True, but there’s one major saving to on-demand renting: parking space. If you don’t own a car, and rely on self-driving Ubers, you don’t need a parking space at home. This can be a major expense in some areas. Even in a typical suburb, the garage + driveway is a non-negligible fraction of your property and building cost.
I’m envisioning a future when self-driving Ubers/taxis are ubiquitous, in which case wait time will be negligible. Presumably you could call one a few minutes before you need one, and it’ll be waiting on the curb when you step out of your house.
The cost of owning a car is also unpredictable - you never know when your car is going to break down and need repairs.
This is a disadvantage of owning a car, because most of us don’t have the resources to buy several cars optimized for each use. So people end up buying a minivan large enough for the whole family, and driving it alone most of the time. Or buying an SUV that can tow your boat, and driving it to work every day. It would be much more efficient if we could just summon an appropriate vehicle anytime.
Presumably, if a dirty car arrives, you flag it as such on your app and summon another one. The operating company will send the car to their facility for cleaning. Perhaps charge the previous user extra for soiling the car, if it’s obvious.
Sure, the per-mile cost of depreciation & wear and tear on the car probably won’t go down dramatically. We do gain all the real estate previously used to store the 300 million cars when they aren’t in use.
Not if there is a flow of unoccupied self-driving cars past the street, any one of which can pick you up immediately.
Why would the car need to stay near the worker?
No, the idea is for the cars to drop people off at work, and loiter in the area (or just stop on an empty street) until someone in the area needs a ride.
I agree with you there. Most self-driving cars will probably be used by a single occupant.
I agree with you there. And I don’t think anyone is claiming that self-driving cars will lead to reduced dependency on cars. What it can do is allow more traffic through existing roads.
Probably to some extent. But there’s still a per-mile cost for using a self-driving car. And not many people will be willing to spend 4 hours a day in a car even if they can spend the time working.
Absolute, I disagree with you that self-driving cars won’t reduce car ownership. You’ve identified all the big issues with car-sharing, such as car-sharers facing the marginal cost of operating the cars, but I think you underestimate the cost savings that car sharers will see. There may be more costs than you realize to own a resting car.
A lot of parts on a car wear out because of time rather than use. Many parts need to be serviced on a shorter of time or mileage basis. Even if you don’t drive it, tires age (the current recommendation is to replace them every five years) as do rubber belts, seals, brake lines, brake fluid, and coolant. Some parts, like exhausts, catalytic converters, and engine sensors seem to wear out in fewer miles when cars are used infrequently. The car needs to be either garaged or washed from time to time or it collects dust and grime. Part of your insurance cost reflects the risk that a car will be stolen, vandalized, broken-into, flooded, hailed on, or crushed by a falling tree even when it’s parked. 100% of those extra costs are paid by an individual owner each year but if five people could share the same car, they would each pay only 1/5 of those expenses.
Car sharers also reduce the finance costs. You assume that a car lasts one person ten years (right now, the average car in America is over 11 years old). If I finance a $30,000 car by myself over ten years at 5%, I pay $8,183 in interest, or an average of $818 per year. Yes, you can finance a car for fewer years, but then you are tying up more capital and facing an opportunity cost. One of the opportunity costs is that you are not lending that money at 5% to someone else who wants to buy a car. If five people can share the car and the financing, a $30,000 car over two years at the same rate will cost them a total of just $1587.40 in interest, or an average of just $158.74 per person per year in interest. If they can get better credit terms because of the shorter term and because five buyers are more creditworthy than one, they can save even more.
Furthermore, you seem to think all the cars will wear out at the same mileage. A lot of cars are scrapped because they are no longer desirable enough to justify repairs when they are old and obsolete. A two-year old car with 200,000 miles will not be as outdated as an individually-owned car that is ten years old with 200,000 miles, so if both cars face a $2,000 repair, the two-year old car is probably worth fixing but the ten-year old car might get scrapped. This means the shared self-driving car will probably have a much longer life in miles than the individually-owned car.
Because I expect the state-of-the-art in self-driving cars to advance more rapidly than manually-driven cars, obsolescence is an even greater factor. Thus, it’s even more likely that people would be willing to pay a small premium (if they must) to share a cutting-edge car rather than save a little but be stuck with a 10-12 year old dinosaur. Perhaps cars could be updated to newer tech over time to stretch their useful lives, but even this means that shared self-driving cars make more sense – there are more people to share the cost of the updates.
Guaranteeing that I have a car at times of peak demand might mean that I want to own my owns self-driving car, but many more people might be willing share and take the occasional risk that they will either have to pay much more to drive home at Thanksgiving or carpool with other people. Self-driving cars make car-pooling much better because no one has to worry about getting a ride with an unsafe driver. Self-driving cars might even have things like separate passenger compartments to make car-pooling even more attractive.
I think the city vs. suburb divide is very important here. Today, you can live in a city easily without owning a car, and just “renting” one on-demand for the occasions you need one (whether that renting is through Uber, Zipcar, a normal cab, or Hertz). This is already a reality today in many cities, and if at some point the cars you “rent” are self-driving, nothing is really substantively different than the situation now (except a bunch of drivers are out of work).
The situation in the suburbs or rural areas is quite different. No one “needs” a garage today, you can just park your car in the driveway. But people like garages because they can load things into the car while being protected from the weather. You don’t need a driveway either, you can park your car on or near the street and carry your groceries into the house by walking across the front lawn. Yet people still spend the money to build garages and driveways. Why should the existence of self-driving cars change that behavior?
The wait times would not be guaranteed to be negligible at times when demand is high (e.g. 5:00 PM on a Friday evening). Either you’d need to put up with high wait times at times of high demand, or you’d need to have enough cars in the system to handle the high demand, and have a huge excess at other times. The latter is basically the situation we have now, and if the economics are equivalent as I argued above, why wouldn’t you want to own a car?
This is certainly true, but newer cars are pretty reliable and come with warranties anyway (and if you can send your self-driving car to the dealer by itself to get serviced, staying on top of regular maintenance is less of a hassle). Personally, I would rather own a car and deal with the chance of having to fix an unexpected problem once or twice a year, then have to worry about “surge pricing” when trying to get a ride home from work every day.
Yeah, but self-driving cars are not going to magically make that possible. The availability of certain vehicle types will still be limited, because the fundamental economics of driving (and car production) won’t change.
If people are mostly happy with a small econobox during the week, but then everyone wants to rent a minivan to bring the family to the beach on a holiday weekend, or everyone wants to rent the 4WD SUVs because your town just got 12" of snow, then either those vehicles won’t be available in a timely manner at those times, or you’ll have to pay extra for them. Self-driving cars don’t magically change the principles of supply and demand. If guaranteed availability of those vehicles is important to you, you’re better off buying one - even if that means you don’t use it’s capabilities the rest of the time.
All it will take for most people is one weekend where they can’t do what they want because the desired “rental” self-driving car is not available or costs 3x what they planned, and they’ll say “Honey, why don’t we just buy our own minivan and not have to worry about this?” Again, the economics are not that different.
Sure, if it really is completely filled with vomit. If it’s just a bit icky and dusty, are you really going to bother? Or are you just going to put up with it, like people put up with
Why would there be a flow of unoccupied cars? They’re just going to drive around for no reason waiting for passengers? That is a tremendous waste of gas.
Also, if you’re shopping at multiple shops, you might want to actually store things in your car, rather than carry them around with you for the entire day.
Because driving across town and then back again every day is a waste.
I know that’s your idea, but I’m not buying that it will be nearly as common as you think. I think people will still end up buying cars to guarantee their availability and features, these cars will still be underutilized, and they will still need to park somewhere and wait for their owners.
I understand that a nice ride through the country is a good thing, however, very little of my driving is like that. Most of it is going from stop sign to stop sign or light to light in a traffic jam.
I would like to have the option of letting the car do this.
The worst thing is when you have “good” drivers who have nowhere to go but are in a hurry to get there.
It’s impractical for the cars to drive clear across town to park, or all the way back home… but how about a parking garage, say, half a mile away from the office? That’s far enough that most people aren’t going to want to walk there and back every day, but still near enough that it’s not going to be a big deal for the car to drive that far. And it would still be enough to allow for much greater efficiency in use of land downtown.
Yes, I would like that option too. What I disagree with is the conjecture that it will be the only option.
Maybe I have a different perspective on things to other city commuters. I work in the city but I live on the outskirts of a small country town. A lot of my driving is in the country. When I drive somewhere with my partner 99% of the time I choose to drive (if it is in my car). Why? Because I like it. Now if I’m declining to have a person I know and trust drive me somewhere, why would I want to have a machine do it? I’m not a luddite, I look forward to all cars having some kind of adaptive cruise control so that everyone on the freeway drives at the same speed and you don’t get wallies slowing down up a hill and speeding up going down it, or passing you then pulling into your lane and subtly slowing down so that you either have to pass them or drive slower than you want to. Removing the human from driving would be good in many ways but I will always want the option of driving myself, whether it be in city traffic or on a country road.
People who drive now must already have some kind of parking option. Surely you’d still use the same car parking you do now? If I owned an SFD I don’t think I’d want it driving around the city without me. What happens if it breaks down or has a crash? What if it hits a pedestrian through no fault of its own?
Or directly to the morgue, as appropriate. It could also put out an emergency broadcast so that all other self-driving cars pull aside and it let it through.
The OP asked about the future of self-driving cars (SDCs) however many of the responses are talking about driver-less cars (DLCs). While the a future of DLCs would necessarily involve a period of SDCs, I believe the two technologies are so far removed that the present interest and advances in SDCs in no way indicates a future for DLCs as there are some significant technological leaps separating them.
Redundancy
A self-driving car requires no more redundancy than currently exists in an automobile. If the self-driving functions fail for some reason the driver just takes over, same as they do now if the self-parking or cruise control doesn’t work properly. The driver can then drive the car until they feel the need to have the problem fixed.
A driver-less car would need enough redundancy so that there were no single points of failure and if there was a failure such that there was a loss of redundancy, the driver-less car would be grounded (garaged?) until it could be fixed. Depending on the failure it may have to be garaged immediately, i.e., it pulls over and stops at a designated emergency park and calls for a driver-less tow truck. At best it would become a self-driving car where the redundancy role is filled by the driver, assuming it has a “control seat”.
Dealing with unexpected and unforeseen circumstances.
A self-driving car has a relatively simple task. It must follow a set of typically clearly marked roadways following rules that govern right of way from point A to point B. It needs some sort of collision detection and avoidance system for routine problems such as another car not following the rules or a pedestrian stepping out on to the road. It also needs some kind of simple “moral” system where it can choose between killing a person or a cat for instance. If all else fails it stops safely and says “your controls!” to the actual driver or the driver just takes over on their own if they see something developing that car won’t be able to handle.
A driver-less car needs to do those simple things but it also needs to do everything else that a human would have taken over for. Importantly it not only has to be able to deal with foreseen circumstances but it must also be able to deal with unforeseen circumstances. What circumstances would they be? I hear you ask. Well if I knew they wouldn’t be unforeseen anymore would they? That’s the thing, it would have to be able to deal with a problem that nobody has programmed it for because it will inevitably come across such a thing and I think it would happen very quickly. So, my opinion is that a driverless car is not possible without true artificial intelligence, and as I understand it, AI poses some significant challenges.
There are also very simple things that I can do with a car that a driver-less car can’t do. When I walk from the shopping centre to the carpark, I can casually survey the parking area and note any obstacles such as sleeping homeless people, shopping trolleys, etc and choose a route out of the carpark that avoids those things. The driver-less car can’t get out of the car to see beyond the limit of its immediate environment.
No, we’re not talking about a driverless car. We’re talking about a car where the driver is a computer. Adding the “redundancy” of allowing a human to take over control of the car would be a huge safety risk. Would you want to drive a car where a passenger could override your driving?
Scenario one does not result in an extra vehicle, it just allows the “driver” to be doing something else. I mean, somebody has to go pick up the kids, does it matter if it is Mom or just the car?
And self-driving cars will spend less time circling looking for parking than humans do. Because the reason we don’t park 3 miles from where we are going is that the driver would then have to walk 3 miles.
Even if you haven’t switched over to some kind of car-sharing/taxi service (that will be MUCH cheaper when they don’t have to pay a human driver), your car can drop you off right outside your office building, then go park itself wherever. And if parking in the city is $10 a day, and operating the car costs $0.25 per mile, you start doing a cost/benefit analysis on how far away from your office you can park.
So you buy a monthly pass at a parking garage 10 miles from your office, because at $50 a month it works out to be cheaper than parking downtown, even after paying for the drive. The only real downside is that if you leave work early, it may take 20 minutes for your car to get to you. In anything short of a building fire, it will probably take you most of that time to pack up to go home from the time you decide you are leaving.
The only people who can use Uber to replace car ownership either travel a remarkably small amount, or can make productive use of their time in a car if they are not driving.
The problem with Uber, or taxis or whatever, is the need to pay a human driver.
I read an article where the author tried Uber for a week, and pronounced it cheaper and better, but what he was comparing it to was renting a car.
Almost nobody who owns a car is paying $350 a week, even counting all the hidden costs. Of course hiring a car for the few times a day you are moving from one place to another is cheaper than renting a car for the week you are in LA.
Don’t get me wrong: I think it will be a LONG time before privately owned vehicles become scarce. And a similarly long time before cars without an autopilot become the minority. But I think that once “hired cars” get the huge drop in cost that comes with eliminating the human driver, their use will grow enormously,
I live in a city with a lot of farmland (yes, within the city: our biggest industry is growing soybeans) and lots of folks here hunt every weekend/
Most of the 4WDs I see belong to people who will freak out if their ride gets muddy.
I have to agree; the vast majority of SUVs are sold to people who want the seating space, not the off-road capability.
A recent study of Virginia Beach, Virginia found that half of the most valuable land in the city (an area about a block-and-a-half wide and a bit over a mile long, bounded on the East by the Atlantic Ocean) is taken up with parking.
It would be a great boon to the city if all of that parking could be moved to cheaper land a few miles away, allowing buildings to be built where parking lots now stand.
We all do.
Fun fact: in Virginia, if your hand is on the steering wheel, you are legally driving.
Which is how an ex-roommate of mine who caused an accident got charged with DUI and driving without a license: he grabbed the steering wheel (he said he thought he saw an oncoming car) and crashed into a guard rail.
But yes, there are 2 main camps on this self-driving software:
One says it should be “Driver assist”, with the human ultimately responsible.
The other views the human occupants as unpredictable variables, and giving them access to vehicle controls as the #1 potential cause of an accident.
Renting also includes the cost of “down time”. That is, the time that the rented item is not being used must also be paid for. If you expect U-Haul to have a 24-foot truck ready whenever you decide that you need one, they are going to have to have a lot of trucks just sitting in parking lots waiting to get rented. Those trucks still need to have their finance payments made, they still depreciate, they still have insurance. So when you rent one, you are paying for all the time it sat waiting too.
So when does renting make sense? When the amount of time you need the object for is small. It may be possible to own a brand-new pickup for a few dollars a day, but if you only need it twice a year (when taking your boat out of the water, and when putting it back in), then spending $20 a day to rent one instead of owning is a good deal.
Some of depreciation is based on mileage, but it is also based on time. Doubling the miles put on a car will make it depreciate faster, but not twice as fast. So sharing your car with another person will actually save you both money.
#1 is actually 2 separate things: your car is always available, and it is exclusive (so you can leave important stuff in the trunk).
The first part is solved with scale. The reason car-sharing hasn’t really caught on is that I could only use such a service if someone who lives fairly near me would finish using their car before I leave for work and they (or someone else) wouldn’t need it until after I got back. And in an ideal world, someone who is near where I work needs a car shortly after I arrive at work, and someone else drops a car off near my work just before I leave.
If the entire city were magically put on car sharing, it would probably work okay. But until you reach critical mass, there just aren’t enough sharable cars around to get people from where they are to where they need to go.
But if I could have car arrive at my house 10 minutes after I asked for it, then that problem is solved. (yes, taxis do that already. And an 8-mile ride in a taxi costs $25. Not really comparable.)
The second part is definitely a perk of owing versus hiring. Each person will need to decide what that is worth to them.
2) and 3) (what I said was basically said by another poster and responded to above, so I have nothing new to add.)
This is a big point in your favor. While the fact is most cabs I have been in were cleaned with remarkable regularity (far better than any car I owned), that relies on the human driver noticing that a passenger just puked on the floor. Or left their keys. Or whatever.
But at worst you are looking at a 10-minute delay for a filthy car: just click the buttons to indicate that this car is unacceptable and you need another one.
(BTW: the only cabs I have seen where all the surfaces were plastic were converted police cruisers. Most cabs are minivans or Town Cars, and the interiors are quite nice.)
Carpooling: I agree with you that people overestimate the impact, but I think you underestimate it.
People don’t reject carpooling because it costs them time, they reject it because it costs them flexability, but they are generally willing to trade flexibility for time.
See, the problem with carpooling is that somebody has to be the driver. Even if you take turns, somebody has to be the driver. And from the perspective of the driver, carpooling is a pretty bad deal. Sure, other folks chip in for gas, but you become a slave to their schedule, having to be at their house at a specific time every day and such. If you are five minutes late leaving the house, that’s 3 other people you made late to work.
But from the view of a passenger, what you are being offered is the free use of the time you spend commuting. A teacher could spend the ride grading papers, or reading the newspaper, or streaming a favorite show over wifi. The passenger is being offered freedom from needing to drive.
And lots of folks think that is a wonderful thing.
As you point out, self-driving cars offer that to everyone. What carpooling then offers is a trade of flexability for sharing hte cost.
I have heard lots of anecdotes of someone who decided to use Uber to commute so that they could use the time to answer emails and such. Almost invariably, they shortly found a few coworkers who lived nearby and now share the ride. People who have decided to pay somebody else to drive have no problem spending the time and loss of flexibility to share the ride (and the cost) with someone else.
The people who hate carpooling are the people who are weighing driving themselves versus driving a bunch of other people too, and see the upside as too small.
So yeah, I think once people get used to being driven instead of driving, a lot of them will be happy to share the ride.
I’m a sci-fi fan, and there is a special place in my heart for the stories from the 1950s where folks are traveling the solar system in a rocket-propelled motor home, but are double checking calculations with a slide rule. Or Arthur C. Clarke’s communications satellite design, where he predicted the ability to build a man-made satellite, but not to built radio equipment that could function without a human to maintain it. Or stuff from the 80s, predicting phones we could wear on our wrist that were nothing but phones. Or the stuff from the 90s predicting that the internet would turn us all into Aristotle.
Point is, we’ve got a pretty terrible record on predicting the future: we tend to miss major implications of advancing technology, and some of what we see looming on the horizon is further away than we thought.
Still, I’ll take a swing at the ball that MegaNazz has pitched at me.
First, I’d like to point out that, as CGP Grey points out in his YouTube video “Humans Need Not Apply”, calling them “self-driving-cars” can limit how we think of their use, much as calling cars “horseless carriages” would. He calls them “Autos, the automated solution to the moving people or things from one place to another problem”. Whether it is long-haul trucks, things that ferry containers around ports, stacks of goods moving themselves around in a warehouse, or how the kids get to school, they are all part of the same tech.
I usually talk about what they mean to cars, so I’d like to come at it from a different angle today: TRANSIT
I used to follow a few blogs from transit designers, and two principles they all seemed to agree on were:
Most people will walk up to a quarter-mile (about 400 meters) to get to (or from) a transit stop. Most people will not walk further. So a transit stop only serves a circle around it with a radius of a quarter mile.
People will use transit for unplanned trips if a vehicle services each stop ever 15 minutes or less, and will not if the interval is longer.
One of the major problems with designing a transit system is the balance between service and demand. If you have a route where the buses run once an hour, would you attract more riders by running more often? How many new riders?
On the one hand, you won’t attract many riders on even your busiest corridor if the system doesn’t go anywhere else, on the other it is hard to justify running service along a corridor where there aren’t enough riders to cover costs.
You are going to attract the most possible riders if your system goes everywhere, and goes there often. But that is expensive. How many riders do you lose if you don’t go to the ballpark? How many if you don’t go to the college? How many if you only serve the mall every half hour, instead of every 10 minutes? How much money will that save you?
The problem is, because of #2 above, it often would be best to serve an area more often with a smaller vehicle: if only 10 riders per hour come from a neighborhood, it could be serviced by a van every 15 minutes instead of a bus every hours.
But smaller vehicles aren’t much cheaper to operate, because they still need a driver
For the agency that operates bus service in most of Hampton Roads, payroll is more than half their budget.
If they were suddenly given a fleet of self-driving buses, they could fire all their drivers. That would free up enough budget to nearly double their number of buses.
Or, it could give them the chance to invest in smaller vehicles that do the same job. And those vehicles would be cheaper to operate than a bus (although likely more expensive on a per-passenger-capacity basis), which would mean that some routes would now be economically viable that weren’t before.
And serving more areas means more riders on the system as a whole.
People who didn’t used to take the bus anywhere because it didn’t come close enough to their home or work now will, and will take it other places as well.
And the “bus” can now run more frequently. The one bus an hour can be replaced by 4 vans, and the more frequent service will mean more riders as people view taking the bus to the store for some cinnamon a lot better than driving there and then parking.
So just one change I see coming as a result of Autos is a significant improvement in mass transit in urban and suburban areas of the country, as operating buses gets much cheaper. This will allow service to more areas and greater frequency in existing areas, which will make transit a more attractive option on the list of ways to get around.
Another advantage of SDC is that they could park like the proverbial “leave your keys” lot, 4 to 5 deep or worse. When your car is called, it simply works its way out like one of those tile shuffle rpoblems, the cars in its way move out of the way, let it out, then pack back in. For slight increase in gas consumption, you can fit more than 50% more cars in the same space. And, as mentioned, it does not need to be that convenient to the drop off / pick up point.
I imagine almost every car (at first) would simply have an “autopilot”. To take over control, you would have to do something to disengage and prove that you have control - tap the brake, or shake the wheel then push the button, perhaps.
The financial calculation depends - many people living in a downtown have almost no need of a car. They would be fine with the occasional rental. As many car rental agnecies and taxi companies show, servicing multiple people serially with one piece of equipment can spread the cost. The downside is when the demand exceeds supply. Perhaps there will be an “Uber Plus” where for a premium price, you guarantee quicker response in busy times. Or as I mentioned earlier, some for of “Budget Uber” where if you agree to ride-share you get a discount, for a slight inconvenience of side trips. It’s the people who live in the distant suburbs, who have hour-long commutes, who probably need to own personal transport; but if there were a commuter service like a train, then a service to pick you up and another at the other end to take you the rest of the way would be handy. Unfortunaely, a lot of jobs now are themselves in remote or suburban areas,. Perhaps someday a commute service will be a job perk, harening back to the days of company chauffeurs and cars - but for a much lower echelon of worker, thanks to the cost. (Or for the bottom of the ladder, company-provided ride sharing…)
As for cleanliness - the same applies nowadays to those car-sharing companies. You find a messy car, you report it, the management sorts it out. Between video surveillance possibilities and use tracking, they can figure out who will pay.
Incidentally, video surveillance also solves the “unattended” problem. Mom can send the car out, be alerted when it’s occupied and face time with Junior all the way home if need be.
Another massive important point - SDC’s can drive off to local charging stations during daytime slow times, so the electric vehicle no longer has to have the all-day range; all the more so if it is a service rather than your own private car.