Here's a statement of the obvious: The opinions expressed here are those of the participants, not those of the Mutual Fund Observer. We cannot vouch for the accuracy or appropriateness of any of it, though we do encourage civility and good humor.
"A lot of legal questions that should have been answered by now. Despite the long time push toward self-driving cars by industry and government, not much has been delivered in the way of determining responsibilities."
Jeeze, Maurice... you sound just like Elizabeth Warren... "determining responsibilities"?
But surely we are talking more useless costly anti-growth liberal deep-state big-government regulation here. Can't have that!!! And let's keep all those damned lawyers out of this too!! Just send the whole thing to arbitration (with Uber getting to choose the arbitrators, of course).
This was bound to happen, but it's worth noting that human-driven cars kill 30,000 people a year. There are some interesting questions regarding self-driving cars, but I don't think safety is the biggest one. I'm fairly certain they will be much safer than human-driven cars. Some of the more important questions to me are: 1. What happens to labor replaced by self-driving cars? 2. Who is legally responsible when an accident occurs? 3. How do self-driving cars solve the ethical trolley problem? 4. What happens when human beings lose another skill--driving--as tech takes over? See film Wall-e to understand this concern. 5. How dangerous is AI overall to humanity?
Latest: Tempe, AZ police chief says Uber vehicle apparently "innocent". It appears that a "bag-lady" pushing a bicycle stepped directly into the oncoming traffic lane from the center divider in a poorly lit area away from the actual pedestrian crossing.
“it’s very clear it would have been difficult to avoid this collision in any kind of mode (autonomous or human-driven) based on how she came from the shadows right into the roadway”. The Uber vehicle was said to have been doing 38 in a 35.
Not to scare everyone that commercial planes typically land on automatic mode. Unless a difficult situation arises as one the Hudson River that a skillful pilot can land it on water safely. Perhaps AI still requires further refinement before they are deployed.
There is video being examined. NTSB is onsite. If she suddenly stepped in front of the car, not much to do. Did you see the photo of the front of the autonomous car? I will let you search for it.
No one said 'expendable'; that was you. Way to work in the presidential uncle! You are a master.
With respect to the area lighting, the police chief is quoted as saying “It is dangerous to cross roadways in the evening hour when well-illuminated, managed crosswalks are available”, which would seem to suggest, by comparison, that the area of the accident was not particularly well-lit.
I'll stipulate that "bag ladies" are certainly much less "expendable" than the current presidential administration.
Re "problems with lawyers": it occurs to me that commercial interests and management are typically the main sources of requirements to use arbitration- not customers or employees.
Keep in mind that next to our kids (then wife...then dog)...our cell phone and autonomous car come in a close 4th or 5th.
All of the Tech that will be necessary to get the autonomous car as ubiquitous as the cell phone needs to be evaluated for their merits with safety being the first hurdle to overcome. We all have gotten over cellphone brain cancer and exploding battery fires right?
Q?: Did the Autonomous car realize that it had just hit a pedestrian and stop, or did the backup human intervene after the crash?
"You expect your computer to paint a page in a second or two. Why can a computer see a person crossing the road, even outside a designated crosswalk, and react in the same amount of time as your PC."
@Maurice- I'm assuming that you meant to say "why can't a computer see", and at first glance that seems to be reasonable. But then consider that at 35 mph a vehicle goes over 50 feet in one second, so the relative distances between the pedestrian and the vehicle become very important. Even with a computer resolution time of 1/2 a second, if you add the actual braking time required, there's still going to be a problem at close distances.
The SF Chronicle article reports that: "From viewing the videos, “it’s very clear it would have been difficult to avoid this collision in any kind of mode (autonomous or human-driven) based on how she came from the shadows right into the roadway,” Moir (the police chief) said". That seems to suggest that the cameras did register the situation, but that the computer was unable to respond in time.
I'm glad that the federal people are getting involved. (That would be some of those those overstaffed, overpaid, totally useless members of the "deep state" that you frequently complain about, right?)
@bee- Wondered that myself. Hard to tell from the info that I've seen so far. The SF Chronicle article reported that “The driver said it was like a flash, the person walked out in front of them,” said Sylvia Moir, police chief in Tempe, Ariz., the location for the first pedestrian fatality involving a self-driving car. “His first alert to the collision was the sound of the collision" which really doesn't answer that point.
I'm going to guess that the car would have to be smart enough to stop in the event of a collision, though.
@Maurice, I cannot tell how your screed has to do with me or anything I wrote. Stepping in front of a car going 38mph is not jaywalking and is going to result in a hit regardless. IF that's what happened.
Visual detection systems are not remotely like 'computer screen drawing', but you know that by referring to cameras. The systems are always multiple:
'Definitive ruling'? Huh? Not how it works. Not how any of this works. But you know that too.
I think lots of things weirdly cloud your responses and your thinking here. Politics is the least of it.
And actually, if a relative of mine, or indeed anyone, stepped right in front of a moving car, stating that it was not responsible would for sure be the third or fourth thing I would say. Unavoidable (dictionary sense) accidents happen all the time.
@davidrmoran- If you have the good sense not to investigate the reader commentary on most WSJ articles you may not be tuned in to an interesting social perspective. There is a fairly large segment of the WSJ readership which is predictably and permanently outraged at quite a number of things. If you want "weird", there's the place to be.
To list a few of the favorites, we have- • planetary warming (not proven, not happening, but if it is, somebody else is doing it) • rapidly melting icecaps (those things have always been melting or something) • renewable electric power (a socialist governmental plot) • autonomous vehicles ( it may be happening but it won't be good) • electric vehicles (yet another tree-hugging impractical boondoggle) • health care (I've got mine, let the other guy take care of himself) • anything and everything that President Obama ever said, accomplished, or tried to accomplish • the judicial system, especially when it rules against the current occupant of the White House
With respect to autonomous vehicles, it's been interesting to watch the grudgingly forced transition of this group over the past few years, from "it will never happen" to "well, it may be happening but it won't be good".
A common theme running through their commentary is more or less "there's no need to change anything... everything was working just fine before all of this new stuff got here". ("New stuff" in this instance seems roughly to be anything after Bush II.)
If you evaluate some of the commentary here on MFO you can pick up echoes of that perspective: those folks neither know nor want to know "how any of this works" because it shouldn't be happening.
As a personal note, I'm not totally sold on a lot of the current high tech stuff myself. I get along just fine without Google, Facebook, Twitter, etc. I do regret and worry about the ever-decreasing right to privacy, and the implications for big-brother government which are sure to follow. And I must confess that I'm very glad that I'm too old to have to worry about the future hailing of an autonomous pick-up truck, rather than just jumping into the one that I own.
Following are excerpts from a current WSJ article* regarding this event.
"A newly released video of the fatal accident involving a pedestrian and an Uber Technologies Inc. self-driving car appears to show the vehicle heading straight into a woman walking her bike across the road without slowing down or swerving to avoid her."
"The video... from cameras inside and outside the Uber vehicle, appears to also show the human safety operator at the wheel was looking down for approximately five seconds until the moment of impact. This person’s role is to take over controls to help prevent accidents or erratic driving from the robot vehicle."
"There is no indication that the Uber car—equipped with laser sensors, radar and camera sensors to detect its surroundings—identified the pedestrian ahead of the collision. If that is the case, this accident could damage the public’s perception of robot vehicles, jolt regulators to take action and threaten the progress of dozens of auto makers and tech giants investing billions of dollars in autonomous vehicle technology."
"The vehicle was traveling at about 40 miles per hour when it struck Ms. Herzberg as she crossed the road with her bicycle carrying bags, according to Tempe police and a review of the video."
* The WSJ link is, unfortunately, available only to subscribers.
Further on this article- for a pleasant change there are a number or serious and well-written comments from the readers of this WSJ article. One reader in particular, David Stanwick, makes some very interesting observations, and here are a few excerpts:
"Now that I've re-watched it about 20 times I think I see the problem. The LIDAR is probably looking further down the road and saw a clear path ahead when the woman entered the Uber lane of travel. It saw a clear road ahead, THEN the bike entered the car's path. This is clearly a defect in the Uber design. There is no reason a large target like this bike + person + belongings could not be detected by LIDAR except that it was outside the field of view."
"Further clarification: in the video the woman suddenly appears quite clearly in the car's headlights. Obviously I wasn't there, but knowing a little bit about the sensitivity of digital cameras and their auto exposure settings, it's very likely that the woman and bike would be faintly visible to the human driver well in time to prevent an accident. Cameras are very sensitive, but the car has powerful headlights, so the camera sensitivity is turned way down. We can tell. That's why we see nothing until we are almost to the point of impact. If that camera sensitivity was turned way up, we would have seen the woman in the video much earlier and near the point of impact we would just see a large white (overexposed) image."
@Maurice: You asked some interesting questions regarding this accident, and you may appreciate the perspective of Mr. Stanwick. I know that I did.
yes, alas, yes, driver video shown tonight shows clear hit of person in road, not stepping out suddenly at all, open and shut, something lidar seriously gone wrong
human driver also looking down and then suddenly aghast
aghast part not shown here maybe, it appears, though I saw it tonight on the news
As a retired radio/electronics tech with many long years of experience my observation is that the existence of a complex electronics system that is guaranteed 100% failure-proof is a myth, right up there with the belief that Facebook or any other "social" engagement app is designed primarily to benefit users.
@Maurice, And you do not know what 'recant' means. Good grief. Amazing how kneejerk adversarial you are, unfailingly. How goes it for you in the workplace, I wonder. Good that you call for gov involvement even though of course it sucks totally.
• Uber Technologies Inc. doesn’t plan to renew its permit to test autonomous vehicles in California, according to a letter from the state’s motor-vehicle department.
"The company will let its testing permit lapse at month’s end, rather than resubmit for renewal and face potential scrutiny amid an investigation into the circumstances of the accident."
“We proactively suspended our self-driving operations, including in California, immediately following the Tempe incident,” said an Uber spokeswoman in a statement. “Given this, we decided to not reapply for a California DMV permit with the understanding that our self-driving vehicles would not operate on public roads in the immediate future.” Uber could seek to renew the permit at a later date."
Additional Articles: • Nvidia Halts Test of Self-Driving Tech Following Uber Crash • Arizona Governor Suspends Uber’s Self-Driving Cars From Roads • Arizonans Rethink Self-Driving Experiment After Uber Crash
Comments
Regards,
Ted
https://www.bloomberg.com/news/articles/2018-03-19/uber-autonomous-car-involved-in-fatal-crash-in-arizona
Jeeze, Maurice... you sound just like Elizabeth Warren... "determining responsibilities"?
But surely we are talking more useless costly anti-growth liberal deep-state big-government regulation here. Can't have that!!! And let's keep all those damned lawyers out of this too!! Just send the whole thing to arbitration (with Uber getting to choose the arbitrators, of course).
1. What happens to labor replaced by self-driving cars?
2. Who is legally responsible when an accident occurs?
3. How do self-driving cars solve the ethical trolley problem?
4. What happens when human beings lose another skill--driving--as tech takes over? See film Wall-e to understand this concern.
5. How dangerous is AI overall to humanity?
“it’s very clear it would have been difficult to avoid this collision in any kind of mode (autonomous or human-driven) based on how she came from the shadows right into the roadway”. The Uber vehicle was said to have been doing 38 in a 35.
https://www.cnet.com/roadshow/news/self-driving-uber-crash-pedestrian-death-not-at-fault/
There is video being examined. NTSB is onsite. If she suddenly stepped in front of the car, not much to do. Did you see the photo of the front of the autonomous car? I will let you search for it.
No one said 'expendable'; that was you. Way to work in the presidential uncle! You are a master.
With respect to the area lighting, the police chief is quoted as saying “It is dangerous to cross roadways in the evening hour when well-illuminated, managed crosswalks are available”, which would seem to suggest, by comparison, that the area of the accident was not particularly well-lit.
I'll stipulate that "bag ladies" are certainly much less "expendable" than the current presidential administration.
Re "problems with lawyers": it occurs to me that commercial interests and management are typically the main sources of requirements to use arbitration- not customers or employees.
All of the Tech that will be necessary to get the autonomous car as ubiquitous as the cell phone needs to be evaluated for their merits with safety being the first hurdle to overcome. We all have gotten over cellphone brain cancer and exploding battery fires right?
Q?: Did the Autonomous car realize that it had just hit a pedestrian and stop, or did the backup human intervene after the crash?
@Maurice- I'm assuming that you meant to say "why can't a computer see", and at first glance that seems to be reasonable. But then consider that at 35 mph a vehicle goes over 50 feet in one second, so the relative distances between the pedestrian and the vehicle become very important. Even with a computer resolution time of 1/2 a second, if you add the actual braking time required, there's still going to be a problem at close distances.
The SF Chronicle article reports that: "From viewing the videos, “it’s very clear it would have been difficult to avoid this collision in any kind of mode (autonomous or human-driven) based on how she came from the shadows right into the roadway,” Moir (the police chief) said". That seems to suggest that the cameras did register the situation, but that the computer was unable to respond in time.
I'm glad that the federal people are getting involved. (That would be some of those those overstaffed, overpaid, totally useless members of the "deep state" that you frequently complain about, right?)
I'm going to guess that the car would have to be smart enough to stop in the event of a collision, though.
Stepping in front of a car going 38mph is not jaywalking and is going to result in a hit regardless. IF that's what happened.
Visual detection systems are not remotely like 'computer screen drawing', but you know that by referring to cameras. The systems are always multiple:
https://www.sensorsmag.com/components/three-sensor-types-drive-autonomous-vehicles
'Definitive ruling'? Huh? Not how it works. Not how any of this works. But you know that too.
I think lots of things weirdly cloud your responses and your thinking here. Politics is the least of it.
And actually, if a relative of mine, or indeed anyone, stepped right in front of a moving car, stating that it was not responsible would for sure be the third or fourth thing I would say. Unavoidable (dictionary sense) accidents happen all the time.
To list a few of the favorites, we have-
• planetary warming (not proven, not happening, but if it is, somebody else is doing it)
• rapidly melting icecaps (those things have always been melting or something)
• renewable electric power (a socialist governmental plot)
• autonomous vehicles ( it may be happening but it won't be good)
• electric vehicles (yet another tree-hugging impractical boondoggle)
• health care (I've got mine, let the other guy take care of himself)
• anything and everything that President Obama ever said, accomplished, or tried to accomplish
• the judicial system, especially when it rules against the current occupant of the White House
With respect to autonomous vehicles, it's been interesting to watch the grudgingly forced transition of this group over the past few years, from "it will never happen" to "well, it may be happening but it won't be good".
A common theme running through their commentary is more or less "there's no need to change anything... everything was working just fine before all of this new stuff got here". ("New stuff" in this instance seems roughly to be anything after Bush II.)
If you evaluate some of the commentary here on MFO you can pick up echoes of that perspective: those folks neither know nor want to know "how any of this works" because it shouldn't be happening.
As a personal note, I'm not totally sold on a lot of the current high tech stuff myself. I get along just fine without Google, Facebook, Twitter, etc. I do regret and worry about the ever-decreasing right to privacy, and the implications for big-brother government which are sure to follow. And I must confess that I'm very glad that I'm too old to have to worry about the future hailing of an autonomous pick-up truck, rather than just jumping into the one that I own.
it is just awful, stupefying, and for me unbearable after a point; makes you lose your faith, until you recall that it is a self-selected group (duh)
I read worse, too, sometimes
"A newly released video of the fatal accident involving a pedestrian and an Uber Technologies Inc. self-driving car appears to show the vehicle heading straight into a woman walking her bike across the road without slowing down or swerving to avoid her."
"The video... from cameras inside and outside the Uber vehicle, appears to also show the human safety operator at the wheel was looking down for approximately five seconds until the moment of impact. This person’s role is to take over controls to help prevent accidents or erratic driving from the robot vehicle."
"There is no indication that the Uber car—equipped with laser sensors, radar and camera sensors to detect its surroundings—identified the pedestrian ahead of the collision. If that is the case, this accident could damage the public’s perception of robot vehicles, jolt regulators to take action and threaten the progress of dozens of auto makers and tech giants investing billions of dollars in autonomous vehicle technology."
"The vehicle was traveling at about 40 miles per hour when it struck Ms. Herzberg as she crossed the road with her bicycle carrying bags, according to Tempe police and a review of the video."
* The WSJ link is, unfortunately, available only to subscribers.
"Now that I've re-watched it about 20 times I think I see the problem. The LIDAR is probably looking further down the road and saw a clear path ahead when the woman entered the Uber lane of travel. It saw a clear road ahead, THEN the bike entered the car's path. This is clearly a defect in the Uber design. There is no reason a large target like this bike + person + belongings could not be detected by LIDAR except that it was outside the field of view."
"Further clarification: in the video the woman suddenly appears quite clearly in the car's headlights. Obviously I wasn't there, but knowing a little bit about the sensitivity of digital cameras and their auto exposure settings, it's very likely that the woman and bike would be faintly visible to the human driver well in time to prevent an accident. Cameras are very sensitive, but the car has powerful headlights, so the camera sensitivity is turned way down. We can tell. That's why we see nothing until we are almost to the point of impact. If that camera sensitivity was turned way up, we would have seen the woman in the video much earlier and near the point of impact we would just see a large white (overexposed) image."
@Maurice: You asked some interesting questions regarding this accident, and you may appreciate the perspective of Mr. Stanwick. I know that I did.
human driver also looking down and then suddenly aghast
aghast part not shown here maybe, it appears, though I saw it tonight on the news
https://www.washingtonpost.com/news/dr-gridlock/wp/2018/03/21/tempe-police-release-video-of-moments-before-autonomous-uber-crash/?utm_term=.0274d8b136c6
And you do not know what 'recant' means. Good grief. Amazing how kneejerk adversarial you are, unfailingly. How goes it for you in the workplace, I wonder. Good that you call for gov involvement even though of course it sucks totally.
https://www.nytimes.com/interactive/2018/03/20/us/self-driving-uber-pedestrian-killed.html
Uber to Let California Self-Driving-Car Permit Lapse
• Uber Technologies Inc. doesn’t plan to renew its permit to test autonomous vehicles in California, according to a letter from the state’s motor-vehicle department.
"The company will let its testing permit lapse at month’s end, rather than resubmit for renewal and face potential scrutiny amid an investigation into the circumstances of the accident."
“We proactively suspended our self-driving operations, including in California, immediately following the Tempe incident,” said an Uber spokeswoman in a statement. “Given this, we decided to not reapply for a California DMV permit with the understanding that our self-driving vehicles would not operate on public roads in the immediate future.” Uber could seek to renew the permit at a later date."
Additional Articles:
• Nvidia Halts Test of Self-Driving Tech Following Uber Crash
• Arizona Governor Suspends Uber’s Self-Driving Cars From Roads
• Arizonans Rethink Self-Driving Experiment After Uber Crash
Regards,
Ted