The SDC would already have the rules for that intersection in its knowledge base.
If you're going to continually assume perfection in perfectly available data, then you might as well just say you assume the self driving car is going to be perfect and leave it at that. No need for arguing about anything, we know the answer because you assume it will be that way. Only morons would produce autonomous vehicles that weren't perfect and weren't able to deal with an imperfect world, and we assume that there are no morons in the management or design teams.
And we assumed that the o-rings in the Challenger solid rocket boosters would deal with frigid temperatures just fine. That turned out ok, I guess. The original designers of the Tacoma Narrows Bridge designed a simply perfect bridge. And the much smaller problem of navigation databases for automotive GPSs is a perfectly solved problem, too. Nobody ever gets hurt by trusting GPS navigation databases today, so much more complex systems will be fine and we should trust them with our lives. The literature is just replete with documented cases of humans designing perfect systems that behave perfectly all the time.
Yes, I welcome our autonomous overlords, but pass me the dikes anyway.
If the car drives itself, then you can relax and enjoy the scenery even more.
Let's see, worry about whether the car will fail and try to kill me by driving off the edge of a mountain road, fail and suddenly say "you're it, start driving human, I can't deal with it", or simply drive myself and enjoy the trip. Hmmm. I can't decide which I'd rather do. What a quandary.
But the I think the point was, roads are not for pleasure, they're for transportation. What scenery am I supposed to enjoy?
No I didn't. I admitted that it would know that one sensor (out of dozens) had failed.
And that one sensor is the one that tells it where it is. Sensor failed. GPS data unreliable. Where are you?
Why would it "forget" where it was?
Because the sensor that tells it where it is is unreliable, and without that data you don't know where you are. You knew five minutes ago because the GPS was working. You don't know now because it isn't. In human terms, we call that "forgetting".
It could just read the inertial sensors to detect any acceleration
Oh. The cars will all have INS to backup GPS. Sorry. Didn't assume this.
read the rotary sensors in the wheels to determine its velocity,
That's why I was very careful to say you've forgotten where you are, not forgotten how fast you are going wherever it is you are going. Even so, I've always found my "rotary sensors in the wheels" to be about 3 MPH off compared to the GPS. You've lost your most accurate speed sensor and are putting your life in the hands of, essentially, your speedometer, and hoping that your tires are inflated to the calibrated value so the rotation into distance calculations are right.
and calculate where it is from its last known position.
How long prior to the "teleportation" was the GPS not working correctly? I've dealt with GPS. High precision RTK systems. There have been times when the GPS says "I'm working fine" despite producing garbage position reports, then it decides there is a problem and says "your guess is as good as mine", and then it usually figures it out. But the GPS was sure it knew what it knew before it knew it didn't know. And then it was just as certain it knew where it was, after it was certain it was somewhere else. I didn't trust that system with my life like you expect us all to trust it when it gets put in a car.
It could also continue to use cameras for lane detection,
"I'm in the middle lane of some road which I think is I65." Ok.
cameras for landmark recognition,
"I see a mountain off in the distance, I think that's Mt. McKinley. Maybe it's Mt. Hood. It's at a bearing of 54 degrees from me, and I'm on I65, I think, so I must be right here." Ok.
and a combination of cameras and radar for collision avoidance.
So it won't collide with someone else deliberately when it decides it needs to stop. Notice I didn't say "slam on the brakes sudden stop" or "just shuts down". Still didn't say that. It will cause a lot of hassles for everyone else on the road, though, and some of those cars may have decided THEY need to stop, too, or handed control over to the human because their GPS failed for the same reason yours did -- blocked GPS. Fascinating news story out of New Jersey, I think it was. Fellow was carrying a GPS blocker in his delivery truck to foil his boss's attempt at tracking him. Screwed up GPS for everyone near the major airport he was driving around.
You're putting a lot of faith in capitalists building cars that don't exist yet creating large scale systems of autonomous computers that may have interesting emergent behaviour. Every time one sensor fails you're hypothesizing another sensor to make up for it, sometimes using databases that don't exist and methods of navigation that are of such low accuracy you can't pilot a vehicle with the answers. It would be prudent to say "let's see how it turns out" instead of "it will be perfection and we'll all love it to death." It's that latter part that causes concern for some of us.
you can be completely compliant with HTML standards, and still have a website that doesn't work/look the same on one browser vs another.
You're talking to someone who writes website code by hand and maintains several. Of course you can have different looks on different browsers -- it's called "customization". Fonts, colors, width of screen, allowing pop-ups/js/cookies, all kinds of things. None of that deals with the issue of one manufacturer fielding an "extension" to his communications protocol that doesn't play well with someone else's extension.
if an Automated Vehicle Communications protocol is ever created, you can bet the government will be involved.
Which government? And how many of them? Will a committee of governments design a consistent camel, or will some of them be one hump and some two? Will the EU camel deal with data differently? Will the French camel surrender whenever it is threatened? Will the Chinese camel be reverse engineered and get it almost right?
because you're comparing something that will be enforced by laws and regulations
Because enforcement of perfect worldwide laws and regulations leads to perfection and full compliance everywhere.
Have you noticed, modern cars are getting full entertainment centers and communications capabilities built in at the same time that many states are cracking down on distracted driving? Laws and regulations are going to save us. Some laws say that you're too distracted if you simply talk on a phone, but a context-sensitive touch-screen menu system on a car audio system is perfectly ok.
You put a lot of faith in laws that don't exist yet, governing cars that don't exist yet.
No it doesn't. Only a complete moron would design it that way. And everyone else on the team, and the full management chain would need to be morons as well.
Two words about software designers and management chains: Windows 8. Need I say more?
I didn't say it just shut down. I said it decides it needs to stop.
You admitted it doesn't know where it is. If the GPS failed, that means the GPS-based course and speed indications are unreliable. Of course the designers may have have thought of their idea of what to do in such a situation. Just like you've had time to think about it and can't see any course of action that would be called "needs to stop" that isn't "just shut down" or "suddenly stop."
You blink, and while your eyes are closed you forget where you are and what direction you are going. Sounds like dementia. Welcome to GPS failure, citizen.
Maybe a few zany-looking cars of the type google uses to create street view will map out all the roads with bulky lidar and high-definition cameras, perform an insane amount of off-line processing, and then humans will hand-verify the location of everything that isn't clear. Cars will download all this as they drive along a new road so they are familiar with the road before they get there.
I seem to recall there being a story here in/. about an airport in Fairbanks Alaska needing to block off a taxiway because a GPS system kept telling people to go that way and cross an active runway to get to the airport. This was with cars where a human was in charge and no autonomous control whatsoever. But if we put a computer using GPS driving instructions in charge, nothing like this would ever happen?
"I'm getting nervous, please take over for me."
Sorry, I'm in the middle of the final episode of Breaking Bad and my hands are filled with roast beef sandwich. And I haven't hand driven for two years. That's your job. What do you expect me to do?
So there you go, google already knows the location of every curb and every speed bump, as of whenever one of their cars last went by.
Well, thank God that roads never change after Google takes pictures of them.
When the car in front of you slams on its brakes, how much context do you need?
Is the lane to the left clear so I can deviate that direction and avoid a collision?
If the lane to my left is oncoming traffic, can I safely cross the double yellow line to avoid serious injury?
Is the shoulder clear so I can go that direction?
If the shoulder is not clear, does it have an obstruction that is less valuable and causes less damage if I run it over instead of colliding with the car ahead of me?
Is there a large truck or other similar vehicle behind me that I need to worry about?
Did the driver of the large vehicle appear to be awake and alert when I passed him a few minutes ago?
Does there appear to be a traction limiting condition directly ahead of me that is not to the left or right? E.g, ice or snow?
Is there a traction limiting condition to the left or right?
Did the car ahead of me slam on his brakes for a transient reason (to miss a squirrel that is now out of the way) or because it was about to run into a stopped semi truck?
Enough context? The fact that you only need one piece of information in this kind of situation tells me you don't drive much, and shouldn't be driving the autonomous vehicle bandwagon.
We only have one data set for real world performance of an autonomous vehicle, 350k miles by Google's prototype. It has been involved in one accident- when a person backed into it. So it has been perfect thus far.
One data point. Many years ago I remember the buzz about a new field of study -- emergent behaviour in clusters of simple robotic devices. For example, you can program a small robot to obey simple rules like when to turn, when to go, when to stop, etc, and predict pretty well what it will do. But when you put a roomful of them together they start doing things that weren't obvious. I wish I could remember the name of the guy who was at the forefront of this. Randy something, I think.
The point being, one computer programmed to do one thing is something we can pretty much understand, and predict how it will react. When you get large collections of systems interacting, you sometimes find miraculous, and sometimes not so miraculous, results. Kind of like, if you have one computer programmed to watch the stock market and make trades you will probably be ok. When you get thousands of them all interacting indirectly (via the stock prices, e.g.) and all acting in milliseconds, you can get havok. That's why I referred to that in my original post.
So, I'm really surprised you would respond with a statement that we know the world filled with autonomous vehicles will be safe and accident free because one autonomous vehicle acting in controlled conditions and in a human-centric environment had only one accident. I suspect that we will find all kinds of unpleasant emergent behaviour as the roads begin to have more and more of these things.
I'm also fascinated that you see this as an issue of cars vs. horses, since in neither case was the human removed from "driving".
I may be ignorant to my states laws but I have never heard of slowing down a requirement for NY.
Oregon. The law says you must slow down or pull over when going past an emergency vehicle on the side of the road. Not every vehicle on the side, just emergency ones. This points out another risk of autonomous vehicles. Will they be programmed with all the laws of all the states they may be driven in, and how will they deal with people who live close enough to a border that there are two sets of laws?
Here's another interesting law. In North Carolina, if you have the wipers on, you must also have your headlights on. Also in NC, there is an arcane twist to the laws regarding stopping for police. I don't recall the details, but I think there are exceptions of when you have to stop for flashing lights from the car behind you (cops), based on a rash of fake cops pulling people over.
In fact I feel it makes things a nuisance and possibly dangerous. Most cops I see position themselves well to mitigate risk.
They do their best. As do road workers. And yet, they get run over by drivers to the point that there are special laws for work zones and now emergency vehicles.
If the GPS suddenly reports that the car has instantaneously transported itself to a new location, I don't think it will be blindly trusted.
You're in the middle lane of I65 going 65MPH surrounded by other cars when YOUR car suddenly decides it doesn't know where it is and needs to stop. Welcome to computer dementia. Happy travels, citizen!
Traffic lights are already designed to be as visible as possible,
When your initial premise is wrong, expect the rest to be wrong, too.
Plus, if these cars were mostly used for regular commutes & shopping trips along the same routes, it would probably remember where it usually stops for red lights,
Judge: "Why didn't you signal your right turn?"
Me: "I always turn right at that intersection. Everyone knows that."
Judge: "Why, not guilty of course. How stupid of me to forget what you 'always do'."
Of course everyone will know where you "usually" stop for a red light and know you're going to stop there again when your computer fails to detect the green light.
And no matter how capable you are, a computer has a faster reaction time.
And that's why when you make a mistake with a computer's help, you can do much much more damage than doing it by hand. It takes a long time to shred 100 paper files, and you can figure out after five minutes that it is a mistake and stop, and still have half of them. If you do that to computer files, five minutes later means you have nothing left but the regrets.
I'd go find a reference to the havok created by computer trading on wall street, where instant reaction times lead to financial catastrophe, but I think we probably are all well aware of these stories. Why we want to ignore those stories and assume that a massive system of independent computers would be perfection personified this time, when they've been problematic before, is a mystery. I sense a lot of money to be made for people shilling autonomous vehicles on an unsuspecting public...
I'll be sure to tell the local and state road departments to take down all the "Scenic Loop" signs and close all the "viewpoint" stops in the mountain roads. Transportation, not pleasure, you hedonistic bastards!
It's too bad in your universe that roads cannot serve more than one purpose.
Well, once self-driving cars fill the roads in significant numbers, if they'll have provisions for mutual communication and data exchange, you can count on them being more polite to each other than human drivers would.
What anthropomorphic hogwash. The cars will not be polite. They will behave as they were programmed to.
But here's the rub. My car speaks Toyota HTML, yours speaks Hyundai HTML, and the next guy's speaks Microsoft HTML. Kinda all the same language, somewhat close to the standards, but each manufacturer does it better by doing it their way just a bit.
I spent a wonderful half hour of my life I'll never get back writing with Word some material I wanted to put on the web. I kept indenting some paragraphs and exporting the document as HTML, but the web browser never showed an indent. Then I realized I was looking at it with Mozilla and not IE. Isn't it great that HTML is a standard? Can we survive when our tools that we depend on to keep us alive don't use exactly the same standards?
The story says it will pass slower traffic. Great. But will it detect that guy you've seen in your rear view mirror switching lanes doing 40mph faster than you? I see a great chance to have a multicar accident just on that one thing.
More important, will it realize that the reason the fast lane is empty and everyone is slowing down in your lane is because there is a state trooper with someone pulled over on the left side of the road and your state's law says you must either pull over or slow down when going past him? Will you get to be the next recipient of a ticket from that trooper because your car endangered him?
what is curious for me is at what point will insurance companies insist on automated assist drive systems in cars especially considering 95% of crashes are due to human error.
Citation? And you do realize, I hope, that 95% (if true) being human error is because there currently is no "autonomous computer failure" category or contestants as regular participants. I expect that the numbers will be vastly different when there are.
Unless, of course, the NTSB uses the same standards for car crashes it does for aircraft mishaps and everything is labelled driver (pilot) error as an overarching cause. "Driver did not properly supervise the autonomous vehicle that wasn't supposed to require his supervision anyway."
Get ready to be annoyed for about 30 years, because "automated" commonly means "more automated than before," not "automated in every conceivable way."
Perhaps that's why the word being used is "autonomous", not "automated".
They do what you do - they also watch for and avoid other cars,
Part of fixing the bug that says "stop at any red light no matter where it is" results in "don't pay any attention to the red lights on the cars in front of you."
pedestrians
In my state, you aren't a pedestrian to be stopped for unless you are IN the crosswalk. Standing on the side of the road looking wistfully at cars as they pass doesn't cut it.
(I guess as a backup it could know the GPS location of stoplights and stop if it doesn't see the light and confirm that it is green).
So we're all expecting that they will stop when they have absolutely no reason to, which will certainly do wonders for improving traffic flow and safety. I love the idea of autonomous cars more and more each time the topic comes up.
But I am sure we will end up with some level of instrumentation on the road such as stoplights that emit at a frequency not obfuscated by sunlight, snow, etc, like visible light is.
We already have instrumentation in the roads that detect when cars are there so they can cycle the traffic lights. Before they came up with one sensitive enough for bikes, it was fun to watch a law-abiding cyclist (yes, there are some) stopped at a red light waving his bike around the road hoping to trigger the signal so he could go.
Sometimes, though, the sensor wasn't sensitive enough for cars, and other than backing up and trying again, there was no way of "waving your car" to try to get it to trigger, so you were left with deciding how long was too long and when you could justifiably run the red light. I can imagine a world where that sensor has a transmitter to tell the car it is at an intersection with a red light so the car will stop, and a world where a few critical transmitters fail and new autonomous not autonomous vehicles blow through a red and get t-boned by a semi.
I for one welcome our new robot overlords. Hand me the dikes, I need to clip a few wires...
Most likely there will be dedicated lanes to encourage people to adopt the tech.
Yeah, we have so many empty lanes now that dedicating one of them to a small percentage of cars that are "not autonomous autonomous vehicles" will be a really smart idea to solve traffic problems.
This "not autonomous autonomous vehicle" automatically stops at red lights? How nice. The local store has a red light in their signage out front and all the cars stop. Good for the store. Bad for the traffic, especially the updated version of the "not autonomous autonomous vehicle" which has that bug fixed and doesn't feel like stopping, and runs into V1.0.
"Hello, Toyota support? My autonomous vehicle computer crashed." "Did you try to reboot?" "No, I mean it crashed into the autonomous vehicle computer in the car ahead of me."
The headline is deliberately misleading. It's not an autonomous vehicle if is isn't autonomous, which even the summary admits.
The analogy (what it takes for a car company to be racist) is yours.
First, you failed to quote all the words of the analogy. I doesn't start with the word "more". Second, it is not an analogy that says what it takes for a car company to be racist, it is an analogy about what DOES NOT MAKE A CAR COMPANY RUN BY A CHURCH.
You read my words as me talking to you. I wasn't talking to you,
You replied to something I posted, so I guessed you were talking to me. You wrote about "my words" without attributing that statement to anyone else, unquoted, presumably referring to yourself. And you talked about an analogy being used against you, which is what I have been doing continually - using it against you. My bad for not reading your mind instead of your words.
You've come up with an analogy, and when that analogy is explored to try to gauge the thresholds to then compare to the original situation, you play dumb (or aren't playing).
Your question asking if Ford did have racists on their board would they be run by racists is not a "threshold", it's asking "if Ford was run by racists would they be run by racists" without dealing with the issue of how those racists got there. It's a worthless question, and I'm not playing the game you want to play. The issue I've been responding to is a claim that BSA is run by the Mormon church because some Mormon churches host BSA troops. That's the game of the day.
Here's the fact. The fact that some Mormon churches host BSA troops does not mean BSA is run by the Mormon church. Period. End of sentence. That's the claim that I replied to, that's what you need to disprove if you want to argue with me about it. I'm sorry the analogies confused you, especially to the point that you need to keep referring to Mormons as racists. I apologize to my Mormon friends for giving you that opening to vent your hate.
What *would* it take for Ford to be racist? You seem to indicate the answer is "nothing can make it racist, I win!".
That's nowhere near what I said and you know it. I "seem" to indicate that "the fact that some Mormons buy Fords does NOT make Ford run by the Mormon Church." "A does not mean B". That you're pretending I argued "A does not mean B thus B can never happen" says more about your logic and reading comprehension than mine.
But your arguments don't make sense.
Perhaps because you're making them up and trying to stuff them into my mouth and you're deliberately trying to not make sense?
If 100% of the directors, executives, and employees and 100% of the buyers were members of the KKK, would that make it racist (presuming the KKK is assumed to be de facto racist)?
Once again, FMC does not appoint its board of directors based on the fact that some racists, (or Mormons in the actual analogy) buy their cars. They appoint directors based on business acumen and management ability. So, in terms of the original issue, you're assuming because Mormon churches host BSA troops that BSA has automatically appointed the Mormon Church as their board of directors, and then asking if that doesn't mean BSA is run by the Mormon Church. If you can't see how that is assuming the result you need to prove, then I can't help you.
They said the malware was a 'highly critical threat to system security and integrity' and could not be detected by any operating system."
Can someone 'splain that, or is it just nonsense? The malware was put into the GPU or whatever by a program running on the OS, why can't another program on the OS detect it? Write Only Memory?
The "liberal" stance is to be the best you can be at what you want to be. If that's the best grunge rocker making $10 a month in tips, so be it.
The "conservative" stance is to be a banker, lawyer, doctor, and be the best at that.
There is no difference between the two. Both statements are trying to define "best person" by the occupation and not by the person himself. And that is why your attempt at bringing politics into this fails miserably, because you are simply wrong. You don't see liberal bankers sitting around wishing they were $10/month grunge rockers, they think about how to make more money. It ain't a liberal/conservative thing.
But I wasn't talking about being best at some specific occupation, because Boy Scouts isn't involved in trying to make anyone the best at a specific occupation, and most ethical systems (what you call "parameters") aren't, either. They deal with the person as a whole and not with what how well they do what they do for a living.
So, I'll ask again, since you managed to bypass the question by changing the context -- do YOU not think that your "parameters" are a basis for YOU being the best person YOU can be? Not the best belly dancer or the best banker or the best musician, but the best person. If so, then it is hypocritical to denounce Boy Scouts for believing their "parameters" are what leads people to be the best they can be. You think your parameters are best, they think theirs are. You can argue with what those parameters are (and now that I've opened the door you will probably try to twist this discussion that direction) but to claim that they are bad people because they think their parameters are best and not realize you do the same yourself is not honest.
What did I say to indicate that one was "hated" and the other "loved"?
Hyperbole. Billions of slashdot posters use it every day. You should be used to it by now. So you don't fully love one and don't fully hate the other. You like one and dislike the other. You smile when one is mentioned and frown for the other.
Pick your level and then deal with the concept that one teaches you to kill and the other to live when faced with limited resources, and that you prefer the one that taught you how to kill more efficiently.
Someone else equated the two, so I shared my experience with both.
Yes, we got that part. Your experience with scouts was that they forced you to camp out. Those bastards. Good thing Kenny wasn't a boy scout, the experience would probably have killed him.
First you claimed that the "racists running Ford" analogy was MY words, now you're claiming that it is your analogy being used against you. Seriously? Have you bothered paying attention to the discussion at all? You didn't seem to know that the claim I was responding to wasn't written by you, now the analogy authorship is slipping your mind.
Slashdot Mirror
Vision is something that's easy for humans but very hard for computers.
Good point. If autonomous vehicles are so great, why aren't spammers buying them to solve Captchas for them?
The SDC would already have the rules for that intersection in its knowledge base.
If you're going to continually assume perfection in perfectly available data, then you might as well just say you assume the self driving car is going to be perfect and leave it at that. No need for arguing about anything, we know the answer because you assume it will be that way. Only morons would produce autonomous vehicles that weren't perfect and weren't able to deal with an imperfect world, and we assume that there are no morons in the management or design teams.
And we assumed that the o-rings in the Challenger solid rocket boosters would deal with frigid temperatures just fine. That turned out ok, I guess. The original designers of the Tacoma Narrows Bridge designed a simply perfect bridge. And the much smaller problem of navigation databases for automotive GPSs is a perfectly solved problem, too. Nobody ever gets hurt by trusting GPS navigation databases today, so much more complex systems will be fine and we should trust them with our lives. The literature is just replete with documented cases of humans designing perfect systems that behave perfectly all the time.
Yes, I welcome our autonomous overlords, but pass me the dikes anyway.
If the car drives itself, then you can relax and enjoy the scenery even more.
Let's see, worry about whether the car will fail and try to kill me by driving off the edge of a mountain road, fail and suddenly say "you're it, start driving human, I can't deal with it", or simply drive myself and enjoy the trip. Hmmm. I can't decide which I'd rather do. What a quandary.
But the I think the point was, roads are not for pleasure, they're for transportation. What scenery am I supposed to enjoy?
No I didn't. I admitted that it would know that one sensor (out of dozens) had failed.
And that one sensor is the one that tells it where it is. Sensor failed. GPS data unreliable. Where are you?
Why would it "forget" where it was?
Because the sensor that tells it where it is is unreliable, and without that data you don't know where you are. You knew five minutes ago because the GPS was working. You don't know now because it isn't. In human terms, we call that "forgetting".
It could just read the inertial sensors to detect any acceleration
Oh. The cars will all have INS to backup GPS. Sorry. Didn't assume this.
read the rotary sensors in the wheels to determine its velocity,
That's why I was very careful to say you've forgotten where you are, not forgotten how fast you are going wherever it is you are going. Even so, I've always found my "rotary sensors in the wheels" to be about 3 MPH off compared to the GPS. You've lost your most accurate speed sensor and are putting your life in the hands of, essentially, your speedometer, and hoping that your tires are inflated to the calibrated value so the rotation into distance calculations are right.
and calculate where it is from its last known position.
How long prior to the "teleportation" was the GPS not working correctly? I've dealt with GPS. High precision RTK systems. There have been times when the GPS says "I'm working fine" despite producing garbage position reports, then it decides there is a problem and says "your guess is as good as mine", and then it usually figures it out. But the GPS was sure it knew what it knew before it knew it didn't know. And then it was just as certain it knew where it was, after it was certain it was somewhere else. I didn't trust that system with my life like you expect us all to trust it when it gets put in a car.
It could also continue to use cameras for lane detection,
"I'm in the middle lane of some road which I think is I65." Ok.
cameras for landmark recognition,
"I see a mountain off in the distance, I think that's Mt. McKinley. Maybe it's Mt. Hood. It's at a bearing of 54 degrees from me, and I'm on I65, I think, so I must be right here." Ok.
and a combination of cameras and radar for collision avoidance.
So it won't collide with someone else deliberately when it decides it needs to stop. Notice I didn't say "slam on the brakes sudden stop" or "just shuts down". Still didn't say that. It will cause a lot of hassles for everyone else on the road, though, and some of those cars may have decided THEY need to stop, too, or handed control over to the human because their GPS failed for the same reason yours did -- blocked GPS. Fascinating news story out of New Jersey, I think it was. Fellow was carrying a GPS blocker in his delivery truck to foil his boss's attempt at tracking him. Screwed up GPS for everyone near the major airport he was driving around.
You're putting a lot of faith in capitalists building cars that don't exist yet creating large scale systems of autonomous computers that may have interesting emergent behaviour. Every time one sensor fails you're hypothesizing another sensor to make up for it, sometimes using databases that don't exist and methods of navigation that are of such low accuracy you can't pilot a vehicle with the answers. It would be prudent to say "let's see how it turns out" instead of "it will be perfection and we'll all love it to death." It's that latter part that causes concern for some of us.
you can be completely compliant with HTML standards, and still have a website that doesn't work/look the same on one browser vs another.
You're talking to someone who writes website code by hand and maintains several. Of course you can have different looks on different browsers -- it's called "customization". Fonts, colors, width of screen, allowing pop-ups/js/cookies, all kinds of things. None of that deals with the issue of one manufacturer fielding an "extension" to his communications protocol that doesn't play well with someone else's extension.
if an Automated Vehicle Communications protocol is ever created, you can bet the government will be involved.
Which government? And how many of them? Will a committee of governments design a consistent camel, or will some of them be one hump and some two? Will the EU camel deal with data differently? Will the French camel surrender whenever it is threatened? Will the Chinese camel be reverse engineered and get it almost right?
because you're comparing something that will be enforced by laws and regulations
Because enforcement of perfect worldwide laws and regulations leads to perfection and full compliance everywhere.
Have you noticed, modern cars are getting full entertainment centers and communications capabilities built in at the same time that many states are cracking down on distracted driving? Laws and regulations are going to save us. Some laws say that you're too distracted if you simply talk on a phone, but a context-sensitive touch-screen menu system on a car audio system is perfectly ok.
You put a lot of faith in laws that don't exist yet, governing cars that don't exist yet.
No it doesn't. Only a complete moron would design it that way. And everyone else on the team, and the full management chain would need to be morons as well.
Two words about software designers and management chains: Windows 8. Need I say more?
But it certainly isn't going to just shut down.
I didn't say it just shut down. I said it decides it needs to stop.
You admitted it doesn't know where it is. If the GPS failed, that means the GPS-based course and speed indications are unreliable. Of course the designers may have have thought of their idea of what to do in such a situation. Just like you've had time to think about it and can't see any course of action that would be called "needs to stop" that isn't "just shut down" or "suddenly stop."
You blink, and while your eyes are closed you forget where you are and what direction you are going. Sounds like dementia. Welcome to GPS failure, citizen.
Maybe a few zany-looking cars of the type google uses to create street view will map out all the roads with bulky lidar and high-definition cameras, perform an insane amount of off-line processing, and then humans will hand-verify the location of everything that isn't clear. Cars will download all this as they drive along a new road so they are familiar with the road before they get there.
I seem to recall there being a story here in /. about an airport in Fairbanks Alaska needing to block off a taxiway because a GPS system kept telling people to go that way and cross an active runway to get to the airport. This was with cars where a human was in charge and no autonomous control whatsoever. But if we put a computer using GPS driving instructions in charge, nothing like this would ever happen?
"I'm getting nervous, please take over for me."
Sorry, I'm in the middle of the final episode of Breaking Bad and my hands are filled with roast beef sandwich. And I haven't hand driven for two years. That's your job. What do you expect me to do?
So there you go, google already knows the location of every curb and every speed bump, as of whenever one of their cars last went by.
Well, thank God that roads never change after Google takes pictures of them.
When the car in front of you slams on its brakes, how much context do you need?
Is the lane to the left clear so I can deviate that direction and avoid a collision?
If the lane to my left is oncoming traffic, can I safely cross the double yellow line to avoid serious injury?
Is the shoulder clear so I can go that direction?
If the shoulder is not clear, does it have an obstruction that is less valuable and causes less damage if I run it over instead of colliding with the car ahead of me?
Is there a large truck or other similar vehicle behind me that I need to worry about?
Did the driver of the large vehicle appear to be awake and alert when I passed him a few minutes ago?
Does there appear to be a traction limiting condition directly ahead of me that is not to the left or right? E.g, ice or snow?
Is there a traction limiting condition to the left or right?
Did the car ahead of me slam on his brakes for a transient reason (to miss a squirrel that is now out of the way) or because it was about to run into a stopped semi truck?
Enough context? The fact that you only need one piece of information in this kind of situation tells me you don't drive much, and shouldn't be driving the autonomous vehicle bandwagon.
We only have one data set for real world performance of an autonomous vehicle, 350k miles by Google's prototype. It has been involved in one accident- when a person backed into it. So it has been perfect thus far.
One data point. Many years ago I remember the buzz about a new field of study -- emergent behaviour in clusters of simple robotic devices. For example, you can program a small robot to obey simple rules like when to turn, when to go, when to stop, etc, and predict pretty well what it will do. But when you put a roomful of them together they start doing things that weren't obvious. I wish I could remember the name of the guy who was at the forefront of this. Randy something, I think.
The point being, one computer programmed to do one thing is something we can pretty much understand, and predict how it will react. When you get large collections of systems interacting, you sometimes find miraculous, and sometimes not so miraculous, results. Kind of like, if you have one computer programmed to watch the stock market and make trades you will probably be ok. When you get thousands of them all interacting indirectly (via the stock prices, e.g.) and all acting in milliseconds, you can get havok. That's why I referred to that in my original post.
So, I'm really surprised you would respond with a statement that we know the world filled with autonomous vehicles will be safe and accident free because one autonomous vehicle acting in controlled conditions and in a human-centric environment had only one accident. I suspect that we will find all kinds of unpleasant emergent behaviour as the roads begin to have more and more of these things.
I'm also fascinated that you see this as an issue of cars vs. horses, since in neither case was the human removed from "driving".
I may be ignorant to my states laws but I have never heard of slowing down a requirement for NY.
Oregon. The law says you must slow down or pull over when going past an emergency vehicle on the side of the road. Not every vehicle on the side, just emergency ones. This points out another risk of autonomous vehicles. Will they be programmed with all the laws of all the states they may be driven in, and how will they deal with people who live close enough to a border that there are two sets of laws?
Here's another interesting law. In North Carolina, if you have the wipers on, you must also have your headlights on. Also in NC, there is an arcane twist to the laws regarding stopping for police. I don't recall the details, but I think there are exceptions of when you have to stop for flashing lights from the car behind you (cops), based on a rash of fake cops pulling people over.
In fact I feel it makes things a nuisance and possibly dangerous. Most cops I see position themselves well to mitigate risk.
They do their best. As do road workers. And yet, they get run over by drivers to the point that there are special laws for work zones and now emergency vehicles.
If the GPS suddenly reports that the car has instantaneously transported itself to a new location, I don't think it will be blindly trusted.
You're in the middle lane of I65 going 65MPH surrounded by other cars when YOUR car suddenly decides it doesn't know where it is and needs to stop. Welcome to computer dementia. Happy travels, citizen!
Traffic lights are already designed to be as visible as possible,
When your initial premise is wrong, expect the rest to be wrong, too.
Plus, if these cars were mostly used for regular commutes & shopping trips along the same routes, it would probably remember where it usually stops for red lights,
Judge: "Why didn't you signal your right turn?"
Me: "I always turn right at that intersection. Everyone knows that."
Judge: "Why, not guilty of course. How stupid of me to forget what you 'always do'."
Of course everyone will know where you "usually" stop for a red light and know you're going to stop there again when your computer fails to detect the green light.
And no matter how capable you are, a computer has a faster reaction time.
And that's why when you make a mistake with a computer's help, you can do much much more damage than doing it by hand. It takes a long time to shred 100 paper files, and you can figure out after five minutes that it is a mistake and stop, and still have half of them. If you do that to computer files, five minutes later means you have nothing left but the regrets.
I'd go find a reference to the havok created by computer trading on wall street, where instant reaction times lead to financial catastrophe, but I think we probably are all well aware of these stories. Why we want to ignore those stories and assume that a massive system of independent computers would be perfection personified this time, when they've been problematic before, is a mystery. I sense a lot of money to be made for people shilling autonomous vehicles on an unsuspecting public...
Roads are for transportation not pleasure.
I'll be sure to tell the local and state road departments to take down all the "Scenic Loop" signs and close all the "viewpoint" stops in the mountain roads. Transportation, not pleasure, you hedonistic bastards!
It's too bad in your universe that roads cannot serve more than one purpose.
Well, once self-driving cars fill the roads in significant numbers, if they'll have provisions for mutual communication and data exchange, you can count on them being more polite to each other than human drivers would.
What anthropomorphic hogwash. The cars will not be polite. They will behave as they were programmed to.
But here's the rub. My car speaks Toyota HTML, yours speaks Hyundai HTML, and the next guy's speaks Microsoft HTML. Kinda all the same language, somewhat close to the standards, but each manufacturer does it better by doing it their way just a bit.
I spent a wonderful half hour of my life I'll never get back writing with Word some material I wanted to put on the web. I kept indenting some paragraphs and exporting the document as HTML, but the web browser never showed an indent. Then I realized I was looking at it with Mozilla and not IE. Isn't it great that HTML is a standard? Can we survive when our tools that we depend on to keep us alive don't use exactly the same standards?
The story says it will pass slower traffic. Great. But will it detect that guy you've seen in your rear view mirror switching lanes doing 40mph faster than you? I see a great chance to have a multicar accident just on that one thing.
More important, will it realize that the reason the fast lane is empty and everyone is slowing down in your lane is because there is a state trooper with someone pulled over on the left side of the road and your state's law says you must either pull over or slow down when going past him? Will you get to be the next recipient of a ticket from that trooper because your car endangered him?
what is curious for me is at what point will insurance companies insist on automated assist drive systems in cars especially considering 95% of crashes are due to human error.
Citation? And you do realize, I hope, that 95% (if true) being human error is because there currently is no "autonomous computer failure" category or contestants as regular participants. I expect that the numbers will be vastly different when there are.
Unless, of course, the NTSB uses the same standards for car crashes it does for aircraft mishaps and everything is labelled driver (pilot) error as an overarching cause. "Driver did not properly supervise the autonomous vehicle that wasn't supposed to require his supervision anyway."
Get ready to be annoyed for about 30 years, because "automated" commonly means "more automated than before," not "automated in every conceivable way."
Perhaps that's why the word being used is "autonomous", not "automated".
They do what you do - they also watch for and avoid other cars,
Part of fixing the bug that says "stop at any red light no matter where it is" results in "don't pay any attention to the red lights on the cars in front of you."
pedestrians
In my state, you aren't a pedestrian to be stopped for unless you are IN the crosswalk. Standing on the side of the road looking wistfully at cars as they pass doesn't cut it.
(I guess as a backup it could know the GPS location of stoplights and stop if it doesn't see the light and confirm that it is green).
So we're all expecting that they will stop when they have absolutely no reason to, which will certainly do wonders for improving traffic flow and safety. I love the idea of autonomous cars more and more each time the topic comes up.
But I am sure we will end up with some level of instrumentation on the road such as stoplights that emit at a frequency not obfuscated by sunlight, snow, etc, like visible light is.
We already have instrumentation in the roads that detect when cars are there so they can cycle the traffic lights. Before they came up with one sensitive enough for bikes, it was fun to watch a law-abiding cyclist (yes, there are some) stopped at a red light waving his bike around the road hoping to trigger the signal so he could go.
Sometimes, though, the sensor wasn't sensitive enough for cars, and other than backing up and trying again, there was no way of "waving your car" to try to get it to trigger, so you were left with deciding how long was too long and when you could justifiably run the red light. I can imagine a world where that sensor has a transmitter to tell the car it is at an intersection with a red light so the car will stop, and a world where a few critical transmitters fail and new autonomous not autonomous vehicles blow through a red and get t-boned by a semi.
I for one welcome our new robot overlords. Hand me the dikes, I need to clip a few wires ...
Most likely there will be dedicated lanes to encourage people to adopt the tech.
Yeah, we have so many empty lanes now that dedicating one of them to a small percentage of cars that are "not autonomous autonomous vehicles" will be a really smart idea to solve traffic problems.
This "not autonomous autonomous vehicle" automatically stops at red lights? How nice. The local store has a red light in their signage out front and all the cars stop. Good for the store. Bad for the traffic, especially the updated version of the "not autonomous autonomous vehicle" which has that bug fixed and doesn't feel like stopping, and runs into V1.0.
"Hello, Toyota support? My autonomous vehicle computer crashed." "Did you try to reboot?" "No, I mean it crashed into the autonomous vehicle computer in the car ahead of me."
The headline is deliberately misleading. It's not an autonomous vehicle if is isn't autonomous, which even the summary admits.
The analogy (what it takes for a car company to be racist) is yours.
First, you failed to quote all the words of the analogy. I doesn't start with the word "more". Second, it is not an analogy that says what it takes for a car company to be racist, it is an analogy about what DOES NOT MAKE A CAR COMPANY RUN BY A CHURCH.
You read my words as me talking to you. I wasn't talking to you,
You replied to something I posted, so I guessed you were talking to me. You wrote about "my words" without attributing that statement to anyone else, unquoted, presumably referring to yourself. And you talked about an analogy being used against you, which is what I have been doing continually - using it against you. My bad for not reading your mind instead of your words.
You've come up with an analogy, and when that analogy is explored to try to gauge the thresholds to then compare to the original situation, you play dumb (or aren't playing).
Your question asking if Ford did have racists on their board would they be run by racists is not a "threshold", it's asking "if Ford was run by racists would they be run by racists" without dealing with the issue of how those racists got there. It's a worthless question, and I'm not playing the game you want to play. The issue I've been responding to is a claim that BSA is run by the Mormon church because some Mormon churches host BSA troops. That's the game of the day.
Here's the fact. The fact that some Mormon churches host BSA troops does not mean BSA is run by the Mormon church. Period. End of sentence. That's the claim that I replied to, that's what you need to disprove if you want to argue with me about it. I'm sorry the analogies confused you, especially to the point that you need to keep referring to Mormons as racists. I apologize to my Mormon friends for giving you that opening to vent your hate.
What *would* it take for Ford to be racist? You seem to indicate the answer is "nothing can make it racist, I win!".
That's nowhere near what I said and you know it. I "seem" to indicate that "the fact that some Mormons buy Fords does NOT make Ford run by the Mormon Church." "A does not mean B". That you're pretending I argued "A does not mean B thus B can never happen" says more about your logic and reading comprehension than mine.
But your arguments don't make sense.
Perhaps because you're making them up and trying to stuff them into my mouth and you're deliberately trying to not make sense?
If 100% of the directors, executives, and employees and 100% of the buyers were members of the KKK, would that make it racist (presuming the KKK is assumed to be de facto racist)?
Once again, FMC does not appoint its board of directors based on the fact that some racists, (or Mormons in the actual analogy) buy their cars. They appoint directors based on business acumen and management ability. So, in terms of the original issue, you're assuming because Mormon churches host BSA troops that BSA has automatically appointed the Mormon Church as their board of directors, and then asking if that doesn't mean BSA is run by the Mormon Church. If you can't see how that is assuming the result you need to prove, then I can't help you.
They said the malware was a 'highly critical threat to system security and integrity' and could not be detected by any operating system."
Can someone 'splain that, or is it just nonsense? The malware was put into the GPU or whatever by a program running on the OS, why can't another program on the OS detect it? Write Only Memory?
The "liberal" stance is to be the best you can be at what you want to be. If that's the best grunge rocker making $10 a month in tips, so be it. The "conservative" stance is to be a banker, lawyer, doctor, and be the best at that.
There is no difference between the two. Both statements are trying to define "best person" by the occupation and not by the person himself. And that is why your attempt at bringing politics into this fails miserably, because you are simply wrong. You don't see liberal bankers sitting around wishing they were $10/month grunge rockers, they think about how to make more money. It ain't a liberal/conservative thing.
But I wasn't talking about being best at some specific occupation, because Boy Scouts isn't involved in trying to make anyone the best at a specific occupation, and most ethical systems (what you call "parameters") aren't, either. They deal with the person as a whole and not with what how well they do what they do for a living.
So, I'll ask again, since you managed to bypass the question by changing the context -- do YOU not think that your "parameters" are a basis for YOU being the best person YOU can be? Not the best belly dancer or the best banker or the best musician, but the best person. If so, then it is hypocritical to denounce Boy Scouts for believing their "parameters" are what leads people to be the best they can be. You think your parameters are best, they think theirs are. You can argue with what those parameters are (and now that I've opened the door you will probably try to twist this discussion that direction) but to claim that they are bad people because they think their parameters are best and not realize you do the same yourself is not honest.
What did I say to indicate that one was "hated" and the other "loved"?
Hyperbole. Billions of slashdot posters use it every day. You should be used to it by now. So you don't fully love one and don't fully hate the other. You like one and dislike the other. You smile when one is mentioned and frown for the other.
Pick your level and then deal with the concept that one teaches you to kill and the other to live when faced with limited resources, and that you prefer the one that taught you how to kill more efficiently.
Someone else equated the two, so I shared my experience with both.
Yes, we got that part. Your experience with scouts was that they forced you to camp out. Those bastards. Good thing Kenny wasn't a boy scout, the experience would probably have killed him.
Oh no, Someone used my own analogy against me!
First you claimed that the "racists running Ford" analogy was MY words, now you're claiming that it is your analogy being used against you. Seriously? Have you bothered paying attention to the discussion at all? You didn't seem to know that the claim I was responding to wasn't written by you, now the analogy authorship is slipping your mind.