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

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

Re: Cars discussion, split from 'other news'

Postby morriswalters » Thu Jul 23, 2015 7:45 pm UTC

ucim wrote:Hijacking wasn't new, but the objective was. For the first (known) time, the objective was to kill people spectacularly, dying in the process. As to controlling the simulators, that's a very basic attack on fundamental freedoms. (Yes, I count the ability to learn to fly as a basic freedom.)
And there you have it. You want what you want, and resent that controlling access to those simulators is a good thing. Yet today it is controlled. Because your desire to be free of regulation ran afoul of deaths from that lack. Auto makers make that very argument, even in the face of existing law.
ucim wrote:It is directly related in the attitude towards failure being displayed.
No. I understand the danger. And it is playing out as I would expect. We will adopt the technology, auto makers will learn. Congress(maybe) will legislate when the auto makers fuck up. And I will do the only thing I can do. Vote with my wallet.
ucim wrote:This is a reason to look for bullets, not to not look for bullets.
And you won't find them soon enough, if the history of software is any indication. And the software in cars is changing as is the hardware. Yearly. Anyway I get it. You don't like self driving cars. Okay.

User avatar
gmalivuk
GNU Terry Pratchett
Posts: 26836
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

Re: Cars discussion, split from 'other news'

Postby gmalivuk » Thu Jul 23, 2015 7:58 pm UTC

One of them was enumerated and extrapolating was left as an exercise for the reader: If all cars came with a device that prevented speeds in excess of 20mph, road fatalities could be reduced *drastically*. And yet, I don't see anyone advocating that.

There are any number of other actions that could be taken to reduce deaths that are similarly so inconvenient that no one's willing to implement them. (I wonder how many accidents that would be fatal today occurred per horse-mile back when that was how most people got around.)
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 » Thu Jul 23, 2015 8:09 pm UTC

gmalivuk wrote:One of them was enumerated and extrapolating was left as an exercise for the reader: If all cars came with a device that prevented speeds in excess of 20mph, road fatalities could be reduced *drastically*. And yet, I don't see anyone advocating that.


Yes. And what is THAT for computer driven cars?

Because the analogy seems to just skip right over that, and I don't see any equivalence.

User avatar
slinches
Slinches get Stinches
Posts: 1036
Joined: Tue Mar 26, 2013 4:23 am UTC

Re: Cars discussion, split from 'other news'

Postby slinches » Thu Jul 23, 2015 8:17 pm UTC

gmalivuk wrote:I wonder how many accidents that would be fatal today occurred per horse-mile back when that was how most people got around.

Probably fewer collisions per mile, but the death rate may be higher than you'd expect. A bunch of panicked horses in close quarters can be awfully dangerous and some can be spooked by almost anything.

Tyndmyr wrote:
gmalivuk wrote:If all cars came with a device that prevented speeds in excess of 20mph, road fatalities could be reduced *drastically*. And yet, I don't see anyone advocating that.
Yes. And what is THAT for computer driven cars?

Rails or tracks instead of roads?

User avatar
gmalivuk
GNU Terry Pratchett
Posts: 26836
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

Re: Cars discussion, split from 'other news'

Postby gmalivuk » Thu Jul 23, 2015 8:33 pm UTC

Tyndmyr wrote:
gmalivuk wrote:One of them was enumerated and extrapolating was left as an exercise for the reader: If all cars came with a device that prevented speeds in excess of 20mph, road fatalities could be reduced *drastically*. And yet, I don't see anyone advocating that.


Yes. And what is THAT for computer driven cars?

...a device that prevents speeds in excess of 20mph.
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: 6896
Joined: Fri Sep 28, 2012 3:23 pm UTC
Location: The One True Thread

Re: Cars discussion, split from 'other news'

Postby ucim » Thu Jul 23, 2015 8:33 pm UTC

Tyndmyr wrote:Why not? If we can prevent 500 deaths due to nutty drivers at a cost of 200 deaths via explosions, isn't that a net gain of 300 lives saved?
It's not the math that I dispute. The math is impeccable. 500-200=300. I can do that. I even agree.

But deaths via random explosions are different from deaths due to driver error.

Why do you think we spend so much time and effort to track down a murderer? The victim's already dead, so why bother? And murders will happen anyway. Why do we care? The answer to that may help you see why I consider random explosions to be a net loss, even though 500-200=300.

morriswalters wrote:You want what you want, and resent that controlling access to those simulators is a good thing.
Controlling access to simulators is not a good thing. For one thing it's not addressing the actual problem. For another thing, there is nothing wrong with wanting to experience "being an airline pilot" without actually flying a jetliner, and there's nothing wrong with wanting to learn stuff. Concentrate on what the "wrong thing" actually is. (Although flight simulators are OT to this discussion, the use of them as an example is still illustrative).

morriswalters wrote:No. I understand the danger. And it is playing out as I would expect.
Is isn't should.

Tyndmyr wrote:Because the analogy seems to just skip right over that...
(huh? What "that"? Computer driven cars?)

The analogy was to illustrate that even now, even you don't value saving lives over everything else, making the whole "saving lives" argument a bit of a red herring. It's not a question of just saving lives. It's a question of what the tradeoffs are for saving those lives. In the case of computer driven cars, they are twofold:

1: It gives Google (for example) and it's "affiliated and unaffiliated companies" access to incredible amounts of very personal information about people who aren't even driving, but just going about their business (even if their business happens to be chasing a duck in a wheelchair. What the duck was doing in a wheelchair is another issue.) I don't know what they will do with this data; I can only extrapolate based on what they have done with such information in the past, and that's not very inspiring.

2: It opens the door (and some people aren't even interested in trying to close this door!) to adverse control of the actual vehicles in question, either individually or en masse, whether parked or in motion. This control could be accomplished from a distance of thousands of miles away, from another country. And this is not (just :)) a hypothetical rant from a crazy person. It has been demonstrated in the field.

Now, do you find (1) or (2) acceptable tradeoffs? I sure don't.

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.

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

Re: Cars discussion, split from 'other news'

Postby Tyndmyr » Thu Jul 23, 2015 9:10 pm UTC

gmalivuk wrote:
Tyndmyr wrote:
gmalivuk wrote:One of them was enumerated and extrapolating was left as an exercise for the reader: If all cars came with a device that prevented speeds in excess of 20mph, road fatalities could be reduced *drastically*. And yet, I don't see anyone advocating that.


Yes. And what is THAT for computer driven cars?

...a device that prevents speeds in excess of 20mph.


No. That is the analogy. It is not the reality. There is no 20mph restriction on computer driven cars.

ucim wrote:
Tyndmyr wrote:Why not? If we can prevent 500 deaths due to nutty drivers at a cost of 200 deaths via explosions, isn't that a net gain of 300 lives saved?
It's not the math that I dispute. The math is impeccable. 500-200=300. I can do that. I even agree.

But deaths via random explosions are different from deaths due to driver error.

Why do you think we spend so much time and effort to track down a murderer? The victim's already dead, so why bother? And murders will happen anyway. Why do we care? The answer to that may help you see why I consider random explosions to be a net loss, even though 500-200=300.


We do it to prevent murders. Both because a killer may kill again, and because knowing you will be caught is a disincentive to committing murders in the first place. In short, we do not wish to be murdered, so we punish murderers. This is consistent with wishing to save net lives.

And, if your objection is to randomness, then I must point out that of those killed by careless drivers, a great many of them are people OTHER than the careless drivers, who are essentially struck at random. If you're on a bike and get hit by someone texting, it is little different than a random explosion from your perspective. Simply a matter of wrong place at the wrong time. Not predictable to you.

ucim wrote:
Tyndmyr wrote:Because the analogy seems to just skip right over that...
(huh? What "that"? Computer driven cars?)

The analogy was to illustrate that even now, even you don't value saving lives over everything else, making the whole "saving lives" argument a bit of a red herring. It's not a question of just saving lives. It's a question of what the tradeoffs are for saving those lives.


I do value other things in addition to lives. Injuries. Property damage. Etc. Lives score pretty high, though. And, generally computer driven cars that are less likely to kill me are going to score better on those metrics as well.

ucim wrote:1: It gives Google (for example) and it's "affiliated and unaffiliated companies" access to incredible amounts of very personal information about people who aren't even driving, but just going about their business (even if their business happens to be chasing a duck in a wheelchair. What the duck was doing in a wheelchair is another issue.) I don't know what they will do with this data; I can only extrapolate based on what they have done with such information in the past, and that's not very inspiring.


What they have done with it in the past is sold advertising. This is not terribly fearsome. It does not rank very highly against "risk of death".

I expect they'll push advertising to me in the future regardless of if I buy a computer driven car.

ucim wrote:2: It opens the door (and some people aren't even interested in trying to close this door!) to adverse control of the actual vehicles in question, either individually or en masse, whether parked or in motion. This control could be accomplished from a distance of thousands of miles away, from another country. And this is not (just :)) a hypothetical rant from a crazy person. It has been demonstrated in the field.


....no. Not thousands of miles away. That's not what was demonstrated. That would be a very different hack indeed.

The hack that WAS demonstrated was not demonstrated on an computer driven car. So, it's not a vulnerability introduced by computer driven cars.

So, what you are talking about is indeed very hypothetical. It relies on very significant inference.

User avatar
Dauric
Posts: 3999
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 » Thu Jul 23, 2015 9:28 pm UTC

Tyndmyr wrote:The hack that WAS demonstrated was not demonstrated on an computer driven car. So, it's not a vulnerability introduced by computer driven cars.


No, but it does demonstrate a caviler attitude on the part of automotive manufacturers to device security. Even if Google and Apple start manufacturing self-driving cars, and we assume for a moment that they have reasonable (to a certain definition thereof) understanding of wireless connection security given their histories working with data devices, it won't be long before the existing auto manufacturers come out with their own competing products.

Unless those manufacturers start demonstrating a better grasp of device security than they do at present, do we really want to be turning over more of their car's functionality to onboard systems that have demonstrably insecure wireless access?
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.

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

Re: Cars discussion, split from 'other news'

Postby Tyndmyr » Thu Jul 23, 2015 9:41 pm UTC

Devices are imperfectly secure yes. This will be the future regardless of computer driven cars or not. At some point, errors will be made, and faulty stuff will be released.

This happened long before computers. The idea that you could just cut someone's brake lines because they were critical and reasonably accessible isn't particularly new.

And yet, that sort of thing is a rounding error on automobile deaths.

commodorejohn
Posts: 1201
Joined: Thu Dec 10, 2009 6:21 pm UTC
Location: Placerville, CA
Contact:

Re: Cars discussion, split from 'other news'

Postby commodorejohn » Thu Jul 23, 2015 9:45 pm UTC

See, the difference is that a compromised PC can't hit a pedestrian at 60 MPH.
"'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: Cars discussion, split from 'other news'

Postby Tyndmyr » Thu Jul 23, 2015 9:49 pm UTC

commodorejohn wrote:See, the difference is that a compromised PC can't hit a pedestrian at 60 MPH.


I'm not comparing to compromised PCs.

I'm comparing to dysfunctional vehicles with human drivers, which totally can hit a pedestrian at 60 mph. This is the apples to apples comparison.

User avatar
Dauric
Posts: 3999
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 » Thu Jul 23, 2015 10:16 pm UTC

Tyndmyr wrote:Devices are imperfectly secure yes.


There's "Imperfectly" secure, which is about the best any secure system will be, and then there's "Leaving your front door unlocked while you go to work" insecure, which the Wired article notes the auto manufacturers knew was a problem for four years, and it continues to be a problem.

Shit happens because someone got the upper hand in the 'measure-countermeasure' race I can understand. Shit happens because a manufacturer is wantonly indifferent to the security on their devices (as these auto manufacturers have already demonstrated themselves to be) I'm decidedly less inclined to place blind faith in their products.
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.

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 » Thu Jul 23, 2015 10:33 pm UTC

Tyndmyr wrote:This happened long before computers. The idea that you could just cut someone's brake lines because they were critical and reasonably accessible isn't particularly new.

And yet, that sort of thing is a rounding error on automobile deaths.


The difference is that it's approximately 10,000 times more difficult to cut the brake lines on 10,000 cars compared to one car. With many attack vectors it is only moderately more difficult to hack 10,000 devices (cars, PCs, whatever) than to hack one. In a word: scalability.


Dauric wrote:No, but it does demonstrate a caviler attitude on the part of automotive manufacturers to device security. Even if Google and Apple start manufacturing self-driving cars, and we assume for a moment that they have reasonable (to a certain definition thereof) understanding of wireless connection security given their histories working with data devices, it won't be long before the existing auto manufacturers come out with their own competing products.

This is actually why I'm very glad that Google appears to be leading the way with self-driving cars.

User avatar
Negated
Posts: 230
Joined: Wed Jan 13, 2010 6:31 am UTC

Re: Cars discussion, split from 'other news'

Postby Negated » Thu Jul 23, 2015 10:48 pm UTC

A lot of potential vulnerabilities from hijacking wireless communication can be resolved by having multi-layer safety features. The cars should have the intelligence to detect imminent dangers based on the sensor readings. As long as the local safety decisions override the wirelessly transmitted commands, it makes hacking on mass scale much less likely.

User avatar
slinches
Slinches get Stinches
Posts: 1036
Joined: Tue Mar 26, 2013 4:23 am UTC

Re: Cars discussion, split from 'other news'

Postby slinches » Thu Jul 23, 2015 11:08 pm UTC

Quercus wrote:
Tyndmyr wrote:This happened long before computers. The idea that you could just cut someone's brake lines because they were critical and reasonably accessible isn't particularly new.

And yet, that sort of thing is a rounding error on automobile deaths.


The difference is that it's approximately 10,000 times more difficult to cut the brake lines on 10,000 cars compared to one car. With many attack vectors it is only moderately more difficult to hack 10,000 devices (cars, PCs, whatever) than to hack one. In a word: scalability.

Is it really that easy to compromise 10,000 cars? Sure, it's easier than it was with no remote access at all, but even if you could attack all of one make and model of vehicle, they would be very widely dispersed and only a small percentage would be moving at any given time. Maybe that could add up to 10k accidents caused for this sort of hacking incident, but that happening once or twice a year would still be "rounding error" since several million auto accidents already occur each year.

By the way, that doesn't mean I excuse poor security practices. I just wanted to give some context with respect to how high the risks of driving currently are.

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 » Thu Jul 23, 2015 11:11 pm UTC

slinches wrote:By the way, that doesn't mean I excuse poor security practices. I just wanted to give some context with respect to how high the risks of driving currently are.

Yeah, that's probably a fair comment

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

Re: Cars discussion, split from 'other news'

Postby morriswalters » Thu Jul 23, 2015 11:12 pm UTC

ucim wrote:Is isn't should.
There is no should morally. And you just spent a significant amount of time arguing the point. And there is no legislation yet, and no certainty about what it might need to say.
ucim wrote:Controlling access to simulators is not a good thing
You can't have it both ways. You want to impose some type of regulation on an industry which you have concerns about, some of which I share. You can't then argue that regulations for access to devices to teach you to fly Boeing Dream Liners are an abridgement of your freedom.
ucim wrote: For one thing it's not addressing the actual problem.
I don't get a clear read of what you think the problem is. In terms of self driving cars the technology is not close, optimistic answers are always unattributed. Think in terms of 15 years. So we're are going to get some answers. This conversation is itself an indication that it is on peoples minds.

Google is approaching a million miles in their testing and has branched to a second site, Austin I believe. And they test to find just these types of things. It's the point of an experimental program. Legislators have to approve the use, if and when they are ready for release. And hackers are going to be trying to penetrate. It's what they do. Talk about lowering the death rate due to accidents is premature, it's based on a theoretical usage pattern where everything is automated. For the foreseeable future that won't be the case. The traffic will be mixed, which in an of itself may make things marginally worse.

The long game is about money. If Google or Ford or Mercedes can make money off this it will happen. They know their reputation will suffer if they don't get it right. They have an example that they don't want to follow in the first jet airliner, the de Havilland DH 106 Comet.

User avatar
Thesh
Made to Fuck Dinosaurs
Posts: 6598
Joined: Tue Jan 12, 2010 1:55 am UTC
Location: Colorado

Re: Cars discussion, split from 'other news'

Postby Thesh » Thu Jul 23, 2015 11:50 pm UTC

slinches wrote:Is it really that easy to compromise 10,000 cars? Sure, it's easier than it was with no remote access at all, but even if you could attack all of one make and model of vehicle, they would be very widely dispersed and only a small percentage would be moving at any given time. Maybe that could add up to 10k accidents caused for this sort of hacking incident, but that happening once or twice a year would still be "rounding error" since several million auto accidents already occur each year.


Well, the Jeep was attacked remotely through their internet connection. Presumably they are all within an easily determinable range of IP addresses, and you can attack them all with a script.
Summum ius, summa iniuria.

User avatar
gmalivuk
GNU Terry Pratchett
Posts: 26836
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

Re: Cars discussion, split from 'other news'

Postby gmalivuk » Fri Jul 24, 2015 12:16 am UTC

Tyndmyr wrote:No. That is the analogy. It is not the reality. There is no 20mph restriction on computer driven cars.
But there could be, and it would certainly make them pretty safe, even with current technology that hasn't been tested in adverse weather conditions.

Why are you so opposed to making cars safer? Huh?

ucim wrote:2: It opens the door (and some people aren't even interested in trying to close this door!) to adverse control of the actual vehicles in question, either individually or en masse, whether parked or in motion. This control could be accomplished from a distance of thousands of miles away, from another country. And this is not (just :)) a hypothetical rant from a crazy person. It has been demonstrated in the field.
....no. Not thousands of miles away. That's not what was demonstrated. That would be a very different hack indeed.
No part of the hack, as I understand it, was limited by distance. The fact that the hackers in the article happened to be just 10 miles away at the time is irrelevant.
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)

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

Re: Cars discussion, split from 'other news'

Postby morriswalters » Fri Jul 24, 2015 2:20 am UTC

From Ars Technica. Read this review from Ars and laugh.

User avatar
ucim
Posts: 6896
Joined: Fri Sep 28, 2012 3:23 pm UTC
Location: The One True Thread

Re: Cars discussion, split from 'other news'

Postby ucim » Fri Jul 24, 2015 3:43 am UTC

morriswalters wrote:From Ars Technica. Read this review from Ars and laugh.
Two things stand out:
Ars Technica wrote:Everyone I've taken for a drive in this car has—completely unprompted—tried to say "OK Google" to it.[...]Our inner geek would love a whole car wired for sound so that anyone could shout "OK Google" like it's the Star Trek computer.
Can you imagine a radio announcer saying "OK Google, Hit the brakes!"

regarding texting while driving using this app, Ars Technica wrote:Stop-and-go traffic and red lights exist—with your foot on the break, those are likely perfectly safe times to read the computer screen or even poke at it.
I guess my first exploit won't work unless the radio announcer says "Ok Google, hit the breaks!" :) But this is the kind of thinking I'm afraid of. Fortunately Google is enforcing certain safety provisions, despite complaints by this reviewer. For now.

Tyndmyr wrote:We do it [(expend lots of money to find murderers)] to prevent murders.
But why prevent murders in the first place? Is this a "leading cause of death"? That money could be spent curing heart disease, or preventing household accidents. Point is, we don't (and shouldn't) spend money based solely on number of lives saved.

The reason we track down murderers is to enforce a societal norm, under which we want to live. And although it's consistent with wanting to save lives, that's not the underlying reason. We do this so we ourselves can feel safe from our neighbors, and although this puts restrictions on murderers, we don't care because we are not murderers. Whether this is Good or Bad is debatable (substitute "troublemaker" for "murderer"), but OT for this topic. This is why we do it.

Tyndmyr wrote:What they have done with it in the past is sold advertising. This is not terribly fearsome.
The lessons of umwelt are lost on you.

Tyndmyr wrote:....no. Not thousands of miles away. That's not what was demonstrated. That would be a very different hack indeed.
They demonstrated it through a medium for which distance was not an object. That is sufficient. If you truely think that this is not sufficient, they you might as well point out that they did not demonstrate it on a blue car, or on a station wagon.

Tyndmyr wrote:The hack that WAS demonstrated was not demonstrated on an computer driven car. So, it's not a vulnerability introduced by computer driven cars.
No, it is a vulnerability that is greatly amplified by computer driven cars. Unless the industry somehow prevents it. Which they do not appear to be doing, Ars Technica notwithstanding.

Tyndmyr wrote:I'm not comparing to compromised PCs.
You should be, because cars nowadays are nothing more than compromised PCs going at highway speed. With seat belts.

Negated wrote:As long as the local safety decisions override the wirelessly transmitted commands, it makes hacking on mass scale much less likely.
Yes, so long as the local system does not interact in any way with the networked system, which includes not being able to update it wirelessly. (Which is a safety issue in its own right)

morriswalters wrote:
ucim wrote:Is isn't should.
There is no should morally. And you just spent a significant amount of time arguing the point. And there is no legislation yet, and no certainty about what it might need to say.
I'm not arguing morals (let alone objectivity in moral systems). I am stating what should happen (you can take that as IMHO, IMEO, or "because if it doesn't, there will be mustard"). You seem to be arguing "It won't happen, so let's be ok with it".

morriswalters wrote:
ucim wrote:Controlling access to simulators is not a good thing
You can't have it both ways. You want to impose some type of regulation on an industry which you have concerns about, some of which I share. You can't then argue that regulations for access to devices to teach you to fly Boeing Dream Liners are an abridgement of your freedom.
Yes I can, and without contradiction. It's important to decide which regulations should be imposed, on whom, and why. The kneejerk simulator (and actual flight training) restrictions fail all those tests. (Not to invoke argument from authority, but I happen to be a pilot, and have actual exposure to these regulations.)

morriswalters wrote:Google is approaching a million miles in their testing
Is that a big number? It's more than I can visualize. A million dollars doesn't buy what it used to, but a million pounds is more than I can bench press even on a good day. For comparison, there were three trillion miles driven in 2012 in the United States alone, making their test the equivalent of one third of a billionth of the yearly US traffic.

morriswalters wrote:The long game is about money. If Google or Ford or Mercedes can make money off this it will happen.
... and money will be made the same way it's being made on software today - by selling our personal information to all comers. That's point (1) of my prior post, and the one referencing the (laughable) privacy policy: For "strictly internal" purposes. To "improve the user experience".
Spoiler:
If I must explain, "improving the user experience" may well include telling the driver about a sale at every Starbucks they pass, and telling Starbucks how many such sales were passed up.
morriswalters wrote:They know their reputation will suffer if they don't get it right. They have an example that they don't want to follow in the first jet airliner, the de Havilland DH 106 Comet.
They have other examples they don't mind following however: The Pinto. The GM key. The airbag fiasco. The Toyota accelerator (not proven though). And just about every major recall that happened, and the ones that haven't happened yet. But they keep doing it, and by "it" I don't mean making mistakes. I mean deliberately covering up known flaws hoping nobody will notice and they can keep getting away with shaving sixty cents on each car they make. There are other examples to follow: The Sony rootkit. Internet Exploder. The Pentium. Flatscreen auto entertainment interfaces. And just about every software release known to man, woman, or whatever. Again not because software is hard, but because release trumps ready.

Jose
edit: quote mustard
Last edited by ucim on Fri Jul 24, 2015 12:39 pm UTC, edited 1 time in total.
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.

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

Re: Cars discussion, split from 'other news'

Postby morriswalters » Fri Jul 24, 2015 9:42 am UTC

Check your quotes, I'm getting more attention than I earned.
ucim wrote:Can you imagine a radio announcer saying "OK Google, Hit the brakes!"
It was what the reviewer felt was important and the general tone that I found amusing. This reviewer will walk into the abyss happily texting on his car infotainment device rather than watching the road. On the rest of your comment I can imagine a lot. However I have been rear ended six times by someone like that fool or his sisters. There lies the genesis of our difference in outlook.
ucim wrote: You seem to be arguing "It won't happen, so let's be ok with it".
I know it may seem that way, but no. First, you've already been run over by the train your trying to avoid. Google is already tracking you, as are any number of others. Your car, assuming it has been purchased within the last 10 years is overloaded with computers. So first, it isn't, will cars get criminally hacked? It's, when will cars get criminally hacked? The maker community and shade tree mechanics have been tinkering with the computer on cars for some time through the on board diagnostics connector. It was inevitable from the day that computers starting controlling the performance variables that this would happen. So we are going to have to fix it after the fact. And we will. But to repeat, the cat is out of the bag, you're late.
ucim wrote: It's important to decide which regulations should be imposed, on whom, and why.
Yes precisely, and thus it has ever been. The industry has been fighting this particular battle for years. It isn't that you shouldn't see it the way that you do, it's that you should understand that the industry doesn't. And it doesn't make them evil, it makes them businessmen.
ucim wrote:Is that a big number? It's more than I can visualize.
Is it? I just read something that quoted the number as 1.7 million miles, for what is currently 23 cars. It indicates that Google and others are doing what engineers do, gathering data, and finding out what works and what doesn't work. And what you need you need to do to improve the process. Would you rather they didn't test it, or would you like to see a ban?
ucim wrote:If I must explain, "improving the user experience" may well include telling the driver about a sale at every Starbucks they pass, and telling Starbucks how many such sales were passed up.
That ship has sailed, they are already doing it. And now they have a new toy. Beacons. :lol:
ucim wrote:They have other examples they don't mind following however
Yes, I know, I've posted a number of them. Are you actually reading what I'm writing? We are monkeys, playing with sticks of dynamite. That is the world I live in. First we light them, and then we run around blowing on the fuses.

elasto
Posts: 3778
Joined: Mon May 10, 2010 1:53 am UTC

Re: Cars discussion, split from 'other news'

Postby elasto » Fri Jul 24, 2015 12:23 pm UTC

ucim wrote:If I must explain, "improving the user experience" may well include telling the driver about a sale at every Starbucks they pass, and telling Starbucks how many such sales were passed up.

I still feel like you need to explain further.

Targeted advertising means I'm most likely only going to get told about a sale at Starbucks if I actually care about buying coffee at Starbucks. If I don't, then I'll get an advert about something else that probably does interest me.

If my car knows I like buying chain-coffee, and knows I am thirsty, and informs me I'm about to pass a Starbucks and they are having a sale, that advert does me a favour.

OTOH, if my car knows I dislike expensive chain-coffee, and knows I'm nearly out of coffee in my home, and informs me I'm about to pass a supermarket that has a sale on the brand of coffee I like, that also does me a favour.

You and I have a different outlook on the world I fear: I am not afraid to be advertised to; Relevant advertising can anticipate my needs and improve my quality of life.

You are fearful of 'umwelt', whereas I think it's never been easier to get accurate information on any given topic in the whole history of mankind...

User avatar
gmalivuk
GNU Terry Pratchett
Posts: 26836
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

Re: Cars discussion, split from 'other news'

Postby gmalivuk » Fri Jul 24, 2015 12:36 pm UTC

It may have never been easier in the past, but it could be made easier still if Google didn't tailor search results in addition to advertising.

Plus, the ease of finding accurate information is not the same as ease of *verifying* accuracy.
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
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 » Fri Jul 24, 2015 1:04 pm UTC

elasto wrote:
ucim wrote:If I must explain, "improving the user experience" may well include telling the driver about a sale at every Starbucks they pass, and telling Starbucks how many such sales were passed up.

I still feel like you need to explain further.

Targeted advertising means I'm most likely only going to get told about a sale at Starbucks if I actually care about buying coffee at Starbucks. If I don't, then I'll get an advert about something else that probably does interest me.

If my car knows I like buying chain-coffee, and knows I am thirsty, and informs me I'm about to pass a Starbucks and they are having a sale, that advert does me a favour.

OTOH, if my car knows I dislike expensive chain-coffee, and knows I'm nearly out of coffee in my home, and informs me I'm about to pass a supermarket that has a sale on the brand of coffee I like, that also does me a favour.

You and I have a different outlook on the world I fear: I am not afraid to be advertised to; Relevant advertising can anticipate my needs and improve my quality of life.

You are fearful of 'umwelt', whereas I think it's never been easier to get accurate information on any given topic in the whole history of mankind...

I think the situation is slightly more nuanced than that for me. I always find that targeted advertising gets very samey very fast. if I've read a few books about spaceflight recently, then my amazon recommendations fill with books about spaceflight, when actually at that point I'm likely to start to crave something completely different, like an historical novel. Targeted advertising doesn't often provide any novelty or surprise. For me a lot of the fun in life is about trying new stuff, but targeted advertising seems determined to keep me doing and buying the same things I often do or buy. I want targeted advertising that sometimes says things like "you go to a lot of Turkish restaurants, why not try this well-reviewed Russian restaurant instead, because your history says you've never had Russian food before and it's about time you tried it".

However, I think it's probably not in companies' commercial interests to take risks in advertising things based on novelty - because it's likely to lead to a lot of failures among the successes - the advertisers know I like Turkish food, but maybe I hate Russian food, who knows. That to me is the problem with targeted advertising - it makes one less likely to discover new stuff unless that new stuff is very similar to the old stuff. It's a similar principle to the "search bubble" that gmalivuk is talking about.

The availability of information bit I do agree with - I love being able to google more in-depth information on things when I'm out and about - If I see an interesting looking sculpture for example, I can find out the history of it there and then, which is wonderful.

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

Re: Cars discussion, split from 'other news'

Postby Trebla » Fri Jul 24, 2015 1:46 pm UTC

Quercus wrote:For me a lot of the fun in life is about trying new stuff, but targeted advertising seems determined to keep me doing and buying the same things I often do or buy. I want targeted advertising that sometimes says things like "you go to a lot of Turkish restaurants, why not try this well-reviewed Russian restaurant instead, because your history says you've never had Russian food before and it's about time you tried it".

However, I think it's probably not in companies' commercial interests to take risks in advertising things based on novelty - because it's likely to lead to a lot of failures among the successes - the advertisers know I like Turkish food, but maybe I hate Russian food, who knows. That to me is the problem with targeted advertising - it makes one less likely to discover new stuff unless that new stuff is very similar to the old stuff. It's a similar principle to the "search bubble" that gmalivuk is talking about.


There's a lot to be said for the adage "Consumers know what they want, but they want what they know." If you like Turkish restaurants, usually you'll be more interested in hearing about the things you like. Sometimes you'll want to try something new which won't be advertised to you the same way non-targeted ads are, but that gives you the power to find something new that interests you rather than seeing something new but not really what you wanted.

You have the added convenience of usually getting what you want, and added inconvenience of having to try a bit harder to find something new. There's just disagreement on whether or not that's worth it. I think most people would approve (other than being creeped out that google seems to read your mind sometimes).

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 » Fri Jul 24, 2015 2:05 pm UTC

Trebla wrote:There's a lot to be said for the adage "Consumers know what they want, but they want what they know." If you like Turkish restaurants, usually you'll be more interested in hearing about the things you like. Sometimes you'll want to try something new which won't be advertised to you the same way non-targeted ads are, but that gives you the power to find something new that interests you rather than seeing something new but not really what you wanted.

You have the added convenience of usually getting what you want, and added inconvenience of having to try a bit harder to find something new. There's just disagreement on whether or not that's worth it. I think most people would approve (other than being creeped out that google seems to read your mind sometimes).

I am aware that I am at risk of sounding douchey and exceptionalist here, but looking at my life I genuinely do believe that I have a very high level of novelty seeking behaviour. I have all the classic symptoms, good and bad - I'm into extreme sports, I'm a jack of all trades, I enjoy changing things for the sake of changing things, I'm fascinated by exploration and experiment and I tend to get bored with career directions after a few years and want to switch to something entirely new. This means that I actually want what I don't know quite a lot of the time. I routinely choose restaurants by looking at what is least like anything I've ever tried before; I sometimes come back from shopping trips with half a dozen weird ingredients that I have no idea how to cook.

Of course I have things that I know and love and I want them quite a lot of the time, but I think the amount of newness that I crave in my life means that I'll probably never be particularly well served by targeted advertising.

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

Re: Cars discussion, split from 'other news'

Postby Tyndmyr » Fri Jul 24, 2015 2:49 pm UTC

Quercus wrote:
Tyndmyr wrote:This happened long before computers. The idea that you could just cut someone's brake lines because they were critical and reasonably accessible isn't particularly new.

And yet, that sort of thing is a rounding error on automobile deaths.


The difference is that it's approximately 10,000 times more difficult to cut the brake lines on 10,000 cars compared to one car. With many attack vectors it is only moderately more difficult to hack 10,000 devices (cars, PCs, whatever) than to hack one. In a word: scalability.


This is unlikely. Different brands of cars, different models, sometimes, and even different years within the same model often use different systems. It isn't like the desktop world where it's much more of a monoculture for operating systems, browsers, etc.

I see as I scan down that this has already been addressed, so I suppose that'll be sufficient. I'm concerned that people are way, way too quick to jump to "think of the DANGERS" without actually demonstrating those dangers. And thus, because of fear of hypothetical dangers that may not be all that large, we live with much greater dangers, simply because they are familiar and comfortable.

gmalivuk wrote:
Tyndmyr wrote:No. That is the analogy. It is not the reality. There is no 20mph restriction on computer driven cars.
But there could be, and it would certainly make them pretty safe, even with current technology that hasn't been tested in adverse weather conditions.

Why are you so opposed to making cars safer? Huh?


So, are you arguing just to argue, or does this actually make an analogy at some point here? Where's the relevance to what was being talked about?

ucim wrote:
Tyndmyr wrote:We do it [(expend lots of money to find murderers)] to prevent murders.
But why prevent murders in the first place? Is this a "leading cause of death"? That money could be spent curing heart disease, or preventing household accidents. Point is, we don't (and shouldn't) spend money based solely on number of lives saved.


The mere fact that we do things does not make things entirely rational. Some parts of the criminal justice system(how much we spend, the lengths we go to to punish instead of fix things) are indeed also irrational.

And your point has never been something I claimed. I have, in fact, explicitly stated that other things are good too. Obviously. Lives just happen to rank fairly highly in safety discussions. Can we drop this strawman already?

They demonstrated it through a medium for which distance was not an object. That is sufficient. If you truely think that this is not sufficient, they you might as well point out that they did not demonstrate it on a blue car, or on a station wagon.


As is traditional for exploits, a lot of stuff has been withheld. While this is responsible on their part, it means that drawing conclusions about the difficulty without significant amounts of information is...guessing, basically. Claims about the ease of the hack without this are basically all bullshit.

And it's STILL not something that is dependent on a computer controlled car. Yes, yes, cars of the future will have more integrated electronic systems. This will happen regardless of if you ban computers from driving or not. Security errors will still be possible.

Advertising and personal infomation concerns are doomed for the same reason. These are *already* pervasive. People already know where you go, because credit card records. And those are timestamped. It's not as if extrapolating your driving habits from this is *hard*, if anyone cares. Really, though, advertisers are much more interested in the purchasing information directly. Most advertising ties back into the quest for money, so your route mostly only matters if someone can make a dollar off it. It's a pretty easy model to understand.

elasto wrote:OTOH, if my car knows I dislike expensive chain-coffee, and knows I'm nearly out of coffee in my home, and informs me I'm about to pass a supermarket that has a sale on the brand of coffee I like, that also does me a favour.

You and I have a different outlook on the world I fear: I am not afraid to be advertised to; Relevant advertising can anticipate my needs and improve my quality of life.


Yup. We get information, and, obviously, others can also more easily get information. That's sort of how things work.

Ya'll probably don't get ads for CNC machines and industrial lasers. That's cool. But I want to buy them, so yeah, I'd rather have ads for something like that than something irrelevant to me, or something I actually hate. There is no viable option for "no ads"*. There are simply better and worse ads.

*Unless everyone's suddenly willing to pay extra for ad-free services, and promptly ditch services the instant they toss on ads. History does not give this strategy good odds.

Quercus wrote:I think the situation is slightly more nuanced than that for me. I always find that targeted advertising gets very samey very fast. if I've read a few books about spaceflight recently, then my amazon recommendations fill with books about spaceflight, when actually at that point I'm likely to start to crave something completely different, like an historical novel.


Yup. And this is because ads are not targetted enough. They lack sufficient context, sometimes, or sufficient information about me. With more data and a better recommendation engine, a sufficiently advanced advertising machine could determine that additional advertising of the same thing has hit diminishing returns, and to start sampling new things likely to interest me.

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 » Fri Jul 24, 2015 3:08 pm UTC

Tyndmyr wrote:
Quercus wrote:I think the situation is slightly more nuanced than that for me. I always find that targeted advertising gets very samey very fast. if I've read a few books about spaceflight recently, then my amazon recommendations fill with books about spaceflight, when actually at that point I'm likely to start to crave something completely different, like an historical novel.


Yup. And this is because ads are not targetted enough. They lack sufficient context, sometimes, or sufficient information about me. With more data and a better recommendation engine, a sufficiently advanced advertising machine could determine that additional advertising of the same thing has hit diminishing returns, and to start sampling new things likely to interest me.

That's a good point. I suppose that if I know that I want to try the most unusual restaurant in town (unless I'm really tired, in which case I want something comforting and familiar) then there's no reason that an advertising engine that knew my eating habits and my sleep patterns couldn't know that too.

I'm also kind of excited about the possibility of telling your "personal advertising bot" things directly - things like "I'm on a diet - don't advertise high calorie products for the next month". I like this, because it makes it easier to stick to my diet, advertisers like this because they can push loads of ads for expensive health foods.


Tyndmyr wrote:I'm concerned that people are way, way too quick to jump to "think of the DANGERS" without actually demonstrating those dangers. And thus, because of fear of hypothetical dangers that may not be all that large, we live with much greater dangers, simply because they are familiar and comfortable.

In my case I'm definitely in favour of "think of the dangers", and also definitely in favour of self-driving cars. In fact I'm in favour of "think of the comfortable and familiar dangers" too. In fact think of all the dangers you can, and all the advantages you can and proceed in as informed a way as possible. There's a line to be trodden between being reckless and being alarmist.

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

Re: Cars discussion, split from 'other news'

Postby morriswalters » Fri Jul 24, 2015 3:31 pm UTC

elasto wrote:I am not afraid to be advertised to; Relevant advertising can anticipate my needs and improve my quality of life.
Ads require my time and attention. Both of which are limited. And unscrupulous people use your bandwidth to serve you ads. Which you may or may not actually see. The value is based on the trust that I place on those who advertise to me.

What I don't want to be is manipulated. It isn't even a question of if it happens. The only question is can you spot it. In that sense I love targeted ads from Amazon, for instance. And they get quite a bit of money from me. But I don't like it when they manipulate their catalog. I want an honest broker. When I look for a book I don't want them to manipulate what I see based on their petty war with publishers.

I see some of the fears based upon the idea that with the current techniques available to say, Amazon that they may be able to influence me to buy what I otherwise might not. That is a real threat. But it has to be balanced against the positives.

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

Re: Cars discussion, split from 'other news'

Postby HES » Fri Jul 24, 2015 5:05 pm UTC

He/Him/His Image

User avatar
gmalivuk
GNU Terry Pratchett
Posts: 26836
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

Re: Cars discussion, split from 'other news'

Postby gmalivuk » Fri Jul 24, 2015 8:28 pm UTC

Tyndmyr wrote:
gmalivuk wrote:
Tyndmyr wrote:No. That is the analogy. It is not the reality. There is no 20mph restriction on computer driven cars.
But there could be, and it would certainly make them pretty safe, even with current technology that hasn't been tested in adverse weather conditions.

Why are you so opposed to making cars safer? Huh?


So, are you arguing just to argue, or does this actually make an analogy at some point here? Where's the relevance to what was being talked about?
Why do you keep looking for an analogy that isn't there?

I am talking about a literal hardwired speed limit in computer-driven cars. It would undoubtedly make them safe, and yet you don't seem to be advocating for it. Meaning there are tradeoffs you're unwilling to make to save lives with computer-driven cars.

The relevance to other parts of this discussion is that people like ucim are unwilling to make some of the tradeoffs you've suggested, and you can't just claim it all comes down to minimizing loss of life.

They demonstrated it through a medium for which distance was not an object. That is sufficient. If you truely think that this is not sufficient, they you might as well point out that they did not demonstrate it on a blue car, or on a station wagon.
As is traditional for exploits, a lot of stuff has been withheld. While this is responsible on their part, it means that drawing conclusions about the difficulty without significant amounts of information is...guessing, basically. Claims about the ease of the hack without this are basically all bullshit.
Details were withheld, sure, but you're implying that what wasn't withheld was fabricated.

They didn't say exactly how the hack worked, but they did say it doesn't require being in the same network cell. Unless you believe that part of the article was a lie (in which case please share with the rest of the class why you think that), it means distance isn't as important as you seem to think.
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)

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

Re: Cars discussion, split from 'other news'

Postby KrytenKoro » Sat Jul 25, 2015 8:26 pm UTC

gmalivuk wrote:Why do you keep looking for an analogy that isn't there?

I am talking about a literal hardwired speed limit in computer-driven cars. It would undoubtedly make them safe, and yet you don't seem to be advocating for it. Meaning there are tradeoffs you're unwilling to make to save lives with computer-driven cars.

The relevance to other parts of this discussion is that people like ucim are unwilling to make some of the tradeoffs you've suggested, and you can't just claim it all comes down to minimizing loss of life.

Let's be generous and assume that you are exempting emergency vehicles from the list of vehicles that are locked down, and that you are exempting all vehicles which transfer time-sensitive material like money, medicine, organs, pregnant women, people fleeing disaster areas, etc. I'm not sure how you could enforce this, but maybe there's an on-star button in every car and you have to justify why you need to go fast, and the network can be assumed to never, ever falter. Any vehicles that would be necessary to allow people to still get to jobs on time without reducing quality of life through decreased sleep time, possibility of losing jobs, etc. All the various ways people's lives can be worsened or ended through the cessation of their employment or company. We'll assume someone can do this calculation and determine which vehicles need to travel at higher than 20mph and which under in order to save the most lives.

How many cars are left? And how does the 4x time spent on the road factor in to excess pollution, road fatigue, etc.? Then, after that, what about the fact that there's still a significant chance of fatalities at 20mph collisions?

Finally, the fact that you're specifying computer-driven cars -- what evidence do you have that limiting them to 20 mph actually produces a significant change in lives saved?

Most importantly:

I am talking about a literal hardwired speed limit in computer-driven cars. It would undoubtedly make them safe, and yet you don't seem to be advocating for it.

The core of the idea of computer-driven cars is that the car chooses the optimal speed that guarantees safety, while also maximizing fuel efficiency and driver comfort, minimizing pollution, etc. Can you actually point to anyone suggesting that google cars "should be allowed to hit people, sometimes, if it gets you to your destination quicker"?

Tyndmyr's not advocating it, I assume, for the same reason he doesn't advocate that air in spaceships contain oxygen -- because "the need for speeds low enough to ensure safety" is so blinkeringly obvious a fundamental requirement for what we're discussing that it goes without saying.
From the elegant yelling of this compelling dispute comes the ghastly suspicion my opposition's a fruit.

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

Re: Cars discussion, split from 'other news'

Postby morriswalters » Sat Jul 25, 2015 8:40 pm UTC

Your misreading what he is saying I think.
gmalivuk wrote:The relevance to other parts of this discussion is that people like ucim are unwilling to make some of the tradeoffs you've suggested, and you can't just claim it all comes down to minimizing loss of life.
This is the money quote. I take this to mean, what are you willing to accept to get what it is you want from a driverless car? He isn't suggesting a 20 mph limit, rather he is using it as and outlier to test the notion that it is solely about saving lives. Certainly an absolute 20 mph limit would achieve a reduced death toll, even if it didn't stop all accidents and deaths.

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

Re: Cars discussion, split from 'other news'

Postby KrytenKoro » Mon Jul 27, 2015 2:49 am UTC

morriswalters wrote:This is the money quote. I take this to mean, what are you willing to accept to get what it is you want from a driverless car? He isn't suggesting a 20 mph limit, rather he is using it as and outlier to test the notion that it is solely about saving lives. Certainly an absolute 20 mph limit would achieve a reduced death toll, even if it didn't stop all accidents and deaths.

If it can be demonstrated that an absolute 20mph limit for computer-driven cars achieve a statistically-significant (as in, it's clear from the numbers that the deaths were due to traveling over 20 mph, rather than freak lightning strike) reduction in death toll, then I would certainly be behind it. I would love a reason to tell my boss I won't be at work until an hour or two later than normal, especially if the car being computer-driven means I can take that time to read or nap.

I take issue with the claim that the 20 mph limit will "obviously" make the type of car we're discussing safer, when the type of car we're discussing has had no reported fatalities as is, or even accidents that were the fault of the computer. Here's where I wave my freedom flag: the burden of proof for new legal restrictions is on those suggesting to change the laws. The government should not be in the habit of mandating behavior because "common sense says it should probably work".
From the elegant yelling of this compelling dispute comes the ghastly suspicion my opposition's a fruit.

User avatar
gmalivuk
GNU Terry Pratchett
Posts: 26836
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

Re: Cars discussion, split from 'other news'

Postby gmalivuk » Mon Jul 27, 2015 11:53 am UTC

If you know how long it takes to get to work, what makes you think you could get there two hours late every day and still keep your job? Is your boss really that unconcerned with when you show up?

More like now you have to leave home two hours earlier.

(And really the same argument cound be made for not-computer-controlled cars. We could definitely reduce fatalities now by making human-controlled cars incapable or less capable of breaking existing speed limits, without needing to wait a decade or more for driverless technology to come fully up to snuff, and yet I don't see you lobbying for anything like making cruise control unavailable for maintaining high speeds.)
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)

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

Re: Cars discussion, split from 'other news'

Postby KrytenKoro » Mon Jul 27, 2015 1:10 pm UTC

gmalivuk wrote:If you know how long it takes to get to work, what makes you think you could get there two hours late every day and still keep your job? Is your boss really that unconcerned with when you show up?

More like now you have to leave home two hours earlier.

What do you care?

(And really the same argument cound be made for not-computer-controlled cars.

It could, but that's not the argument you made and not the one I took issue with.

I took issue with the one you made.

yet I don't see you lobbying for anything like making cruise control unavailable for maintaining high speeds.)

Seriously, stop it. I took issue with the words you actually said.
From the elegant yelling of this compelling dispute comes the ghastly suspicion my opposition's a fruit.

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

Re: Cars discussion, split from 'other news'

Postby Tyndmyr » Mon Jul 27, 2015 2:31 pm UTC

KrytenKoro wrote:I take issue with the claim that the 20 mph limit will "obviously" make the type of car we're discussing safer, when the type of car we're discussing has had no reported fatalities as is, or even accidents that were the fault of the computer. Here's where I wave my freedom flag: the burden of proof for new legal restrictions is on those suggesting to change the laws. The government should not be in the habit of mandating behavior because "common sense says it should probably work".


This.

In fact, in an environment with both computer and human drivers, being capped at twenty miles per hour would probably be outright dangerous in many areas.

In an environment with *only* computer drivers, well...we don't really have extensive data on that. We might get there one day, but the small amount of data we have so far doesn't portray computer drivers as very dangerous, or suddenly becoming dangerous at speeds over twenty mph.

Even if it were, yes. There are tradeoffs. Duh. I've said this like five times so far. That's obvious. So what?

gmalivuk wrote:
They demonstrated it through a medium for which distance was not an object. That is sufficient. If you truely think that this is not sufficient, they you might as well point out that they did not demonstrate it on a blue car, or on a station wagon.
As is traditional for exploits, a lot of stuff has been withheld. While this is responsible on their part, it means that drawing conclusions about the difficulty without significant amounts of information is...guessing, basically. Claims about the ease of the hack without this are basically all bullshit.
Details were withheld, sure, but you're implying that what wasn't withheld was fabricated.

They didn't say exactly how the hack worked, but they did say it doesn't require being in the same network cell. Unless you believe that part of the article was a lie (in which case please share with the rest of the class why you think that), it means distance isn't as important as you seem to think.


No, I'm not implying they lied. They're simply talking about that specific step, which is cool, sure...but without more complete information, distance could be limited by other factors.

Also, media is notorious for over-representing scare stuff, and being really shitty at understanding technical details. So, I'm not putting a great deal of weight in anything other than what they said, which right now is obviously incomplete for good reason. Understanding how the hack worked is kind of important for fully understanding the impact and risk.

KrytenKoro wrote:
I am talking about a literal hardwired speed limit in computer-driven cars. It would undoubtedly make them safe, and yet you don't seem to be advocating for it.

The core of the idea of computer-driven cars is that the car chooses the optimal speed that guarantees safety, while also maximizing fuel efficiency and driver comfort, minimizing pollution, etc. Can you actually point to anyone suggesting that google cars "should be allowed to hit people, sometimes, if it gets you to your destination quicker"?

Tyndmyr's not advocating it, I assume, for the same reason he doesn't advocate that air in spaceships contain oxygen -- because "the need for speeds low enough to ensure safety" is so blinkeringly obvious a fundamental requirement for what we're discussing that it goes without saying.


Yeah. Some limitation system will exist. It is not obvious that a hard limit of 20mph is ideal, or even superior to a sometimes faster system. Or why gmalvaik cares about it. It is blindlingly obvious that there are multiple factors that need to be considered when comparing things. Such a trivial statement need not be said, and fixating on it after the point has been granted several times just seem strange.

User avatar
ucim
Posts: 6896
Joined: Fri Sep 28, 2012 3:23 pm UTC
Location: The One True Thread

Re: Cars discussion, split from 'other news'

Postby ucim » Mon Jul 27, 2015 5:35 pm UTC

Tyndmyr wrote:It is not obvious that a hard limit of 20mph is ideal, or even superior to a sometimes faster system.
It is however obvious that a hard limit of 20mph on all vehicles will eliminate virtually all high speed collisions. The attendant savings of lives is also pretty evident (although there may be a few mitigating circumstances, it's hard to imagine it not being a net plus).

However, this is not a tradeoff I would be willing to make in exchange for those lives. I suspect most of this thread agrees.

It's about tradeoffs.

There is a very real danger that self-driving cars could be turned against us the same way internet surveillance has been. It's potentially far more intrusive. And it's not about what ads you see - that's a red herring. It's about what editorial content you see, and how you see it. It will happen slowly, so we get used to it, like a frog in hot water. But the time to remove the pot from the stove is now, before the fire comes on.

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.

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

Re: Cars discussion, split from 'other news'

Postby Tyndmyr » Mon Jul 27, 2015 5:58 pm UTC

ucim wrote:
Tyndmyr wrote:It is not obvious that a hard limit of 20mph is ideal, or even superior to a sometimes faster system.
It is however obvious that a hard limit of 20mph on all vehicles will eliminate virtually all high speed collisions. The attendant savings of lives is also pretty evident (although there may be a few mitigating circumstances, it's hard to imagine it not being a net plus).

However, this is not a tradeoff I would be willing to make in exchange for those lives. I suspect most of this thread agrees.

It's about tradeoffs.

There is a very real danger that self-driving cars could be turned against us the same way internet surveillance has been. It's potentially far more intrusive. And it's not about what ads you see - that's a red herring. It's about what editorial content you see, and how you see it. It will happen slowly, so we get used to it, like a frog in hot water. But the time to remove the pot from the stove is now, before the fire comes on.

Jose


Customized editorial content because my car drives itself?

No, no, customized editorial content will happen as a result of tracking and data collection, not due to the options on my car. In short, it will happen regardless.


Return to “News & Articles”

Who is online

Users browsing this forum: No registered users and 28 guests