Sure, maybe; but let's think creatively about private vehicle ownership at least for urban areas. In the city, a private automobile sits most of the time on a driveway, residential, work or commercial business or shopping centre, or on the street parked. It is driven how many hours a week?
Let's have AVs, but only as a per-use trip option or people can have tiered plans on vehicle types and hours used, like phone data plans were at the beginning. And, at the same time, let's phase-out private vehicle ownership; this would solve a number of problems common to cities: reduce air pollution, GHGs, noise pollution, traffic congestion, and reduce the need for homes to have garages (People will have to get by with less stuff, another benefit for our planet.)
Perry, yes. Think of all the expense of owning a car. With your idea, there would still be transportation costs, but my guess is that it would be cheaper than owning a car. Thanks for your thoughtful comment.
Great article! It had a lot of pros and cons I hadn’t thought about. Another con I did read about is AVs heading back out of the city center after dropping owners off at work in order to find cheap or free parking. Or just circling while the owners do an errand in a place with little parking. So this could double the miles driven on these regular trips.
It seems like everyone owning their own AV could easily be more wasteful than the current situation. The true promise lies in robotaxis. Instead of parking a car for 20+ hours per day, each car could get maximal use, resulting in fewer cars needing to be built. Many people who are currently forced to own a car could opt out of ownership. The objection to this used to be peak-hour demand, but maybe the new flexibility in work places and schedules alleviates this problem somewhat.
Larry, thanks for your thoughtful comment. Yes, a friend mentioned to me the notion of robotaxis. That would reduce parking needs. The idea is that parking lots could be reduced in size. Streets would be cleared of all the cars that just sit out in front of people’s homes just waiting to be driven. The next question is who would own the fleet of taxis? Lyft? Uber? One great advantage would be no more parking tickets!
Paul- Thanks for sharing this. I saw a self-driving car in San Francisco recently. And it got everyone in our car talking: what is the responsibility of ownership in the case of an accident? And the discussion was long and inconclusive to say the least. But all I can say was that when my kids saw that car with no one driving it, they got down and hid from the window. I asked them what's wrong and this is what they said: "It's scary."
Thalia, thank for your input. Yes! I'm just like your kids. I have a lot of trepidation over the idea of getting into a self-driving car. And you raise a good point. The technology might be there, but are people willing to use it? Over 15 years ago, Segways were supposed to transform our cities and neighborhoods. But few people bought them. I think part of the reason is that it always looks like the person on one is about to fall off!
Responsibility. yes. that is another conundrum. My guess is that with self-driving cars the burden of responsibility would shift to the manufacturer of the cars. Even if the number of crashes diminishes, that could seriously curtail or end their development. Companies don't want lawsuits. Thanks for your input!
I think you are correct, except I don't believe AVs are, or ever can be, safer than human drivers. This is because you can't solve for unusual situations, no matter how much data you chew through in training the AI models.
Humans possess abductive reasoning, and driving requires this. AI can only use deductive or inductive reasoning, which is very brittle. This means they will always be 'nearly there', but never quite safe enough to let loose at scale. They sort of work in certain, very well mapped out neighbourhoods if they can go slowly. Even then, they cause obstructions and are vastly more unpredictable than people.
So, frankly not only would the cons outweigh the pros of AVs, even if they ever did become all they're promised to be, but there is no way they will ever become ubiquitous enough for the problems you've brought up to be a meaningful concern.
Dan, thanks for your thoughts. You raise a good point, and you might be right, that the thousands of observations needed and the judgement needed to safely move a car down the street is beyond the computational powers of a computer that can fit on a car. I think we might wind up with a hybrid system, one in which human drivers navigate in city traffic, in snow and rain, but a car's self-driving capabilities are used on freeways, finding parking, doing tasks that people suck at like parallel parking.
I mean, it's not really about the size of the computer. Abductive reasoning is as different in kind from what AIs do as machine learning is from what a calculator does. In general I can't see this really taking off unless regulators get incredibly complacent. It just doesn't make sense.
Dan, my gut feeling is that you're right. But I already know folks getting around, at least partly, in self-driving cars. I am, for the most part, doubtful about the promises of "the next big advance." My doubts were proven wrong with cell phones; they were confirmed with Segways.
Self-driving cars are the future but not the near-term future. I don't trust my Tesla to drive itself at all. (I won't get another Tesla when my lease expires this year because Elon.) On Tesla's version of adaptive cruise control I must stay ever vigilant because for some reason the car wants to stop at green lights. So, it's all more hype than reality, in my opinion--which doesn't mean it's not coming one day.
Joel, yes. I don't believe that we will have self-sufficient self-driving cars in the next five years. In the next 20, I think they are a real possibility. Personally, I'm not sure if it's a technology that really needs to be developed. If all the brainpower that has gone into self-driving cars had been deployed into developing a breath mint that could also brush my teeth after I stopped for a quick taco, I'd be happy.
Think of all the time people waste in traffic. To me, making this time productive (even if that just means watching a movie or reading a book--preferably one of mine!) is the use case for self-driving cars.
A new acronym: ZOV, for Zero Occupant Vehicle. The goal is fewer vehicle miles traveled (VMT), but self-driving cars may be driving around carrying nobody, whether dead-heading from ride to ride or being sent to a remote parking space.
Sure, maybe; but let's think creatively about private vehicle ownership at least for urban areas. In the city, a private automobile sits most of the time on a driveway, residential, work or commercial business or shopping centre, or on the street parked. It is driven how many hours a week?
Let's have AVs, but only as a per-use trip option or people can have tiered plans on vehicle types and hours used, like phone data plans were at the beginning. And, at the same time, let's phase-out private vehicle ownership; this would solve a number of problems common to cities: reduce air pollution, GHGs, noise pollution, traffic congestion, and reduce the need for homes to have garages (People will have to get by with less stuff, another benefit for our planet.)
Perry, yes. Think of all the expense of owning a car. With your idea, there would still be transportation costs, but my guess is that it would be cheaper than owning a car. Thanks for your thoughtful comment.
Great article! It had a lot of pros and cons I hadn’t thought about. Another con I did read about is AVs heading back out of the city center after dropping owners off at work in order to find cheap or free parking. Or just circling while the owners do an errand in a place with little parking. So this could double the miles driven on these regular trips.
It seems like everyone owning their own AV could easily be more wasteful than the current situation. The true promise lies in robotaxis. Instead of parking a car for 20+ hours per day, each car could get maximal use, resulting in fewer cars needing to be built. Many people who are currently forced to own a car could opt out of ownership. The objection to this used to be peak-hour demand, but maybe the new flexibility in work places and schedules alleviates this problem somewhat.
Larry, thanks for your thoughtful comment. Yes, a friend mentioned to me the notion of robotaxis. That would reduce parking needs. The idea is that parking lots could be reduced in size. Streets would be cleared of all the cars that just sit out in front of people’s homes just waiting to be driven. The next question is who would own the fleet of taxis? Lyft? Uber? One great advantage would be no more parking tickets!
I have similar thoughts. :)
Paul- Thanks for sharing this. I saw a self-driving car in San Francisco recently. And it got everyone in our car talking: what is the responsibility of ownership in the case of an accident? And the discussion was long and inconclusive to say the least. But all I can say was that when my kids saw that car with no one driving it, they got down and hid from the window. I asked them what's wrong and this is what they said: "It's scary."
Thalia, thank for your input. Yes! I'm just like your kids. I have a lot of trepidation over the idea of getting into a self-driving car. And you raise a good point. The technology might be there, but are people willing to use it? Over 15 years ago, Segways were supposed to transform our cities and neighborhoods. But few people bought them. I think part of the reason is that it always looks like the person on one is about to fall off!
Responsibility. yes. that is another conundrum. My guess is that with self-driving cars the burden of responsibility would shift to the manufacturer of the cars. Even if the number of crashes diminishes, that could seriously curtail or end their development. Companies don't want lawsuits. Thanks for your input!
I think you are correct, except I don't believe AVs are, or ever can be, safer than human drivers. This is because you can't solve for unusual situations, no matter how much data you chew through in training the AI models.
Humans possess abductive reasoning, and driving requires this. AI can only use deductive or inductive reasoning, which is very brittle. This means they will always be 'nearly there', but never quite safe enough to let loose at scale. They sort of work in certain, very well mapped out neighbourhoods if they can go slowly. Even then, they cause obstructions and are vastly more unpredictable than people.
So, frankly not only would the cons outweigh the pros of AVs, even if they ever did become all they're promised to be, but there is no way they will ever become ubiquitous enough for the problems you've brought up to be a meaningful concern.
Dan, thanks for your thoughts. You raise a good point, and you might be right, that the thousands of observations needed and the judgement needed to safely move a car down the street is beyond the computational powers of a computer that can fit on a car. I think we might wind up with a hybrid system, one in which human drivers navigate in city traffic, in snow and rain, but a car's self-driving capabilities are used on freeways, finding parking, doing tasks that people suck at like parallel parking.
I mean, it's not really about the size of the computer. Abductive reasoning is as different in kind from what AIs do as machine learning is from what a calculator does. In general I can't see this really taking off unless regulators get incredibly complacent. It just doesn't make sense.
Dan, my gut feeling is that you're right. But I already know folks getting around, at least partly, in self-driving cars. I am, for the most part, doubtful about the promises of "the next big advance." My doubts were proven wrong with cell phones; they were confirmed with Segways.
Self-driving cars are the future but not the near-term future. I don't trust my Tesla to drive itself at all. (I won't get another Tesla when my lease expires this year because Elon.) On Tesla's version of adaptive cruise control I must stay ever vigilant because for some reason the car wants to stop at green lights. So, it's all more hype than reality, in my opinion--which doesn't mean it's not coming one day.
Joel, yes. I don't believe that we will have self-sufficient self-driving cars in the next five years. In the next 20, I think they are a real possibility. Personally, I'm not sure if it's a technology that really needs to be developed. If all the brainpower that has gone into self-driving cars had been deployed into developing a breath mint that could also brush my teeth after I stopped for a quick taco, I'd be happy.
Think of all the time people waste in traffic. To me, making this time productive (even if that just means watching a movie or reading a book--preferably one of mine!) is the use case for self-driving cars.
A new acronym: ZOV, for Zero Occupant Vehicle. The goal is fewer vehicle miles traveled (VMT), but self-driving cars may be driving around carrying nobody, whether dead-heading from ride to ride or being sent to a remote parking space.