But the alternative is not having a taxi, it is having my own car ready to go at a moments notice whenever I need it, not subject to any demand, at a cost I’m quite happy to pay.
That might work, or it might not. If there’s a big festival or some event on there may be a shortage of cars when I need one. If I have my own car parked at the office it is ready to go, I know it is there, there is no chance that I will have to wait to get home.
Or I could just drive myself, in my own car that is ready and waiting. I live in a small town of 15000, the hospital is five minutes away, and the traffic is minimal.
Your solutions are adequate but they don’t demonstrate how ride sharing would benefit me in any way. In general I would suffer a loss of freedom to go where I want when I want. I wouldn’t want to arrange for a ride any more than I do now. Right now, if one of my biking buddies offered to pick me up on the way to the MTB trails, I’d decline, I’d rather take myself.
I don’t doubt that in big congested cities such as London having autonomous taxis would be a good thing, but I don’t really see them replacing private ownership.
YMMV of course, but in my experience most people own the most expensive car they can afford or justify to themselves at the time of purchase. No one is looking around saying to themselves “I can afford this Ford Focus with the cruise control and leather seats but I’m going to get a Hyundai with fewer features because it comes closer to meeting my minimum needs.” No, they set themselves a budget and then get the best car they can within their budget.
I don’t know where you live, but where I live there is a vast difference between the number of cars on the road between 7:00 am and 9:30 am and say 11 - 1 pm. A ride-share company is going to have to have enough vehicles for peak travel and charge congestion pricing. And mostly have newer cars, while if you own a car you can still drive it after it is paid off and fully depreciated.
Plus, if they make two or three trips per rush hour (which might be pushing it) they are going to have to return empty, since there are definite counter-commute directions here. Livermore to San Jose at 8 am - jammed. San Jose to Livermore - empty.
Car pooling? Someone is going to have a longer trip Not to mention riding with random people without any adult supervision. Is the ride sharing company going to do background checks on all its passengers?
Will probably work in cities - but not in suburbia.
Who needs a Lexus except maybe a realtor? Vast majority, maybe not, but still lots and lots of people.
Good luck with that once it is proven that self-driving cars are safer than manually-driven ones. I’m sure lots of people enjoyed driving without their seatbelts on, too. Where did that get them? The “won’t somebody please think of the children” argument always prevails.
Define “most expensive they can afford”. Everyone has some necessities, and some discretionary income above what they need for those necessities. People use some of that discretionary income to live in a nicer home, or to eat better food, or to drive a better car, than they need. They use some of it on hobbies that don’t directly build on their needs. They use some to improve their quality of life in all sorts of different ways.
You look at anyone who’s driving what you call “the most expensive car they can afford”, and you’ll find that they could drive a car that’s even more expensive, if they were willing to cut back on some other luxuries. And in fact, some people value nice cars more than they value those other luxuries, and so do just that. Others, meanwhile, will care about other luxuries more than their car, and so will spend “as much as they can afford” on those other things, while keeping their car cheap.
“Most expensive they can afford or justify to themselves at the time of purchase.”
Obviously some people spend more of their discretionary income on cars than others, but I think most people do not get the cheapest car that barely meets their needs, they will instead spend more on a car to get some stuff they don’t need at all. Whether it be a more expensive brand, bigger car, nicer options, etc. We are not all driving around in a Hyundai Getz when many of us could be.
Uh, no. Why would autonomous vehicles be responsible for a million more lives than manually driven ones?
If we are confident that autonomous vehicles are even twice as safe as a typical human driven version, it’s a win, and I’m pretty sure it will be a far far greater level of safety than that within the next five to ten years, at which point I fully expect for there to be millions of semi-autonomous cars (fully autonomous will likely lag behind in numbers for several years) on the road.
Not the car, the algorithm controlling the car. You might have the same piece of software responsible for every car in the US, if it had a serious unknown problem it could possibly cause a lot of deaths. Which just means the software has to be very rigorously tested (not by doing a DMV test obviously.)
Again, my beef was with the statement that we had to be a million times more confident. I can provide links if you want, but surprisingly, they are rigorously testing various algorithms as we speak. Shocking!
Of course there is always the possibility that thousands of people will die and no one will actually notice the problem until we finally reach a million deaths.
Computers with far less than 16 years of age land airliners. You are engaging in pedantery that is far beneath you.
I am pretty sure given the proper arrangements a car manufacturer can get a DMV tester to come to a test facility to try out prototype cars for safety purposes. It would also be childs play for Google or a car manufacturer to hunt down someone who used to do that kind of work and hire them to put the cars through testing to where it will ace any standard driving test. You don’t need to test every single car any more than you need to proofread every copy of a book.
If it tests out better than the average driver, its already a win. I think I see what you are trying to say, but as long as the average car autopilot system is better than the average driver every single self driving car in service will make the road a little bit safer for everyone. It doesn’t need to be a million times better than a person, if its 10% better than the average person thousands of lives and tens of thousands serious injuries a year will be avoided.
They are also not just going to make a driving system and walk away like “we have created the ultimate everything” they will constantly be incorporating new data from cars that are involved in various incidents to help make those systems better.
Well yes of course. The rigorous testing that is currently under way is so that when it comes time to use the things for real we are a million times* more confident that it will work correctly.
Put it this way, if there is a fatal accident that is found to be the fault of the autodrive algorithm, you can expect all cars with that software to be recalled for a fix. Therefore even just from a commercial perspective the manufacturers must be significantly more confident in their software than we are in people.
People are not going to accept a computer driver that is slightly better than the average person, it must be better than the best person.
If a self-driving car is 10% better than the average person then all of the people who are >10% better than the average person will be better off driving themselves.
If you want mass acceptance of this, then it has to better than a person and it has to be a lot better. Most people don’t think about risk analytically.
*Or whatever amount, significantly more confident than we are in people.
I disagree. I’m a brilliant driver, naturally, so if I am injured on the road through someone’s fault it is going to be the fault of someone else. No doubt the same is true for you.
Therefore the salient issue is not whether the computer driver is a better driver than me - how could that be even possible? - but whether it is a better driver than the average of all those other schmoes out there, which of course is a considerably lower bar. If I’m confident that you are raising the average competence of drivers as a whole, I’ll feel safer because I feel that they, not me, are the substantial risk to me. And I’ll feel that this is true for every other road user as well.
Yes, there is a slightly increased risk to me because I am now exposed to having my car driven by a machine that is less naturally gifted a driver than I am. But that is more than offset by the fact that other cars are driven by machines that are vastly more competent that the schlubs who have been behind the wheels up to now.
In the EU, and also I think in the USA, we have type testing. When a new car model is produced, it has to pass a number of stringent tests before it can be sold. All those crash test dummies. Once that is done, there is no requirement to test every car as it comes off the production line.
I can see no reason why the manufacturers of computer driven cars, having passed all the tests, would not be able to sell them to the public. Naturally the tests would have to be modified to include tests of the computer systems.
Of course, VW demonstrated the weakness of such a system, but one hopes that this is unlikely to happen again.
First just speaking statistically, if the computer is slightly better than the average driver then the safest outcome is to have the average and below average drivers driven by a computer while the better drivers drive themselves.
Second, taking into account people’s perceptions, if people generally think they drive better than average then they will naturally want other people to be driven by a computer but not themselves. Therefore in general the people actually buying the car will not be inclined to get a self-driving car unless they can drive it themselves.
The way around these two things, if you want mass acceptance, is to either mandate that people must own an autonomous car and that the computer must drive it, which will piss a lot of people off, OR convince people that the computer is better than them. Each person buying a car must be convinced that the computer is better than they are, not that the computer is better than other people.
The best way to convince people that the computer is better than them is to actually make it a whole lot better than them. So that after accounting for all of the misconceptions and biases and so on, everyone is still convinced that the computer is better than them.
Say I want to buy a car. It’s easy enough to convince me that I’d be safer if other people were driven by computers. But I’m not buying a car for other people, I’m buying one for me, I need to be convinced that I’d be safer if I was driven around by a computer.
I think the testing issue is a red herring. The cars won’t appear on the market unless they’ve had adequate testing.
Obviously not: We want as many automated cars as possible (ideally, all of them) and people keep things in their cars, which they wouldn’t be able to do if they were all ride-share services.
Some automated cars should be ride-shares. Most should most certainly not be. I’m shocked anyone could be confused on this point.
That hasn’t been the case for me in so long I’m feeling old. Back when I was in college I’d have this conversation:
“Head to Bonanova and when we get closer I’ll give you directions.”
“[long suffering sigh optional]Just give me the address, I’ll know where it is.”
“OK. Esperanza, 5.”
“Uhm?” Takes out the map book. Searches. Searches some more. Gives up. “Uhm, so… Bonanova, right?”
The current version involves the GPS and a different address, but still, if I give them “Provença with Roger de Flor” they know the way (pretty popular intersection), if I give them the exact address they need to ask the GPS.
Ah, then what we need is a new driving test. Instead of a test which asks “can you reverse around a corner without mounting the kerb? Can you say what this sign signifies?” we need a test which measures your skills and knowledge against those of a computer driver. If and only if you can match or outperform the computer driver do you get a licence.
Perhaps. Or you just make it really good and win the public over.
I think it will be very good if/when it becomes available to the public. I don’t think any manufacturer is going to release something that’s “a bit better than most people” and lets face it, it’s a computer, it should be able to do the standard driving things far better than any human can.
While doubtless there are a lot of people out there who think they’re much better drivers than they are, that’s still not everyone. I, for one, expect that I am probably less skilled than most drivers. I might be safer than average, but that’s just because I’m not stupid enough to try to text while driving or the like. And, a key point, even if I thought that I was a little safer than the computer, I might still accept the computer driving, just because it’s more convenient. Safety isn’t the end-all be-all, and we accept tradeoffs that decrease our safety all the time.