Driverless cars' effect on LCF
#76
Senior Member
Join Date: Jan 2013
Location: Atlanta, GA. USA
Posts: 3,804
Bikes: Surly Long Haul Disc Trucker
Likes: 0
Liked 1 Time
in
1 Post
One of the biggest blockers for driverless cars will be hackers trying to disable the cars or even make them crash. Why? There are mean people out there. And driverless cars will put a lot of people out of work and they'll be mad. Attacks on driverless cars may be very sophisticated. Or it might just be bored teenagers with a high power laser to confuse the sensors.
The car companies are already overwhelmed with the challenges of driverless AI and just as typically happens, the bleeding edge of technology will probably fail to account for various easy attacks that are possible. That's only natural - it's everything the engineers can do just to make it work under normal conditions. Then throw a world of confusing clues designed to confuse the car in there and you have a real mess. Once those attacks cause a few accidents, that may be enough to stop the deployment of driverless cars for many years. Driverless will be all over the place in factories and other private properties but not in general use.
It's also a problem that deep learning/neural networks are not programmed in a traditional sense where you have a specific set of instructions that can be referred to and refined over time. The fact is that when neural networks misbehave there's a lot of head scratching and vodoo involved in fixing the problem and little assurance that the fix always works and didn't introduce new problems.
https://www.technologyreview.com/s/6...e-heart-of-ai/
The car companies are already overwhelmed with the challenges of driverless AI and just as typically happens, the bleeding edge of technology will probably fail to account for various easy attacks that are possible. That's only natural - it's everything the engineers can do just to make it work under normal conditions. Then throw a world of confusing clues designed to confuse the car in there and you have a real mess. Once those attacks cause a few accidents, that may be enough to stop the deployment of driverless cars for many years. Driverless will be all over the place in factories and other private properties but not in general use.
It's also a problem that deep learning/neural networks are not programmed in a traditional sense where you have a specific set of instructions that can be referred to and refined over time. The fact is that when neural networks misbehave there's a lot of head scratching and vodoo involved in fixing the problem and little assurance that the fix always works and didn't introduce new problems.
https://www.technologyreview.com/s/6...e-heart-of-ai/
#77
Senior Member
You can even track soldiers on patrol Afghanistan. Twenty years ago, who would have thought this technology would become so cheap and widespread?
#78
One of the biggest blockers for driverless cars will be hackers trying to disable the cars or even make them crash. Why? There are mean people out there. And driverless cars will put a lot of people out of work and they'll be mad. Attacks on driverless cars may be very sophisticated. Or it might just be bored teenagers with a high power laser to confuse the sensors.
#80
So you think that anger against driverless cars is the only potential source of domestic terrorism? That's ridiculous. If people wanted to go around clipping brake lines with a bolt cutter, there would be plenty of other reasons.
The good thing about driverless cars is they will be programmed to slow down to manageable speeds in the event of a disruption or lack of sufficient information. That is better than humans who go on driving fast when, for instance, visibility is reduced due to fog/rain because the driver thinks that the posted speed limit is how fast they're supposed to drive regardless of conditions.
The good thing about driverless cars is they will be programmed to slow down to manageable speeds in the event of a disruption or lack of sufficient information. That is better than humans who go on driving fast when, for instance, visibility is reduced due to fog/rain because the driver thinks that the posted speed limit is how fast they're supposed to drive regardless of conditions.
#81
Senior Member
Join Date: Jan 2013
Location: Atlanta, GA. USA
Posts: 3,804
Bikes: Surly Long Haul Disc Trucker
Likes: 0
Liked 1 Time
in
1 Post
That said, I see the risk to self driving technology that's been deployed so far. But by the time it made itself onto public roads I'm not sure the risk would be there anymore. Putting the self driving algorithms on the internet and suitable to automatic updates will be highly compelling. But the associated hacking risks might succeed in holding that at bay - I could also see self driving software put behind a very secure firewall in the car that can be circumvented only with a physical connection - this would require people to visit their dealer to get software updates.
Last edited by Walter S; 02-03-18 at 04:01 AM.
#82
Sophomoric Member
That will not be necessary 100% of the time. It will smooth the flow of traffic. Not necessary for accident avoidance. And the specific means of intercommunication is not a settled mater and will almost certainly not involve the cell network. The information exchange is only to other cars within a few hundred meters. And it needs to be WAY faster than a cell phone can place a call. And “who” would the car call??
__________________
"Think Outside the Cage"
#83
Sophomoric Member
Right now today my Garmin tracks my progress while cycling or walking way faster that a cell phone. Right now today a device the size of silver dollar can extrapolate my route after losing one to three of the satellites sending information to that device. A computer in a car will have a lot more AI than my cycling hiking computer. Plus if I go to Strava I can track any of my friends to see if they were on the same route as I rode or walked or when no matter what part of the day I am researching. We could be closer than many believe.
In sports cycling the old saying of “Pictures or it didn’t happen.” Has changed to,”Strava or it didn’t happen.” For those times when someone brags about how far or how fast they ride.
In sports cycling the old saying of “Pictures or it didn’t happen.” Has changed to,”Strava or it didn’t happen.” For those times when someone brags about how far or how fast they ride.
__________________
"Think Outside the Cage"
#84
Senior Member
Join Date: Jan 2013
Location: Atlanta, GA. USA
Posts: 3,804
Bikes: Surly Long Haul Disc Trucker
Likes: 0
Liked 1 Time
in
1 Post
My understanding is that the cars must AT LEAST communicate with the mapping programs, which must be very up-to-date and accurate. For example, the car needs to "know" exactly when and where a road is closed for construction or even for an accident. And yes, it will have to communicate very quickly. From what I've read, this is a major stumbling block.
Hey, it wasn't me it was the car!
#85
Of course not. But hacking computers is a well known threat that billions are spent on every year. The psychology around it is well studied. It is attractive in part because of the technical challenges and the anonymity and remoteness of the hacker and the number of devices he can impact all in one move. Now computers will presumably be in in control of automobiles and there will be significant numbers of computer geeks who'll find that irresistible to mess with. There are enough parallels in modern society, that this is by no means a stretch.
That said, I see the risk to self driving technology that's been deployed so far. But by the time it made itself onto public roads I'm not sure the risk would be there anymore. Putting the self driving algorithms on the internet and suitable to automatic updates will be highly compelling. But the associated hacking risks might succeed in holding that at bay - I could also see self driving software put behind a very secure firewall in the car that can be circumvented only with a physical connection - this would require people to visit their dealer to get software updates.
That said, I see the risk to self driving technology that's been deployed so far. But by the time it made itself onto public roads I'm not sure the risk would be there anymore. Putting the self driving algorithms on the internet and suitable to automatic updates will be highly compelling. But the associated hacking risks might succeed in holding that at bay - I could also see self driving software put behind a very secure firewall in the car that can be circumvented only with a physical connection - this would require people to visit their dealer to get software updates.
#86
Prefers Cicero
I have an older Garmin watch. It tracks signals from satellites to plot where I am, but it doesn't broadcast back to them in real time - you upload the data at home. I don't know if newer ones broadcast your location live.
#87
Sophomoric Member
Self-driving cars aren’t good at detecting cyclists. The latest proposed fix is a cop-out.
Relevant article:
https://slate.com/technology/2018/02...he-answer.html
(I haven't had time to read it yet.)
Relevant article:
https://slate.com/technology/2018/02...he-answer.html
(I haven't had time to read it yet.)
__________________
"Think Outside the Cage"
#88
Senior Member
Join Date: Oct 2015
Posts: 15,929
Bikes: 2015 Workswell 066, 2017 Workswell 093, 2014 Dawes Sheila, 1983 Cannondale 500, 1984 Raleigh Olympian, 2007 Cannondale Rize 4, 2017 Fuji Sportif 1 LE
Liked 3,933 Times
in
2,053 Posts
#89
Prefers Cicero
Self-driving cars aren’t good at detecting cyclists. The latest proposed fix is a cop-out.
Relevant article:
https://slate.com/technology/2018/02...he-answer.html
(I haven't had time to read it yet.)
Relevant article:
https://slate.com/technology/2018/02...he-answer.html
(I haven't had time to read it yet.)
#90
Senior Member
Join Date: Oct 2015
Posts: 15,929
Bikes: 2015 Workswell 066, 2017 Workswell 093, 2014 Dawes Sheila, 1983 Cannondale 500, 1984 Raleigh Olympian, 2007 Cannondale Rize 4, 2017 Fuji Sportif 1 LE
Liked 3,933 Times
in
2,053 Posts
Calling putting a chip in bicycles a "cheat" is an emotionally laden label intending to disparage the whole idea of AI cars. It is like calling an airbag a cheat because the bumper and the crumple zone is not sufficient.
it could well be that no matter How good AI ever gets, having chips in bikes is the most cost-effective measure in that the chips would be another layer of safety, could help locate stolen bikes, could help find lost or injured riders, and would make the "call of shame" much easier than trying to send names of cross streets. (Or maybe not. Unlike some other posters, I am honest enough to admit that I cannot predict the future and will not blindly cleave to my prejudices to the exclusion of reason.)
The goal is to prevent accidents---- so, "good" is preventing accidents and "bad" is any action or lack of action which causes accidents. In that context "cheating" would be an action which dishonestly caused more accidents---like claiming there was a sensor when there wasn't, or claiming the AI detected bikes when it didn't .... fraud resulting in more accidents would be cheating.
AI may rock or AI may suck .... but posters who ignore common decency and respect of keeping the discussion honest .... yeah. Enough said.
As I said elsewhere, the discussion has devolved to "I like the idea of AI cars" or "I hate the idea of AI cars" with no reference---because it has not yet been tested---to real-world, wide-spread applications.
"Smartphone, smartphone, smartphone ... "
#91
Prefers Cicero
I would argue that it is also a safety issue, similar to having a Lojack on a car, or OnStarr.
Calling putting a chip in bicycles a "cheat" is an emotionally laden label intending to disparage the whole idea of AI cars. It is like calling an airbag a cheat because the bumper and the crumple zone is not sufficient.
it could well be that no matter How good AI ever gets, having chips in bikes is the most cost-effective measure in that the chips would be another layer of safety, could help locate stolen bikes, could help find lost or injured riders, and would make the "call of shame" much easier than trying to send names of cross streets. (Or maybe not. Unlike some other posters, I am honest enough to admit that I cannot predict the future and will not blindly cleave to my prejudices to the exclusion of reason.)
The goal is to prevent accidents---- so, "good" is preventing accidents and "bad" is any action or lack of action which causes accidents. In that context "cheating" would be an action which dishonestly caused more accidents---like claiming there was a sensor when there wasn't, or claiming the AI detected bikes when it didn't .... fraud resulting in more accidents would be cheating.
AI may rock or AI may suck .... but posters who ignore common decency and respect of keeping the discussion honest .... yeah. Enough said.
As I said elsewhere, the discussion has devolved to "I like the idea of AI cars" or "I hate the idea of AI cars" with no reference---because it has not yet been tested---to real-world, wide-spread applications.
"Smartphone, smartphone, smartphone ... "
Calling putting a chip in bicycles a "cheat" is an emotionally laden label intending to disparage the whole idea of AI cars. It is like calling an airbag a cheat because the bumper and the crumple zone is not sufficient.
it could well be that no matter How good AI ever gets, having chips in bikes is the most cost-effective measure in that the chips would be another layer of safety, could help locate stolen bikes, could help find lost or injured riders, and would make the "call of shame" much easier than trying to send names of cross streets. (Or maybe not. Unlike some other posters, I am honest enough to admit that I cannot predict the future and will not blindly cleave to my prejudices to the exclusion of reason.)
The goal is to prevent accidents---- so, "good" is preventing accidents and "bad" is any action or lack of action which causes accidents. In that context "cheating" would be an action which dishonestly caused more accidents---like claiming there was a sensor when there wasn't, or claiming the AI detected bikes when it didn't .... fraud resulting in more accidents would be cheating.
AI may rock or AI may suck .... but posters who ignore common decency and respect of keeping the discussion honest .... yeah. Enough said.
As I said elsewhere, the discussion has devolved to "I like the idea of AI cars" or "I hate the idea of AI cars" with no reference---because it has not yet been tested---to real-world, wide-spread applications.
"Smartphone, smartphone, smartphone ... "
#92
Senior Member
Join Date: Oct 2015
Posts: 15,929
Bikes: 2015 Workswell 066, 2017 Workswell 093, 2014 Dawes Sheila, 1983 Cannondale 500, 1984 Raleigh Olympian, 2007 Cannondale Rize 4, 2017 Fuji Sportif 1 LE
Liked 3,933 Times
in
2,053 Posts
As i said elsewhere, this discussion is not so much a discussion any more. it could be a poll. People who think AI has a future explain why they think that, and people who hate the idea invent stuff about why it couldn't possibly happen.
it might be impossible--no one knows. But if all we are doing is inventing reasons ... fabricating stuff to back up our positions ... well, if this is a creative writing contest the writing quality is unacceptably low.
I don't see this as a discussion ... it is more like people from different religions and a couple atheists all trying to prove they are right.
I would prefer it if people actually stuck to facts, and maybe researched claims .... I don't know, bothered to read articles they posted as "proof" that they were right, before posting them as proof? You know, sort of discuss this the way intelligent well-mannered children might, instead of like ignorant adults (thew worst type of people on the planet sadly, when it comes to using reason in discussions. Talk radio is popular because it resonates .... )
I don't "Know " that my spiritual beliefs are "Right" and I don't know if AI cars will ever come to pass. It seems some others here cannot accept the simple fact that they might not work ... or might work better than what we have now.
If we cannot base a discussion in fact ... whatever, I guess. have fun.
At the moment biking and walking are freely available to us and are also more or less anonymous pursuits unless we choose to make our location and route data available; so the idea that we may soon be expected (read: forced) to carry technology that is designed to make somebody else less responsible for the safety of their driving or their cars' driving, and which might also put us under increased surveillance, is kind of offensive to me.
You have a cell phone? Take it on your rides? Trackable. Every use a GPS program? Trackable if anyone cared to track you. Ever ride outdoors? Satellite tracking.
I recall many years ago when I was invited to sensitive meetings, that everyone had to take the batteries out of their phones because it turns out that the phones could be used as listening devices even if they were shut off. And that is not tinfoil-hat stuff. That's the real stuff.
Makes sense to me that TVs could do the same. Thing is most people aren't worth the effort of tracking.
Funny, though, how more often nowadays cyclists show up in court using Strava as proof that the driver who hit them was at fault.
But yeah, I get that some people see a chip in their bikes as "invasion of privacy." (I wonder if they realize their cars will all have that and most of them already do?)
I guess there would be a market in older bikes, hand-crafted bikes ("Our company has been chip-free since our founding" would be a sales slogan ... or maybe it would be a wink-and-nudge deal.) But I would bet that most people simply wouldn't care---how many people ride without a phone, anyway?
But that is a privacy issue ... and not necessarily part of the AI discussion because .... IT ISN'T REAL. It is something people are discussing ... and sorry you live in America, but if most consumers want it and most manufacturers want it, it will be offered. It might be mandated. It has not been yet.
I am sure there are people who think driver's licenses are an illegal infringement on privacy too ("sovereign citizens," they call themselves. And if you think that is a joke, there was a guy in Arizona a few weeks back busted after a high-speed chase where he rammed a couple cars----claimed he didn't have to pull over when a cop hit the lights and siren for speeding because as a sovereign citizen he had to agree that the cop had a right to stop him, and he didn't agree.)
Thing is, in a representative republic where we are all free individuals and also share this nation ... we sometimes have to compromise individual freedom for the greater good. If, in order to increase road safety, I need to run lights at night ... I need to use hand signals .... and if, in order to help ensure the safety of my fellow citizens I need to have chip in my bike ... I think that is a fair trade for having access to millions of miles of paved roads, and law enforcement officials who will come to my rescue and protect my rights to use them.
You cannot yell "fire" in a crowded theater, you cannot try to incite people to violence or harm ... and if you want to Share the Roads you have to follow the Rules of the Road.
If you Really cannot ride a bike with a chip in it, there will be ways around it. (I bet all kinds of frames will be shipped out of China without chips.)
The idea that "Big Brother is Watching" is pretty ridiculous, you know. Simply think about it (something so few do.) It would take two people one eight-hour shift each to track on person through one day. That means there would need to be two government employees on duty for every non-government citizen. You see the basic logistical impossibility?
Even thought right now the government Could track any citizen through every second of every day ... it simply cannot in reality. The manpower and the expense are prohibitive.
That is why I don't freak out knowing that whenever I use the internet, the phone, watch TV, or ride with my phone using a GPS program, I am trackable. I don't care and no one else does either. Shoot ... are you afraid to respond on this forum? Do you stop to think that you are possibly being tracked, even being monitored, through your ISP data?
As I said ... this is not specifically an AI cars issue ... and for the vast majority of cyclists it is already being done voluntarily through Strava and RidewithGPS and MapMyRide.
(The idea of chips in bikes is a speculation raised in a few articles ... pardon me if I don't have a lot of respect for people who tell ghost stories and then get scared by their own fabrications. )
By the way, I have heard many cyclists on other forums on this site Ask for some kind of tracking chip in their bikes as a Lojack-style anti-theft device. It could come to pass regardless of whether AI cars ever happen ... and it could be a positive thing in the eyes of many consumers, a selling point for people buying $4K bikes. What then?
What then, for the people who are so concerned that they leave their Mobile phone at home whenever they go mobile .... is to buy old bikes and sell them to other privacy aficionados. This is America ... where is your entrepreneurial spirit? Fill a warehouse with chip-free frames and get rich!
If there is sufficient demand, there will be a supply.
It is funny though ... no one here has raised the issue of Cars being tracked. I know most of us use cars at least now and then.
Whatever. I think the whole "Big Brother is Watching" thing is a little .... misguided and counterproductive (as most fears are) and I also think there will be options for those who really want them.
I also think that in America, we have to Voluntarily sacrifice some pure freedom for the benefits of collective living. Roads, fire and police departments, emergency responders, environmental and safety standards .... I do not necessarily like everything being an American entails, but on the whole it is Way better than a lot of other options, and I choose it.
No one is Forced to stay here. If it is really intolerable, you have options.
The thing is, and for me this is one of the Big things about living in America ..... if it Really sucks so bad, we have ways to change it.
Start now gaining support for your lobby against chips in bikes. If it is That much of an issue to you, act. Talk is free, here in America .. . we can say some truly awful stuff and get away with it. But some folks mean what they say and walk their talk.
Now Is the Time. Get a jump on the issue and start organizing now if you want to stop chips from being implanted in bikes. I am one hundred percent serious. It is not my issue, but if it is Really an big thing for you, you need to get counter-traction before the idea gains traction.
I hate to say it but chips in bikes will be an easy sell to most people. If you want even the Option of not having a chip in your bike, get busy.
Look at this---motorcyclists are required by law to wear helmets---except in some states where they lobbied and got the laws changed. It is Not impossible. It will be a lot easier if you get in there early, before the laws are written and the various people with vested interests start adding up the profits ....
Now is the time. A GoFundMe page, a petitions.com petition, a website ... eventually late-night TV commercials (those 11=p.m.-6 a.m. timeslots can be had pretty cheap.) If it Really matters ... get down, brother, and may your efforts bring you satisfaction.
But yeah. back on topic ... chips in bikes are Not part of the AI cars discussion because AI cars are only speculative, and the chips are speculations based on speculations. Using the fantasy of chips in bikes as a reason not to like AI cars is a bit of a stretch. i could equally posit that there will be AI gunships patrolling the street shooting people driving their own cars. Let's leave science fiction to a different website, shall we?
Let me know when you get your anti-chip website running.
Last edited by Maelochs; 02-06-18 at 11:13 AM.
#93
At the moment biking and walking are freely available to us and are also more or less anonymous pursuits unless we choose to make our location and route data available; so the idea that we may soon be expected (read: forced) to carry technology that is designed to make somebody else less responsible for the safety of their driving or their cars' driving, and which might also put us under increased surveillance, is kind of offensive to me.
What if the chips are not GPS-monitorable, but only get sensed by the cars within a certain range, e.g. 100ft? What if the tech is just an extra device, like a bike light, that broadcasts at a generic frequency, which triggers driverless cars to slow/stop if they don't register anything on their radar that corresponds with the presence of the signal?
In other words, what is wrong with cyclists/pedestrians being able to signal their presence to cars using a certain designated frequency? It wouldn't absolve the cars of the responsibility to sense pedestrians/cyclists and avoid collisions, but it would give us something to do to make certain they are aware of our presence.
#94
Prefers Cicero
Well, that hurts.
The idea that "Big Brother is Watching" is pretty ridiculous, you know. Simply think about it (something so few do.) It would take two people one eight-hour shift each to track on person through one day. That means there would need to be two government employees on duty for every non-government citizen. You see the basic logistical impossibility?...Even thought right now the government Could track any citizen through every second of every day ... it simply cannot in reality. The manpower and the expense are prohibitive.
(more to follow)
Last edited by cooker; 02-06-18 at 01:40 PM.
#95
Been Around Awhile
Join Date: Oct 2004
Location: Burlington Iowa
Posts: 30,119
Bikes: Vaterland and Ragazzi
Liked 1,650 Times
in
1,111 Posts
"What we have here ... (say it with me ...) "is failure to communicate."
As i said elsewhere, this discussion is not so much a discussion any more. it could be a poll. People who think AI has a future explain why they think that, and people who hate the idea invent stuff about why it couldn't possibly happen.
As i said elsewhere, this discussion is not so much a discussion any more. it could be a poll. People who think AI has a future explain why they think that, and people who hate the idea invent stuff about why it couldn't possibly happen.
There are the thinkers who agree with you and explain their thoughts, opinions or view of life, and there are haters who don't agree or hold differing thoughts, opinions or views of life and justify such alleged aberrant thoughts with invented stuff.
#96
Senior Member
Join Date: Oct 2015
Posts: 15,929
Bikes: 2015 Workswell 066, 2017 Workswell 093, 2014 Dawes Sheila, 1983 Cannondale 500, 1984 Raleigh Olympian, 2007 Cannondale Rize 4, 2017 Fuji Sportif 1 LE
Liked 3,933 Times
in
2,053 Posts
#97
Prefers Cicero
Wow, that is a single semicolon-divided sentence!
What if the chips are not GPS-monitorable, but only get sensed by the cars within a certain range, e.g. 100ft? What if the tech is just an extra device, like a bike light, that broadcasts at a generic frequency, which triggers driverless cars to slow/stop if they don't register anything on their radar that corresponds with the presence of the signal?
In other words, what is wrong with cyclists/pedestrians being able to signal their presence to cars using a certain designated frequency? It wouldn't absolve the cars of the responsibility to sense pedestrians/cyclists and avoid collisions, but it would give us something to do to make certain they are aware of our presence.
What if the chips are not GPS-monitorable, but only get sensed by the cars within a certain range, e.g. 100ft? What if the tech is just an extra device, like a bike light, that broadcasts at a generic frequency, which triggers driverless cars to slow/stop if they don't register anything on their radar that corresponds with the presence of the signal?
In other words, what is wrong with cyclists/pedestrians being able to signal their presence to cars using a certain designated frequency? It wouldn't absolve the cars of the responsibility to sense pedestrians/cyclists and avoid collisions, but it would give us something to do to make certain they are aware of our presence.
Yes, it is possible that the device could be a generic, non-identifiable electronic beacon that cars could detect when they are close, much like drivers can now see your flashing tail light. I don't know if it is easy for sensors to then match that to your visual image so they "see you" and track your movement to avoid hitting you, but certainly that would be cheaper and less invasive than some kind of personalized electronic ID.
Last edited by cooker; 02-06-18 at 03:04 PM.
#98
Long sentences aren't automatically a bad thing. (spoiler alert!) I read an article on how you could condense the plot of Macbeth either into 7 sentences, or into one power sentence with the same information in way fewer words, and the one sentence was way more reader-friendly than the seven. It went something like this: "Spurred on by his wife, MacBeth realized his ambition and fulfilled the witches' prophesy by killing Duncan and replacing his as King".
#99
Sophomoric Member
I will be glad to wear a sensor to make myself visible to machines--provided the car companies pay me very well to do it, and agree to pay me if I get injured because the sensor doesn't work 100% of the time.
__________________
"Think Outside the Cage"
#100
Senior Member
Join Date: Oct 2015
Posts: 15,929
Bikes: 2015 Workswell 066, 2017 Workswell 093, 2014 Dawes Sheila, 1983 Cannondale 500, 1984 Raleigh Olympian, 2007 Cannondale Rize 4, 2017 Fuji Sportif 1 LE
Liked 3,933 Times
in
2,053 Posts
You surviving family members will be grateful---funeral expenses are higher every day.