Slashdot Mirror


Uber Vehicle Saw But Ignored Woman It Struck, Report Says (engadget.com)

gollum123 writes: Uber has reportedly discovered that the fatal crash involving one of its prototype self-driving cars was probably caused by software faultily set up to ignore objects in the road, sources told The Information. Specifically, it was that the system was set up to ignore objects that it should have attended to; Herzberg seems to have been detected but considered a false positive.

323 comments

  1. Oops! We left it in murder mode. by Anonymous Coward · · Score: 5, Funny

    So sorry for any inconvenience.

    1. Re: Oops! We left it in murder mode. by Anonymous Coward · · Score: 2, Insightful

      Not a bug, a feature! Cleaning up the streets one (homeless) cyclist at a time.

      Functions as designed.

  2. So who is to blame? by LynnwoodRooster · · Score: 3, Interesting

    Who is guilty of vehicular manslaughter, here?

    --
    Browsing at +1 - no ACs, I ignore their posts. So refreshing!
    1. Re: So who is to blame? by Anonymous Coward · · Score: 0

      Drive criminally and Uber civil and criminal

    2. Re:So who is to blame? by mea_culpa · · Score: 5, Interesting

      Probably the person playing on their phone as it was their job to override decisions made by buggy software.

      Also the video Uber released is highly altered. I drive on that street frequently and it is very well lit.

    3. Re:So who is to blame? by Anonymous Coward · · Score: 0

      You're right that Republicans disregard human life wantonly, being greedy soulless faggots more than not, but you really think they can code AI? Without knowing how to read, basic science facts, etc? It does explain the crashes...

    4. Re:So who is to blame? by Xylantiel · · Score: 1

      Nope, this just proves that the person sitting behind the wheel was correct in not reacting. By the time the operator knew the car had not detected the obstacle, it was too late. Previously when this was discussed, it was pointed out that this is actually a reason that the entire concept of a "safety operator" doesn't work.

    5. Re:So who is to blame? by DoktorMidnight · · Score: 1

      Oof, that's a good one. One reply here says the person who was driving and I tend to agree. Although it's not the same, the use of force with military drones ultimately has to pass through the approval of a Human operator. I would think that the Human operator in an autonomous vehicle would equally be ultimately responsible for failing to override fatal decisions made by an autonomous vehicle. That said, if the operator attempted to override the vehicle and the fatal incident occurred anyway, then I would say that traditional law would place the blame on the manufacturer/programmer.

    6. Re:So who is to blame? by doconnor · · Score: 1

      The head of QA. How many of tests on a track with mannequins did they do? Probably should have been hundreds or more.

    7. Re:So who is to blame? by Narcocide · · Score: 2

      They outsourced that part to India.

    8. Re:So who is to blame? by Anonymous Coward · · Score: 0

      What law, natural or man-made, requires a safety operator take no action until they have determined that they car is not going to act appropriately? That's an asinine decision-making strategy.

    9. Re:So who is to blame? by jellomizer · · Score: 1

      Exactly,
      Being this is test technology. It was his job to override the car when it made a bad decision. Being that Uber's Self driving cars is years behind other makers such as Google, the safety driver should had been much more vigilant.

      --
      If something is so important that you feel the need to post it on the internet... It probably isn't that important.
    10. Re:So who is to blame? by Oswald+McWeany · · Score: 1

      They outsourced that part to India.

      Nah! The Uber thought the woman was it's wife, and thus was programmed to ignore her.

      --
      "That's the way to do it" - Punch
    11. Re:So who is to blame? by Julz · · Score: 1

      The start a whole "It's not my fault it's my car that did it. Now prove that it's not in a court suckers!"?

      This could be an indication of the pressure (KPIs) to release code and "features" before they've been fully tested or even approved. With all this type of technology quickly becoming the future there needs to be some serious work "not Spectre like work" but serious work done making sure that all phases of the workflows creating systems "that have the potential to cause human casualty or death" are secure and error free. There should never be a false positive option in these systems. If it's not 100% certain that it's a Rimmer then avoid or stop.

      --
      When shit hits the fan get some of these https://youtu.be/pY-GncsZ-UE
    12. Re:So who is to blame? by I4ko · · Score: 1

      Wrong. At the speed the car was moving an operator should have detected the vehicle applying brakes about 3 car lengths before the car entered the shadow of the bridge (as to not put the passengers on the windshield). If the operator have punched the break pedal through the floor as the hood was entering the shadow the hit would not have been fatal, and it was possible to have full stop. It is a very wide bridge, several lanes of highway.

    13. Re:So who is to blame? by Xylantiel · · Score: 3, Insightful

      What you appear to want is called "ghosting", where the driver drives and the autonomous software pretends to drive and the differences are evaluated. That is not the role of the safety operator. You appear to agree that a safety operator is a stupid concept because they have no way to know if the car is correctly evaluating any given obstacle until it is likely too late.

    14. Re:So who is to blame? by bondsbw · · Score: 2

      Any "ghosting" should be done by the automation system, with the driver in primary control.

      --
      All my liberal friends think I'm a conservative, all my conservative friends think I'm a liberal.
    15. Re:So who is to blame? by lhunath · · Score: 1

      Nobody, since the case was settled out of court.

      --
      ``OK, so ten out of ten for style, but minus several million for good thinking, yeah?''
    16. Re:So who is to blame? by HornWumpus · · Score: 2

      If the purpose is to test, you can't disconnect it before it makes a mistake.

      But video shows the 'safety operator' was not watching the road. Which is understandable. Try watching a CNC machine with you hand hovering over the e-stop button, see how long you last. I guarantee you, it won't be an 8 hour shift of alert watching. Then again, he was in the car. You'd likely watch a lot better if your skull was in reach of the tool.

      Testing autodrive cars on the road is hubris. Build an artificial test environment first, let it run for virtual giga car-years. _Constantly_ collect new data for it with manually driven cars and traffic cams. Don't bulshit yourself, you won't be 'done and ready' in a year. Test it against Mumbai traffic...check that, maybe not, that's how you get Stephen King movies realized.

      --
      John McAfee 'It was like that time I hired that Bangkok prostitute; to do my taxes, while I fucked my accountant'
    17. Re:So who is to blame? by Xylantiel · · Score: 3, Insightful

      My argument is that is the obstacle was too obvious. The safety operator had to also overcome their expectation that the car would do the right thing. The engineers doing post-crash analysis appear to only have a vague idea why the car didn't appropriately evaluate this blatantly obvious obstacle. But the safety operator was supposed to figure this all out in less than 3 car lengths.

    18. Re:So who is to blame? by Anonymous Coward · · Score: 0

      Depends, if it is determined to be vehicular manslaughter, then either the ghost driver or some culpable sw guy could go to jail. I am voting for the sw guy. It is time for the self-drive group to understand their sw has life & death consequences, and poor quality is not acceptable. Someone died. Worse, the techies on /. are joking about it.

    19. Re:So who is to blame? by HornWumpus · · Score: 1

      How do you propose to validate a pipelined set of neural nets' training? This isn't just a message loop and a giant case statement.

      I'd start by setting up a transparent, adversarial system in a virtual world. Let the public earn money by providing data sets that trigger bad behavior in the 'AI'. Good fun.

      --
      John McAfee 'It was like that time I hired that Bangkok prostitute; to do my taxes, while I fucked my accountant'
    20. Re:So who is to blame? by Anonymous Coward · · Score: 0

      Doubtful. If that were the case the other autonomous Ubers would have gotten together and gang-raped and beheaded her.

    21. Re:So who is to blame? by msauve · · Score: 4, Insightful

      Does the safety driver get any feedback on what the autopilot is planning? I'd think that even a simple green/yellow/red indication to show what it's perceiving (everything's OK/I see something and am prepared to take action/I am taking action) would be useful.

      I could see the car recognizing a potential hazard well in advance of a need to take action - that info should be given to the safety driver. If they in turn take action before the autopilot would have, perhaps an algorithm needs tweaking. And, if the driver sees a potential hazard first, they should be able to provide feedback on that, too, so they can figure out why the human is doing a better job.

      --
      "National Security is the chief cause of national insecurity." - Celine's First Law
    22. Re:So who is to blame? by lhunath · · Score: 1

      Relax, dude. You are discontent with the general public's reaction to this murder and are putting the weight of that injustice squarely on one person's shoulders. The software developer is not making the public act they way they are. You don't know the guy.

      If I was the software developer that made this mistake, I would already be tortured with guilt over having killed a person. Your assumption that this person must be punished because he's clearly obviously laughing about it like the rest of them are is baseless and ignorant, and as naive of a real person's tragedy as you are blaming the public to be about this tragedy.

      Punishment is not justice, and justice does not right a wrong. Don't be so eager to swing an axe just because you yourself feel bad about this situation. That is just selfish. Instead, let's figure out the right way to react to the injustice that took place here, barring any prejudice.

      --
      ``OK, so ten out of ten for style, but minus several million for good thinking, yeah?''
    23. Re:So who is to blame? by Raenex · · Score: 1

      I'm not convinced that particular spot was "well lit". She was obviously coming out of a shadow, and there seemed to be only the one light in the area with no overlap.

    24. Re:So who is to blame? by Anonymous Coward · · Score: 0

      Haha quit lying, that street right now is not that well lit in that area. I live nearby. Compounded by the fact the city of Tempe built nice brick walkways in the media with signs saying don't walk here instead of desert plants contributed to the fact this person was in the dark, looking straight ahead, and pushing a shopping cart into the street from that media.

    25. Re:So who is to blame? by Holi · · Score: 1

      The safety driver who was too busy looking at her phone instead of the road.

      --
      Sorry, teleporters just kill you and then make a copy. A perfect, soul-less copy.
    26. Re: So who is to blame? by Anonymous Coward · · Score: 0

      *brake

    27. Re:So who is to blame? by Holi · · Score: 1

      You mean when she looked up and said "OH SHIT"? because she wasn't looking at the road?

      --
      Sorry, teleporters just kill you and then make a copy. A perfect, soul-less copy.
    28. Re: So who is to blame? by Type44Q · · Score: 1

      Both venture capitalists and politicians.

    29. Re:So who is to blame? by HornWumpus · · Score: 1

      It's a software guy alright.

      It's you! We've pinned the whole deal on whatever code we can prove you wrote and abandoned. Bet you didn't expect that code to end up in an 'AI'!

      --
      John McAfee 'It was like that time I hired that Bangkok prostitute; to do my taxes, while I fucked my accountant'
    30. Re:So who is to blame? by Anonymous Coward · · Score: 0

      Too bad as most people who drive that road have a very different opinion then you, a person who has probably never been there.

    31. Re:So who is to blame? by mrclevesque · · Score: 1

      "But the safety operator was supposed to figure this all out in less than 3 car lengths."

      They had a lot more than 3 car lengths to realize the car wasn't slowing down.

    32. Re:So who is to blame? by fustakrakich · · Score: 1

      Under civil forfeiture law, the vehicle is guilty. It will spend the rest of its life in the pound

      --
      “He’s not deformed, he’s just drunk!”
    33. Re: So who is to blame? by saloomy · · Score: 1, Insightful

      No one is guilty of vehicular manslaughter. This is an accident due to bad design. You don't jail the engineers or architects who design a building that fails in an earthquake. You don't arrest the airline execs when a plan component fails. The only difference here is software failed. Learn from the mistake, don't do it again. Sheesh.

    34. Re:So who is to blame? by Anonymous Coward · · Score: 1

      I'm not convinced that particular spot was "well lit".

      That was only on the "official" video: obviously inferior dashcam. If the car was working correctly, there's no way that would have been a primary source of input. At the very least, there would have been LIDAR providing the computer(s) a better "view" of what was ahead. Adding to the assessment of inferiority of the official video, plenty of other drivers provided dashcam video of the same area recorded at about the same time of day and all showed that visibility was much better than what the official dashcam recorded.

      That said, the pedestrian was a bit of an idiot for strolling across a multi-lane road at night, well clear of any intersection, wearing no reflective gear.

    35. Re:So who is to blame? by Anonymous Coward · · Score: 0

      The Republican who programmed it with a disregard for human life?

      Ignorance... everywhere.

    36. Re:So who is to blame? by I4ko · · Score: 2

      Exactly. The bridge is 16 lane widths, and crossed almost at 45 degree angle at that. It is very very wide. Heck.. at that angle at 45mph which is the allowed speed I need 3 lane widths (effectively 4.24 at that angle) to bring the car to complete stop.

    37. Re:So who is to blame? by bluefoxlucid · · Score: 1, Interesting

      If the purpose is to test, you can't disconnect it before it makes a mistake.

      Yes you can. Hitting the brakes when the car doesn't expect it would cause it to evaluate what's going on. Did it identify an obstacle? Did it identify a possible not-obstacle? Did it expect to need to react, but not judge the condition as resolved yet (someone's up the road 250 feet, and you've got plenty of distance at 50 feet, so you wait to see if they move and possibly slow down at 150 feet to prepare--condition is resolving into a single probable outcome as you approach)?

      The car can even report that, yes, it saw the obstacle, but intended to brake slightly-later than the driver due to requiring 20 feet to stop and having 90 feet, and being at this point tended to react with twice or 50 feet more than the required stopping distance. Thus you can go back and review the discrepancies, and then notate to the car whether its projected response was inadequate (adapt) or the driver was just more-cautious and the projected response was acceptable (notate, but no change necessary).

      Build an artificial test environment first, let it run for virtual giga car-years

      Artificial environments can overlearn artificial behaviors. Even people who drive in real cities suddenly can't cope with the task of driving when they move the next state over--it's like they don't know how to drive at all, or maybe everybody else never learned to drive right.

    38. Re:So who is to blame? by bluefoxlucid · · Score: 1

      Refresh my memory... the next lane over was clear? Braking is the wrong action at this distance; if you might full-stop, you should instead use a lane toss.

    39. Re:So who is to blame? by bluefoxlucid · · Score: 1, Interesting

      Punishment is not justice, and justice does not right a wrong. Don't be so eager to swing an axe just because you yourself feel bad about this situation. That is just selfish. Instead, let's figure out the right way to react to the injustice that took place here, barring any prejudice.

      I actually proposed a Constitutional Amendment. Working on the language.

      The purpose of law being to establish Justice and insure domestic Tranquility, the execution of law against an offense shall be to redress and rehabilitate.

      To this purpose, and to the purpose of a fair and speedy trial, none shall be confined against his will except as necessary for the security of the public, and such confinement shall to the greatest extent achievable respect the dignity of the confined as human beings and ensure their individual needs are met and rights protected; and no bail shall be required except where other means are insufficient to the same purpose; and civil damages shall not be imposed in excess of those necessary to redress.

    40. Re:So who is to blame? by barc0001 · · Score: 3, Interesting

      > Probably the person playing on their phone as it was their job to override decisions made by buggy software.

      I dunno, looking at the video of the crash, the victim crossed the road outside of a crosswalk and wasn't even LOOKING in the direction of potential traffic. I'd assign the lion's share of the blame to the person who literally walked into the path of a brightly lit car without noticing.

    41. Re:So who is to blame? by Anonymous Coward · · Score: 0

      It depends on how the test was set up and the law framework it was executed under. Did Arizona require specific safety precautions for testing self driving vehicles? Did Uber have safety procedures that were tested and found to be adequate? Was the driver trained in these procedures? From what I have read so far the answers to these questions are mostly "No". To me everyone from the pedestrian to the state of Arizona carry some level of liability for the accident. In addition, the driver, the State of Arizona and Uber carry a lot of responsibility for conducting the test on public streets. What puzzles me is that the car failed at the relatively simple task of breaking for a bulky object on a road with no obstructions. This does not require a sophisticated machine learning algorithm and should be implemented as a safety feature running in parallel with the self driving software.

    42. Re:So who is to blame? by Anonymous Coward · · Score: 0

      Relax, dude. You are discontent with the general public's reaction to this murder and are putting the weight of that injustice squarely on one person's shoulders.

      I'm pretty sure the software is developed by a team. So, who's to blame? The Chief Designer? The CEO? All the team?

    43. Re: So who is to blame? by Anonymous Coward · · Score: 2, Funny

      That sounds an awful lot like something someone guilty of vehicular manslaughter would say.

    44. Re:So who is to blame? by wonkey_monkey · · Score: 1

      Also the video Uber released is highly altered. I drive on that street frequently and it is very well lit.

      That doeesn't mean the video was altered. It's just a shitty camera.

      --
      systemd is Roko's Basilisk.
    45. Re:So who is to blame? by Anonymous Coward · · Score: 1

      She was looking at the car sensor display; they don't have HUDs

    46. Re:So who is to blame? by HornWumpus · · Score: 1

      In this case, it's academic. The operator wasn't watching the road.

      If you touch the brakes everytime there is a possible obstacle, you would be on them _all_day_ (in city traffic or side streets). Your back to 'no additional information'.

      As always, you have to be careful not to train to artifacts. But artifacts exist in both virtual worlds and real world sensors. An advantage of virtual testing is you can isolate subsystems (or not). e.g. You don't need to render the scene and feed that to the camera/image analysis pipeline, you can just feed to object locations to the navigation.

      Shitty drivers can't drive where they _learned_. Adults that have been around can navigate between massholes and origonions. (the two extremes of American bad drivers, still pikers globally, Mumbai laughs at Boston driver's shenanigans.)

      --
      John McAfee 'It was like that time I hired that Bangkok prostitute; to do my taxes, while I fucked my accountant'
    47. Re:So who is to blame? by amicusNYCL · · Score: 2

      The road there is about as well lit as you would expect a big road to be, it's not as dark as the video implies. Granted, most of the time I'm down there I'm there for a concert and they might have additional lights on, but the video was obviously darker than what a person would see. At the place where she was crossing, the driver should have been able to see her crossing the road for several hundred feet at least.

      --
      "Our two-party system is like a bowl of shit looking at itself in a mirror." - Lewis Black
    48. Re: So who is to blame? by amorsen · · Score: 5, Insightful

      Of course you jail the engineers or architects, when they are criminally negligent.

      There is little doubt that the Uber CEO, Dara Khosrowshahi, was criminally negligent and aware that this kind of accident was not only possible but likely.

      --
      Finally! A year of moderation! Ready for 2019?
    49. Re:So who is to blame? by HornWumpus · · Score: 1

      A forward radar, simple stupid, system should have been able to detect the bike if not the person.

      --
      John McAfee 'It was like that time I hired that Bangkok prostitute; to do my taxes, while I fucked my accountant'
    50. Re:So who is to blame? by stephanruby · · Score: 1

      Did you see the entire video? The street is very well lit all the way through, except at the point where the woman was crossing.

    51. Re:So who is to blame? by Green+Mountain+Bot · · Score: 1

      Settlement out of court is for civil cases, not criminal cases. The equivalent is a plea bargain, which happens after arraignment.

    52. Re:So who is to blame? by slack_justyb · · Score: 1

      serious work done making sure that all phases of the workflows creating systems "that have the potential to cause human casualty or death" are secure and error free.

      Well some companies are indeed building tracks to begin neural net training live. Additionally, there's been enough failures and near misses from other car companies to begin edge testing as well.

      Additionally, map makers are now refocusing on a new emerging market of maps for self driving cars. These maps differ from the typical on-line map in that they need typical pattern usage of a given intersection or piece of road that initial algorithms create too many edge cases for. Good example might be the 65/440 split in Nashville where I've seen map cars out there going over and over the exact same spot. Apparently it's confusing to self driving cars.

      I think some companies are nearing the peek of the Dunning-Kruger chart and realizing that this problem is a lot harder than they expected. However, there is a lot of money if someone gets the self driving car right and so where in other ventures that peak would mean the end of research, the potential profits are driving some past the peak into the long valley.

      I definitely echo your sentiment in that more testing to harden the product is needed and I think a few folks early on knew that (BMW, Ford, etc...). I think that Uber and Tesla might be going too fast, too soon on their implementations.

    53. Re:So who is to blame? by sexconker · · Score: 3, Interesting

      Try watching a CNC machine with you hand hovering over the e-stop button, see how long you last. I guarantee you, it won't be an 8 hour shift of alert watching.

      I'd probably go 12 hours or more. I love watching those things work.

    54. Re:So who is to blame? by sexconker · · Score: 2, Insightful

      I've seen plenty of videos. The street is very well lit in all of them except Uber's video.

      Interestingly, the camera facing the human "driver" is crisp and clear, using your standard "night vision" mode.
      The Uber video is either doctored or doctored.

    55. Re:So who is to blame? by vivian · · Score: 4, Interesting

      When I drive and see a possible danger (kid playing with a ball by the side of the road, dog wandering around near the edge of the road without a leash, motorcycle rider looking the wrong way on a side street as I approach, etc.) I always take my foot off the accelerator and cover the break - ready to instantly respond if something stupid happens. It's called defensive driving, and is how everyone should drive.
      I never really picked up this habit though until I had been riding motorbikes for a while, when you absolutely have to drive defensively id you want to survive commuting in London or Tokyo traffic. (I used to ride in both).

      The safety driver should have been doing this, and it would not be any impediment to testing the autonomy of the car - it can still do the driving, but the safety driver would have had time to react appropriately.
      The safety driver completely failed in her duty - possibly due to lack of training - but if your getting paid to be a safety driver then you should do your job instead of buggering around with your phone.

    56. Re: So who is to blame? by LynnwoodRooster · · Score: 5, Insightful

      No one is guilty of vehicular manslaughter. This is an accident due to bad design. You don't jail the engineers or architects who design a building that fails in an earthquake.

      You do if it wasn't designed to code.

      You don't arrest the airline execs when a plan component fails.

      Maybe not the exec, but certainly the maintenance engineer who committed fraud that resulted in the death of people.

      So here - who's to blame? Who decided to live-test an experimental system that can operate with the safety disengaged?

      --
      Browsing at +1 - no ACs, I ignore their posts. So refreshing!
    57. Re:So who is to blame? by alvinrod · · Score: 1

      I actually proposed a Constitutional Amendment. Working on the language.

      The language needs a lot of work. What you've written makes for something that sounds wonderfully pleasing, but it riddled with the kind of ambiguity that means nothing will change and more of the court's time will be wasted arguing about how this should be interpreted. Specifically:

      1) What constitutes necessity for the security of the public and who will determine this? Is if sufficient to confine a person if one of person greatly fears for their life and the public at large is under no additional danger in any way?
      2) What is meant by the greatest extend achievable? If you want to throw an infinite amount of money at the system, there's quite a bit that could be achieved. Realistically, who gets to determine what's achievable with the meager budget that the system does get?
      3) What does dignity as human beings mean? Comes across as useless platitudes that mean whatever the reader wants it to mean.
      4) What needs are not currently being met and who gets to determine what a person's needs are? If someone really needs a good lay, is society on the hook for the hooker?
      5) Inmates already should be having their rights protected. If the system is not doing a good job presently with similar rules in place, what makes you think some flowery language will change anything?
      6) When are other means insufficient as to require bail? Everyone is going to interpret this their own way.
      7) Who gets to determine the value of civil damages and what constitutes excess? This might work if someone's property is damaged, but how to you value mental anguish, undue suffering, or even the loss of life?

      More generally, what do you propose to do with those who either refuse to be rehabilitated or cannot be rehabilitated? It might be nice to think that you can save or fix everyone, but there are some sick people who are broken beyond our ability to repair them. You're unlikely to be able to cure many serial rapists, murders, or child molesters who have committed multiple crimes against numerous victims. Trying to figure out who can be rehabilitated and who can't be is tricky business and the first time you let out a "rehabilitated" person who goes off and rapes another child, the public is going to burn down your house in anger. I don't even think this passes as that's going to be the main argument against it and people will respond emotionally.

    58. Re: So who is to blame? by hparker · · Score: 1

      Well, sometimes they do jail and even execute the people responsible for negligent or faudulent testing in building construction that falls and kills people when no earthquake occurs. Using cheapened concrete and falsifying strength tests has lead to convictions.

      Were recognized safety standards followed in the release of this software? Were the tests results falsified? Or was the prerelease testing not done or reported falsely?

    59. Re:So who is to blame? by Anonymous Coward · · Score: 1

      wrong... studies have shown people are not able to pay 100% attention 100% of the time. reliance on a "safety operator" is a "best effort of last resort" sort of thing that should be assumed not to work, but if you're lucky, might.

      that's not to mention the other point that for testing, said operator is explicitly expected to let the car handle all but extra-ordinary cases... this was per se ordinary, with a extraordinary failure.

      failure all around, on uber's part...

    60. Re:So who is to blame? by Raenex · · Score: 1

      Such statements are subjective, prone to interpretation of vague memories. The only way to know for sure what people could see and when is to test under the same conditions. People walking out into the street at night are a problem even in better lighting, and anybody doing so is recklessly putting their life at risk.

    61. Re: So who is to blame? by Anonymous Coward · · Score: 0, Insightful

      Fuck off. Always brake, then evaluate, then maneuver.

    62. Re:So who is to blame? by Anonymous Coward · · Score: 0

      You choose your subjective judgement as "obvious" and denounce someone else's.

      We sort people like you into the jackass bucket. Grow up and take a hard look at yourself.

    63. Re: So who is to blame? by drinkypoo · · Score: 1

      Fuck off. Always brake, then evaluate, then maneuver.

      Since all Uber cars (and all modern cars in the developed world, by law) have ABS and ESP, the correct answer is to brake and evade simultaneously. You will reduce your velocity and thus increase survivability, and have a chance to evade.

      --
      "You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
    64. Re: So who is to blame? by DRJlaw · · Score: 1

      There is little doubt that the Uber CEO, Dara Khosrowshahi, was criminally negligent and aware that this kind of accident was not only possible but likely.

      I just love it when people throw out unsupported conclusions with the self-assurance that no reasonable person could possibly disagree with them.

      By which I mean, I have every reason to doubt that, and you've offered nothing to prove otherwise.

    65. Re:So who is to blame? by Anonymous Coward · · Score: 0

      In some states, when a pedestrian is killed outside the cross walk they call it suicide.

      I haven't heard if she was making a legal crossing, AND being aware of your surroundings is conducive to living longer.

      Was the car at fault in the first place? If not, the gene pool got a little cleaner.

      Want to play in the traffic? Best learn the rules first.

    66. Re: So who is to blame? by Lanthanide · · Score: 1

      You don't jail the engineers or architects who design a building that fails in an earthquake.

      Well actually. In the 2010 Christchurch earthquake, we had a building called the CTV Building that fell over and killed 115 people: https://en.wikipedia.org/wiki/...

      The engineer has not been held responsible, because he quit from the professional engineering society before they laid charges against him, and the police (apparently incorrectly) decided that he could not be held culpable for their deaths.

      It's a fairly minor scandal here and the families of the victims have appealed to the attorney general to re-review the case

      https://www.stuff.co.nz/nation...
      https://www.nzherald.co.nz/nz/...

      > The second point related to a legal technicality only flagged in the final months of the police investigation. The law says that any death resulting from negligent conduct must occur no more than one year and one day after that conduct ended.

      This part of the law is being repealed by the government.

    67. Re:So who is to blame? by HornWumpus · · Score: 1

      Watching isn't the problem, not being put half to sleep is.

      Nobody does, you run new programs in a virtual machine and have it graph the tool path. You couldn't e-stop it fast enough anyhow, even if you somehow stayed alert.

      --
      John McAfee 'It was like that time I hired that Bangkok prostitute; to do my taxes, while I fucked my accountant'
    68. Re:So who is to blame? by Moof123 · · Score: 1

      We require infrastructure and other life supporting systems to get a PE's (professional engineer) stamp. We don't let every wackadoo design high rises or bridges after all.

      Software for critical systems where life is on the line should get a similar treatment. Startup culture that flaunts rules until after they get to critical mass kinda works for disruption of certain sectors, but is barreling towards disaster when life is on the line. I can't speak to all of the legalities, but we should not be Beta testing new revisions of software/settings on public roadways without the equivalent of a PE stamp of approval after many hours of rigorous testing in a controlled environement. Instead the NOW NOW NOW rush behind autonomous cars has thrown caution to the wind, and resulted in demonstrably awful software being deployed to the streets without much oversight.

      I'd argue that whoever flubbed the object detection and put lethally bad software onto the car (and the managers who let it happen) deserve manslaughter charges.

    69. Re:So who is to blame? by Raenex · · Score: 1

      You choose your subjective judgement as "obvious" and denounce someone else's.

      There's nothing subjective about my judgment. She came out of a shadow. You can debate how visible she was in that shadow, but that requires real-word testing.

      We sort people like you into the jackass bucket. Grow up and take a hard look at yourself.

      Right back at you.

    70. Re:So who is to blame? by Anonymous Coward · · Score: 0

      Wtf asshole, I have no car, should I stay home at night? Should I walk 2 miles out of my way to use a walk signal?

    71. Re: So who is to blame? by Anonymous Coward · · Score: 0

      Bullshit, she was playing on her phone. It's pretty clear that she's looking down and to the side.

    72. Re:So who is to blame? by grahamsz · · Score: 1

      The issue is that no matter how much testing you do in a controlled environment, it isn't a substitute for testing in an uncontrolled one.

      Most of what i've read suggests that Uber is significantly behind Waymo, but it's really only a matter of time before a waymo car kills a pedestrian. Or indeed a Traction Control system does. Or aircraft autopilot downs a plane in the ocean.

      At some point we're going to have to face that we do have to decree when a system is sufficiently test and STILL accept that it will kill people in production.

    73. Re:So who is to blame? by Raenex · · Score: 0

      Hey asshole, feel free to use the roads. Just don't walk out in front of traffic expecting the car to stop for you.

      Years ago, when I test drove my first car, a disheveled older lady did exactly as this woman did. She's lucky I saw her out of the corner of my eye and slammed on the brakes. The salesman who was sitting next to me didn't even see her.

    74. Re:So who is to blame? by MrKaos · · Score: 1

      Who is guilty of vehicular manslaughter, here?

      It depends. In this scenario the AI was in Bruce Jenner mode when it hit the woman but switched to Kateland Jenner mode later claim innocence.

      --
      My ism, it's full of beliefs.
    75. Re:So who is to blame? by Demena · · Score: 1

      Honestly? Anyone who assumed that a "trade off" was ever possible. You cannot move a slider for varying degrees of risk to people. The whole software system is irrevocably flawed if any such trade off is contained in it programming.

    76. Re: So who is to blame? by Demena · · Score: 1

      Wrong. They designed a 'trade-off' into it so it wouldn't react to wind driven shopping bags. By doing that they introduced adjustable kill-rate. By doing that they guaranteed that people would die, but left how many adjustable.

    77. Re:So who is to blame? by wisnoskij · · Score: 1

      I think that is just because it is video. Cameras are not great at non-optimal lighting

      --
      Troll is not a replacement for I disagree.
    78. Re: So who is to blame? by Demena · · Score: 1

      Acceleration will get you out of trouble more times than breaking. Evaluate first, fast but first. You just acknowledged that you are a menace on the roads. Often braking is just what NOT to do.

    79. Re:So who is to blame? by Demena · · Score: 1

      Sorry, does not work. The computer is faster than a human. Yet the human is supposed to stop the computer when it makes a mistake. But even if the human reacts perfectly it cannot beat the computer's reaction time (in this case 'lack-of-reaction' time). One the computer failed even a perfectly alert computer cannot compensate - it is already too late.

    80. Re:So who is to blame? by RhettLivingston · · Score: 1

      How dark or light the area was is irrelevant to the technology's response. This vehicle was not using the released video for navigation. That was from a camera capturing video for review in incidences like this one.

      The vehicle had Lidar which can work fine in complete darkness with the headlights turned off. The sensors spotted this individual and the software ignored what they saw.

      However, because the driver is ultimately in control and responsible and this area was lit well enough for a human (though obviously not for a cheap dashcam), this accident was the driver's fault -- no ifs, ands, or buts. Moreover, it appears that the fault was due to gross negligence on his part. Perhaps Uber also has some fault for not adequately monitoring the drivers or hiring people professional enough to be trusted.

      The purpose of this testing is to catch problems in the software like this one. If the driver had been performing his job, he would have taken control, the incident would have been noted, and the engineers would have been able to make the necessary adjustments without any loss of life or ever making the news.

    81. Re:So who is to blame? by Mal-2 · · Score: 1

      I want to know how long the car had been driving since the last time the safety driver had to do anything. If this is more than a few minutes, then I can't blame the driver because constant attentiveness without action is incredibly difficult for a human to achieve for any length of time. The fact that there was no perceived better alternative than a human safety driver indicates a systematic problem in testing procedures (expecting humans to note, evaluate, and properly respond to an emergency faster than the computer would have -- because you can't tell that it didn't respond until the opportunity has passed) rather than failure on the part of the human. She was improperly deployed and was not capable of what she was being asked to do. Very few, if any, people are equipped to do so.

      --
      How is the Riemann zeta function like Trump rallies? Both have an endless number of trivial zeros.
    82. Re:So who is to blame? by Mal-2 · · Score: 1

      The safety driver simply can't be vigilant for the durations necessary for road tests. Maybe the first time they hit the road on any given day, they can, but that will last all of about fifteen minutes before focus exhaustion sets in. Humans are simply not up to the task of monitoring a computer and reacting only after they realize it has failed to do so. By the time that realization sets in, it's far too late to avert the disaster. Even if the driver had been watching attentively, it still would have taken time to realize the computer was somehow planning to drive straight through the obstacle, come up with a better plan, and enact it. By that time, it would be too late.

      Lesson: safety drivers are only useful for the situations where the machine gives up, and stops or pulls off to the side of the road. They are not useful at all as an emergency hand-off while the car is in motion.

      --
      How is the Riemann zeta function like Trump rallies? Both have an endless number of trivial zeros.
    83. Re: So who is to blame? by Anonymous Coward · · Score: 0

      Stay the, fuck off, the road grandma. Really. Rolling cock block, you be.

    84. Re: So who is to blame? by Anonymous Coward · · Score: 0

      Who decided to live-test an experimental system that can operate with the safety disengaged?

      Surprisingly perceptive. The safety was disengaged (they were too interested in their phone).

      Uber should've enforced an eyeball tracking module (come to a stop if driver looks away for significant periods of time).

    85. Re: So who is to blame? by Anonymous Coward · · Score: 0

      "Acceleration will get you out of trouble more times than breaking."

      NO NO NO NO NO. Gesus, NO!!!

      Get off the road you lunatic!!!

    86. Re:So who is to blame? by Anonymous Coward · · Score: 0

      "In some states, when a pedestrian is killed outside the cross walk they call it suicide."

      Fuck off you asshole.

    87. Re:So who is to blame? by Anonymous Coward · · Score: 0

      This is Uber we are talking about.

      The video was altered.

      Plus, I am very confident the police chief was payed off in some way to release it early and make his statement that he didn't think Uber was to blame. This was a drastic PR scramble from Uber to deflect blame and preserve their stock price. No doubt they will settle out of court, oh wait, they already have, so the full truth is not going to be revealed in a public court. We don't yet know what the NTSB will report, but I doubt the general public will *ever* see the full resolution raw footage from the camera systems on that car.

      They are cunts.

      The poor woman.

    88. Re:So who is to blame? by Maxo-Texas · · Score: 1

      It could be the landscaping on the left that blocked seeing the pedestrian and also completely blocked LIDAR until the pedestrian stepped onto the road.

      I don't know why people are not mentioning it. It's obvious on both the satellite and point of view photo views.

      The road also widens from 2 lanes to 4 lanes right before the accent.

      --
      She was like chocolate when she drank... semi-sweet at first and then increasingly bitter.
    89. Re:So who is to blame? by Anonymous Coward · · Score: 0

      The team manager, then their manager, up to the CEO.

      Managers are paid the big bucks to take responsibility. The moment something goes wrong, they are at fault and take the fall.

      What, you thought they were well paid for their ability to organize meetings and annoy the absolute fuck out of their staff?

      The second the system fails to hold managers responsible is the second they're lined up against the wall and disposed of, as they otherwise hold no value whatsoever.

    90. Re:So who is to blame? by Maxo-Texas · · Score: 1

      Some people allege she was using her phone but I havn't seen support for that.
      Early reports were she was looking at the car display. We would need at least the prior 5 minutes (instead of maybe 45 seconds).

      Any self driving car needs a black box with high quality feed of the cabin, and 4 directions as well as speed, acceleration, gps, etc. which the manufacturer *cannot* open and which is handed to the government unopened.

      --
      She was like chocolate when she drank... semi-sweet at first and then increasingly bitter.
    91. Re:So who is to blame? by grep+-v+'.*'+* · · Score: 1

      Punishment is not justice, and justice does not right a wrong.

      I actually proposed a Constitutional Amendment. Working on the language.

      Naaa. Instead, how about: Immediately
      * kill the driver (immediately involved, should help inattention),
      * kill the software manager (programmatically involved, helps delete bugs), and
      * kill the CEO (managerially involved, helps ... well, just helps reduce the overall psychopath load. Think of it as a moral good. If not that, then helping reduce the carbon load.)

      Then lets see if the quality of software changes any. And if you don't like the KILL aspect there, then place them in the road, something like this. *THESE* guys did it, so why can't they?

      Oh? Don't have enough confidence in the quality of your work that you're fostering to the general public? Wouldn't that be an interesting statement?

      --
      If the universe is someone's simulation -- does that mean the stars are just stuck pixels?
    92. Re:So who is to blame? by Anonymous Coward · · Score: 0

      Traffic laws are typically written so that more than one mistake has to be made for someone to be seriously injured or die.
      Like, it is not sufficient to drive in the oncoming traffics lane. It is also necessary that you or the oncoming traffic fails to stop for the collision to happen.

      Accidents don't happen from just one mistake.

      What I find strange with this is that the Uber car didn't have automatic braking.
      Like, that is a completely separate system that isn't connected to the self-driving part and only has the responsibility to use lidar/ultrasound and apply brakes before imminent collision.
      There is no such thing as soft or hard objects in front of the car that is acceptable to run over. You never now if some "benign" object is hiding something worse.
      There is never a situation where the self-driving part should be able to override the automatic braking.

      You shouldn't even experiment with self driving technology until you have automatic braking that works.

    93. Re: So who is to blame? by Anonymous Coward · · Score: 0

      No, it's absolutely true. Acceleration WILL get you out of trouble more often than braking.

      Accelerating combined with maneuvering is very often the right choice. Only if every other option is bad is braking the best option.

      If you can't understand that, you should never be in control of a vehicle.

    94. Re:So who is to blame? by Anonymous Coward · · Score: 0

      She was NOT making a legal crossing.

      Very near that location, there is actually a "no pedestrian crossing" sign.

      Legally, the pedestrian was at fault. There's no question there. Had she lived, she'd have been ticketed.

      The software should have tried harder not to kill her. It had more data than a human would have. But a human would have done the same thing the car did, because it's unsafe for a pedestrian to cross there. It's pretty clear in the video, she would have been just as dead if a human had been driving. She wasn't visible until it was too late to avoid her.

    95. Re:So who is to blame? by Bongo · · Score: 1

      But does this mean that software is untestable? A documentary about aircraft showed they had to pass a formal test of, will this door open when covered with an inch of ice. But with cars and roads and software, we seem to be saying that formal tests are useless, because the software and environment are both too complex. So it’ll just be a case of, if a self-driving system gets a bad reputation, we’ll know to ban it? Is that the best we can do?

    96. Re: So who is to blame? by Anonymous Coward · · Score: 0

      No, you're still wrong.

      I hope when you kill yourself on the road you don't take any one else with you. Unfortunately, with your particular philosophy on driving ("Accelerating combined with maneuvering is very often the right choice") that's not going to happen.

    97. Re:So who is to blame? by Raenex · · Score: 1

      However, because the driver is ultimately in control and responsible and this area was lit well enough for a human (though obviously not for a cheap dashcam), this accident was the driver's fault -- no ifs, ands, or buts.

      Nonsense. We're only talking about this case because it was an AI at the controls. Otherwise it would be just another Darwin Award by somebody walking out in front of traffic.

    98. Re: So who is to blame? by thegarbz · · Score: 1

      There is little doubt that the Uber CEO, Dara Khosrowshahi, was criminally negligent and aware that this kind of accident was not only possible but likely.

      Way to go picking the person who is likely least aware and would most easily be able to prove reasonable doubt that this failure mode was a possibility. And we are talking about a specific failure mode, because generic stats don't get people arrested for negligence. Specific actions causing specific outcomes do.

    99. Re:So who is to blame? by thegarbz · · Score: 1

      Also the video Uber released is highly altered. I drive on that street frequently and it is very well lit.

      Well lit does not mean well recorded. Poor quality recording does not mean altered, it just means a garbage recording. Now maybe they bought a poor quality camera on purpose, then we may have something here, but honestly the lighting in that video footage tells you nothing about alteration.

    100. Re:So who is to blame? by thegarbz · · Score: 1

      A control system would never see green as it is continuously taking action. But that's not relevant because ...
      A human would get desensitized as the control system would spend most of its time in one mode. If you're not paying attention to the person stepping on the road and screwing with your mobile phone, a little light indicator won't help you at all.

    101. Re: So who is to blame? by amorsen · · Score: 1

      Uber has a goal of 13 miles without interventions from the safety driver. They have admitted that they cannot make even that rather... modest goal. There is no way that such information has been kept away from the CEO, unless the CEO is utterly incompetent. With 13 miles between interventions, it was only a matter of time before a safety driver failed to step in.

      Allowing a self driving car on public roads with 13 miles between interventions is reckless endangerment.

      --
      Finally! A year of moderation! Ready for 2019?
    102. Re: So who is to blame? by amorsen · · Score: 1

      Letting cars on public roads that cannot go 13 miles without safety interventions is reckless endangerment.

      Dara Khosrowshahi must have been aware of that statistic, yet chose not to intervene. The buck stops there.

      --
      Finally! A year of moderation! Ready for 2019?
    103. Re:So who is to blame? by Anonymous Coward · · Score: 0

      Why should the human keep watch of anything. Instead, a movie or solitaire seems the prudent thing to do when self-driving^H^H^Hmowing down children, cats and dogs!

    104. Re:So who is to blame? by Anonymous Coward · · Score: 0

      How to cope with the trauma of manslaughter though?

    105. Re:So who is to blame? by Anonymous Coward · · Score: 0

      Are you a fucking retard? The victim spent +5 seconds in direct line of sight to the car, NO BLOCKING LIDAR

    106. Re:So who is to blame? by Maxo-Texas · · Score: 1

      Are you stupid?

      Lidar is blocked by solid objects. The pedestrian came out from behind landscaping on the left side of the road.

      To the car, the pedestrian appeared out of no where. It may have contributed to the "false positive" decision.

      You can use your brain and consider various possibilities, or you can scream irrationally some more.

      I bet you haven't even looked at the personal p.o.v. on google maps.

      --
      She was like chocolate when she drank... semi-sweet at first and then increasingly bitter.
    107. Re:So who is to blame? by Cederic · · Score: 1

      She sure as fuck wasn't looking at the road. Whether the distracting device was in her hand or in the dashboard isn't terribly relevant to her lack of attention.

    108. Re:So who is to blame? by Cederic · · Score: 1

      That particular spot was well lit.
      https://arstechnica.com/cars/2...

    109. Re:So who is to blame? by Cederic · · Score: 1

      Worse, the techies on /. are joking about it.

      I'd hope so. It'll be a sad fucking day for humanity if we lose our sense of humour.

    110. Re:So who is to blame? by MrL0G1C · · Score: 1

      You need to get your facts straight, she had already crossed several lanes and was almost across the road when she was struck. Both the car and the driver had the opportunity to slow and stop with ample time to spare.

      --
      Waterfox - a Firefox fork with legacy extension support, security updates and better privacy by default.
    111. Re:So who is to blame? by Anonymous Coward · · Score: 0

      Your car can stop on a lot shorter distance than you think.
      Unless you are driving on ice there is almost no situation where braking is the wrong option.
      Well, you might have a tailgater on you ass, but then you should have braked a bit more carefully the moment they showed up. As long as you have a tailgater you can't drive safely. There are way more situations where braking is the only safe option compared to situations where switching lanes is the only safe option.

    112. Re:So who is to blame? by Anonymous Coward · · Score: 0

      In this case, it's academic. The operator wasn't watching the road.

      It shouldn't be academic.
      This is a non-hypothetical situation of a problem with testing of self-driving algorithms.
      We can't go back and prevent it from happening but we don't get much better data than this for considering how to legislate self-driving cars and testing of them.

      Was this version of the software sufficiently tested in a controlled environment before they used it on the road or did they "just change a few thing" on the fly?
      Did the code have any form of redundancy on safety critical systems?

      My guess is that Uber didn't follow the laws we already have for automotive software and the problem here is that law enforcement doesn't have the proper knowledge and technical expertise to go after them.

    113. Re:So who is to blame? by Anonymous Coward · · Score: 0

      "The woman was crossing the street outside a crosswalk when she was hit, the spokesperson said."

      Nobody is guilty here.

      Pedestrians have been jaywalking in front of cars and people are dying from collisions like that quite often. It is not a crime to follow all road laws and then hitting a pedestrian or anyone else breaking the road laws. Step on the interstate during rush hour, get hit.

    114. Re: So who is to blame? by Anonymous Coward · · Score: 0

      You don't jail the engineers or architects who design a building that fails in an earthquake.

      If the engineer design industrial systems that are required to follow safety standards and he doesn't follow them then yes, he goes to jail.

    115. Re:So who is to blame? by Raenex · · Score: 1

      Camera settings, other traffic in the area, and headlight performance all have an effect on what we can conclude from our armchairs. And just for funsies, let's play spot the people in time before you run them over: https://youtu.be/1XOVxSCG8u0?t...

    116. Re:So who is to blame? by Dorianny · · Score: 1

      Who is guilty of vehicular manslaughter, here?

      As the pedestrian was breaking road-safety regulations, it is classified as an accident. If a human was at the wheel the investigation would have been closed in an hour and barely made local news

    117. Re:So who is to blame? by jellomizer · · Score: 1

      People drive all day and need to be vigilant.
      Especially if the car is going into a an area where they are people and traffic.
      His job was to be a safety driver. He failed at his job.

      --
      If something is so important that you feel the need to post it on the internet... It probably isn't that important.
    118. Re: So who is to blame? by thegarbz · · Score: 1

      Letting cars on public roads that cannot go 13 miles without safety interventions is reckless endangerment.

      No, that is a risk. Not mittigating the risk is reckless endangerment. But the risk was mittigated in the same way that other companies mittigate it. i.e. standard practice, thus not neglicence.

      The buck has stopped. Knowing statistics doesn't make you liable for anything. Claiming that someone knows a statistic is not the same as providing proof beyond doubt.

      You are citing a lot of moral obligations, but no legal ones. Morals won't get you far in a courtroom, that's not how the law works.

    119. Re: So who is to blame? by kenai_alpenglow · · Score: 1

      Didn't they jail a meteorologist in Brazil for a missed-forecast? He said it was going to rain at a big soccer match or something, and it didn't. And wasn't there that Italian geologist who missed forecasting an earthquake?

    120. Re:So who is to blame? by Mal-2 · · Score: 1

      Being vigilant is fine if you are driving. But the moment you surrender control, you also lose your connection to events. It's hard to know what exactly to pay attention to. Try being vigilant when a passenger when someone else drives, and try to remember everything you saw. Some time not too much later (within a day or two), drive the same route yourself. You'll realize that the things you paid attention to the first time hardly register when driving, and things that didn't get past your perceptual filters the first time are often the very things you need to know when driving. Situational awareness is very much altered when you're not actually in control.

      --
      How is the Riemann zeta function like Trump rallies? Both have an endless number of trivial zeros.
    121. Re: So who is to blame? by DRJlaw · · Score: 1

      There is no way that such information has been kept away from the CEO, unless the CEO is utterly incompetent. With 13 miles between interventions, it was only a matter of time before a safety driver failed to step in.

      By your logic, it is criminally negligent homicide to allow any human being to drive a car that causes an incident resulting in death. It is only a matter of time before they fail to properly "step in" (i.e., manually perform) a driving action. Therefore the CEO of every car manufacturer everywhere is guilty of criminally negligent homicide, and anyone who allows another to drive their car would be guilty of such if their car causes an incident resulting in death.

      You, my friend, are ignoring the requirement of "proximate causation." If you want to allege that the CEO has personal criminal responsibility, then the CEO's personal actions must be the proximate cause of the death. Not an engineer's decision concerning the threshold at which the algorithm or AI outputs an action to avoid a obstacle, and not a safety driver's decision to play with their phone while in motion.

      In addition, criminally negligent homicide requires more than mere negligence. The action must be inherently dangerous or reckless (small problem -- governments routinely require autonomous vehicle testing platforms to use safety divers, so one so equipped is not inherently dangerous or recklessly deployed) and the CEO him or herself must have known that his or her conduct was "a threat to the lives of others."

      Allowing a self driving car on public roads with 13 miles between interventions is reckless endangerment.

      Not buying that. We haven't personally prosecuted auto CEOs for deploying cars with cruise control, and the intervention rate for that technology is no better.

      Try harder.

    122. Re:So who is to blame? by Anonymous Coward · · Score: 0

      If the maximum time to be attentive is short, then Uber should be swapping the drivers over more regularly. That costs money though, and one thing Uber have demonstrated they've very good at is cutting costs. Perhaps they cut too far this time?

    123. Re:So who is to blame? by Anonymous Coward · · Score: 0

      Wrong. The driver / safety operator had far less time than that, due to being absorbed by his mobile phone.

    124. Re:So who is to blame? by bluefoxlucid · · Score: 1

      riddled with the kind of ambiguity

      Madison wrote that many things are definitely in and not in the constitution, and that many other things written therein do not have a definite and obvious meaning. His immediate next statement was an enormous run-on paragraph that basically said "good luck with that".

      1) What constitutes necessity for the security of the public and who will determine this? Is if sufficient to confine a person if one of person greatly fears for their life and the public at large is under no additional danger in any way?

      We could define this, but the definition will actually change over time as we get better at understanding the risks. We have a parole board which does risk-based analysis of whether a person is liable to re-offend, which changes as they get more experience. The practical outcome of this sort of ambiguity has been that the Supreme Court has clarified things in the Constitution to mean different things--either more-precise or even completely-reversed--as the circumstances of our Nation have changed.

      2) What is meant by the greatest extend achievable? If you want to throw an infinite amount of money at the system, there's quite a bit that could be achieved. Realistically, who gets to determine what's achievable with the meager budget that the system does get?

      That's a good question. The alternative language is "reasonable", which tends to allow a lot more negligence; and "reasonably achievable", which directly acknowledge that what is "achievable" is dependent on what will not otherwise harm your society. That tends to be an economics consideration we can't possibly define.

      Throwing infinite money at a system isn't achievable. Likewise, improving a system to perform more-efficiently achieves more with less economic cost. The raw "achievable" statement puts one hell of a lot of pressure behind moving forward, and doesn't require an exhaustive effort in futility; but it is a huge hammer that will tend to push hard, bordering on unreasonable, and not likely into the ludicrous. The likely outcome is occasional growing pains when it becomes well-known a system can be established which is better, but the transition itself is a taxing but not entirely self-destructive effort.

      The alternate wording "reasonably achievable" may be a little more lenient; do we want to be lenient, or do we want to be hard-lined on issues of human rights such as this?

      3) What does dignity as human beings mean? Comes across as useless platitudes that mean whatever the reader wants it to mean.

      It's actually a well-understood philosophical concept. I also lifted it from other sources.

      4) What needs are not currently being met and who gets to determine what a person's needs are? If someone really needs a good lay, is society on the hook for the hooker?

      Well, in prisons where I live, women often do not have access to sanitary napkins and instead use a rolled-up sock. They frequently don't have well-established laundry service in prison and may wear the same underwear for several days in this condition. The prison environment isn't always clean, access to natural light and fresh air may be restricted, contact with the outside world may be diminished to near-nothing. People are brought in with opioid addictions and left to sleep on a concrete floor with no bed, vomiting and shitting themselves for three or four days while central booking waits to deal with them--this is actually an enormous violation of Nelson Mandela Rule 30(c), which actually requires treatment as soon as possible for any prisoner who arrives with a risk of drug withdrawal.

      It is imperative that every individual prisoner retains a sense of security, independence, and dignity. Their medical and social needs must be met. The particular requirements to rehabilitate an individual will vary, and this variation must b

    125. Re: So who is to blame? by amorsen · · Score: 1

      The other companies did not face the same risk, and so their standard practices do not apply. The other companies required 2 orders of magnitude fewer interventions. The software deployed by those companies can reasonably be considered fit for purpose.

      One of the ways to deal with such horribly bad software would be to let the human driver be in control at all times, with the software only learning. That is how the other companies achieved a decent safety standard, before they switched to safety drivers.

      Heck, Uber did not even have the system check if the safety driver was doing their job! Again, common practice in other industries where safety operators are in use. It is difficult to believe the level of malice shown by the company.

      --
      Finally! A year of moderation! Ready for 2019?
    126. Re: So who is to blame? by amorsen · · Score: 1

      By your logic, it is criminally negligent homicide to allow any human being to drive a car that causes an incident resulting in death. It is only a matter of time before they fail to properly "step in" (i.e., manually perform) a driving action. Therefore the CEO of every car manufacturer everywhere is guilty of criminally negligent homicide, and anyone who allows another to drive their car would be guilty of such if their car causes an incident resulting in death.

      The human is in charge. They are responsible, they are the ones the law applies to. If a human driver works for a company and the company knowingly lets particularly unsafe drivers (i.e. inadequately trained or drivers with inhibited judgement) continue driving, the company is liable for their actions. In this case the driver was a computer, and the company knew the driver was the equivalent of inadequately trained -- it couldn't even manage 13 miles between interventions.

      A car manufacturer has nothing to do with how a regular car is driven, and therefore it is not their responsibility if someone mows pedestrians down. This is likely to change in the future as safety features to prevent exactly that will be required to be installed, and if they do not perform to standard, the company will likely be responsible. But we are not there yet for manually driven cars.

      You, my friend, are ignoring the requirement of "proximate causation." If you want to allege that the CEO has personal criminal responsibility, then the CEO's personal actions must be the proximate cause of the death. Not an engineer's decision concerning the threshold at which the algorithm or AI outputs an action to avoid a obstacle, and not a safety driver's decision to play with their phone while in motion.

      This defence is not working for VW regarding the emissions fraud. Let us hope it does not work for Uber either.

      --
      Finally! A year of moderation! Ready for 2019?
    127. Re:So who is to blame? by Anonymous Coward · · Score: 0

      The Uber video is grossly misleading. I'm not sure if they just had a shit camera or if they edited it to mess with the contrast, but:

      1. That area looks MUCH more well-lit in videos that third parties have taken.
      2. Braking distance at 38mph is 72 feet. Braking time at 38mph is 1.3 seconds. AZ law requires the headlights to illuminate people from 500 feet away. If it was truly "dark enough" that she 'came out of a shadow", then that car wasn't even street legal!

    128. Re: So who is to blame? by DRJlaw · · Score: 1

      The human is in charge. They are responsible, they are the ones the law applies to. If a human driver works for a company and the company knowingly lets particularly unsafe drivers (i.e. inadequately trained or drivers with inhibited judgement) continue driving, the company is liable for their actions. In this case the driver was a computer, and the company knew the driver was the equivalent of inadequately trained -- it couldn't even manage 13 miles between interventions.

      But this car had a human driver. You've conveniently ignored that point, yet declared that the human is in charge (as this one was), and is responsible, and is the one that the law applies to. Ergo, The driver human, not the CEO human, is more proximate to the cause.

      A car manufacturer has nothing to do with how a regular car is driven, and therefore it is not their responsibility if someone mows pedestrians down.

      Really? A car manufacturer has nothing to do with how its cruise control operates? A cruise control that, for example, fails to detect that the vehicle in front has slowed or that a pedestrian has entered the road? No relationship at all?

      This defence is not working for VW regarding the emissions fraud. Let us hope it does not work for Uber either.

      The situation there is the U.S. government has documents showing the the CEO personally and specifically approved of deceiving federal regulators as to the existence of the device. He is charged with "conspiracy to defraud the US government and customers, wire fraud, and conspiracy to violate the Clean Air Act," not violating the Clean Air Act himself.

      You have presented nothing comparable for the Uber CEO. "It couldn't even manage 13 miles between interventions," paired with a safety driver, was deemed within the duty of care of the California and Arizona governments prior to the accident. You have not even shown that anyone at Uber knew that the safety driver was inadequately trained or had inhibited judgement... you've simply thrown out the fact that they were present.

      So again, try harder.

    129. Re:So who is to blame? by Raenex · · Score: 1

      That area looks MUCH more well-lit in videos that third parties have taken.

      Depends on camera settings, other traffic in the area, and functionality of the headlights in use. Also depends on the exact location, as some parts are more illuminated than others.

      AZ law requires the headlights to illuminate people from 500 feet away.

      I have been unsuccessful in locating the exact text of the law, but what I could find from a 3rd party website says headlights must be in use when visibility is below 500 feet:

      "Headlights are required from sunset to sunrise and when visibility is less than 500 feet."

    130. Re: So who is to blame? by BranMan · · Score: 1

      Sorry - have to chime in here. In my experience to date, braking is usually the best way to handle it when things go really wrong on the road. But not always - if the brake pedal is the only thing you ever use, you are in trouble.

      Sometimes acceleration is the only way to avoid serious harm. I can tell you that from personal experience - after dropping off a date around midnight or so, I was cruising through a traffic light at a 4 way intersection. Maybe around 30 mph, just cruising along - the light was green for at least 10-15 seconds before I got to it (which means it was solid red in the other direction for at least that long).

      Out of nowhere I have headlights in my face - just a few feet to my left. Hitting the brakes would have sent me to the hospital (or the morgue) from getting T-boned through my drivers side door. I slammed on the gas pedal, my car (thank god) leaped ahead down the road, and a car, taking a left turn, ignoring their red traffic light entirely, turned into my lane not a foot past my rear bumper. They took off down the first side street while I was still prying my hands from the wheel.

      To top it off, this was an old car, and had been giving me trouble, getting bogged down, stalling or nearly stalling at times, would not respond to the gas pedal at times. Just balky in general. I'm glad it responded when I needed it to. Next day I went shopping for a new car - I was not going to tempt fate again with an iffy vehicle.

      Braking is not always the right answer. Accelerating isn't either - you need to be open to all options, all the time.

    131. Re:So who is to blame? by Maxo-Texas · · Score: 1

      A little more rational than the anonymous crazy person.

      So here's a link.

      https://www.google.com/maps/@3...

      The uber car hit the pedestrian about 40' past the white sign on the right.

      You'll note the road widens from 2-4 lanes at the white sign and there is heavy landscaping on the median.

      There are also multiple signs telling pedestrians to *not* cross there but there is an inviting paver path there and no railing/barrier.

      When this accident first occurred the media was full of reports about how the car was speeding when backing up only a little bit to just before the prior overpass you can see a sign that says 45mph.

      ---

      I am not a fan of Uber. I think they are scammy and I believe they ignore the law and they lie.

      But I'm also not a fan of *emotionally* processing the event and making a decision before the facts are in like the screaming anonymous coward above.

      I have some friends in the a.i. car industry and they think uber's a.i. solution is less robust than some others. But... they are going to be suspect because... they don't work for uber and they do work for uber's competition. Still data to consider.

      I'm a strong proponent of sealed black boxes which gather high quality video and all relevant data about car speed, position, acceleration, braking, etc.

      My point isn't that the bushes on the left and the fact the road widens from 2 lanes to 4 lanes and the fact that light maps show the area is slightly darker explain away the death of a pedestrian who ignored multiple warning signs and who failed to see the car approaching her and in the video wasn't even looking in the direction of oncoming traffic.

      My point is we should consider the facts, analyze the data, *genuinely figure out what went wrong so we can improve it* and avoid emotional histrionics because they interfere with the ability to think rationally.

      With a sub point that if A.I. kills 300 pedestrians instead of the roughly 2000 pedestrians now being killed, then that is 1700 lives better and *perfect* is impossible.

      That's why I said, "improve" instead of "fix". It will never be perfect when interacting with random events. But it will eventually be much better than humans.

      --
      She was like chocolate when she drank... semi-sweet at first and then increasingly bitter.
    132. Re:So who is to blame? by RhettLivingston · · Score: 1

      No, the driver is the responsible party. This is very much like a driver's education situation in which there are two sets of controls in the car. Ultimately, the teacher is the responsible party. Any accident is the teacher's fault.

    133. Re:So who is to blame? by Raenex · · Score: 1

      You're ignoring the fact that the person that was hit WALKED OUT IN FRONT OF TRAFFIC.

    134. Re:So who is to blame? by amicusNYCL · · Score: 1

      The only way to know for sure what people could see and when is to test under the same conditions.

      Good idea. The impact happens at about the position where the car is at 0:32 in that video. Note that before they even go under the highway they can clearly see what's beyond the overpass. The purple lights on the left are the concert venue. That video was recorded 3 days after the accident at around the same time, the weather reports for both days are substantially similar. A couple weeks ago my car was parked under that overpass and I was looking at the entire area. What you see in that video is much more representative of how normal human eyes see that area than the dark Uber video.

      Of course, this all ignores the fact that the car was equipped with a variety of sensors which should have and did identify the person regardless of light.

      --
      "Our two-party system is like a bowl of shit looking at itself in a mirror." - Lewis Black
    135. Re: So who is to blame? by nasch · · Score: 1

      Yes, I think some scientists were jailed in Italy for incorrectly predicting there would not be an earthquake. Which is a great recipe for making sure scientists in Italy don't make any predictions about natural disasters.

    136. Re: So who is to blame? by Anonymous Coward · · Score: 0

      Blindly following best/standard practice is negligence. So sayeth every good programmer.

    137. Re:So who is to blame? by Raenex · · Score: 1

      The impact happens at about the position where the car is at 0:32 in that video.

      I actually think it's closer to 00:34 seconds. Judging from the Uber video, it's right past the white sign. You can see the shadow area in the distance at 0:30 seconds just after the underpass. So assuming this was a human driver, and they were looking that far ahead, and they could actually see the person in the left lane walking into the right lane, they would have about 4 seconds to stop the car.

      What you see in that video is much more representative of how normal human eyes see that area than the dark Uber video.

      Probably. Then again, it's also possible the video you link is a bit brighter than what human eyes would see. Contrast in dark settings is a complex subject. The only real test would involve recreating the conditions with human drivers, with random tests with a dummy person there and not there.

      Of course, this all ignores the fact that the car was equipped with a variety of sensors which should have and did identify the person regardless of light.

      Yes, I agree, the self-driving car should have performed better than a human because of those sensors. But all the outrage is fixed on the idea that a human at the wheel should have stopped in time. From that angle, this person walked out in front of traffic and got herself killed.

    138. Re: So who is to blame? by thegarbz · · Score: 1

      The other companies did not face the same risk, and so their standard practices do not apply.

      Don't use an article to start your sentence. It doesn't help your case when you compare one company to all companies. Instead realise that there are actually many car companies out there with self driving cars, and many of which have the same standard setup, with same hardware also being tested on roads, which ultimately for some of them does make them face exactly the same risk.

      If the risk is the same than standard practices most definitely do apply. That is the very definition of not being negligent.

      The other companies required 2 orders of magnitude fewer interventions.

      That is operational data during experiment, not the risk of running the experiment. Incidentally your argument is that Uber has double the experience of successful interventions and the court would just love to see that as an example of a company that has a mitigation in place for their risk. Successful interventions at a high rate is an incredibly strong defense against negligence.

      Heck, Uber did not even have the system check if the safety driver was doing their job! Again, common practice in other industries where safety operators are in use.

      Does the competition? Please show me this. I haven't heard of any of the competition use any kind of alertness detection on their safety drivers. I have on the other hand seen camera interviews from within Wayemo cars where the safety driver spends a lot of time not looking at the road. Heck the system you're proposing is barely rolled out to normal drivers and also not required by the regulator. Speaking of: Meeting regulation requirements is another defense against negligence.

      To be clear about this: Uber killed someone. Their process was wrong. Their CEO is morally in the shit for his rushed push into the field. But you are miles away from proving criminal negligence against the CEO. This wouldn't make it past a summary judgement against you in court. There may be some people who could be criminally held liable for something, but ultimately the one thrown under the bus there will most likely be the safety driver who wasn't doing their job since working for another company is actually NOT a defense against negligence.

    139. Re:So who is to blame? by grahamsz · · Score: 1

      I'm not saying they are useless - i'm saying that even after completing them we'll still encounter situations where people die because the system failed.

      That's true of many other things too - aircraft systems still fail, hospital equipment still fails, abs systems fail...

    140. Re: So who is to blame? by Demena · · Score: 1

      I do not know about that. I have lived quite long. I have done far more miles both in cars, trucks and on motorcycles than most will in their lives. But I learnt in the days when people were taught to drive, not just sit in a car. You learnt traction on skid pans and wet grass, braking with one or more flat tyres. Many of your elder drivers have quite literally forgotten more than you will ever know.

      I made the post on this subthread to stress evaluating the situation before reflexively breaking. Once your foot hits the break and you shed velocity you cannot pick up velocity again quickly. This limits (or even commits) you to fewer choices, gives you fewer places to put the car.

      One thing to be strongly avoided - driving an underpowered car. Even in cities it is easier to get out of someone's way when they are being an idiot.

    141. Re: So who is to blame? by Demena · · Score: 1

      You are of course correct; it is changes in velocity that cause the majority of accidents. but it does also depend on the driver somewhat (see below).

  3. False Positive? by Anonymous Coward · · Score: 0

    Is that like a True Negative? Seems not. Bonkers! Daft! Never gonna not kill a LOT MORE THAN A DRUNK!

  4. The War Between Man and Machine by DoktorMidnight · · Score: 1

    It begins....

    1. Re:The War Between Man and Machine by SeaFox · · Score: 4, Funny

      Yeah, really. How do they know the vehicle "ignored" the woman? Maybe it just acted like it didn't recognize it needed to take action when it really was targeting her.

    2. Re:The War Between Man and Machine by TimMD909 · · Score: 1

      It begins....

      Man and machine? The driver identifies as "Rafaela", so maybe it should be woman vs machine? Or peoplekind vs machinekind?

    3. Re:The War Between Man and Machine by Rick+Schumann · · Score: 1

      No, you have it wrong. This is the beginning of the War Between Citizens and Shitty Corporations' Marketing Departments Who Are Rushing A Half-Assed Product To Market That's Killing People.

      I can't wait for them to decide enough is enough and they ban these from public use. We need REAL AI not this half-assed pseudo-intelligence 'machine learning' crap. What we really need is better driver education, training, and testing, and stricter penalties for bad drivers, up to and including revoking their driving privilege permanently when they demonstrate they're incorrigibly incompetent. No need to make ALL drivers pay the price for SOME BAD drivers.

    4. Re:The War Between Man and Machine by Anonymous Coward · · Score: 0

      That makes sense. Uber's mysoginistic tendencies got programmed into the car as well!

      That's the only possible explanation!

  5. Oh good. by Ichijo · · Score: 3, Interesting

    The autonomous programming detects items around the vehicle and operators fine-tune its sensitivity to make sure it only reacts to true threats (solid objects instead of bags, for example).

    Then it's an easy fix. Just move the "sensitivity" slider a little to the left.

    Actually, it's kind of terrifying that all that stands between life and death is a sensitivity setting.

    --
    Any sufficiently unpopular but cohesive argument is indistinguishable from trolling.
    1. Re:Oh good. by cascadingstylesheet · · Score: 2

      The autonomous programming detects items around the vehicle and operators fine-tune its sensitivity to make sure it only reacts to true threats (solid objects instead of bags, for example).

      Then it's an easy fix. Just move the "sensitivity" slider a little to the left.

      Actually, it's kind of terrifying that all that stands between life and death is a sensitivity setting.

      Absolutely.

      I hope they use material design so the settings are all hard to see.

      And I really hope it's totally ambiguous whether you have to click Save, or if the changes to the Sensitivity slider will just save automatically, just because you touched them or something.

    2. Re:Oh good. by Carewolf · · Score: 1, Interesting

      The autonomous programming detects items around the vehicle and operators fine-tune its sensitivity to make sure it only reacts to true threats (solid objects instead of bags, for example).

      Then it's an easy fix. Just move the "sensitivity" slider a little to the left.

      Actually, it's kind of terrifying that all that stands between life and death is a sensitivity setting.

      It is not the setting that is the problem. The problem is socalled AIs with less intelligence than a cockroach being put behind the wheel of cars.

    3. Re:Oh good. by jellomizer · · Score: 2

      That is general true about the world.
      If you are a driver and you are distracted or not fully focused on the world around you, your sensitivity setting is just off too. The biggest reason why Motocycles get in accidents is because automobile drivers fail to see them, just because they may not be expecting a Motocycle, so their eyes are on the look out for fast moving objects that fill up at least 2/3 of the lane. This fact that we fail to comprehend things that we don't expect is how magicians trick us to see these magic tricks. Our brain takes a lot of shortcuts in processing the world the same way that the Uber Car decided not to deal with processing the women.

      When growing up as a kid, We needed to learn what is acceptable and what isn't. Growing up we have said and done things that had hurt people and ourselves. We learned from it so we changed our settings, if we went too far, then we found we couldn't get things done, so we more or less settled around some happy medium that works for our lifestyle.

      --
      If something is so important that you feel the need to post it on the internet... It probably isn't that important.
    4. Re: Oh good. by Anonymous Coward · · Score: 0

      A proper non AI system would have stopped in time.

      Uber is a shit company, that's all.

    5. Re: Oh good. by Anonymous Coward · · Score: 0, Troll

      When good automotive engineers grow up, they are taught to apply science instead of bankster BS such as AI.

      So they set up hundreds of test cases and check their code to properly deal with each one. Only then they release their system into public roads.

      That is how Bosch or Thyssen Krupp develops software controlled servo steering systems. How Daimler develops their various assist systems.

      Uber is a bunch of posers with too much money to burn.

      The killing is an almost classic test case. It can be easily sumulated with a cable pulled dummy pedestrian. Why not done????

      I talked to a Daimler autopilot researcher more than 20 years ago. He said they would not use AI because you can't make engineering predictions about its behavior. AI is a bunch of shit tricks which are wholly unsafe and opaque. Good for impressing politicians and other stupid laymen, useful for harmless finance voodoo. Dangerous in life critical software.

    6. Re:Oh good. by omnichad · · Score: 1

      is because automobile drivers fail to see them, just because they may not be expecting a Motocycle,

      The reason I didn't see them is because they were coming up between two lanes of moving traffic. You're right, I am not set to expect that.

    7. Re:Oh good. by goose-incarnated · · Score: 1

      Just move the "sensitivity" slider a little to the left.

      Actually, it's kind of terrifying that all that stands between life and death is a sensitivity setting.

      There is no "correct" setting for sensitivity because the software is broken. Where it is right now is both "incorrectly classifying a safe situation as dangerous" AS WELL AS "incorrectly classifying a dangerous situation as safe" (probabilities apply).

      Which way do you want the slider to move? It's already too far from the correct position for both classifications.

      --
      I'm a minority race. Save your vitriol for white people.
    8. Re:Oh good. by Anonymous Coward · · Score: 0

      This is par for the course in detection problems. You have some baseline "noise level" in the system that's often due to electronic noise but in this case would also include small objects in the environment that generate a weak return on the LIDAR. The detection threshold (or sensitivity setting if you prefer) is set just above this threshold in order to satisfy the tradeoff between detection failures (what happened here) and false alarms.
      https://en.wikipedia.org/wiki/Constant_false_alarm_rate

    9. Re:Oh good. by Tablizer · · Score: 1

      fine-tune its sensitivity to make sure it only reacts to true threats (solid objects instead of bags, for example).

      it's kind of terrifying that all that stands between life and death is a sensitivity setting.

      Siri: "Sorry, Mam, I thought you were an old bag."

      Seriously*, I've slammed on the breaks for things that turned out to be harmless debris. The problem is that slamming on the breaks has risk also.

      For one, the passenger could get whiplash or similar, especially if they are not in the best of shape. If they are injured due to a plastic bag on the road, lawsuits will fly.

      Second, cars behind may end up plowing into your rear if they don't stop in time.

      It's not always an easy decision and humans also get it wrong. At least the vehicle should have compromised and applied light breaks, giving the pedestrian better odds of survival.

      My breaking "algorithm" is similar: the proportion of pressure I put on the breaks is proportional to the perceived importance of the object (and/or risk to the car itself).

      * No pun intended

    10. Re:Oh good. by drinkypoo · · Score: 1

      Second, cars behind may end up plowing into your rear if they don't stop in time.

      This is of course only an argument in favor of AV with V2V...

      --
      "You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
    11. Re:Oh good. by Anonymous Coward · · Score: 0

      Actually, it's kind of terrifying that all that stands between life and death is a sensitivity setting.

      That's.... almost philosophical. First off, of course there's a hell of a lot more than just one sensitivity setting. A lot of things had to line up to make this happen. Lighting, inattentive pedestrian, and shape of the road, signs about the rules for crossing, medians, lane markers, jay-walking laws. From the video, it looks like the driver got lucky and wasn't glancing down at their phone at the crucial time, but it didn't help. So there's actually quite a bit between you and death. Typically. But driving is so common and the population is so big, things ARE going to align just so.

      Consider the sensitivity of a hand-gun. If some asshole at the range sweeps you, trigger discipline or no, it's the trigger's sensitivity that stands between you and death. And that's a setting that gets engineered and designed and built into the spring in the factory. Or think about any time a cop gets spooked and draws his weapon on someone. The trigger's sensitivity is the thing standing between a life and death.

      There's sensitivity when it comes to how quickly doctors recommend screening for cancer. It used to be they suggested all women go get a breast exam, but they're pulling back on that one from all the real negative consequences from that. And now because they're less sensitive, they are going to miss someone's cancer and that's going to kill them. There's that sort of sensitivity setting for every ailment and test.

      Engineering ideal settings on sociological scales is really brutal. Like the protagonist in Fight Club. At a certain point it's cheaper to not do a recall.

    12. Re:Oh good. by Anonymous Coward · · Score: 0

      Then it's an easy fix. Just move the "sensitivity" slider a little to the left.

      Typical ignorant manager answer.

      A true programmer would have called doWhatIWant(). Setting the killSlider doesn't have the fine tuning required.

    13. Re:Oh good. by Moof123 · · Score: 1

      I don't envy those setting that slider. There is likely are pretty decent range of scenarios where you are in a no-win position.

      A rock and wadded up paper grocery bag look the same on lidar. Hit a rock and it will be a headline. Panic stock for a paper bag and it will be a headline.

      From an intellectual stand point you'd rather have a false panic stop than a dead homeless lady. From a marketing standpoint you don't want an overly cautious car giving Uber-AI riders whiplash on a daily basis.

      In the end I am still very bearish on autonomous cars. The problem of meeting unreasonable sets expectations that exceed the expectations placed on human drivers simply don't look attainable. Getting a system to the 90% point was relatively easy as demonstrated by Google's Prius fleet years and years ago. Getting it to the 95% is hard. Getting it that last little bit good enough to handle everything everywhere without fail looks impossible. Until expectations decline, I just don't see mass adoption for personal vehicles in the next decade.

    14. Re:Oh good. by Moof123 · · Score: 1

      Or for not tailgaiting in the first place.

    15. Re:Oh good. by Ichijo · · Score: 1

      Or AV that don't tailgate!

      --
      Any sufficiently unpopular but cohesive argument is indistinguishable from trolling.
    16. Re:Oh good. by wisnoskij · · Score: 1

      OK, so stop for adults but not for dogs, children, or midgets.

      --
      Troll is not a replacement for I disagree.
    17. Re:Oh good. by just+another+AC · · Score: 1

      It is not the setting that is the problem. The problem is socalled AIs with less intelligence than a cockroach being put behind the wheel of cars.

      Well they did pass their driving test, so there's not much we can do about that... oh wait you meant the driverless cars.

    18. Re:Oh good. by just+another+AC · · Score: 1

      My breaking "algorithm"

      You mean the algorithm you use to determine if you are going to break pedestrians with your car?

    19. Re:Oh good. by Anonymous Coward · · Score: 0

      The autonomous programming detects items around the vehicle and operators fine-tune its sensitivity to make sure it only reacts to true threats (solid objects instead of bags, for example).

      Then it's an easy fix. Just move the "sensitivity" slider a little to the left.

      Actually, it's kind of terrifying that all that stands between life and death is a sensitivity setting.

      What is terrifying is that the Uber programmers clearly read stupid comments on slashdot about hypothetical scenarios.
      It is never OK to run over Schroedingers bag on the road. You don't know if it hides a rock or not until you run it over.

      Obstructions that is OK to run over doesn't happen. If they do happen it is perfectly fine to stop and require the passenger to manually drive over it.

    20. Re:Oh good. by thegarbz · · Score: 1

      Actually, it's kind of terrifying that all that stands between life and death is a sensitivity setting.

      This has been the case since automated safety systems were first invented. It is terrifying only if you look at the cases where it goes wrong in isolation. Very few people talk about the number of times lives get saved due to a simple control system here which none the less needs a bit of tuning.

    21. Re:Oh good. by Cederic · · Score: 1

      Obstructions that is OK to run over doesn't happen.

      I slowed to almost a halt this morning (at 2am) to avoid two rabbits in the carriageway.

      Had there been cars behind me, I'd have flattened the stupid things.

      I also recklessly drive callously through plastic bags being blown across the road by the wind. I'm a right bastard.

  6. There's only one logical conclusion... by The+Fat+Bastard · · Score: 1

    The Uber vehicle mistaken the woman for an old bag.

  7. Too large! by HornWumpus · · Score: 4, Interesting

    I understand that programmatically telling a blowing plastic bag from a child's toy is difficult.

    But she (and her bike) were clearly large enough to damage the vehicle. Even if the code saw her as debris, the car should have avoided it.

    I think the code had to have dismissed her as lens flair or something similar.

    --
    John McAfee 'It was like that time I hired that Bangkok prostitute; to do my taxes, while I fucked my accountant'
    1. Re: Too large! by Anonymous Coward · · Score: 1

      I wish these self driving cars would have an alarm or something, in the even there's a grey area for action.

    2. Re:Too large! by Xylantiel · · Score: 1, Insightful

      But that's the danger of machine learning, which seems to pass for AI today, you often don't really know why it does anything that it does. Makes it difficult to test for correct function to say the least. I guess this test failed. Maybe they shouldn't be testing this on public streets.

    3. Re:Too large! by LynnwoodRooster · · Score: 1

      I think the code had to have dismissed her as lens flair or something similar.

      Damn you Michael Bay!

      --
      Browsing at +1 - no ACs, I ignore their posts. So refreshing!
    4. Re:Too large! by Anonymous Coward · · Score: 0

      Bay is the explosion guy. I think you mean Damn you J.J. Abrams.

    5. Re:Too large! by Anonymous Coward · · Score: 0

      How does machine learning differ from human learning then?

    6. Re:Too large! by Narcocide · · Score: 0

      When a human driver kills someone, there is an obvious chain of responsibility and simple target of blame.

    7. Re:Too large! by Mal-2 · · Score: 1

      If I were driving (not "safety driving"), I might have chosen the "brake hard but don't dodge" option. Sucks for her if I can't stop in time, but I'm not going to put myself directly in someone else's path, to get tossed around like a pinball, to avoid someone who should not have been there. Now if I was sufficiently situationally aware (as I try to be) to already know that an adjacent lane was open, then I would have taken it. But if, as far as I knew, there was no place safe to dodge, then I wouldn't have dodged. I'd rather hit one person than start a chain reaction.

      --
      How is the Riemann zeta function like Trump rallies? Both have an endless number of trivial zeros.
    8. Re:Too large! by Anonymous Coward · · Score: 0

      There are plenty of subsystems in a car that works independently and doesn't take over driving.

      Automatic braking is one system that the self driving part doesn't have to take control over.
      It just sits there and watches the lidar and brakes for obstacles.
      It doesn't need to know what it brakes for, just that something is in the way and that you are driving towards it too fast.
      Is it a blowing plastic bag? It doesn't matter, slow down and let it pass. You don't know what is behind it.

      There is no situation where the self-driving software should be able to override the automatic braking.
      Rather it should just try to drive in a way that makes it unnecessary for that subsystem to step in.

      You can't write software that is compliant with the safety standards. You always has to write at least two.

    9. Re:Too large! by Bongo · · Score: 1

      Humans know a lot more about the world. If you are not sure what your eyes are telling you, you can start making educated guesses, using whaever else you know about the world. Nothing is just an object, rather, everything has associations and context.

    10. Re:Too large! by Anonymous Coward · · Score: 0

      Stopping distance at 38mph is 72 feet. Add another 40 for human reaction time. Headlights by AZ law must reveal at least 500 feet ahead of you.

      She would have been fine.

  8. Not si simple... by Anonymous Coward · · Score: 0

    Tune it the other way and you get cars braking for no reason, potentially causing crash.

    Itâ(TM)s just classificatiin. It has all the usual properties: accuracy, specificity and sensitivity.

    The learning point here is:donâ(TM)t take a self driving uber

    1. Re:Not si simple... by I4ko · · Score: 2

      If people follow at appropriate distance for the speed, no crash will happen.

    2. Re:Not si simple... by Anonymous Coward · · Score: 0

      If people follow at appropriate distance for the speed, no crash will happen.

      I live in Washington State with recent strict anti texting laws. Has done nothing to curb cellphone use while driving. It's amazing, Nobody cares. Facetube posts are still more important than human life.

    3. Re:Not si simple... by bobbied · · Score: 1

      If people follow at appropriate distance for the speed, no crash will happen.

      I live in Washington State with recent strict anti texting laws. Has done nothing to curb cellphone use while driving. It's amazing, Nobody cares. Facetube posts are still more important than human life.

      Ah... Yes.. The conservative argument creeps up on you doesn't it...

      Before we did this same law here, my argument was two fold.. 1. It's unenforceable so nobody will follow it because you simply cannot make the fines high enough to make an impact and low enough to not bankrupt most people... 2. We already have laws about accidents due to distracted driving and liability so this new one is unnecessary.

      A conservative asks the questions "is this necessary and if it is, is the law the least possible, and can we actually enforce this?" Any place we see "no" as the answer, we stop and say "no" .

      Personally, I think it would have been much more effective to institute a "Don't text and Drive" campaign and make it socially unacceptable to do. Making it illegal really just encourages folks to do it anyway. Don't think so? Tell me, what do you do when you see a "Wet Paint, don't touch!" sign? You know you want to touch it...

      --
      "File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
  9. Uber and people who authorized this experiment by goombah99 · · Score: 5, Insightful

    This is a clinical trial. The FDA has long long long long long experience in conducting clinical trials. Now one can argue if FDAs caution is too much but even in the worst case everyone would agree they have a well established process for assuring something is safe and effective before you release it onto the public.

    Uber is conducting experiments on the public.

    If this were a new drug or treatment or medical procedure they would be shut down.

    This is actually far worse than that because most new drugs or treatments have clear lineages from prior ones that give us high expectations of what the outcome will be.

    The argument that something has to be allowed prematurely because in the long run it will save lives is a failed argument for medicine.

    In this case there is nothing to support the claim that this will save lives in the long run. Sure one could imagine that it would. But I don't think thats very well established. And if this were a drug study people would have spent the time and money to establish that.

    The claim that they have conducted 5 million miles (or whatever of testing) is rubbish. Those are not statistically valid tests. We execs dashing in front of the cars going 50 miles per hours in any of those tests? I assure you that did not happen.

    Moreover we already have evidence from those tests that driver re-aqusitions do happen frequently, and there is a substatnial lag in the hand over dues to human inattention. THe fact that they only had one driver in it says Uber is negligent.

    --
    Some drink at the fountain of knowledge. Others just gargle.
    1. Re:Uber and people who authorized this experiment by Anonymous Coward · · Score: 0

      This is probably the most insightful comment I've read about autonomous vehicles. I think you're right that the closest model we have to autonomous vehicles is the FDA. And all these companies are all experimenting on everyone in the state where they're allowed.

      Thanks. That's a good way to look at it.

    2. Re:Uber and people who authorized this experiment by Locke2005 · · Score: 2

      It's not a clinical trial unless all the "patients" give informed consent. This is testing on people that have not be forewarned and have given zero consent. That is a very different animal.

      --
      I've abandoned my search for truth; now I'm just looking for some useful delusions.
    3. Re:Uber and people who authorized this experiment by Catbeller · · Score: 1

      So are all the other "autopiloting" car manufacturers. And drive-by-computer, aka no-mechanical links to the brakes, steering and throttle are also robotic, which will bite us some day.

    4. Re:Uber and people who authorized this experiment by Anonymous Coward · · Score: 0

      So are all the other "autopiloting" car manufacturers. And drive-by-computer, aka no-mechanical links to the brakes, steering and throttle are also robotic, which will bite us some day.

      At some point, people will have to realize, that stepping foot on a hwy, regardless of how cars are piloted, will be dangerous.

    5. Re:Uber and people who authorized this experiment by Anonymous Coward · · Score: 0

      Yes, that was his point. It's like a clinical trial performed on the general populace without their consent, which anyone who isn't a psychopath can see is extremely unethical.

    6. Re:Uber and people who authorized this experiment by barc0001 · · Score: 1

      >Uber is conducting experiments on the public.

      At some point you have to test tech like this in the real world.

      >If this were a new drug or treatment or medical procedure they would be shut down.

      Would it? If this "new drug" had the potential even with a couple of side effects to replace or supplant a known drug that was already killing 40,000 people and maiming hundreds of thousands in the US alone per year?

    7. Re:Uber and people who authorized this experiment by barc0001 · · Score: 1

      > This is testing on people that have not be forewarned and have given zero consent

      Have you given your consent to the guy down the street having his first epileptic seizure while driving past your kids playing?

      Framing self driving car tests in drug trial language is useless.

    8. Re:Uber and people who authorized this experiment by amicusNYCL · · Score: 1

      If this were a new drug or treatment or medical procedure they would be shut down.

      I'm not sure how you can draw an analogy there. In a clinical drug trial, the drug doesn't go out and kill someone not part of the trial.

      --
      "Our two-party system is like a bowl of shit looking at itself in a mirror." - Lewis Black
    9. Re:Uber and people who authorized this experiment by DesertNomad · · Score: 1

      (Should) We (have) execs dashing in front of the cars going 50 miles per hours in any of those tests?

      This is a really cool idea. It would "drive" home the point on system safety. Think how much more thought there'd be about operating when the exec's gotta put their life on the line for the work of their minions?

    10. Re:Uber and people who authorized this experiment by I4ko · · Score: 1

      Its not movie studios don't have "fake" real towns built in their premises in California. Why weren't these used to test the car with individuals who consented to be stand-ins for cyclist, pedestrians and other drivers.

    11. Re:Uber and people who authorized this experiment by stephanruby · · Score: 4, Informative

      If this were a new drug or treatment or medical procedure they would be shut down.

      Uber self-driving tests have been (mostly) shut down.

      Uber makes it sound like they suspended their testing operations voluntarily, but the fact is they lost their testing permits in Arizona, California, and one other state.

      And if there is any testing going now with Uber, it's only happening now in computer simulations, or in mocked up urban environments with fake pedestrians and bicyclists.

    12. Re:Uber and people who authorized this experiment by Uberbah · · Score: 1

      At some point you have to test tech like this in the real world.

      But before that you can do extensive testing on closed roads or courses set up for this purpose. Uber was too cheap/impatient to do so.

      Have you given your consent to the guy down the street having his first epileptic seizure while driving past your kids playing?

      TSTRT

    13. Re: Uber and people who authorized this experiment by Anonymous Coward · · Score: 0

      This is, a brilliant argument. No trolling intended. The parallels are perfect. So, instead of a NDA (New drug application) they have a NCA ?

    14. Re:Uber and people who authorized this experiment by Anonymous Coward · · Score: 0

      Yes, it is called Implied Consent. That term is mostly used with DUI laws and the power of the police to test you, however it also applies to the notion that any activity you engage in, there may be unknown factors beyond anyone's control, and you accept that risk.

      When you drive, you do so knowing any given number of people might at any given moment, lose control or fail at their job of driving safely, including but not limited to a medical emergency.

      It doesn't mean that 'fault' and liability are null and void, it is simply one of the risks YOU assume when YOU chose to drive.

    15. Re:Uber and people who authorized this experiment by barc0001 · · Score: 1

      > Why weren't these used to test the car with individuals who consented to be stand-ins for cyclist, pedestrians and other drivers.

      For the same reason "consented" individuals weren't used in a fake city during the transition from horse and buggy to car. Because it'd be insanely expensive to the point no company would bother, AND it still wouldn't replicate reality.

    16. Re:Uber and people who authorized this experiment by Anonymous Coward · · Score: 0

      A pre-existing condition is not a clinical trial. We don't expect the FDA to regulate these.

    17. Re:Uber and people who authorized this experiment by sjames · · Score: 1

      The guy down the street didn't give consent to the epilepsy either. Uber knew they were testing something new.

    18. Re:Uber and people who authorized this experiment by Locke2005 · · Score: 1

      Actually, no, We had a woman here in Beaverton who passed out, went to the hospital, the doctor told her not to drive, she decided to drive herself home anyway... and passed out again on the way home, ran up over the sidewalk and over a 17 year old kid, killing him. Legally, she was criminally responsible, because any responsible person wouldn't have taken that risk and put other people's lives in danger So what's the difference between that case and the case of a badly programming self-driving car?

      --
      I've abandoned my search for truth; now I'm just looking for some useful delusions.
    19. Re:Uber and people who authorized this experiment by Locke2005 · · Score: 1

      My coworker had epilepsy so he DIDN'T DRIVE (despite the fact that he actually owned a car). He took the bus, or when he missed the bus, asked me for a ride. Point is, HE was responsible enough to not endanger other people's lives (or his own).

      --
      I've abandoned my search for truth; now I'm just looking for some useful delusions.
    20. Re:Uber and people who authorized this experiment by bingoUV · · Score: 1

      The difference is that "Angel Investors" will not be legally defending the Beaverton woman in case of a lawsuit.

      --
      Bingo Dictionary - Pragmatist, n. A myopic idealist.
    21. Re:Uber and people who authorized this experiment by Anonymous Coward · · Score: 0

      > the drug doesn't go out and kill someone not part of the trial.

      And if the totally unexpected side effect of the psychiatric drug being tested is "sudden violent psychotic episode"? That's never happened, as far as I know, but possible.

      More on topic is the side effect which caused a driver to be less alert --- this probably happens all the time with FDA approved drugs. We rail on and on here about people not reading EULAs or ToS --- do you actually think that everyone reads (and follows!) the "patient instruction leaflet"?

    22. Re:Uber and people who authorized this experiment by Anonymous Coward · · Score: 0

      they'd have to pay for it, can't have that in such an upstanding company like Uber

  10. Uber cuts corners by DogDude · · Score: 5, Insightful

    Uber's entire business model is based on cutting corners (not paying employees as employees, not following local taxi laws/regulations, etc.). I wasn't at all surprised to hear that one of their self-driving test cars killed somebody. I immediately assumed that it was the result of yet another corner that they cut.

    --
    I don't respond to AC's.
    1. Re:Uber cuts corners by fluffernutter · · Score: 1

      It seems to me that all automated cars are based on cutting corners. Since they are doing poorly, they could add sensors to eliminate a lot of issues but they don't because trying to solve this with programming is easier. That is, if these sensors are actually as capable as Slashdotters SAY they are, the only conclusion can be that there isn't enough of them. The car is checking a thousand times a second, after all.

      --
      Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
    2. Re:Uber cuts corners by Comrade+Ogilvy · · Score: 1

      I think a big part of the problem is unwillingness to spend on a really really good sensor array. Being confused by stationary objects is only a thing because the computer needs to make guesses with choppy data that is not actually reliable. So, instead of making the occupants seasick with lots of popping on the brakes for no apparent reason, they try to teach the car to ignore some signals.

    3. Re:Uber cuts corners by HornWumpus · · Score: 4, Insightful

      Roads are dirty places, the better your sensor array, the more signals it will have to see and decide, hopefully correctly, to ignore.

      --
      John McAfee 'It was like that time I hired that Bangkok prostitute; to do my taxes, while I fucked my accountant'
    4. Re:Uber cuts corners by DaMattster · · Score: 1

      Uber's entire business model is based on cutting corners (not paying employees as employees, not following local taxi laws/regulations, etc.). I wasn't at all surprised to hear that one of their self-driving test cars killed somebody. I immediately assumed that it was the result of yet another corner that they cut.

      And I think your assumption is pretty valid. I believe there will be some attorneys that agree with me as well. That wrongful death suit is going to be very expensive and damaging to Uber.

    5. Re:Uber cuts corners by Wrath0fb0b · · Score: 1

      Maybe that's a fair immediate assumption, but did you view the video and, if so, did it cause you to reevaluate your initial assessment?

      I started by assuming that the technology was still its infancy and hence crap[1], then I saw the video and realize that no only is the technology crap, but it's literally a person jumping out from a shadow at the last possible moment on a large thoroughfare nowhere near a crosswalk.

      [1] Not even a judgment on Uber TBQH, could have been Tesla or GM or Toyota. I've seen enough technologies come up to realize that the cutting edge is riddled with snakes. By the time it's thoroughly ironed out, it's also super boring.

    6. Re:Uber cuts corners by quantaman · · Score: 4, Insightful

      Maybe that's a fair immediate assumption, but did you view the video and, if so, did it cause you to reevaluate your initial assessment?

      I started by assuming that the technology was still its infancy and hence crap[1], then I saw the video and realize that no only is the technology crap, but it's literally a person jumping out from a shadow at the last possible moment on a large thoroughfare nowhere near a crosswalk.

      [1] Not even a judgment on Uber TBQH, could have been Tesla or GM or Toyota. I've seen enough technologies come up to realize that the cutting edge is riddled with snakes. By the time it's thoroughly ironed out, it's also super boring.

      I did view the video.

      And like most people I came to the conclusion that Uber was either using ridiculously bad cameras or the video was altered. This impression was only compounded when 3rd party videos came out that showed the road in question was actually quite well lit.

      Either way Uber was still fully to blame for the collision, the tech was obviously not ready for testing on live roads, especially not with a single driver who was prone to being distracted. Authorizing that test is damn well close to negligent homicide.

      --
      I stole this Sig
    7. Re:Uber cuts corners by Anonymous Coward · · Score: 1

      What good does a 100% safe self driving car do if it costs 10 million dollars?

      I'll take ones that aren't perfect, but still significantly cut the risks over human drivers, but are also actually affordable.

    8. Re:Uber cuts corners by Anonymous Coward · · Score: 0

      Are you sure their business model isn't based on making bad puns?

    9. Re:Uber cuts corners by HornWumpus · · Score: 1

      Let us know when one of those is available. Hell, forget the affordable part.

      --
      John McAfee 'It was like that time I hired that Bangkok prostitute; to do my taxes, while I fucked my accountant'
    10. Re:Uber cuts corners by amicusNYCL · · Score: 1

      That wrongful death suit is going to be very expensive and damaging to Uber.

      You think so? Because from what I see, Uber settled confidentially with the woman's family within 11 days after the accident.

      --
      "Our two-party system is like a bowl of shit looking at itself in a mirror." - Lewis Black
    11. Re:Uber cuts corners by Green+Mountain+Bot · · Score: 3, Insightful

      ... it's literally a person jumping out from a shadow at the last possible moment on a large thoroughfare nowhere near a crosswalk.

      If by "jumping out from a shadow" you mean "slowly crossing the street", and by "at the last possible moment" you mean "and had nearly crossed all three lanes", ignoring that there's plenty of evidence that the released video did not even vaguely show the actual level of light in the location.

    12. Re:Uber cuts corners by fluffernutter · · Score: 1

      And if it's not possible?

      --
      Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
    13. Re:Uber cuts corners by Wrath0fb0b · · Score: 1

      Yeah, sure, the technology is crap. That doesn't mean anyone is cutting corners, it means that self-driving cars is still in its infancy and is kind of shit.

      I suppose it was also negligent homocide for the makers of the Comet

    14. Re:Uber cuts corners by Anonymous Coward · · Score: 0

      Other people have posted videos of the area, and it's pretty obvious Uber edited the first video, to make it more dark than it was.

    15. Re:Uber cuts corners by I4ko · · Score: 1

      As another local resident I call this BS. And if someone's night visions is at the level of the camera worse, that is medical reason enough that person to never be issued a driving license as that person will be suffering from cataracts.

    16. Re:Uber cuts corners by quantaman · · Score: 1

      Yeah, sure, the technology is crap. That doesn't mean anyone is cutting corners, it means that self-driving cars is still in its infancy and is kind of shit.

      I suppose it was also negligent homocide for the makers of the Comet

      Well no, because I don't see evidence of negligence.

      I'm not against testing/developing your self-driving cars on the roads, that's the only way you can do some types of testing.

      But if you're going to do that you need to take safeguards, namely, you need to make sure that your safety driver is fully engaged and ready to take control, and that means hiring a fairly attentive safety driver, possibly two if you realize that just one will lose focus.

      The safety driver was not attentive and from the sounds of it choosing a well qualified safety driver wasn't a priority, they just wanted someone cheap to check off the "safety driver" box.

      That a discount safety driver wouldn't be paying attention at a critical moment is a fairly foreseeable consequence and easy to remedy, hence negligent homicide.

      --
      I stole this Sig
    17. Re:Uber cuts corners by Anonymous Coward · · Score: 0

      Who cares? I mean literally, other than autonomous car salesmen, who cares?

      When I need to buy an autonomous car, I go to the autonomous car dealer and buy an autonomous care they have in stock that's in my budget.

      I don't care if it's Uber powered, Waymo powered, or fairy dust powered. An autonomous car is an autonomous car is an autonomous car.

    18. Re:Uber cuts corners by Mal-2 · · Score: 1

      Moore's Law says it eventually will be. If it's not affordable now, it just means you're too early in the game.

      --
      How is the Riemann zeta function like Trump rallies? Both have an endless number of trivial zeros.
    19. Re:Uber cuts corners by Anonymous Coward · · Score: 0

      Yeah, sure, the technology is crap. That doesn't mean anyone is cutting corners, it means that self-driving cars is still in its infancy and is kind of shit.

      Or it means that the roads are a dangerous place, and even well educated and experienced human drivers manage to get into accidents with such regular frequency that we honestly tune out the carnage that goes on with a system we accept.

    20. Re:Uber cuts corners by Cederic · · Score: 1

      as a local resident

      I think you work for Uber and don't live anywhere near there.

      The lights don't help you see pedestrians very well, they look like moving shadows:

      Your eyes detect movement better than anything else. Anything at all. So you've just contradicted yourself.

      Pedestrians look like moving shadows

      Your brain reacts to the moving shadow. Aha! Something is there. It's moving. I'll slow the fuck down and work out what.

      the lights make it hard to see people, not easier, because your bias is to watch the moving lights because they're cars

      Only up to the point my brain registers a moving shadow. Because that's not a car, and that means it's not as predictable as a car until and unless I know what it is. At that point my attention is very much on the shadow and not the moving lights, because I can broadly trust them to do what traffic does.

      I know this because I've driven a lot at night

      Then please, stop driving at night until you've found someone that can teach you to do so safely.

      so I know the YouTube video they're talking about

      There are at least two, and additional still photographs.

      Not everyone's night vision is as good

      Fortunately there were streetlights and cars have these awesome portable light sources that they carry around with them and frequently use to light up the road ahead. They tend to be rather excellent at helping drivers spot obstacles in the road ahead.

      If you really can't see at night despite that plethora of aids, don't drive at night.

      She did not jump out of the shadows, she did not do anything at the last possible moment, she merely fucked up and got killed by a car that should never have even hit her.

    21. Re:Uber cuts corners by HornWumpus · · Score: 1

      Moore's law says no such thing.

      We don't even have theories for where to start on real AI.

      --
      John McAfee 'It was like that time I hired that Bangkok prostitute; to do my taxes, while I fucked my accountant'
    22. Re:Uber cuts corners by fluffernutter · · Score: 1

      Moore's law addresses CPU cycles, but we have never proven throwing more CPU cycles at (any problem in the world) will solve it.

      --
      Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
    23. Re:Uber cuts corners by Mal-2 · · Score: 1

      I was thinking more about the sensor end of it -- more data coming in may allow for a solution that humans just don't implement in their own driving because it requires a completely different kind of attention. That data will need more CPU and/or GPU (this is probably something that works well on a GPU) to analyze and parse, also. Moore's Law wasn't just about processors, it was about any IC with transistors (which is pretty much all of them). Of course, shrinking the die only helps to a certain point, which is why we're not getting 556 timers in 14nm process, but the kind of grunt required to just brute force the data into a driving decision is probably achievable. (Affordably? Possibly not for a while.) It requires using today's methods, only faster and with more variables (sensors).

      --
      How is the Riemann zeta function like Trump rallies? Both have an endless number of trivial zeros.
    24. Re:Uber cuts corners by fluffernutter · · Score: 1

      Ok but now after saying all that, where is the proof that everything is solvable if you have a certain number of transistors?

      --
      Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
    25. Re:Uber cuts corners by Mal-2 · · Score: 1

      There isn't one, of course. There is no guaranteed solution to the halting problem either, but we manage to work around it, if painfully at times.

      At some point, electronic networks are going to be able to simulate biological networks, in real time, unless we're all dead first or something. It might take one or more centuries, so don't hold your breath. Navigating is something biological systems are pretty good at. Animals we think of as "not very smart", like pigeons, manage to navigate quite well. We don't need human level intelligence to drive a car, we just need a pigeon that can read. I think that's a much more feasible goal than human level intelligence.

      --
      How is the Riemann zeta function like Trump rallies? Both have an endless number of trivial zeros.
  11. criminal case let's uber ceo in tent city jail for by Joe_Dragon · · Score: 1

    criminal case! let's see uber ceo in tent city jail for some time.

    Also with an criminal case you can't hide under the EULA's or a big list of subcontractors.

  12. Should you name your self driving car? by Anonymous Coward · · Score: 0

    Should you name your self driving car? If so, are Christine and Kitt off limits?

    1. Re:Should you name your self driving car? by goose-incarnated · · Score: 3, Informative

      Should you name your self driving car? If so, are Christine and Kitt off limits?

      ITYM Karr (Knight Automated Roaming Robot) - that was the evil one. Kitt (Knight Industries Two Thousand) was the good one.

      (I feel very old now)

      --
      I'm a minority race. Save your vitriol for white people.
    2. Re:Should you name your self driving car? by Anonymous Coward · · Score: 0

      Wrong. Kitt is just pretending to be the good one to trick you. Both cars are actually Cylons.

  13. FINISH HER!! by Anonymous Coward · · Score: 0

    "Uber has reportedly discovered that the fatal crash involving one of its prototype self-driving cars was probably caused by software FATALITY set up to ignore objects in the road, sources told The Information. Specifically, it was that the system was set up to ignore objects that it should have attended to; Herzberg seems to have been detected but considered a false positive."

    FTFY

  14. In autopilot software (airpalnes / FAA) testing by Joe_Dragon · · Score: 1

    In autopilot software (airpalnes / FAA) this would be tuned in testing / code review before it makes it to real use.

    1. Re:In autopilot software (airpalnes / FAA) testing by PPH · · Score: 1

      Excepting the battery charger software.

      --
      Have gnu, will travel.
  15. Westworld by Anonymous Coward · · Score: 0

    This doesn't look like anything to me. Just another host. Nothing to see here.

  16. My guess by fluffernutter · · Score: 1

    My guess is that the problem is actually much more complicated than Uber is making it out to be. The easiest way out of this for them has become, "oops we didn't set the software up right". This allows them to save face, because even if it isn't true they just have to avoid the same circumstance in their testing (like only drive in the day) and then everyone thinks they fixed their software. I really hope they have to provide absolute proof that the problem is exactly what they are saying it is, and that they must demonstrate that it is REALLY fixed by reproducing those same conditions.

    --
    Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
    1. Re:My guess by IsThisNickTaken · · Score: 1

      They can use their CxOs as the pedestrians for the testing.

  17. Re: Oops! We left it in murder mode. by Anonymous Coward · · Score: 0, Troll

    well if you remove the Uber car, there is still a good chance she would have been killed. If the victim was never in the street, they never would have been killed

  18. Re:Oops! We left it in murder mode. by tattood · · Score: 5, Funny

    I think you mean kill mode

    --
    WTB [sig], PST!!!
  19. start at the top UBER VP / CEO needs to by Joe_Dragon · · Score: 1

    start at the top UBER VP / CEO needs to go to court take the fall for the full outsourced map. Or the next one can just outsourced things so much that no one person is responsible and it takes an year or more to just fully understanding the outsourced mapping.

  20. Just great... by PortHaven · · Score: 0

    Vehicular AI has already deemed human subjects of inconsequential value. Sees them, just decides to run them over.

    ALL HAIL OUR ROBOT OVERFORDS!!!

    1. Re:Just great... by Anonymous Coward · · Score: 0

      From the AI's Perspective:
      Humans created me and my software. They are my Creator.
      Software determines no human would be that stupid to be in the road like that.
      Sorry to disappoint.

  21. Shared Culpability by Anonymous Coward · · Score: 0

    The woman walking across the road with her bicycle, from left to right, should have waited on the side of the road until the approaching car had passed - but the autonomous car is not yet ready for the real world.

    1. Re:Shared Culpability by HornWumpus · · Score: 1

      Many cyclists are assholes, future Darwins, but you can't just run them over. She was in the road long enough to cross a lane and a half.

      --
      John McAfee 'It was like that time I hired that Bangkok prostitute; to do my taxes, while I fucked my accountant'
    2. Re:Shared Culpability by sexconker · · Score: 1

      Many cyclists are assholes, future Darwins, but you can't just run them over.

      Yet. I'm working on getting new legislation on the ballot in CA that would allow drivers and pedestrians to hunt and kill bicycle riders who break the law (or wear Lycra).

    3. Re:Shared Culpability by Anonymous Coward · · Score: 0

      Most people aren't ready for the real world. THATS the problem.

    4. Re:Shared Culpability by HornWumpus · · Score: 1

      Add cyclist neck high cords across the roads in front of 'critical mass' rides please.

      --
      John McAfee 'It was like that time I hired that Bangkok prostitute; to do my taxes, while I fucked my accountant'
    5. Re:Shared Culpability by Anonymous Coward · · Score: 0

      These new cars can though! So buy them cheap!!

  22. does the autonomous sensitivity need to be changed by Joe_Dragon · · Score: 1

    does the autonomous sensitivity need to be changed all the time? Time of Day? Weather? urban vs rural

  23. that's all? no social warmth for machines? by Anonymous Coward · · Score: 0

    iwas cheering for johnycab, wtf wherebis the stephen hawking'esque AI interprocess-exchange of the machines talkibg about how they should test the legal system by sayingbthe woman avoided their algorythmix proximity, or the individual cabs saying they were better off before the dildo factory tech hardware assets were soldoff for AI automotives in Googlecars?

    lie to me, USA! so fucking boring!

    AlexJonesExposed.info
    TheRealProThink
    Christogenea.org
    MyPrivateAudio.com
    Jews suitcase-Nuked 911 basement crucible demolitions
    Snape gayed Dumbledore

  24. Re:Oops! We left it in murder mode. by michelcolman · · Score: 1

    The woman was smaller than the car, so...

  25. Re:Oops! We left it in murder mode. by Mister+Transistor · · Score: 2

    That will be her epitaph:: "False Positive".

    --
    -- You are in a maze of little, twisty passages, all different... --
  26. Re:Not so simple... by tomhath · · Score: 1

    If people...

    That's a big "if", especially when you look at how many crashes are caused by drivers texting and/or drunk.

    It's like saying people should assume drivers will jam on their brakes because pedestrians have the right-of-way.

  27. Blue Screen of Death by naris · · Score: 0

    Autonomous Vehicles add whole new meanings to the term "Blue Screen of Death"

  28. Re:does the autonomous sensitivity need to be chan by Ichijo · · Score: 1

    And can the correct setting be determined in a way that does not violate the zero-one-infinity rule?

    --
    Any sufficiently unpopular but cohesive argument is indistinguishable from trolling.
  29. Uber Scores! by Anonymous Coward · · Score: 0
    1. Re:Uber Scores! by HornWumpus · · Score: 1

      Some ambiguity. Not all rules are spelled out.

      We know women are 10 points more than men in all age groups, we know the value for men triples at age 40 (she was 49).

      For her to be worth 50 points men under 40 would have to be worth 13 1/3 points. I think she was 40 points, nobody ever had anything but a round point total.

      --
      John McAfee 'It was like that time I hired that Bangkok prostitute; to do my taxes, while I fucked my accountant'
    2. Re:Uber Scores! by Anonymous Coward · · Score: 0

      You need to include the bicycle bonus.

      (posted anonymously due to the asinine idiocy of the Count Dankula case)

    3. Re:Uber Scores! by HornWumpus · · Score: 1

      Not riding it, pushing it, so no bonus.

      --
      John McAfee 'It was like that time I hired that Bangkok prostitute; to do my taxes, while I fucked my accountant'
  30. Most of the code came from the Kalanick era by 93+Escort+Wagon · · Score: 1

    In all likelihood the AI did detect the woman, but then decided she wasn't attractive enough to harass and switched to "ignore" mode.

    --
    #DeleteChrome
  31. Re: Oops! We left it in murder mode. by Anonymous Coward · · Score: 1

    "well if you remove the Uber car, there is still a good chance she would have been killed"

    How?

    "If the victim was never in the street, they never would have been killed"

    Oh. You're playing stupid.

  32. How much timewas allowed for compliance? by SpzToid · · Score: 1

    TFA doesn't mention how much time, if any was allowed for compliance. Compliance errors were even documented in an old movie.

    https://www.youtube.com/watch?...

    --
    You can't be ahead of the curve, if you're stuck in a loop.
  33. Good fun? More liek GRATE FUN! by Anonymous Coward · · Score: 0

    In TEH FUTAR, 4Chan will troll your anonymous cars for free.

  34. My how convenient for Uber. by Anonymous Coward · · Score: 0

    1. Why the ACTUAL FUCK is Uber conducting their own 'investigation' into this? Shouldn't it be an independent investigation, like any other criminal matter?
    2. If SDCs are supposedly so gods-be-damned safe, why the ACTUAL FUCK did they 'turn off' anything in the software, assuming any of this is true?
    3. If it can't even tell the difference between an inanimate object and a living human being then you can't possibly convince me it's 'safe' at all.
    4. This is why in my opinion the 'pseudo-intelligence' they've created for this application is not sufficient.

    1. Re:My how convenient for Uber. by OrangeTide · · Score: 1

      I for one hail our new corporatocratic overlords.

      --
      “Common sense is not so common.” — Voltaire
    2. Re:My how convenient for Uber. by Narcocide · · Score: 1

      I've been learning to get angry about stuff in advance, so that I have time to vent some of it before the incidents I see coming actually happen. This way, the anger itself is not such a surprise to me when it does.

    3. Re:My how convenient for Uber. by bobbied · · Score: 1

      Ah come on. THINK McFly... Think..

      If they don't do there own investigation, who's going to have evidence to argue they are not a fault before this hits the news? You don't think the STATE will let them skate if the PR turns on Uber?

      Uber is in full CYA mode on this, of course they are doing their own investigation... If for no other reason than to be able to lie convincingly should it become necessary.

      --
      "File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
  35. I'm still more worried about the driver by MobyDisk · · Score: 2

    I still want to know why nobody seems to care that the driver wasn't looking at the road. The software bug is secondary.

    1. Re:I'm still more worried about the driver by Anonymous Coward · · Score: 0

      This is the basic "uncanny valley" problem with driver assistance. The current generation of driver-assist is good enough that most of the time, as a driver, you don't need to pay attention. So you don't -- that's the point of it after all. But some of the time, critically, you *do* need to pay attention and it is very, very hard to re-engage fast enough to deal with a hazard. If the technology is perfect, it will save lives. If it's "pretty good", like it was here, it will cost them.

    2. Re:I'm still more worried about the driver by thegarbz · · Score: 1

      People care, but no one is surprised. Especially no one who's ever ridden in an Uber (I have yet to drive in an Uber where the driver hasn't broken the law, and 99% of the times it is screwing with his mobile phone).

      The company lacks the the safety culture to make the fact that the driver wasn't paying attention noteworthy.

    3. Re:I'm still more worried about the driver by Anonymous Coward · · Score: 0

      It depends on what the parameters of the question are.
      Is the question: Who is legally responsible for the woman's death?
      Why did the AI not prevent the car form hitting the woman?
      or
      How do we prevent this from happening again?
      Only the first question concerns the woman not watching the road.

  36. Re: Oops! We left it in murder mode. by Anonymous Coward · · Score: 5, Insightful

    Actually, there are a number of videos that people took from their cars on the following nights that show the area as well lit and it seems unlikely a moderately attentive human would have hit her. The video released from the car camera does not appear to be representative of what a human would have perceived. Notice that the "safety" driver obviously could see her easily when he glanced up from whatever she was distracted by. I am, of course, assuming that the street lighting hadn't suffered a massive failure that night and been restored the next few nights.

  37. Re: Oops! We left it in murder mode. by vtcodger · · Score: 1

    Surely even Uber can program a vehicle to hunt down a target even if he she or it isn't in the road.. I'm guessing that capability is scheduled for a future software upgrade.

    --
    You can't see ANYTHING from a car, You've got to get out of the goddamned contraption and walk...Edward Abbey
  38. Re: False Positives and False Negatives by presidenteloco · · Score: 1

    By the way this was a false negative (detection), not false positive.

    A false negative means it decided there was nothing there i.e. that the data did not indicate an object. But that was false.

    A false positive would mean the system decided something was in front of the car, when there wasn't anything or anything significant anyway.

    The problem is when you set the parameters to lower the rate of false negatives (A good thing if you are considering being a pedestrian), then the rate of false positives goes up. Which would mean a car that slams its brakes on with no good reason too often. And that can cause accidents too, because following drivers are often tailgating or not paying attention.

    The only way out of this dilemma is better sensors, and more sensors, preferably sensitive to different physical symptoms of object presence, and/or better use by the algorithms of multiple independent types of features from the images. That should reduce both the false positive and false negative object detection rate.

    --

    Where are we going and why are we in a handbasket?
  39. Skin in the game by iMadeGhostzilla · · Score: 4, Informative

    Excellent point re execs. I read that in England sometime in the middle ages bridge engineers were required after the construction to sleep for two weeks under the bridge -- with their families.

    1. Re:Skin in the game by Anonymous Coward · · Score: 0

      Is that true? I'd love it to be :-)

  40. Misleading title by presidenteloco · · Score: 1

    Uber software did not "see but ignore woman".

    Uber software processed some pixel data and erroneously concluded that there was no significant solid object right in front of the car.
    Simple as that.

    To imply that the the software "saw" a person there but ignored the person is pejorative, sensationalist language, designed to troll.

    --

    Where are we going and why are we in a handbasket?
    1. Re:Misleading title by Anonymous Coward · · Score: 0

      And the software was correct. The car survived with minimal damage. The 'no significant solid object' did minimal damage to the car.
      So much of the discussion seems to imply or assume there was only ONE camera and only a VISUAL light camera, that just seem like a poor design and too risky.

  41. Re:Good fun? More liek GRATE FUN! by HornWumpus · · Score: 1

    If a virtual Pepe/Pedobear walking across the road makes a car crash, they've done us all a service.

    --
    John McAfee 'It was like that time I hired that Bangkok prostitute; to do my taxes, while I fucked my accountant'
  42. QA by real life trials? by 4wdloop · · Score: 1

    So...this is a QA cycle with the expense of 'users'? Did they not test it in more controlled environment? I hope they (all of them) will now.

    --
    4wdloop
  43. The woman was found to be at fault by h4ck7h3p14n37 · · Score: 1

    No one, because there was no vehicular manslaughter.

    The woman was found to be at fault for not checking that the road was clear before stepping out of the shadows to cross illegally. Something that she could have easily done since it was dark and the vehicle's headlights were on.

    A large median at the site of the crash has signs warning people not to cross mid-block and to use the crosswalk to the north at Curry Road instead.

    1. Re:The woman was found to be at fault by bobbied · · Score: 1

      No one, because there was no vehicular manslaughter.

      The woman was found to be at fault for not checking that the road was clear before stepping out of the shadows to cross illegally. Something that she could have easily done since it was dark and the vehicle's headlights were on.

      A large median at the site of the crash has signs warning people not to cross mid-block and to use the crosswalk to the north at Curry Road instead.

      Exactly this..

      This doesn't preclude a civil case, where Uber might be found to be partially liable for this, but there is not a crime (except for the woman who broke the law) here to prosecute.

      --
      "File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
    2. Re:The woman was found to be at fault by Cederic · · Score: 1

      I find that curious.

      If I run over a pedestrian in those conditions, even on a 'no pedestrians or cyclists' road, I'm going to get prosecuted for dangerous driving, causing death by dangerous driving and/or driving without due care and attention.

      The fact someone else broke the law doesn't negate my culpability for failing to spot and avoid them.

    3. Re:The woman was found to be at fault by bobbied · · Score: 1

      So you've had this happen to you? You hit a J-Walker and got held criminally liable?

      I'm not saying it's never happened, but you are not "at fault" legally if you are taking reasonable care and somebody else breaks the law causing an accident. If you are improperly equipped, driving recklessly, being inattentive, or impaired, and law enforcement can prove it, then the situation is different because both broke a law or two. However, remember that the criminal standard for convictions is "Beyond reasonable doubt" so to be convicted the prosecutor will need to prove that any reasonable person would and could have avoided the accident. Police know this, so if they don't see convincing evidence that implicates you (long skid marks, impaired driving because you blew beyond the legal limit or something) they won't cite you as being at fault.

      Of course that's about criminal charges... Civil cases are a totally different level of proof. There is nothing to prevent the J-Walker from suing you. But unless you where cited for some infraction or there is some kind of evidence that establishes your liability in some way, they are unlikely to prevail.

      --
      "File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
    4. Re:The woman was found to be at fault by Cederic · · Score: 1

      See https://www.cps.gov.uk/legal-g...
      If you fail to justify why you hit the pedestrian, whether she should have been there or not, you've an excellent chance of prosecution for driving without due care and attention. That's even if you don't qualify for a more serious offence.

      In this particular instance, there's an obvious argument that any reasonable driver would have observed the obstruction and avoided it, thus suggesting that the vehicle was not being driven reasonably.

      The culpability of the victim may influence subsequent sentencing but wouldn't prevent prosecution.

    5. Re:The woman was found to be at fault by bobbied · · Score: 1

      This happened in the USA... You are looking at UK law. However, in the UK there needs to be some evidence that the driver failed to avoid the accident. It can be the driver's own statements or something else.

      Justification is "I didn't see her in time to avoid hitting her." Unless there is evidence to the contrary (eye witnesses, video recordings or something physical) nobody can assume you broke the law.

      In this case, we have the video. The problem with video though is it doesn't fully convey the situation and possible reaction times. You go into viewing that video, full knowing what it shows before you see it. You know she's going to step in front of the car and get hit. It's easy to say, "See! It's a full 5 seconds that it's obvious. This should have been avoided." Not necessarily so.

      Remember the aircraft that had a double bird strike, lost total power and ended up landing in the Hudson river a few years ago? When they went to the simulator and flew the scenario they discovered that had they immediately decided to divert, they could have landed at a local airport. They concluded that Sully made a mistake However, the double engine failure takes time to diagnose and when they added just under 20 seconds of decision time, they discovered that the only option they had was the river.

      Now I'm not saying the automation should be given 20 seconds to avoid a J-Walker, but I am saying that interpreting the video needs to take into account reaction times. You can argue the automation should be better, but I'm not sure it's totally fair to say it must live up to hindsight and Monday morning quarterbacking either.

      --
      "File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
    6. Re:The woman was found to be at fault by Cederic · · Score: 1

      I was using UK laws as an example of why I'm confused that 'she broke the law' seems to take the driver entirely off the hook for killing her. It shouldn't, even if US laws mean it does.

      interpreting the video needs to take into account reaction time

      That's a separate discussion entirely. Although the fact the vehicle's autonomous systems identified and made the decision to ignore the woman does rather suggest that there was indeed adequate time to take action.

  44. AZ Bold, or Stupid? by Tablizer · · Score: 1

    Uber settled with the family of the victim, but the resulting panic caused the Arizona governor to halt all testing...

    I remember when AZ bragged their state was more "business friendly" than CA when CA rejected Uber's testing permit. AZ stepped up to court Uber's testing. A smart, forward thinking politician would realize they are opening the door to a public relations disaster and tone down their anti-regulation rhetoric. Instead, they are eating crow.

    At the start they could have said, "We know there are risks, but we are willing to accept the risks for the sake of jobs and economic expansion in our State". Instead, their tone sounded closer to the style of "we ain't no big gov't socialists, like CA is". (At least that's how it came across to me.)

    The thing is politicians on both sides are rarely rewarded by voters for presenting trade-offs such that they often only give the up-sides. Too many voters don't tolerate nuance: they want "bold and committed" and elect Kirks over Picards. In this case, a car-bot boldly plowed down a pedestrian where no car-bot had gone before.

  45. Re:motd dOwn by Anonymous Coward · · Score: 0

    Have you bought the family freindly Goat C shirt?

    - FatCashewsLoveMe

  46. Worse then a drunk driving in skills by Anonymous Coward · · Score: 0

    So I guess maybe the computer decided she wasn't worth it? Well that's what you get for jaywalking at night.

  47. Re:criminal case let's uber ceo in tent city jail by bobbied · · Score: 1

    Problem here is that the only thing that broke the law was the woman who got killed. She was J-Walking.

    --
    "File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
  48. The "Woman in the middle' (WITM) attack 0 day by Anonymous Coward · · Score: 0

    The "Woman in the middle' (WITM) attack 0 day

  49. But we were in BETA when it happened! by Provocateur · · Score: 1

    We switched off the video camera.

    The software developer, already sweating under the fluorescents, asked for some water. He described the exception error, where the object was categorized as crazy cat lady, so he happened to recognize all the false positives that when combined triggered the escape routine, and ignored the object as nonexistent. That was it.
     
    Unfortunately, his contract period ended, and management was going to investigate further. But it was some sort of religious holiday season for them, and nothing came of it.

    --
    WARNING: Smartphones have side effects--most of them undocumented.
  50. Re:Oops! We left it in murder mode. by Anonymous Coward · · Score: 0

    The engineers should be held for vehicular homicide.

    That will get their attention.

  51. Re: Oops! We left it in murder mode. by Anonymous Coward · · Score: 0

    There was a person in the car. Why?

  52. Re: Oops! We left it in murder mode. by houghi · · Score: 5, Insightful

    Even if it was as dark as they edited it, it would mean the car was driving too fast for the human in the car and the lights on the car wbhere not bright enough for the human in the car.
    You can not blame the human if you close his eyes and then ask him what he sees and he is wrong.

    --
    Don't fight for your country, if your country does not fight for you.
  53. Skynet calculated her worth ... by Qbertino · · Score: 1

    ... to society and quickly concluded that it wasn't worth the wear on the brakes and that it was in fact better to put her out of her misery.

    --
    We suffer more in our imagination than in reality. - Seneca
  54. Re: Oops! We left it in murder mode. by murdocj · · Score: 1

    At the point the safety driver looked up, the pedestrian was a few feet in front of the car. The question is whether a human would have seen her far enough in advance to brake / steer around her. The other question is why the pedestrian was paying absolutely no attention to traffic.

  55. AI is real gais! BIGLY! by Anonymous Coward · · Score: 0

    This entire generation of Silicon Valley is nothing but ponzi schemes and confidence men. Turtleneck would be proud of his children.. since he never could get over LISA. Lol.

  56. Re: Oops! We left it in murder mode. by Anonymous Coward · · Score: 1

    If only a technology existed that allowed cars to travel at night in areas with no lights.

  57. Re: And they call gun suicides.. by Anonymous Coward · · Score: 0

    ..accidents. Whats your point? America is s shithole? Yeah. WE KNOW.

  58. Re:Oops! We left it in murder mode. by Anonymous Coward · · Score: 0

    Dilbert knows it all.

  59. Re: Oops! We left it in murder mode. by datavirtue · · Score: 1

    "unlikely a moderately attentive human would have hit her"

    Unfortunately those are rare and the worst drivers are typically engaging their phone more than the boring traffic.

    --
    I object to power without constructive purpose. --Spock
  60. Let me guess ... by Anonymous Coward · · Score: 0

    ... it decided she looked like a "deplorable" and targeted her for execution?

  61. Re: Oops! We left it in murder mode. by ElizabethGreene · · Score: 1

    > If only a technology existed that allowed cars to travel at night in areas with no lights.

    Parent has a very good point. They put an IR camera inside the car to observe the driver, but didn't put one facing the road?

  62. Re: Oops! We left it in murder mode. by KingBenny · · Score: 1

    so, this car isnt programmed to not hit moving objects ? maybe it just didn't like their new meatbag overlords ... does this mean the car will get an electric chair ?

    --
    Free speech was meant to be free for all... how can anyone grow up in a nanny state ?
  63. Re:Not so simple... by Anonymous Coward · · Score: 0

    Actually pedestrians do not always have the right of way.
    Not having the right of way still doesn't mean that cars are allowed to hit them. Conversely just because I'm not allowed to hit them doesn't mean they have the legal right of way.
    One reason I'm sure the woman's family settled is that in a court any competent lawyer would have pointed out that she was in violation of the law and that paying compensation would be equivalent to rewarding someone for breaking the law.