First, if my car is self driving, the experience of using a self driving service is nearly identical to the experience of using my own car. I’m not interacting with a stranger, making small talk, listening to their radio, or experiencing their preferred cabin temperature. It’s me, in the car, alone, just like in my own car.
Second, today’s car sharing services require convincing a person to spend their day driving other people around. With self drive, you only need to convince a person to block off some time in their day for their car to go out and earn money for them. It’s like SETI@Home, but with a car earning money for you instead of a computer looking for aliens. It’s a far easier case to make than for being an Uber driver.
Automated car sharing can increase the utilization of the average automobile, reducing the total cost of ownership, again, it’s a good business case vs. individual ownership.
Early adopters would be those who simply don’t use their cars all that much, their costs would reduce significantly, while their inconvenience increases little, if at all. After that, it’s just a matter of equilibrium, how low cost can the car sharing services get, and how does that compare against individual ownership.
Most of us don’t have a car-sharing lot within walking distance from home. And you still have to park it at the destination, and pay for the car usage the whole time you are at the destination (or drive it to another lot and then walk to the destination).
Quoth a boss of mine:
“I used to think rear cameras in cars were silly. Then I got a rental which had one!”
Anything that makes driving easier is likely to be considered desirable by a large amount of the market. Automatic shifts have been an exception in a lot of the world because they went against a key point of our driving training; people from majority-manual markets tend to get really nervous when the motor goes to very low revs, as our trained instincts say “it’s about to choke!”
And well, anything which can help stop certain humanoids from driving is a good idea. Like that boss I had (not the previous one) who drove us to a restaurant for lunch once and when it was time to go back we took the keys from him, refusing to ever let him drive any of us again. Dude’s concept of “driving” was almost as terrifying as his sense of entitlement (he was using his wife’s car because his own was in the garage after an accident and his company car had been impounded by the police for “driving in a way the cops found scary”).
Keep in mind that “car sharing service” refers to two different business models. There’s the Uber/Lyft model, which isn’t really car sharing but more like a smartphone-enabled taxi service. And there’s the Zipcar model, which also uses a smartphone but to let people rent cars quickly and easily but drive themselves around.
The count of cars in my vicinity is mostly based on one car per driver. The idea seems to be to have a much smaller number of cars available, and share them. That means increased wait times almost inevitably.
The government has decided it isn’t cost-effective to set up public transport to get me from my house to my work, even with buses that carry lots of people on a cheaper per-rider basis. Apparently having a lot of driverless cars instead of one bus is going to make that cost-effective. I don’t see how that follows.
Then those 2-3 other people have to wait for up to 2 hours, don’t they? Time is a non-zero cost.
Apparently I need to hope there is a driverless car less than three blocks away.
WADR this is saying we don’t need parking lots - just street parking. If the number of driverless cars is much lower than private ownership causes now, that might work. But then, again, the number of driverless cars is going to be much lower, and therefore it is going to be the case that the number of *available *driverless cars is going to be lower.
Maybe that’s something I will have to put up with, because the cost of owning my own car is so high compared to driverless Lyft. But if the cost is anything close to equal, I am going to buy my own car.
Yes, if you don’t make any effort to manage your inventory.
Mass transit requires people to walk significant distances, or change mass transit vehicles to get to their destination. People don’t like doing that, so mass transit is viable only where the population is dense enough to support it. *
Conceptually, if a car was going to ferry 3 passengers home over a 3 hour long “rush hour”, the first passenger would be someone who leaves work early, the second passenger a person who leaves a bit later, and the third passenger a person who leaves rather late. That way, nobody has to wait 2 hours to be picked up.
Seriously, if the argument is that ride sharing companies will be run by mentally deficient orangutans, then yes, they will be a dismal failure. The successful ride share will manage its inventory to balance out utilization and customer wait time, just like every successful service company in the last 100 years.
*I used to live in NYC. If my normal way around the city was having a car pick me up at my front door and deliver me to my destination, I wouldn’t have taken the subway or bus for anything. Mass transit sucks, you have to walk 10 minutes to the stop, wait 5-10 minutes for the thing to show up, have it stop 6 times on the way to your destination, maybe change trains, and STILL walk another 5-10 minutes to get where you’re going.
Define exact. It can know what road you’re on but it’s not good enough to know what lane you’re in, or where you are in that lane (left side vs. right side); additionally GPS frequently go off track. It’s well known in the running & cycling communites (ie. people who use Garmin watches) that it doesn’t always track you perfectly down the path; no, I didn’t cut that corner or run thru the edge of the pond, etc. This is especially true in cities where tall buildings play havoc with the GPS signal getting to the receiver on the ground.
Where I am is currently under a flood watch. I can’t say with certainty that Road X, which frequently floods will flood this time, deep enough that it’s not driveable, & if so from exactly what time to what time.
Under normal conditions there’s white & yellow lines on the road. When it’s snowed a few inches I can make out where the curbs are because the snow it a bit higher there. However, lots of roads around here don’t have a curb. There may or may not be a shoulder to the road, & there may or may not be a guardrail, fence, hill, mailboxes or trees alongside the road. Even when I can’t ‘see’ the road, I use all of these clues to keep me driving on the road. I don’t believe AI is there yet.
Also, what about flash flooding? I’ve come up to roads when I noticed that I could see less & less of the guardrail up ahead of me; my brain told me the road was flooded; fairly deep. I stopped & turned around. Can the AI sensors handle that yet? I can also look at the ground & know if I can turn around on that two lane street. Is the shoulder paved or just dirt/grass? If it’s rained enough to create flooding, I’m not putting my tires into the soft stuff but if it’s a paved shoulder, I can drive a bit past the white edge line to make my U-turn.
I’m not arguing against self-driving cars. I think they’d be awesome, and I look forward to having one.
What I’m saying is that I’m not at all convinced that a switch to self-driving cars and a car-sharing arrangement is inevitable for the majority of drivers.
Personally, I think that the envelope of those for who convenience meets cheap will expand, but I don’t see people giving up owning their own vehicles within any kind of reasonable time frame. IF that ever happens, what’ll drive that is increasing urbanization and population density- basically the things that make taxis and Ubers very attractive in highly dense cities like NYC and San Francisco.
But there’s nothing inherent to self-driving cars or car-sharing services that make this inevitable, and that’s what I’m arguing.
Yep. All of your post really. I have very detailed knowledge of the road that I have lived on for 26 years. A road where I’m the only full time resident. So a self driving car is going to have limited networking resources with other cars to know about the road (if that ever becomes a reality). Some times, in deeper snow, I will drive past my driveway and turn around at the end of the road. This gives me a straighter shot and speed/momentum to make it up my driveway instead of making the tight turn off the road at the bottom. This is all based on the tons of knowledge/experience I have and current input I’m getting. Will I be able to tell the self driving car to make such a move?
A self-driving car may not need to make that maneuver. It might be capable of making it up your driveway without doing so. And even if you’re the only full-time resident of your road, it should be able to build a detailed 3D map of the road and curbs over multiple trips.
Now I may be overestimating how smart these cars can be, but I assume they will be designed to learn from experience.
My point is that a car service, driverless or not, is going to have to show me that they can manage their inventory such that either I don’t have significantly more wait time and inconvenience than owning my own car, driverless or not, or that the cost-benefit analysis of the car service shows that it is significantly better. And I haven’t seen how making every Uber car driverless will accomplish this.
This assumes the 3 passengers want to get home at different times. I don’t think you can assume this automatically. This is like car pooling - a great idea in theory, as long as it is better overall.
No, I don’t think it will be run by orangutans. OTOH, if inventory can be managed such that customers prefer it to owning their own car, and/or costs less, then Uber and Lyft will show us how it can be done, and out-compete private car ownership all by themselves.
I’ve used Uber, and I have ridden taxis, mostly in NYC while traveling. And it was an acceptable taxi service. Maybe making it driverless will make it even better. But so far, as a replacement for owning your own car, outside of some special circumstances, it costs twice as much and you still have to wait.
90%+ of my driving is [ul][li]Get into the car to go to work. AFAIK I am the only person in my neighborhood who works where I work. As a matter of fact, AFAIK I am the only one in my city who works where I work. I leave between 6:00 and 6:45am.[/li][li]Drive to work.[/li][li]Park in the huge parking lot. [/li][li]Work.[/li][li]Occasionally I go out to lunch. There are lots of restaurants within a few miles. There are none within walking distance, and I usually take 45 minutes for lunch. Assume I have to wait 10 minutes for a car either way. That’s kind of a short lunch break - I might as well eat in the cafeteria, and I don’t like their food, which is expensive.[/li][li]Leave work to go home. Usually it is about 4:00pm. Sometimes it is 6:00pm. Once or twice it was 4:00am the next morning.[/li][li]Drive home.[/ul]On weekends, I go to church, and I am the only person in my suburb who goes to my church. I also do the grocery shopping. That might work - lots of people in my suburb shop where I shop. Or my wife and I go out, or we go to see friends, or we host dinner parties. [/li]
So that is the kind of scheduling/inventory management that a driverless Uber has to match, either on cost, on convenience, or both.
Look, I get it. One driver per car is inefficient, in terms of fuel and congestion and so forth. Maybe I should be willing to sacrifice for the environment. But I probably won’t unless I have to, and I doubt that I am the only one who feels that way.
‘Might’. And if it can’t and needs to run down to the end of the road so it can have more momentum?
I highly doubt a self driving car will ever have the capabilities a human will have to handle bad weather conditions. The head of Google’s self driving car division agrees with me.
(can’t seem to copy the link from the previous page)
I’ll buy this. You would need to have the ability to tell it that for this area, you will need higher resolution. Understand that this is a poorly maintained mountain gravel road. No signs no nothing.
And I’m sure GPS will improve with time, but for a moving vehicle, it’s not there yet.
With all the talk of the ‘advantages’ of not having to own and maintain a car, but rent or ride share instead, I am wondering what the suburban homes and streets of neighborhoods might look like without every home needing a garage or carport.
In some places this might look just like a townhome on a car-lined street of parked cars, but in suburban single-family residences, what would this look like? Would some homes have garages and driveways and others not? Would neighborhoods all be required to have a small one-car garage simply for junk storage or car storage (whichever you choose)?
Sure, let’s take advantage of you 1% of all drivers with your need for status, for ego, for being the coolest nerdy first adopter on your block! Let’s let you pay the price economically for the development of this technology by paying high prices for these vehicles, and life and limb for the inevitable "oopsies"https://en.wikipedia.org/wiki/List_of_self-driving_car_fatalities that must come with any leading edge technology. But with your selfless dedication, we should be able to eliminate a particularly irritating speed-bump on the road to fully efficient capitalism. Right now we’re stuck doing it the old-fashion way, pushing people to their human limits- to the point we have a name for it- ADS (Abused Driver Syndrome) (I’d link to something on YouTube about it, but its been removed). You’ve probably seen stuff on the news about drivers at port facilities being asked to drive an ungodly amount of hours.
The point is, the real “demand” isn’t for self-driving cars. It is for self-driving semi-tractor trailers. I don’t know how exactly what percentage of all freight goes through trucking, but I do know it is huge. And I do know the “human element” (e.g., “truckers” is the greatest single expense. So sit back, relax, and imagine that 80,000 lbs of live pigs one lane over being driven by the same kind of software that lost all your tax records for the past 3 years, or that decided to delete your photos of your children from age 1 to age 6. Or that database with all your personal information that was hacked by a 15 yr old. Of course, there’s NO chance anyone might hack self-driving tractor/trailers to have a bit of fun. That sort of thing just doesn’t happen. Not in this day and age- we have our best people on that. But there’s no mystery where the the demand is coming from, and that demand will be satisfied. It’s just that when that much steel and rubber moving at those speeds makes a boo-boo, really bad things happen.
Merry Christmas!
They would only need to be in a radius where they could drive to you faster than you walk three blocks.
Theoretically, there would only need to be as many driverless cars as there are cars on the road at a single time, which would be a massive decrease in cars total. In practice, there will likely be cars above that because distribution would have to be perfect for the theoretical case, but I would be surprised if you couldn’t get perfect coverage at much less than 1 car per current driver.
True. As long as one is within that radius, and available, it would work.
It depends on what you mean by “perfect coverage”. If by “perfect”, you mean the average trip time, door to door, is the same or less than the average trip time for a self-owned car, I have my doubts.
The advantage of self-owning a car vs. a car service is the same as its disadvantage - most of the time, the car is parked somewhere waiting to fulfill one specific set of requirements. Any time the car is out fulfilling some other specific set of requirements, it is mostly unavailable to fulfill my set of requirements. I then have to wait until it is available. Making sure it is only available to me means it isn’t available to anyone else. Making it available to anyone else means it wouldn’t be available to me. Yes, that’s inefficient, and it comes at a cost - most of the time, my car is parked somewhere not driving anybody anywhere. It maximizes the utility of the car to me, and minimizes it to everybody else. Making it available to everyone else increases its utility to everyone except me. I already have 100% access. Extending that access to anyone else reduces the access that I already have.
Currently, it’s worth it to me to have a car that mostly just sits there waiting for me to drive it. It is certainly theoretically possible to have a system where the overall cost, both in money and in wait time, makes it worthwhile for me to use a driverless Uber all or most of the time. But I haven’t seen it.
That’s the point I was making about the bus. Currently it isn’t efficient to set up a bus to run from my front door to where I work, or even to within a mile of where I work. And a lot more people can ride a bus at one time, so the cost per rider is lower. I don’t see how a car that can hold four people can be managed to run efficiently when a bus that holds fifty can’t.
As I said, I would like to own a driverless car. I hate driving. I would love to be able to press a button, have my driverless car pick me up, tell it “go home” and then sit and surf the 'Net or read a book while it takes me home safely. Hell, I would install a wet bar in the back seat and start my weekend early. And I can even envision it where I own the car, but lease it out to Lyft while I am at work and accept that I will have to wait if I have to leave work unexpectedly.
If a car service can be made to work, then Uber and Lyft will be able to bring it off. Let them get a bunch of driverless cars, and prove they can get me door to door cheaper, faster or as fast, and safer than I can do myself.
Never? It may be decades away but I can’t believe that. Humans aren’t all that great at it anyways, and the problem is almost everyone thinks they are. Accident rates can easily double with even a modest snow storm.