This is to go along with threads like this one here. With all the talk of whether or not self-driving cars are “possible,” I never hear anyone talk about who really WANTS one. It’s always treated as some inevitability, as if just because such are are available everyone will just go get one.
I find that hard to believe. Cars and driving have been treated as symbols of status, freedom, and power almost since their invention. They’ve been marketed as such for decades. Are we now supposed to believe that a majority of the public will toss aside such things that they have been trained to think/enjoy in favor of a glorified bus ride?
What kills me is when people say the real benefits of such cars will be when a “majority” of cars are driverless, because then they will act in a networked fashion. This is clueless engineer-think at its finest: If only we could take humans out of everything it would all be more efficient! But the world doesn’t exist to be efficient, and no one will be interested in networking their car for the “greater good” unless they see some immediate obvious and measurable personal benefit, especially when the security risks of such things become widely understood. I think driverless cars are the next version of Iridium satellite phones.
I don’t have anything to link to, but the various economic and tech podcasts I listen all point to companies (ride share, taxi, freight) as the driving (heh) force behind the market for autonomous vehicles, not consumers.
This is just my opinion of course - but imagine being able to take a very, very long driving trip and not having to be awake for all of it. I just drove to Vermont and back. It’s a (long) one-day trip, but at that point I was not able to drive at night, so I had to break it up into two days - which cost another hundred bucks each way in hotel and meals.
Imagine being stuck in nasty traffic, and being able to let the car handle at least part of the driving.
I imagine a self-driving car would do far better on the long-haul bits than a 5-lane highway during rush hour. but I would be thrilled with either sort of help.
If my car is self-driving, I assume I’d have the ability to control it when wanted or needed. Also, having my own car would still give me more freedom than if I had to rely on a cab (self-driving or otherwise).
Would cars actually be networked - taking data linked directly from other cars. or from a central networking service? Or would they be self-contained and make decisions based on their immediate surroundings?
America has an aging population that would love to be able to still travel by themselves even if they no longer have the capability to drive a car.
But on a bigger perspective, the American automobile industry pretty much invented the idea of marketing. They know how to invent a product and then create a demand for that product among consumers.
And it’s an easy sell. Self-driving cars won’t be sold as the equivalent of riding public transportation. Justifiably, because a car where you choose the route is not the equivalent of taking a bus. They’ll be sold as the equivalent of a chauffeur-driven limo. You’re the boss who tells the car where you want to go and it takes you there while you relax and enjoy the ride.
Despite the fact that some people love driving, vanishingly few people love the sorts of driving that most people spend the most time doing: slogging through a traffic-filled commute.
But the real reason that the demand will be huge is that self-driving cars will be vastly cheaper than personal vehicle ownership in most places. Personal vehicles are an extremely inefficient allocation of capital: most of them sit idle most of the time. But it’s worth it for the convenience of being able to go where you want when you want.
Taxis and other car services can provide that convenience, but at the additional cost of paying a driver. Mass transportation can provide lower cost because one driver can transport more people, but at the expense of routes and schedules that might not be convenient. Driverless fleet vehicles will provide the benefits of both.
I can’t point to market research, only provide a personal anecdote. My car does not have true self driving, just autosteer for lane holding and auto stop/go for speed control. It is a huge stress reducer in heavy traffic even though I still pay attention, and am constantly ready to take control if necessary. When I’m forced to drive my truck in traffic, it quickly becomes apparent how much better even the limited self driving is than full manual driving. As far as I know, everybody I’ve ever talked to who has said self driving cars are pointless, has never actually used one on a daily basis.
Anyway, the point of all of that is my personal opinion that the main job of marketing will be to get people over their resistance to change, and fear of giving up a bit of control. I think once a critical mass of people actually experience it, self driving cars will be accepted and embraced by most everybody else.
I’m really looking forward to the day when the car behind me is under computer control, because what I’m seeing in my rear view mirror is the top of a head due to something really interesting happening in the drivers lap.
My Wife’s new Subaru has autosteer for lane holding. Now I haven’t had much experience with it, but the SOP seems to be - Drift over towards one lane or the other, Correct, drift, correct, drift, correct.
Instead of driving down the middle as a person would.
I tried it on a long road trip. No thanks, it’s maddening. And frankly, more stressful than just driving myself.
I would argue that at least initially, it’ll be huge only among the taxi / limo industries.
The example you use, about how owning your own car causes it to sit idle most of the time, isn’t affected directly by the self-driving situation. If I own a car, it’s sitting around most of the time whether I’m driving it or it’s driving itself.
But you’re very right about the cost. I work from home - so I don’t have a daily commute, and there are days where I just need to pop out for the occasional errand, and other days where I don’t leave the house at all. If I pay 10 bucks for a Lyft to the grocery store, and 10 bucks back, and do that twice a week, that’s 40 dollars a week or 160/month. If I do any other errands, it adds up to more - but consider a car payment of 400 a month and insurance / taxes of another 100. That’s 500 a month.
More than casual use of Lyft would add up to more than that, especially with multiple household members.
On the other hand: right now, we have 4 adults and 2 cars in the household. Most of the time, that’s enough - there’s almost always one car in the garage. It takes a little coordination, AND the occasional Lyft. That’s certainly cheaper than a 3rd car.
Now, one car is going off to college soon, so we’ll be 1 car / 3 adults (one of whom works regularly some distance away, so he needs that car). At that point, Lyft (at current prices) would fairly quickly add up to more than the cost of a second car.
If Lyft / Uber were significantly cheaper (because self-driving), the equation tilts back, at least for some people. I
I would argue that a combination approach (ownership + use of car as a service) might become the most prevalent scenario, at least for a while. You have your own car available for the household to share, supplemented with a service when that car is not available. Some household adjustments would be needed - e.g. I can drop my husband off at the Metro station and he takes transit if I need the car, or I run my errands on evenings / weekends. Sure, for perfect convenience we’d each have our own car - but with one car generally available if needed, and the occasional Lyft, we can get by with owning fewer vehicles.
Cost is heavily affected by where you live: if I lived in a big city, the cost of keeping a car goes up a lot (it’s one of the many reasons I think someone having a car in NYC is clinically insane :D), and there is a lot more walkable or bus-able. whereas in the suburbs parking is free and you HAVE to have wheels.
I’d argue that Lyft / Uber won’t be that much cheaper with self-driving cars than the current situation. There’s the cost of purchasing and maintaining those cars - which will be more expensive than today’s regular cars. There’s the infrastructure to develop the communications: if I book on Lyft right now, a request goes through their servers and out to a driver… but the only communication directly with the car is through the driver’s hands and feet, and uses existing technology. There would need to be an interface to the car from the company’s servers to tell it where to go. You won’t be tipping the robodriver - but I’d bet the extra cost for the hardware/software to route the vehicle will be similar to what you would have tipped.
I’m assuming that these cars would be owned by the services. Another option might be nifty: let’s say I owned a car with self-driving capability. Could I go to work, then tell the service “OK - you can rent my car for a while”? Let it work for me while I’m working at my job? There’d need to be limits on what it could do - because I’d want it back in time for me to get home or whatever.
I know people in the industry who are working on this. When autonomous vehicles can talk to each other and/or a central database, they can react to things other than the vehicle immediately in front of them. They’ll know a mile ahead of time when they need to change lanes, start decelerating, or reroute completely. Cars will be able to safely pack in nose-to-tail at highway speed. This is dangerous with human drivers, who (at best) take a 0.5-1 seconds to react to what the vehicle in front of them does, but with cars sharing information beyond their visual range, this becomes a great way to fit more cars into less space on the road, mitigating traffic jams and dangerous traffic “waves”. Cars will make space for other cars to change lanes, instead of cutting them off or simply being oblivious to other vehicles’ needs. Cars will be able to safely draft each other for better fuel economy, and adjust their speed to best cooperate with existing traffic conditions (in their vicinity and well down the road) to get better fuel economy.
TL,DR: Level 4-5 autonomous vehicles will be extremely difficult to achieve, but they will have such immense benefits in terms of safety and productivity that we would be foolish not to advance technology in that direction. Vehicle interconnectivity can help autonomous vehicles deliver even better results than they would if they were operating completely independently.
First, I think there’s serious reason to doubt a large proportion of self driving cars any time in the future near enough to predict anything without looking foolish. For technical reasons. Cars which really driven themselves in all or nearly all conditions, safely enough to overcome the natural heavy bias to emphasize cases where they cause accidents (even if the counter-factual cases of avoiding accident are arguably much more numerous) might be several decades away. Whatever a thread here ‘concluded’ on that, I believe it’s at least doubtful to talk about the technical feasibility of cars being generally self driven even if that’s what people did want.
And it’s not two independent things, the timing of the technology and demand. If true SD cars were really going to arrive imminently, we could take that as a single ‘delta’ change in the way the economy and society work with all else equal, except changes flowing from SD cars. But in the real situation where it might be many years or some decades, all kinds of other stuff could change.
The suggestion in Holman Jenkins’ column in the WSJ Nov 30 is that maybe SD car boosters are underestimating how much tech will reduce people’s desire to be road mobile (eg. Netflix eating into movie theater revenue, Peapod/Amazon v supermarkets, as that trend gains momentum). It’s not IOW necessarily just why would people want SD v non-SD but why they’d want ways to get around on roads. And it could tilt the desire for personal cars more to driving enjoyment.
Personally I wouldn’t mind if a future car had a self driving feature I could really trust so I could stretch to get to home or next motel when sleepy, set the car to handle it if I doze off. But I mainly want a car to drive it. I’m older generation, could be that that’s generation specific, but also younger generations might just live in ways that greatly reduce demand for cars by the time SD cars are a 100% reality.
In one way, yes, I agree with you. The engineers and computer geeks are running the asylum for now. The same way that they go wild with excitement over every new version of Windows, and expect the public to join them. "Look…New tech!!! it has to be good!!! it’s new!!! "
Google’s engineer geeks made this mistake when they created Google Glass. A pair of glasses that was constantly turned on,filming everything you looked at,and displaying the internet in your eyes at the same time… It’s good tech, but it failed because it was too much, too soon, and just plain too creepy. They could have made a slimmed-down version, marketed to specific professions,(for example : pilots, surgeons) which would have been useful and practical. But instead, they tried to force it on the general public, when most people arent ready for it. But “hey, it’s new”.
And then Google did the same thing with their self-driving.They proudly removed the steering wheel, because, hey, it’s TECH!!! IT HAS TO BE GOOD!!!
But after the technical bugs get solved*, then the humans will take over again. Instead of the engineering department it will be the marketing department in charge.
And that’s when self-driving cars will take off. The marketers will realize that people want both options. A steering wheel for fun driving, and a self-driving option for the daily grind in heavy traffic, or for tedious long-distance drives. And there will be a huge demand for a car that’s fun to drive, but also lets you sleep on the way to work.
*(which will take decades. Right now, Waymo’s autonomous cars are a joke: they can’t even handle parking lots, and they can’t change lanes aggressively enough to go where you want. – they remain in the center lane, and pass by the ramp where you wanted to exit.)
Don’t forget about insurance costs. Once the insurance companies realize just how much better drivers the computers are, the insurance will go down for computer-driven cars and up for us meatbags.
And yes, there are some people who enjoy driving, for its own sake. And there are also a lot of us who don’t. For me, the value of a car is in getting from Point A to Point B, and the power, freedom, etc. from having one’s own car comes from being able to set the schedule, route, and so on. Actually sitting at the controls is, from my perspective (and that of a lot of other people) a necessary evil, to be put up with for the sake of the benefits. Get rid of that, while still keeping the freedom, and we’d be all over it.
this alone will be a huge market. My mother recently had to give up driving, and it’s having a large impact on her life. She can still do things, but everything is more difficult. She doesn’t want to have to ask someone for a ride every time she needs to go to the store, but at the same time, she doesn’t want to get taxis every time for a ride that would only pay $5-10. A self-driving car would be perfect for her.
And it’s not just the elderly. Anyone with a physical disability that makes it impossible to drive, but who are otherwise capable of living alone, will benefit.
And families will find them far more useful. Mama Zappa discusses this a bit, but consider: The car can drive itself, which means it doesn’t need a human in it to drive. Dad can take the car to the train station, and then tell it to drive home, so Mom can go to work. Mom could send the car to pick the kids up after school, because the kids won’t need to be able to drive to use the car. Imagine a car with parental controls, that kids can use by themselves to go to certain locations - school, baseball practice, the mall, whatever the parents deem proper.
It will also revolutionize long-distance travel. Most places in North America are within a two-day drive, if you can just drive continuously. Imagine a self-driving RV, you hop in the back, tell it to drive you to San Diego or New York, and then just hang out reading, watching TV or whatever. Two days later you’re there. No security lines at the airport, no paying $75 for a checked bag, no being crammed in next to a few hundred other people.
Add to this, the elimination of most offenses related to drinking and driving. Leave the bar, get in the back seat, tell the car to drive you home.
And that’s all in addition to eliminating the boring parts of driving as others have pointed out.
There will be more benefit when most cars are driverless, but there are plenty of benefits for small-scale adoption of driverless cars. It’s already clear that lots of people want to stare at smartphones on their way to work, for example.
I can understand people who like to be able to take an exciting drive around some curvy, scenic roads on the weekends. But I simply do not believe that people hold the first option above to be their highest priority for how they wish to commute, which is typically where we drive most of our miles. Seriously, if someone chooses number one, I’ll just think you are a liar.
Yes, this is getting lost here. Some of the older generation may still have their love affair with their cars, but that isn’t the experience of the younger generations. Cars are just getting from point A to point B, and Lyft or Uber can do the same thing (at some point owning a car is more cost effective than ride share apps, though). And younger folks are more into mass transit options as well. I think as the tech for self-driving cars gets better and better, you’ll see those under 40 crowd definitely embrace it.
I think you’re underestimating how much of the cost of a car service is the driver.
The IRS estimate for the cost of running a personal vehicle is around $0.50 a mile. That includes maintenance and fuel and capital depreciation. Multiplying this out by the lifetime of a vehicle, it seems reasonable. A sedan that costs $25,000 new and gets 30ish miles to the gallon has a useful service life of 100,000 miles, which means you’ll spend ~$12,000 on fuel in that time, and probably another $10,000 on maintenance.
But how much would it cost to take a taxi 100,000 miles? Right now, it would cost something like $250,000. The difference between the $50,000 capital cost and the
$250,000 retail cost is the labor cost of a human sitting in the car for hours a day, profit, and various rent-seeking regulatory things. Profit’s still going to exist. Rent seeking is probably still going to exist. But the labor cost goes away. At average-ish city-driving speeds, it takes 5,000 hours of work to drive that 100,000 miles. Really, more, because taxis are not 100% utilized. Some of that time the driver is wandering around looking for a fare, or deadheading back from a long drop-off. So it’s probably ~3 years of full-time labor.
Even if you assume that the capital cost of a self-driving car is twice that of a current car (it won’t be. I think it will actually be less pretty soon. Cars designed for fleet use can be much more efficient at the 90% of driving people actually do because they don’t have to handle the last 10%), there’s a lot of room for the price to drop.
How many people would just not own cars if cab rides were $0.75 a mile rather than $2.50?
The amount of money to cover the infrastructure to send routes to the cars will be a rounding error. A few cents per ride. Nowhere near the few dollars you’d tip a human driver.