Cars discussion, split from 'other news'

Seen something interesting in the news or on the intertubes? Discuss it here.

Moderators: Zamfir, Hawknc, Moderators General, Prelates

cphite
Posts: 1370
Joined: Wed Mar 30, 2011 5:27 pm UTC

Re: Cars discussion, split from 'other news'

Postby cphite » Tue Jun 30, 2015 3:17 pm UTC

Thesh wrote:If Google's self-driving cars hadn't already been driving around safely for several years now, putting on well over a million miles, you might have a point, but their record is better than the average driver.


Under controlled conditions, yes. They've mostly driven in good weather, on routes that were planned out to avoid unexpected problems, and they're under constant surveillance. So the fact that their record is better isn't surprising. If you took a group of human drivers and made sure they had the same conditions over the same period of time, it wouldn't be surprising if they had better than average records.

What we haven't seen yet is how these things perform in completely random traffic, dealing with bad weather and unexpected conditions. Currently there is no industry standard on how to secure the navigation systems against outside interference - deliberate or incidental. There is no information on how reliable the navigation and control systems are; even the best computer systems are susceptible to hardware and software problems. What happens, exactly, if the computer crashes or freezes when you're going 65mph? What backup controls are in place?

User avatar
Quercus
Posts: 1810
Joined: Thu Sep 19, 2013 12:22 pm UTC
Location: London, UK
Contact:

Re: Cars discussion, split from 'other news'

Postby Quercus » Tue Jun 30, 2015 3:31 pm UTC

morriswalters wrote:In an ideal sense autonomous cars may achieve that goal, even for people who can't see that sunset, but want to enjoy it with their friends.

Absolutely. I am all in favour of autonomous cars. I do enjoy manual driving, but I'd give that up in a heartbeat if it meant greater safety for the public and more choices for people who can't drive manually for whatever reason. What I'd eventually like to see is ubiquitous autonomous car sharing - if you want to do more shopping than you can carry on the train/bus, rent a car for an hour or two, if you want to spend a month in a cabin up in the mountains, rent a car for a month (or even better, rent it for a day to take you and your supplies to the cabin, then send it back on its own; then rent another car to come and pick you up at the end of your stay).

Having the use of a car lets you do things you couldn't do otherwise, owning a car is mainly just expense and hassle, at least in my experience.

cphite
Posts: 1370
Joined: Wed Mar 30, 2011 5:27 pm UTC

Re: In other news... (humorous news items, etc)

Postby cphite » Tue Jun 30, 2015 3:41 pm UTC

commodorejohn wrote:
cphite wrote:No, instead we'll only have to worry about network connectivity, accurate maps, and a perfectly functioning computer. For example, the missus and I went out driving this past weekend and several times her GPS told us to turn when there wasn't actually a road. Not a big deal when it's just a phone saying "turn right here" because you can ignore that... a bit of a potential mess when the computer is actually driving. You may trust Google to perform flawlessly at highway speeds; I'm not quite there yet.


What he said. It really does astonish me how easily people can completely blow off these concerns. It's like nobody has ever encountered buggy software before, or like they think that it's just some kind of literal inevitability that all problems with technology must have viable solutions that will inevitably be discovered. I guess that's the power of Futurism, you can just insist that some day (probably Real Soon Now!) all the problems will be solved and trust in the fact that it's unfalsifiable so long as you keep kicking it further down the road into the indefinitely near future...


And that's exactly it... right now, Google isn't even at the point where it can tell you, the driver, where you should be going with 100% accuracy... and now you're going to trust it to drive? Aside from occasionally wanting you to turn where there isn't a road, it also isn't uncommon for any GPS unit to place you at the wrong spot, or to lose you for a moment. And these things are a Really Big Deal if the system is controlling the car.

And yes, obviously there are going to be other systems in place. The car will have sensors to detect obstacles, for example. But those need to function perfectly. One of the arguments I keep seeing is that computers are so much faster than humans... but that's only true when the computer is functioning properly. What will be in place to account for a frozen computer? Or a sudden loss of power? Or malfunctioning sensors? What is in place to prevent malicious or even accidental interference?

Some folks have brought up auto-pilot in planes... but there are two critical differences between auto-pilot and what we're talking about here. The first is that, as was already pointed out, there is simply a lot less stuff to consider in an aircraft. It may seem counter-intuitive but it's true. There is lot more room between you and other aircraft. There are a lot less obstacles to keep track of. There are a lot fewer surprises.

The second thing is this: In a plane, you have a pilot. You have a trained and experienced pilot who is ready to take over at a seconds notice. If the auto-pilot stops working or is doing something wrong, you have a pilot to take back control. You cannot count on that in a car. In a car, you have a random person who may or may not even be paying attention; and who, even if they notice something is wrong, may or may not actually have the presence of mind to do something about it.

I'm not against technology; I'm not even against this technology. I just think folks are jumping the gun a bit. There are still some pretty significant problems to work out, and it seems like folks are just blindly assuming these problems will be worked out because they'd really like it to be that way.

User avatar
PeteP
What the peck?
Posts: 1451
Joined: Tue Aug 23, 2011 4:51 pm UTC

Re: Cars discussion, split from 'other news'

Postby PeteP » Tue Jun 30, 2015 4:08 pm UTC

I assume they will be worked out because I have no reason to believe they fall into the category of problems we (as in humanity) can't solve without some new major breakthrough. I can't predict how many years that will take exactly, but when exactly they will be ready for market doesn't seem particularly relevant to this thread.

User avatar
HES
Posts: 4896
Joined: Fri May 10, 2013 7:13 pm UTC
Location: England

Re: Cats discussion, split from 'other mews'

Postby HES » Tue Jun 30, 2015 4:14 pm UTC

The way I see it, in ~5 years time driverless vehicles will be commonplace, ~20 years the majority, ~50 years humans will be banned.

I think a lot of the arguments against self-driving vehicles either:
a) underestimate how good the technology is, or will be.
b) overestimate the quality of human drivers.

Driverless cars might crash, they might get hacked, they might get lost. But I bet my life they'll do it less than driverful cars do today.
He/Him/His Image

User avatar
PeteP
What the peck?
Posts: 1451
Joined: Tue Aug 23, 2011 4:51 pm UTC

Re: Cars discussion, split from 'other news'

Postby PeteP » Tue Jun 30, 2015 4:28 pm UTC

Hacking is the thing I'm most worried about. Coordinating with other cars for instance has advantages. Getting information from a central place and sending infos back has it's uses. Being able to tell your car to search for a parking place itself and then call it back to you would be useful. Car sharing where it automatically drive to where someone needs is has it's uses. And every communication is a place where there could be a security weakness. Also with systems where you use automatic cars as an unmanned taxi service attackers also might be able to get physical access.

Derek
Posts: 2181
Joined: Wed Aug 18, 2010 4:15 am UTC

Re: Cars discussion, split from 'other news'

Postby Derek » Tue Jun 30, 2015 4:41 pm UTC

Quercus wrote:A world where people can't decide at midnight that they want to go to the sea and watch the sunrise is a duller, less magical world, and I would rather work towards a world where everyone can do things like that, instead of no-one being able to.

You can still do this with self-driving cars. You just won't be the one driving at 3 AM.

Tyndmyr
Posts: 11443
Joined: Wed Jul 25, 2012 8:38 pm UTC

Re: Cars discussion, split from 'other news'

Postby Tyndmyr » Tue Jun 30, 2015 4:50 pm UTC

gmalivuk wrote:Sure, there can be real joy in driving. People genuinely enjoy lots of dangerous liesure activities that have been restricted or banned in many places in the interests of public safety.


We do not necessarily need to outright ban driving. And in any case, surely tracks, etc will exist where those who wish to drive for thrills can enjoy that. Racing is fun, but you probably shouldn't drag race on the freeway.

And...one of those big dangers is inattentiveness, be it due to diminished awareness/reflexes from booze, tiredness, cell phones, or whatever else. Driving for pleasure isn't those things. Plus, a roadway filled with automated vehicles will be predictable. If the tech proliferates, you don't really need to ban the alternative. Being more convenient is reason enough for something to dominate.

Sure, there may be niche oldsters who enjoy taking the wheel themselves for a sunday drive. Whatever. They'll diminish to irrelevance with time. We didn't need to ban rotary phones to effectively eliminate them, and automated driving will be similar.

User avatar
Quercus
Posts: 1810
Joined: Thu Sep 19, 2013 12:22 pm UTC
Location: London, UK
Contact:

Re: Cars discussion, split from 'other news'

Postby Quercus » Tue Jun 30, 2015 5:02 pm UTC

Derek wrote:
Quercus wrote:A world where people can't decide at midnight that they want to go to the sea and watch the sunrise is a duller, less magical world, and I would rather work towards a world where everyone can do things like that, instead of no-one being able to.

You can still do this with self-driving cars. You just won't be the one driving at 3 AM.

I realise this, I was responding to Sardia's assertion that the freedom offered by car ownership (or, if we're being precise, the ability to use long range personal transportation) is artificial and could be entirely replaced by better public transportation - my response to that was basically "that depends on where you want to go, and when".

Tyndmyr wrote:Sure, there may be niche oldsters who enjoy taking the wheel themselves for a sunday drive. Whatever. They'll diminish to irrelevance with time. We didn't need to ban rotary phones to effectively eliminate them, and automated driving will be similar.

I suspect that that analogy isn't quite accurate, as I don't think I've ever come across someone describing using a rotary phone as fun. I would say a closer analogy would be sailing. Plenty of people sail, even though it's far more efficient to use a motorboat and has been for over one hundred years. I feel that a few percent of drivers will still drive manually as long as they are allowed to, even among generations born after automated cars become ubiquitous. Whether that generates enough problems to justify banning it remains to be seen.

Derek
Posts: 2181
Joined: Wed Aug 18, 2010 4:15 am UTC

Re: Cars discussion, split from 'other news'

Postby Derek » Tue Jun 30, 2015 5:11 pm UTC

I don't think I've ever come across someone describing using a rotary phone as fun

I think using a rotary phone is fun. My parents still have one, and as I kid I would often make a point to use it instead of one of the cordless phones, just because it's fun to spin the wheel.

morriswalters
Posts: 7073
Joined: Thu Jun 03, 2010 12:21 am UTC

Re: Cars discussion, split from 'other news'

Postby morriswalters » Tue Jun 30, 2015 5:44 pm UTC

PeteP wrote:Hacking is the thing I'm most worried about. Coordinating with other cars for instance has advantages. Getting information from a central place and sending infos back has it's uses. Being able to tell your car to search for a parking place itself and then call it back to you would be useful. Car sharing where it automatically drive to where someone needs is has it's uses. And every communication is a place where there could be a security weakness. Also with systems where you use automatic cars as an unmanned taxi service attackers also might be able to get physical access.
I share your concerns about hacking, but as much of a problem as it will be we are close to the point where that fear may be already realized. Adaptive cruise control, the ability to keep you in lane, and automated parking are hackable now and at least in theory a car could be commandeered today via some hack or another.
Quercus wrote:Having the use of a car lets you do things you couldn't do otherwise, owning a car is mainly just expense and hassle, at least in my experience.
I never travel in my own cars any more, I rent one, convenience trumps sanity. I'm meeting friends this weekend at a National Park, much like your scenario, courtesy of a rental company.

Trebla
Posts: 387
Joined: Fri Apr 02, 2010 1:51 pm UTC

Re: In other news... (humorous news items, etc)

Postby Trebla » Tue Jun 30, 2015 6:15 pm UTC

cphite wrote:I'm not against technology; I'm not even against this technology. I just think folks are jumping the gun a bit. There are still some pretty significant problems to work out, and it seems like folks are just blindly assuming these problems will be worked out because they'd really like it to be that way.


I think it's worth pointing you that the most significant hurdles have already been overcome... the "general ability to drive autonomously" was the far greater challenge involving sensory inputs, spatial recognition and even advanced infrastructure mapping and all kinds of other things beyond the layperson's understanding (i.e., my understand). The remaining problems "driving in bad weather," "identifying different surface types," "more accurately accounting for potholes" are relatively minor if not outright insignificant. There are just a slew of these minor problems that all kind of add up, but each is surmountable with a (again relatively) small amount of effort when compared to the general problem.

KrytenKoro
Posts: 1487
Joined: Tue Apr 05, 2011 2:58 pm UTC

Re: In other news... (humorous news items, etc)

Postby KrytenKoro » Tue Jun 30, 2015 6:55 pm UTC

cphite wrote:And that's exactly it... right now, Google isn't even at the point where it can tell you, the driver, where you should be going with 100% accuracy... and now you're going to trust it to drive?

I would be overjoyed to meet the human who could tell me where I should be going with the accuracy that GPS accomplishes.

Aside from occasionally wanting you to turn where there isn't a road, it also isn't uncommon for any GPS unit to place you at the wrong spot, or to lose you for a moment. And these things are a Really Big Deal if the system is controlling the car.

This as well. They must have a better breed of human where you live.

And yes, obviously there are going to be other systems in place. The car will have sensors to detect obstacles, for example. But those need to function perfectly. One of the arguments I keep seeing is that computers are so much faster than humans... but that's only true when the computer is functioning properly. What will be in place to account for a frozen computer? Or a sudden loss of power? Or malfunctioning sensors? What is in place to prevent malicious or even accidental interference?

More systems than there currently are to prevent the same in humans?

The second thing is this: In a plane, you have a pilot. You have a trained and experienced pilot who is ready to take over at a seconds notice. If the auto-pilot stops working or is doing something wrong, you have a pilot to take back control. You cannot count on that in a car. In a car, you have a random person who may or may not even be paying attention; and who, even if they notice something is wrong, may or may not actually have the presence of mind to do something about it.

Planes are safer than cars, yes.

I'm not against technology; I'm not even against this technology. I just think folks are jumping the gun a bit. There are still some pretty significant problems to work out, and it seems like folks are just blindly assuming these problems will be worked out because they'd really like it to be that way.

Who, exactly, is blindly assuming they'll just be worked out?

The most generic argument I've seen above is that these computer-driven cars are demonstrably better than the average human. No one's saying they're perfect now, and I'm not seeing anyone say that they will ever be completely unassailable. The argument you're running with is "well, they aren't better than a hypothetical perfect human, so let's not be enthusiastic about perfecting this technology," and that seems just massively unrealistic.
From the elegant yelling of this compelling dispute comes the ghastly suspicion my opposition's a fruit.

User avatar
gmalivuk
GNU Terry Pratchett
Posts: 26823
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

Re: In other news... (humorous news items, etc)

Postby gmalivuk » Tue Jun 30, 2015 7:17 pm UTC

cphite wrote:And that's exactly it... right now, Google isn't even at the point where it can tell you, the driver, where you should be going with 100% accuracy... and now you're going to trust it to drive? Aside from occasionally wanting you to turn where there isn't a road, it also isn't uncommon for any GPS unit to place you at the wrong spot, or to lose you for a moment. And these things are a Really Big Deal if the system is controlling the car.
That system, by itself, isn't controlling the car. No one's talking about giving control of your car entirely to your phones Google Maps app.

The second thing is this: In a plane, you have a pilot. You have a trained and experienced pilot who is ready to take over at a seconds notice. If the auto-pilot stops working or is doing something wrong, you have a pilot to take back control. You cannot count on that in a car. In a car, you have a random person who may or may not even be paying attention; and who, even if they notice something is wrong, may or may not actually have the presence of mind to do something about it.
And with close-to-modern technology, in the car you will also have a driver, trained and experienced and ready to take over.

No one is denying that there are still problems. They just aren't as insurmountable as you people are making it out.
Unless stated otherwise, I do not care whether a statement, by itself, constitutes a persuasive political argument. I care whether it's true.
---
If this post has math that doesn't work for you, use TeX the World for Firefox or Chrome

(he/him/his)

User avatar
ucim
Posts: 6890
Joined: Fri Sep 28, 2012 3:23 pm UTC
Location: The One True Thread

Re: Cars discussion, split from 'other news'

Postby ucim » Tue Jun 30, 2015 7:35 pm UTC

gmalivuk wrote:And with close-to-modern technology, in the car you will also have a driver, trained and experienced and ready to take over.
I don't think this is actually the case. If the car is actually driving itself, the "driver" is probably not paying much attention, lulled into a sense of security by the computer. When driving in ordinary conditions, you are often only a fraction of a second away from disaster, should "something go wrong". The driver organic systems manager needs to be on constant alert, and if he or she is, they may as well be driving themselves.

Most likely, they'll be texting, reading the newspaper, eating breakfast, and when "something goes wrong", they won't know what hit them.

Jose
Order of the Sillies, Honoris Causam - bestowed by charlie_grumbles on NP 859 * OTTscar winner: Wordsmith - bestowed by yappobiscuts and the OTT on NP 1832 * Ecclesiastical Calendar of the Order of the Holy Contradiction * Heartfelt thanks from addams and from me - you really made a difference.

KrytenKoro
Posts: 1487
Joined: Tue Apr 05, 2011 2:58 pm UTC

Re: Cars discussion, split from 'other news'

Postby KrytenKoro » Tue Jun 30, 2015 7:47 pm UTC

Most likely, they'll be texting, reading the newspaper, eating breakfast, and when "something goes wrong", they won't know what hit them.

Yes, but what about when the computer is driving the car?
From the elegant yelling of this compelling dispute comes the ghastly suspicion my opposition's a fruit.

User avatar
Dauric
Posts: 3998
Joined: Wed Aug 05, 2009 6:58 pm UTC
Location: In midair, traversing laterally over a container of sharks. No water, just sharks, with lasers.

Re: Cars discussion, split from 'other news'

Postby Dauric » Tue Jun 30, 2015 7:57 pm UTC

KrytenKoro wrote:
Most likely, they'll be texting, reading the newspaper, eating breakfast, and when "something goes wrong", they won't know what hit them.

Yes, but what about when the computer is driving the car?

The point was in case of computer failure the human driver takes over, so we're talking about being suddenly sans-autopilot because of a failure. Problem is the vast majority of drivers, when presented with an autopilot, will flip out their smart phones and text or surf the web, or generally be -less- attentive than if they were actively driving the car, and thus less aware of a computer fault, or an unexpected circumstance the computer isn't handing in a safe manner.

Even without distractions, someone actively engaged in an activity is more aware of their surroundings than someone who is observing but not engaged in the activity. It's a problem with security posts that monitor a lot of closed-circuit cameras. They have to rotate guards at the cameras frequently or else they run the risk of effectively falling asleep watching nothing of interest happen on the monitors.
We're in the traffic-chopper over the XKCD boards where there's been a thread-derailment. A Liquified Godwin spill has evacuated threads in a fourty-post radius of the accident, Lolcats and TVTropes have broken free of their containers. It is believed that the Point has perished.

Derek
Posts: 2181
Joined: Wed Aug 18, 2010 4:15 am UTC

Re: Cars discussion, split from 'other news'

Postby Derek » Tue Jun 30, 2015 8:00 pm UTC

ucim wrote:
gmalivuk wrote:And with close-to-modern technology, in the car you will also have a driver, trained and experienced and ready to take over.
I don't think this is actually the case. If the car is actually driving itself, the "driver" is probably not paying much attention, lulled into a sense of security by the computer. When driving in ordinary conditions, you are often only a fraction of a second away from disaster, should "something go wrong". The driver organic systems manager needs to be on constant alert, and if he or she is, they may as well be driving themselves.

Most likely, they'll be texting, reading the newspaper, eating breakfast, and when "something goes wrong", they won't know what hit them.

Jose

Given sudden adverse situations (ex, the guy in front of you slams his breaks, or someone runs a red light), the automated system will probably respond faster and better than a human possibly could. Given a situation that the automated system simply can't handle well, like rain or construction, the system will usually have plenty of time to alert the driver and slow down if necessary so they can take control.

User avatar
Dauric
Posts: 3998
Joined: Wed Aug 05, 2009 6:58 pm UTC
Location: In midair, traversing laterally over a container of sharks. No water, just sharks, with lasers.

Re: Cars discussion, split from 'other news'

Postby Dauric » Tue Jun 30, 2015 8:21 pm UTC

Derek wrote:Given sudden adverse situations (ex, the guy in front of you slams his breaks, or someone runs a red light), the automated system will probably respond faster and better than a human possibly could. Given a situation that the automated system simply can't handle well, like rain or construction, the system will usually have plenty of time to alert the driver and slow down if necessary so they can take control.


I think the problem with computer control is less one of external circumstances, those are being fairly well handled by the guys doing the engineering at this point, but rather one of computer fault/breakage. I work in I.T. support, everything breaks. Now if most of the computers in a modern car fail (which they will do on occasion, everything can fail at one time or another) the direction and speed of the vehicle is still under driver control. Get some kind of component failure at 60 mph in steering or speed control without an attentive driver and it's more of a problem.

And sure, many systems have internal diagnostics for detecting faults, those diagnostics themselves can be subject to faults (and oh isn't it fun when the bits that are supposed to be telling you something's broke are telling you either nothing is wrong, or something is wrong with a component that isn't broken...)

Tests on sunny California blacktop are one thing, but I'd like to know they run their systems through some physical stress testing, like washboard dirt roads at 40 MPH like you find in rural U.S. Simulate an hour or two of that every day as part of a daily commute and check circuit boards and sensors for cracks and bent/broken mounting brackets.

commodorejohn
Posts: 1198
Joined: Thu Dec 10, 2009 6:21 pm UTC
Location: Placerville, CA
Contact:

Re: In other news... (humorous news items, etc)

Postby commodorejohn » Tue Jun 30, 2015 8:27 pm UTC

KrytenKoro wrote:I would be overjoyed to meet the human who could tell me where I should be going with the accuracy that GPS accomplishes.

Assuming we read "Google Maps" for "GPS," I'll direct you to the clerk at Goin' Postal in Superior, Wisconsin, who helped me figure out that the FedEx processing center wasn't actually where Google said it was, on the edge of the downtown area, and was instead out in the big-box district several miles away.

As for the notion that computer-driven cars can just tell the driver to take over in problem situations, I've yet to meet the human being who can context-switch fast enough for that to be at all feasible, let alone safe.
"'Legacy code' often differs from its suggested alternative by actually working and scaling."
- Bjarne Stroustrup
www.commodorejohn.com - in case you were wondering, which you probably weren't.

cphite
Posts: 1370
Joined: Wed Mar 30, 2011 5:27 pm UTC

Re: In other news... (humorous news items, etc)

Postby cphite » Tue Jun 30, 2015 9:35 pm UTC

gmalivuk wrote:
cphite wrote:And that's exactly it... right now, Google isn't even at the point where it can tell you, the driver, where you should be going with 100% accuracy... and now you're going to trust it to drive? Aside from occasionally wanting you to turn where there isn't a road, it also isn't uncommon for any GPS unit to place you at the wrong spot, or to lose you for a moment. And these things are a Really Big Deal if the system is controlling the car.


That system, by itself, isn't controlling the car. No one's talking about giving control of your car entirely to your phones Google Maps app.


No; we're talking about a system that depends on Google Maps or something like it as input. The point is, Google Maps isn't always accurate; and even when it's accurate, it isn't always timely. So whatever system that is driving your car needs to account for the fact that the positioning information it's getting is wrong, or delayed. If I can't count on Google Maps (or whatever vendor) to provide good positional information for me to use, I can't count on it giving that same information to my car to use. Bad data can be far worse than no data.

The second thing is this: In a plane, you have a pilot. You have a trained and experienced pilot who is ready to take over at a seconds notice. If the auto-pilot stops working or is doing something wrong, you have a pilot to take back control. You cannot count on that in a car. In a car, you have a random person who may or may not even be paying attention; and who, even if they notice something is wrong, may or may not actually have the presence of mind to do something about it.


And with close-to-modern technology, in the car you will also have a driver, trained and experienced and ready to take over.


Maybe. Or, you might have a person who's never driven before, or has rarely driven; because why bother if the car does it for you? Or, you might have a driver who isn't paying attention at all to what is happening. Part of the allure of self-driving cars is the ability to do other things - read a book, play on your phone, work on your laptop - do you seriously think that the average person is going to be able to drop whatever they're doing and take back control if they're prompted somehow, in the space of a few seconds? Having possibly no idea what's happening to require manual control? The average person, I am sorry to tell you, is going to panic; and either freeze up and do nothing, or startle and make things worse.

You think people are bad drivers when they're actively driving? Wait and see what happens when "driving" is something they only do when they hear a warning and have to take over to avoid some disaster that might happen in a matter of seconds.

Even assuming the best case scenario where the driver is trying to pay attention; people tend to lose focus when they aren't actively involved in a task.

No one is denying that there are still problems. They just aren't as insurmountable as you people are making it out.


Based on what? The only people who've done any extensive testing is Google, and even if we give them an enormous benefit of the doubt, there is a clear conflict of interest in them being the primary source of information when it comes to safety and reliability.

Don't get me wrong, it's impressive how far they've come with these things... but I keep hearing people talk about how various issues are "minor" or "inconsequential" based on nothing more than a vague, shrugging assumption that someone will figure it out. And that simply doesn't cut it when it comes to something that will be controlling a two-ton vehicle around people.

Nobody is saying that these problems are insurmountable; just that they're bigger and more significant than some people want to believe. Especially given the timelines that companies like Google are shooting for when it comes to releasing these things into the wild.

morriswalters
Posts: 7073
Joined: Thu Jun 03, 2010 12:21 am UTC

Re: Cars discussion, split from 'other news'

Postby morriswalters » Tue Jun 30, 2015 9:59 pm UTC

It's already in the wild. And more people are testing. And Google may be arrogant, but they aren't stupid, you won't see a product until they have one that won't ruin them by killing people. And they are aware as anyone what it means when the car is driving and you aren't ready to take control. Which is why their latest prototype doesn't have a steering wheel.
cphite wrote:No; we're talking about a system that depends on Google Maps or something like it as input.
No part of the problem is that they are using maps that make Google Maps pale in comparison. They actually account for curb cuts. A much higher level of detail.

Whatever you think of Google Maps it's only a matter of time before the Google reduces the error bars lower than they are now. Paper maps are no better and Google maps the world. And losing a GPS signal doesn't send civilization falling down, assuming the car knows where it is at the failure point. At the worst it stops, whistles, hey rube, to wake the driver and the human takes over.

And engineers, bless em, know machines break and software engineers know that code can be buggy and break. Yet the world totters on. To make that point a plane just fell out of the sky because of a borked update. Shit happens. Fix it and move on.

User avatar
gmalivuk
GNU Terry Pratchett
Posts: 26823
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

Re: In other news... (humorous news items, etc)

Postby gmalivuk » Tue Jun 30, 2015 10:04 pm UTC

commodorejohn wrote:
KrytenKoro wrote:I would be overjoyed to meet the human who could tell me where I should be going with the accuracy that GPS accomplishes.

Assuming we read "Google Maps" for "GPS," I'll direct you to the clerk at Goin' Postal in Superior, Wisconsin, who helped me figure out that the FedEx processing center wasn't actually where Google said it was, on the edge of the downtown area, and was instead out in the big-box district several miles away.
Wow, you have an anecdote about one person giving you directions one time Google messed up. That's definitely what KrytenKoro was talking about.
Unless stated otherwise, I do not care whether a statement, by itself, constitutes a persuasive political argument. I care whether it's true.
---
If this post has math that doesn't work for you, use TeX the World for Firefox or Chrome

(he/him/his)

Tyndmyr
Posts: 11443
Joined: Wed Jul 25, 2012 8:38 pm UTC

Re: Cars discussion, split from 'other news'

Postby Tyndmyr » Tue Jun 30, 2015 10:17 pm UTC

Google Maps and GPS are not *exactly* the same.

GPS is pretty accurate and timely, and getting more so as we improve the network. Sure, in a parking garage or tunnel, you might have crappy reception, but that's not really a huge navigation problem. You just keep going until you get out of the tunnel/garage, and reacquire signal. That's not a particularly crazy thing to expect an automated car to do.

commodorejohn
Posts: 1198
Joined: Thu Dec 10, 2009 6:21 pm UTC
Location: Placerville, CA
Contact:

Re: In other news... (humorous news items, etc)

Postby commodorejohn » Tue Jun 30, 2015 10:36 pm UTC

gmalivuk wrote:Wow, you have an anecdote about one person giving you directions one time Google messed up. That's definitely what KrytenKoro was talking about.

All I know is, it woulda saved me half an hour of frustration and hair-pulling if Google knew the area as well as J. Random Clerk did.
"'Legacy code' often differs from its suggested alternative by actually working and scaling."
- Bjarne Stroustrup
www.commodorejohn.com - in case you were wondering, which you probably weren't.

Tyndmyr
Posts: 11443
Joined: Wed Jul 25, 2012 8:38 pm UTC

Re: In other news... (humorous news items, etc)

Postby Tyndmyr » Tue Jun 30, 2015 10:40 pm UTC

commodorejohn wrote:
gmalivuk wrote:Wow, you have an anecdote about one person giving you directions one time Google messed up. That's definitely what KrytenKoro was talking about.

All I know is, it woulda saved me half an hour of frustration and hair-pulling if Google knew the area as well as J. Random Clerk did.


I'm sure that's true. But it's no more a condemnation of the computer's ability to drive than your lack of knowledge of the area is a condemnation of yours.

commodorejohn
Posts: 1198
Joined: Thu Dec 10, 2009 6:21 pm UTC
Location: Placerville, CA
Contact:

Re: Cars discussion, split from 'other news'

Postby commodorejohn » Tue Jun 30, 2015 10:41 pm UTC

Except that nobody's proposing making me the default navigator for anybody.
"'Legacy code' often differs from its suggested alternative by actually working and scaling."
- Bjarne Stroustrup
www.commodorejohn.com - in case you were wondering, which you probably weren't.

User avatar
addams
Posts: 10331
Joined: Sun Sep 12, 2010 4:44 am UTC
Location: Oregon Coast: 97444

Re: In other news... (humorous news items, etc)

Postby addams » Tue Jun 30, 2015 10:44 pm UTC

Tyndmyr wrote:
commodorejohn wrote:
gmalivuk wrote:Wow, you have an anecdote about one person giving you directions one time Google messed up. That's definitely what KrytenKoro was talking about.

All I know is, it woulda saved me half an hour of frustration and hair-pulling if Google knew the area as well as J. Random Clerk did.


I'm sure that's true. But it's no more a condemnation of the computer's ability to drive than your lack of knowledge of the area is a condemnation of yours.

Very good point.

But...It is fun to talk to J. Random Clerk.
People never cease to amaze me.
Life is, just, an exchange of electrons; It is up to us to give it meaning.

We are all in The Gutter.
Some of us see The Gutter.
Some of us see The Stars.
by mr. Oscar Wilde.

Those that want to Know; Know.
Those that do not Know; Don't tell them.
They do terrible things to people that Tell Them.

Tyndmyr
Posts: 11443
Joined: Wed Jul 25, 2012 8:38 pm UTC

Re: Cars discussion, split from 'other news'

Postby Tyndmyr » Tue Jun 30, 2015 10:49 pm UTC

commodorejohn wrote:Except that nobody's proposing making me the default navigator for anybody.


Humans already are the default navigator, or were, until we invented GPS.

Now, memorizing directions in tedious detail is going the way of memorizing phone numbers. Mostly only an incidental thing, because automating that was way easier and more reliable.

commodorejohn
Posts: 1198
Joined: Thu Dec 10, 2009 6:21 pm UTC
Location: Placerville, CA
Contact:

Re: Cars discussion, split from 'other news'

Postby commodorejohn » Tue Jun 30, 2015 10:57 pm UTC

Tyndmyr wrote:Humans already are the default navigator, or were, until we invented GPS.

For myself, sure. But nobody else has to rely on my sense of direction or knowledge of area geography. It'd be a different thing entirely if I were, say, working as a navigator/driver for somebody else's car.
"'Legacy code' often differs from its suggested alternative by actually working and scaling."
- Bjarne Stroustrup
www.commodorejohn.com - in case you were wondering, which you probably weren't.

morriswalters
Posts: 7073
Joined: Thu Jun 03, 2010 12:21 am UTC

Re: Cars discussion, split from 'other news'

Postby morriswalters » Wed Jul 01, 2015 12:39 am UTC

commodorejohn wrote:All I know is, it woulda saved me half an hour of frustration and hair-pulling if Google knew the area as well as J. Random Clerk did.
You do realize this is a you problem and not a Google one don't you. The world tottered along without the benefit of Google for centuries. Maps have always been outdated as they were printed, since new streets go in almost daily. The world isn't static and sometimes the person who is suppose to check these things fucks up. Google isn't any different than any other map maker in that respect.

User avatar
gmalivuk
GNU Terry Pratchett
Posts: 26823
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

Re: In other news... (humorous news items, etc)

Postby gmalivuk » Wed Jul 01, 2015 1:02 am UTC

commodorejohn wrote:
gmalivuk wrote:Wow, you have an anecdote about one person giving you directions one time Google messed up. That's definitely what KrytenKoro was talking about.

All I know is, it woulda saved me half an hour of frustration and hair-pulling if Google knew the area as well as J. Random Clerk did.
Yes, that is definitely comparable to the collective billions of hours saved by drivers using GPS instead of memorizing a route or looking for paper maps of every relevant area.

(Or to be more analogous, the hours saved by not needing to track down someone who happens to know the area well enough to give directions.)
Unless stated otherwise, I do not care whether a statement, by itself, constitutes a persuasive political argument. I care whether it's true.
---
If this post has math that doesn't work for you, use TeX the World for Firefox or Chrome

(he/him/his)

User avatar
ucim
Posts: 6890
Joined: Fri Sep 28, 2012 3:23 pm UTC
Location: The One True Thread

Re: Cars discussion, split from 'other news'

Postby ucim » Wed Jul 01, 2015 1:56 am UTC

Derek wrote: Given a situation that the automated system simply can't handle well, like rain or construction, the system will usually have plenty of time to alert the driver and slow down if necessary so they can take control.
Yes, the system will slow down as necessary so that the driver won't be needed. That's not the point. No matter what the circumstance, something will happen that demands that the driver take over NOW. I'm saying the driver won't be available NOW. It will be: "Huh, what?" CRASH! Because that's how fast things happen when they go mustard.

cphite wrote:No; we're talking about a system that depends on Google Maps or something like it as input.
I don't think it will be like that. It will use Google (or some other) Maps for strategy, and use its cameras and sensors for tactics, just like a human driver would.

Jose
Order of the Sillies, Honoris Causam - bestowed by charlie_grumbles on NP 859 * OTTscar winner: Wordsmith - bestowed by yappobiscuts and the OTT on NP 1832 * Ecclesiastical Calendar of the Order of the Holy Contradiction * Heartfelt thanks from addams and from me - you really made a difference.

User avatar
sardia
Posts: 6813
Joined: Sat Apr 03, 2010 3:39 am UTC

Re: Cars discussion, split from 'other news'

Postby sardia » Wed Jul 01, 2015 3:13 am UTC

ucim wrote:
Derek wrote: Given a situation that the automated system simply can't handle well, like rain or construction, the system will usually have plenty of time to alert the driver and slow down if necessary so they can take control.
Yes, the system will slow down as necessary so that the driver won't be needed. That's not the point. No matter what the circumstance, something will happen that demands that the driver take over NOW. I'm saying the driver won't be available NOW. It will be: "Huh, what?" CRASH! Because that's how fast things happen when they go mustard.



Jose

Now we can measure this(or at some point will be able to). Show me the accident rate, death rate/etc etc of humans only vs the so called glitch rate on a computer controlled car. Based on that, if one has a much lower rate than the other, it might warrant going with that system. If they have the same rate, it might warrant going over to computers just for the added convenience. Any other consideration is pretty artificial, like not 'feeling' you are in control.

commodorejohn
Posts: 1198
Joined: Thu Dec 10, 2009 6:21 pm UTC
Location: Placerville, CA
Contact:

Re: Cars discussion, split from 'other news'

Postby commodorejohn » Wed Jul 01, 2015 3:23 am UTC

morriswalters wrote:You do realize this is a you problem and not a Google one don't you. The world tottered along without the benefit of Google for centuries. Maps have always been outdated as they were printed, since new streets go in almost daily. The world isn't static and sometimes the person who is suppose to check these things fucks up. Google isn't any different than any other map maker in that respect.

It's a me problem for me in my own life, yes. It becomes a Google problem as soon as Google offers something that's supposed to have reliable navigation and doesn't, especially when that something is a thousand-plus pounds of vehicle tooling along at 30-60 MPH.

ucim wrote:That's not the point. No matter what the circumstance, something will happen that demands that the driver take over NOW. I'm saying the driver won't be available NOW. It will be: "Huh, what?" CRASH! Because that's how fast things happen when they go mustard.

This is the key point. Any significant circumstances the computer can't handle are potentially highly unsafe, because human beings can't context-switch instantaneously, and in driving the acceptably safe reaction time goes down as the potential danger of the hazard goes up. A "take the wheel, I can't deal with this washboard gravel road" message is not a huge problem, because the couple of seconds it would take our hypothetical passenger-driver to put down his bagel and coffee and start driving the car is basically acceptable. An "OHMYGOD THIS DEER IS A SUICIDAL IDIOT" message or an "AHHHH FUCK FUCK BLACK ICE" message? Not something you can spare a few seconds for, that shit needs to be dealt with now.
"'Legacy code' often differs from its suggested alternative by actually working and scaling."
- Bjarne Stroustrup
www.commodorejohn.com - in case you were wondering, which you probably weren't.

Chen
Posts: 5580
Joined: Fri Jul 25, 2008 6:53 pm UTC
Location: Montreal

Re: Cars discussion, split from 'other news'

Postby Chen » Wed Jul 01, 2015 3:30 am UTC

Dauric wrote:I think the problem with computer control is less one of external circumstances, those are being fairly well handled by the guys doing the engineering at this point, but rather one of computer fault/breakage. I work in I.T. support, everything breaks. Now if most of the computers in a modern car fail (which they will do on occasion, everything can fail at one time or another) the direction and speed of the vehicle is still under driver control. Get some kind of component failure at 60 mph in steering or speed control without an attentive driver and it's more of a problem.


As I mentioned before, the aircraft industry already has plenty of rigorous standards in place for developing reliable software. A catastrophic event (many fatalities and/or loss of aircraft) needs to meet a reliability of 10^-9/flight hour. This is usually handled via redundant systems. So if your computer failure rate is 10^-5 you need two of them to meet that reliability. Similarly you have software development/verification standards like DO-178B/C which, depending on the criticality of the software have various levels of rigor you need to go through before you can use them on aircraft. Presumably various givernment organizations will require similar standards to apply to these autonomous car softwares.

BattleMoose
Posts: 1993
Joined: Tue Nov 13, 2007 8:42 am UTC

Re: Cars discussion, split from 'other news'

Postby BattleMoose » Wed Jul 01, 2015 4:42 am UTC

As a cyclist I eagerly await the advent of autonomous cars. The, complete disregard for the life of cyclists in Australia is boarder line and sometimes is actually criminal. A computer will value my safety more than most humans.

Derek
Posts: 2181
Joined: Wed Aug 18, 2010 4:15 am UTC

Re: Cars discussion, split from 'other news'

Postby Derek » Wed Jul 01, 2015 4:54 am UTC

This is the key point. Any significant circumstances the computer can't handle are potentially highly unsafe, because human beings can't context-switch instantaneously, and in driving the acceptably safe reaction time goes down as the potential danger of the hazard goes up. A "take the wheel, I can't deal with this washboard gravel road" message is not a huge problem, because the couple of seconds it would take our hypothetical passenger-driver to put down his bagel and coffee and start driving the car is basically acceptable. An "OHMYGOD THIS DEER IS A SUICIDAL IDIOT" message or an "AHHHH FUCK FUCK BLACK ICE" message? Not something you can spare a few seconds for, that shit needs to be dealt with now.

None of these are situations where you would even want to give control to a human driver. Humans are terrible at handling these kinds of situations. Their reactions are already too slow, and when they do react they often do the wrong thing.

Can you provide a situation where human intervention would be beneficial and there isn't time to switch between self-driving and manual?

commodorejohn
Posts: 1198
Joined: Thu Dec 10, 2009 6:21 pm UTC
Location: Placerville, CA
Contact:

Re: Cars discussion, split from 'other news'

Postby commodorejohn » Wed Jul 01, 2015 5:21 am UTC

People keep repeating "humans are bad at this and that" with the implication that computers are necessarily better at it, but do we have any reason to believe that to be the case? As has been noted, the majority of the testing on this stuff has taken place in essentially optimal conditions, and frankly I have trouble buying that a computer would be better at dealing with the erratic-as-hell behavior of a spooked whitetail than a human being.
"'Legacy code' often differs from its suggested alternative by actually working and scaling."
- Bjarne Stroustrup
www.commodorejohn.com - in case you were wondering, which you probably weren't.

KrytenKoro
Posts: 1487
Joined: Tue Apr 05, 2011 2:58 pm UTC

Re: Cars discussion, split from 'other news'

Postby KrytenKoro » Wed Jul 01, 2015 5:26 am UTC

Dauric wrote:The point was in case of computer failure the human driver takes over, so we're talking about being suddenly sans-autopilot because of a failure. Problem is the vast majority of drivers, when presented with an autopilot, will flip out their smart phones and text or surf the web, or generally be -less- attentive than if they were actively driving the car, and thus less aware of a computer fault, or an unexpected circumstance the computer isn't handing in a safe manner.

The point was a bad point. People already do all of those things simply because they feel confident from muscle memory. Show me the data of a driver in an autopilot car actually doing these things more frequently than your average driver, or stop making up groundless horror stories.

It's a me problem for me in my own life, yes. It becomes a Google problem as soon as Google offers something that's supposed to have reliable navigation and doesn't, especially when that something is a thousand-plus pounds of vehicle tooling along at 30-60 MPH.

If I ask you for directions to the emergency room, or Google, and you both get it wrong, I don't really give a fuck whose ""fault" it is. The person I was racing to get there is already dead. The fault occurred, people got hurt because of it.

If Google makes that fuckup less than you (or anyone else does), then I'm going to ask it for directions by default even if it isn't right 100% of the time. The people getting injured in car crashes don't give so much of a fuck (at least, until it's time for the lawsuit) whose fault it is, just that there was a fault. They want less faults. Being afraid of "who's going to get blamed" is irrational and completely irrelevant to the decision.

ucim wrote:
That's not the point. No matter what the circumstance, something will happen that demands that the driver take over NOW. I'm saying the driver won't be available NOW. It will be: "Huh, what?" CRASH! Because that's how fast things happen when they go mustard.

This is the key point. Any significant circumstances the computer can't handle are potentially highly unsafe, because human beings can't context-switch instantaneously, and in driving the acceptably safe reaction time goes down as the potential danger of the hazard goes up. A "take the wheel, I can't deal with this washboard gravel road" message is not a huge problem, because the couple of seconds it would take our hypothetical passenger-driver to put down his bagel and coffee and start driving the car is basically acceptable. An "OHMYGOD THIS DEER IS A SUICIDAL IDIOT" message or an "AHHHH FUCK FUCK BLACK ICE" message? Not something you can spare a few seconds for, that shit needs to be dealt with now.

It's the same perfect solution fallacy that's been repeated too many times in this thread, and it's as irrational as it was the first time.

There are, right now, people dying because they weren't able to figure out how to solve a driving problem, like black ice or deer, "RIGHT NOW". This is not a problem that the computers are introducing. Therefore, the only way it is relevant is if it occurs at a greater rate than the problems already occurring do.
From the elegant yelling of this compelling dispute comes the ghastly suspicion my opposition's a fruit.


Return to “News & Articles”

Who is online

Users browsing this forum: No registered users and 22 guests