That's why I said that cultural issues, and not technical ones will be the main obstacle to full one large scale autonomous/self driving vehicle.
But, like I said, on long haul roads, with few exits, they could dedicate the fast lanes to this kind of traffic and just let it rip at 100 MPH. This
would mean a lot less traffic on the rest of the road. They could even separate the traffic completely. Put the high speed lanes in the middle of the freeway.
Every movement in and out of this lane would be automated until goes back to the normal freeway which would then probably have 1/2 the traffic it has now. So, risk of accidents there would also diminish.
Even other cars would have a lot of driving/piloting assistance functions; it would not be possible for a vahicle to get into the center lane accidentally piloting/driving assistance would protect from that. Which is a good thing considering how fast cars would be moving there : 100mph in long road trains.
Of course, fair weather states like California/Arizona/Nevada will be amongst the first to get these.
I agree with you that this makes a lot of sense. I have to say that regarding the autonomy of driving I think it will come in a quite foreseeable manner. I am actually more curious how such cars will be powered. Currently, I can only foresee a mix of electricity and hydrogen, maybe some bio-fuels. And mix means in all ways, vehicle to vehicle with "pure" tech, as well as hub rids using more than one tech.
I am not saying it will work from day one, and there is a clear distinction between autonomous and self-driving. Self-driving requires the environment (=toher vehicles, roads etc.) be i n line with this concept. The most dangerous period is the one of mixed population of cis: assisted, autonomous and self.driving. After that, it is just a question of getting used to that paradigm. If you look at the root cause statistics for car accidents you will find that the overwhelming majority is caused by human error. Even if (which will obviously not be allowed), self.driving cars would cause deaths, it would be less than "user-driven" vehicles actually do. No toothpick involved here.
I agree. It's the user override that will get people killed. Eventually, I can see them banning manual drive cars from the freeways, and ultimately city streets. Not sure where the classic car enthusiasts will go to drive them. However, I don't think you can take away control mechanisms altogether since there will be times where the software will need to be reset possibly while driving. And then there are times when the car is not being driven, but being towed, or pulled out of being stuck, or moved off the road when the batty runs out, where human control is just more practical or easier.
The biggest issue I see with the self-driving car, is what kind of algorythm is in place to handle accidents? Let's say you're driving down the road, and a tire blows out. Before the car can safely adjust the car ends up with three choices, none of which avoid an accident -- straight ahead and the driver and passengers are likely to be killed, turn right and a pedestrian pushing a baby carriage will likely be killed. To the left and a head on collision which will likely kill the oncoming driver, and cause other death and injuries in chain-reaction crashes. Now the car can react much faster than the driver, so the override won't be an option here, but the car is also now presented with several options from which it must decide a course of action. And depending on how it's programmed, who is ultimately responsible?
I agree. It's the user override that will get people killed. Eventually, I can see them banning manual drive cars from the freeways, and ultimately city streets. Not sure where the classic car enthusiasts will go to drive them. However, I don't think you can take away control mechanisms altogether since there will be times where the software will need to be reset possibly while driving. And then there are times when the car is not being driven, but being towed, or pulled out of being stuck, or moved off the road when the batty runs out, where human control is just more practical or easier.
The biggest issue I see with the self-driving car, is what kind of algorythm is in place to handle accidents? Let's say you're driving down the road, and a tire blows out. Before the car can safely adjust the car ends up with three choices, none of which avoid an accident -- straight ahead and the driver and passengers are likely to be killed, turn right and a pedestrian pushing a baby carriage will likely be killed. To the left and a head on collision which will likely kill the oncoming driver, and cause other death and injuries in chain-reaction crashes. Now the car can react much faster than the driver, so the override won't be an option here, but the car is also now presented with several options from which it must decide a course of action. And depending on how it's programmed, who is ultimately responsible?
Who's responsible when a train crashes? I think that the responsibility question can be answered quite easily. Much less easy is the answer to your other questions, basically giving me software a choice of who may live and who will not. Still, it would be based on objective criteria, such as highest survival chances and much better than a human judgement. I don't want to even think about the tinkering of such systems through hacking or conscious manipulation. An app crashing and a car crashing are two very different animals.
maybe not but he's still just a software biz guy. advanced energy engineering isn't his wheelhouse. if this stuff worked yet, we'd have it.
Well, people are so irrational about nuclear, that even if the tech to deploy cheap safe one existed (and eventually it will), there would still be a massive inertia and resistance to it.
What would then make the US change its mind would be strategic reason, say China puts those reactors all over the place giving them a huge strategic advantage, I don't think the US would not follow suite.
Stop dreaming! Let's be realistic.
Do I still have to buy automotive insurance for a self-driving car?
The only thing good for this endeavor is economy. Companies pouring money into research creating numerous high paying jobs.
Insurance should be MUCH lower since losses will be drastically reduced.
EDIT: Here's an industry paper discussing the upcoming effect of autonomous vehicles on insurers. It will likely be major. https://goo.gl/mGHFAA
Those thorium breeder reactors seem to be a pretty decent implementation of new-q-lar.
It's funny, Bill Gates has been talking them up for years. Like him or not he ain't no dummy.
maybe not but he's still just a software biz guy. advanced energy engineering isn't his wheelhouse. if this stuff worked yet, we'd have it.
Quote:
Originally Posted by foggyhill
Quote:
Originally Posted by NolaMacGuy
maybe not but he's still just a software biz guy. advanced energy engineering isn't his wheelhouse. if this stuff worked yet, we'd have it.
Well, people are so irrational about nuclear, that even if the tech to deploy cheap safe one existed (and eventually it will), there would still be a massive inertia and resistance to it.
What would then make the US change its mind would be strategic reason, say China puts those reactors all over the place giving them a huge strategic advantage, I don't think the US would not follow suite.
Notably the "Future Plants" section. I actually found a lot on that wiki page pretty informative and interesting.
Also foggyhill I think you might be right.
As China and the rest of the world moves forward, North America sits around burning coal, oil and natural gas for power. Things might change. But when people hear nuclear, they think of bombs, Chernobyl, Three Mile Island, and toxic waste with a half life so long that is hard to comprehend. Breeders largely avoid those issues.
I can imagine a near future (~20 years) where China is exporting both nuclear fuel and electricity, and has largely cleaned their skies of coal pollution.
There is more serious question. It is a dream that software without bugs. Software is constantly updated. We have iOS 9.0, 9.0.1, 9.0.2 and now 9.1 in how long? If a serious bug is causing many accidents and fatalities, who should be responsible?
I think Steve Jobs once made an interesting comparison of Microsoft to Auto companies. Do you remember?
There is more serious question. It is a dream that software without bugs. Software is constantly updated. We have iOS 9.0, 9.0.1, 9.0.2 and now 9.1 in how long? If a serious bug is causing many accidents and fatalities, who should be responsible?
I think Steve Jobs once made an interesting comparison of Microsoft to Auto companies. Do you remember?
The software providers will obviously need to cover liabilities with big policies, perhaps lessening the punch insurers will take from fewer and much lower priced private and commercial insurance policies.
The only universe where a Google car with no user inputs would be even feasible is in environments designed specifically around such a vehicle. This concept is utterly IMPOSSIBLE on real roads for at least the next couple of decades. Just like most things that Google does, that gets endlessly hyped out in blogs and in the media (Google Glass anyone?) these cars are nothing but a tech demo. Unless Google is planning to build their own Google Roads for these vehicles alone, they have zero practical value to consumers in the forseeable future. It's just a "look at us!" product with no go to market strategy.
The only universe where a Google car with no user inputs would be even feasible is in environments designed specifically around such a vehicle. This concept is utterly IMPOSSIBLE on real roads for at least the next couple of decades. Just like most things that Google does, that gets endlessly hyped out in blogs and in the media (Google Glass anyone?) these cars are nothing but a tech demo. Unless Google is planning to build their own Google Roads for these vehicles alone, they have zero practical value to consumers in the forseeable future. It's just a "look at us!" product with no go to market strategy.
Just because you can go fully autonomous doesn't mean the cars won't have ways for actual drivers to intervene. And why are you focusing on Google. They're far from the only company with autonomous cars in testing. For instance BMW says "“Technically (we) could release our driverless cars now. We have tested them in Munich where the human engineer is completely passive, like a passenger,”.
This could have nothing to do with Apple Car or everything. At this point we are jumping to conclusions. Machine learning skills could be needed for Siri or many other projects at Apple. I see no evidence that this hire is for anything specific.
You're the scene in Police Squad with the fireworks exploding in the background, exclaiming there's nothing to see here.
Comments
That's why I said that cultural issues, and not technical ones will be the main obstacle to full one large scale autonomous/self driving vehicle.
But, like I said, on long haul roads, with few exits, they could dedicate the fast lanes to this kind of traffic and just let it rip at 100 MPH. This
would mean a lot less traffic on the rest of the road. They could even separate the traffic completely. Put the high speed lanes in the middle of the freeway.
Every movement in and out of this lane would be automated until goes back to the normal freeway which would then probably have 1/2 the traffic it has now. So, risk of accidents there would also diminish.
Even other cars would have a lot of driving/piloting assistance functions; it would not be possible for a vahicle to get into the center lane accidentally piloting/driving assistance would protect from that. Which is a good thing considering how fast cars would be moving there : 100mph in long road trains.
Of course, fair weather states like California/Arizona/Nevada will be amongst the first to get these.
I agree with you that this makes a lot of sense. I have to say that regarding the autonomy of driving I think it will come in a quite foreseeable manner. I am actually more curious how such cars will be powered. Currently, I can only foresee a mix of electricity and hydrogen, maybe some bio-fuels. And mix means in all ways, vehicle to vehicle with "pure" tech, as well as hub rids using more than one tech.
I agree. It's the user override that will get people killed. Eventually, I can see them banning manual drive cars from the freeways, and ultimately city streets. Not sure where the classic car enthusiasts will go to drive them. However, I don't think you can take away control mechanisms altogether since there will be times where the software will need to be reset possibly while driving. And then there are times when the car is not being driven, but being towed, or pulled out of being stuck, or moved off the road when the batty runs out, where human control is just more practical or easier.
The biggest issue I see with the self-driving car, is what kind of algorythm is in place to handle accidents? Let's say you're driving down the road, and a tire blows out. Before the car can safely adjust the car ends up with three choices, none of which avoid an accident -- straight ahead and the driver and passengers are likely to be killed, turn right and a pedestrian pushing a baby carriage will likely be killed. To the left and a head on collision which will likely kill the oncoming driver, and cause other death and injuries in chain-reaction crashes. Now the car can react much faster than the driver, so the override won't be an option here, but the car is also now presented with several options from which it must decide a course of action. And depending on how it's programmed, who is ultimately responsible?
maybe not but he's still just a software biz guy. advanced energy engineering isn't his wheelhouse. if this stuff worked yet, we'd have it.
Who's responsible when a train crashes? I think that the responsibility question can be answered quite easily. Much less easy is the answer to your other questions, basically giving me software a choice of who may live and who will not. Still, it would be based on objective criteria, such as highest survival chances and much better than a human judgement. I don't want to even think about the tinkering of such systems through hacking or conscious manipulation. An app crashing and a car crashing are two very different animals.
Stop dreaming! Let's be realistic.
Do I still have to buy automotive insurance for a self-driving car?
The only thing good for this endeavor is economy. Companies pouring money into research creating numerous high paying jobs.
maybe not but he's still just a software biz guy. advanced energy engineering isn't his wheelhouse. if this stuff worked yet, we'd have it.
Well, people are so irrational about nuclear, that even if the tech to deploy cheap safe one existed (and eventually it will), there would still be a massive inertia and resistance to it.
What would then make the US change its mind would be strategic reason, say China puts those reactors all over the place giving them a huge strategic advantage, I don't think the US would not follow suite.
EDIT: Here's an industry paper discussing the upcoming effect of autonomous vehicles on insurers. It will likely be major.
https://goo.gl/mGHFAA
Those thorium breeder reactors seem to be a pretty decent implementation of new-q-lar.
It's funny, Bill Gates has been talking them up for years. Like him or not he ain't no dummy.
maybe not but he's still just a software biz guy. advanced energy engineering isn't his wheelhouse. if this stuff worked yet, we'd have it.
maybe not but he's still just a software biz guy. advanced energy engineering isn't his wheelhouse. if this stuff worked yet, we'd have it.
Well, people are so irrational about nuclear, that even if the tech to deploy cheap safe one existed (and eventually it will), there would still be a massive inertia and resistance to it.
What would then make the US change its mind would be strategic reason, say China puts those reactors all over the place giving them a huge strategic advantage, I don't think the US would not follow suite.
NolaMacGuy, check out https://en.wikipedia.org/wiki/Breeder_reactor
Notably the "Future Plants" section. I actually found a lot on that wiki page pretty informative and interesting.
Also foggyhill I think you might be right.
As China and the rest of the world moves forward, North America sits around burning coal, oil and natural gas for power. Things might change. But when people hear nuclear, they think of bombs, Chernobyl, Three Mile Island, and toxic waste with a half life so long that is hard to comprehend. Breeders largely avoid those issues.
I can imagine a near future (~20 years) where China is exporting both nuclear fuel and electricity, and has largely cleaned their skies of coal pollution.
Insurance should be MUCH lower since losses will be drastically reduced.
EDIT: Here's an industry paper discussing the upcoming effect of autonomous vehicles on insurers. It will likely be major.
https://goo.gl/mGHFAA
There is more serious question. It is a dream that software without bugs. Software is constantly updated. We have iOS 9.0, 9.0.1, 9.0.2 and now 9.1 in how long? If a serious bug is causing many accidents and fatalities, who should be responsible?
I think Steve Jobs once made an interesting comparison of Microsoft to Auto companies. Do you remember?
Sell your car insurance company stocks while you can! Sell! Sell! Sell! ????????????????????
The only universe where a Google car with no user inputs would be even feasible is in environments designed specifically around such a vehicle. This concept is utterly IMPOSSIBLE on real roads for at least the next couple of decades. Just like most things that Google does, that gets endlessly hyped out in blogs and in the media (Google Glass anyone?) these cars are nothing but a tech demo. Unless Google is planning to build their own Google Roads for these vehicles alone, they have zero practical value to consumers in the forseeable future. It's just a "look at us!" product with no go to market strategy.
http://phys.org/news/2015-10-german-autonomous-car-mexico.html
http://www.popularmechanics.com/cars/a9541/the-12-most-important-questions-about-self-driving-cars-16016418
https://www.mercedes-benz.com/en/mercedes-benz/innovation/autonomous-long-distance-drive/