/table in the grassy area near the tree that obscures the road, and I'll also bet that they can tell me that the woman took off on her bike from that table, went into the road, and got hit immediately. I'm not going to actually assert that all of that is true, because I don't have evidence of that, but from the evidence I do have it looks highly likely. What I do have evidence of is where the car and the bike came to a stop. You, I don't know what hopes and dreams are pushing all of your assumptions that, for some reason, after well over a decade of development, the sensors on an automated car can't detect a person. I don't know why you're assuming that the sensors on a car designed specifically to see people in any condition from a distance are lower quality than the automatic door sensors on a Wal-Mart, but for some reason that's the hill you've chosen to die on. It's a stupid argument, this entire discussion is completely stupid. I'm trying to force evidence down the throat of someone who doesn't give a shit about evidence. Listen, if you want to go on your stupidity hunger strike and refuse to eat anything based on reality or evidence, then I'll let you do that. There's no reason to fight stupidity, it's like playing chess with a pigeon. It's just going to shit all over the board and strut around like it won.
It's hard to win an argument with a smart person. It's damn near impossible to win an argument with a stupid person.
The driver is in a car that is driving itself, she was probably not attentive as a result.
OK, once again, do you have any evidence for anything you're saying, or do you just feel things in your gut and decide to spew them out? Seriously, any evidence at all?
How big is this freaking tree?
Fuck, if only I had linked to a picture of it. If only I had done that, then you could not ask stupid questions with obvious answers. If only...
Why was the car driving faster than it takes a person with a shopping cart loaded with bags to walk from one side of a tree to the other?
Shopping cart? What shopping cart? Are you inventing things again? Did you just completely ignore everything I wrote? Is it necessary for the woman to traverse the entire width of the tree for some reason, have you added that as a requirement? You're having a really hard time keeping this discussion on the rails, you know that?
Ok so that changes the number to 0.000024 accidents per year per licensed driver.
That means absolutely nothing, it's literally a meaningless statistic.
Is automation looking better now?
Yes it is, every day. Every day the systems improve, and eventually we will get the point when private ownership of vehicles is completely unnecessary for the vast majority of the population. There's no reason to stop that progress because we broke an egg. That's a reason to improve the process.
We might be living near the historical peak of private vehicle ownership. The future is in fact automation and shared resources, there's going to be a day when a garage is just another room where the whole wall opens.
Sorry, but we aren't going back, there's no reason to. We aren't where we want to be yet, but that doesn't mean we stop trying to get there.
The thing is so far nobody has made any credible claim anyone broke any laws as related to this! Not the Obama campaign, not the Trump campaign, not even Cambridge Research.
Well, maybe we can take the opportunity to pass some legislation which would make this kind of thing illegal.
Or, we can sit here and yell about political parties. Whatever you prefer.
What's the point of this message? Should we allow everyone to do it, since Obama did it, is that your point? Should we get in a political partisan pissing contest and argue amongst ourselves while nothing gets done about the actual issue? Is that what you're suggesting? If Obama did anything to break the law, great, let's go after him for that. Otherwise, let's focus on individual data privacy while it's actually being talked about and on everyone's minds instead of some stupid political bullshit. Maybe we can take the opportunity to pass some legislation protecting individual data, or are you trying to derail that and get everyone to start slinging mud and shit at each other.
They allowed a researcher access for research purposes and that guy violated the terms of his agreement and sold the data.
From what I hear, that's not true. It sounds like one of the apps that CA developed, some survey app, was used by 270,000 people. Per the terms (and capabilities) of Facebook's Social Graph API, CA then had access to all of the data for every "friend" of those 270,000 people, which got them to the 50 million number. And those people never agreed to the terms of the app or necessarily had anything to do with CA, they're just Facebook's product. It sounds like CA were doing exactly what Facebook allowed at the time, and the damage control is Facebook trying to convince everything that they don't do that now.