Post-car culture: broader implications to a driverless society
January 15, 2016 12:45 PM   Subscribe

How driverless cars will change cities: a 9 minute video from CBC News' The National, covering Morgan Stanley's 2014 prediction for the end of car culture as we have known it (PDF) due to the convergence of driverless cars and "sharing culture" technology like Uber (who will be testing their system soon), and idea that people would be less inclined to own their own driverless vehicle, gien that personally owned passenger vehicles are parked 95% of the time, all of which will mean potentially significant changes in how urban spaces are planned and designed.

Other topics touched upon: driverless public buses that are currently operating around the world, and police forces would change, as more than 85% of vehicle stops are traffic-related. Or as boldly stated elsewhere, driverless cars will put half our cops out of work. There will also be an impact to fire and EMT services due to decreases in auto collisions, and driverless ambulances will mean changes in EMT services, too.

Not mentioned: how autonomous on-demand taxi-like vehicles would reduce urban congestion that results from people circling blocks looking for a place to park (that article includes some interesting video simulations of this issue), and if cars really are an on-demand service, there could be a significant drop in the sheer number of cars that are made and used.

Alternatively: if people still opt to own their own driverless car, congestion could get much, much worse, or at least the amount of miles driven would increase as cars are sent to and from home to park while idle. Also to consider: as congestion eases, the number of trips may also increase as people are less dissuaded from traveling on a congested route. And if autonomous vehicles are a more comfortable driving experience, people may take longer and more frequent trips.
posted by filthy light thief (232 comments total) 43 users marked this as a favorite
 
What's it all mean for transportation planning professionals?
posted by notyou at 12:55 PM on January 15, 2016 [2 favorites]


What I always wonder is, what are our cities going to look like with all the newly unemployed? No more bus drivers, no more Uber drivers, no more taxi drivers, far fewer cops (which, hooray in most respects), a dramatically different car sales and repair system, dramatically different gas station system - there's going to be a huge surge in unemployment, which means that competition for other jobs will get worse and wages will fall.

I'm not looking forward to this, since it seems likely to produce a car-free utopia for the usual rich people and more precarity for the rest of us.

I also wonder how the costs are going to shake out for really poor people who drive old beaters. Let's say you have a lousy job and live far away from it; there's no public transit so you drive a $2000 car that's fifteen years old, skimp on insurance and get your repairs from Joe down the street. Are you really going to be able to afford a driverless cab back and forth every day?
posted by Frowner at 12:55 PM on January 15, 2016 [41 favorites]


A couple of things I haven't heard anyone address regarding ownership:

1) A significant number of people use their car not just as a point-to-point people mover, but as a kind of mobile locker/office/station. So, sure, I see self-driving cars maybe eventually making up a big portion of work commuter traffic, but series-of-errands or out-with-the-family traffic? Still a skeptic at this point.

2) What happens when I want a barely itinerated trip to the desert? The mountains? Up the coast? Self-driving is nice, but unless the rental economics change significantly, can't see the appeal of owning for this reason going away, and my guess is this is more true the lower the population mass of a given area is.
posted by namespan at 1:07 PM on January 15, 2016 [9 favorites]


Can we stop with the driverless cars hype please. It just won't be happening at any scale.
posted by GallonOfAlan at 1:08 PM on January 15, 2016 [13 favorites]


It just won't be happening at any scale.

Maybe not, but driverless trucks will.
posted by Floydd at 1:11 PM on January 15, 2016 [5 favorites]


Can we please stop with the heavier-than-air flight hype pls
posted by sandettie light vessel automatic at 1:18 PM on January 15, 2016 [93 favorites]


What I always wonder is, what are our cities going to look like with all the newly unemployed? No more bus drivers, no more Uber drivers, no more taxi drivers, far fewer cops (which, hooray in most respects), a dramatically different car sales and repair system, dramatically different gas station system - there's going to be a huge surge in unemployment, which means that competition for other jobs will get worse and wages will fall.

I'm not looking forward to this, since it seems likely to produce a car-free utopia for the usual rich people and more precarity for the rest of us.


If you're tallying up the suffering, you should also account for the dramatic drop in car-related deaths and injuries that will likely take place in a driverless world. We right now live in a world where more than 2 million people are injured in car accidents every year just in the US (per the NHTSA, although their numbers only go up to 2013). That's basically 1% of licensed drivers every single year, and I've heard that means that 1 in 3 people in the US will be injured in a car accident in their lifetime. Even reducing that number by half (and, frankly, I'm guessing that it would have to be a lot lower than that to become a common thing given the trepidation that most people seem to have about the idea every time it comes up) would be an incredible improvement in the quality of life for an enormous number of people, whether or not they even used a driverless car.
posted by Copronymus at 1:19 PM on January 15, 2016 [27 favorites]


Until someone invites an inflatable, portable carseat, anyone with kids under the age of 8 or so is going to need their own car.
posted by soren_lorensen at 1:24 PM on January 15, 2016 [11 favorites]


I, for one, welcome our new autonomous car overlords. The challenge will probably lie mostly during the transition but the science already developed for these cars indicates that they are safer. The trucking industry has the most to gain initially but people with long commutes will be next and once HOV lanes get priority use for these cars we can expect real momentum. I don't expect that it will benefit everyone's life immediately or in uniform fashion and it will be hard to absorb the job losses it creates but from the safety standpoint I think it is a huge win.

It unfortunately may not come quick enough to spare me the harrowing challenge of teaching my kids to drive and the nail biting experience I'm sure that entails.
posted by dgran at 1:24 PM on January 15, 2016 [3 favorites]


There are some substantial technical/financial hurdles that must be overcome before the utopia of on-demand cheap driverless cars available anywhere is realized. I'm not going to say that these hurdles will never ever be overcome, but there's a long road between here and there, and it is certainly possible that we won't get to that utopia in our lifetimes or ever. I really hope that the promise of future driverless cars doesn't motivate disinvestment in the bus, rail, and subway transportation systems that we have today and that we know work, however imperfectly.
posted by burden at 1:24 PM on January 15, 2016 [8 favorites]


I wish a fraction of all this sweet, sweet VC money could be spent on maintaining and expanding our existing public transportation infrastructure. Because let's face it, the janitors coming into town from the far end of the suburbs to clean your VC office won't be able to afford those fancy-ass Googlemobiles anytime soon.
posted by monospace at 1:29 PM on January 15, 2016 [47 favorites]


What happens when I want a barely itinerated trip to the desert? The mountains? Up the coast?

No worries! Our new app, Itineratr, will be there to plan your trip for you, record the route you travelled in case you want to repeat it, and point out some great shops and restaurants along the way!
posted by Greg Nog at 1:30 PM on January 15, 2016 [12 favorites]


Not mentioned: how autonomous on-demand taxi-like vehicles would reduce urban congestion that results from people circling blocks looking for a place to park

Note that there have been decades worth of studies relating to PRT (Personal Rapid Transit), where van-sized vehicles run on-demand trips on a fairly complex and flexible route network. Simulations/models of these networks did not indicate that they would appreciably reduce congestion (particularly in contrast to the cost of building such a network).

Optimizing these networks so they would perform better basically produces something that looks an awful lot like a traditional bus system (larger vehicles, and an uncomplicated network of fixed routes).

Basically, what I'm saying is that something that looks very similar to "driverless car revolution" has already been modeled, and it doesn't appear like it would bring about the benefits that are currently being touted. However, it does appear to be a very convenient excuse to neglect funding for public transit today.

On the other hand? Driverless buses on fixed routes? Technology as an extension to public transit, rather than a replacement for it? That's an idea that actually could work. The economies of scale are much better, and labor costs are the biggest obstacle to expanding public transit in most places.

But nobody is interested in it, because car culture is very much still a thing, no matter what these pundits are saying. All of this hype seems to be driven by people who still want to sell more cars, replace your car with a subscription service, and replace public transit with a private enterprise.

As for me? I enjoy living near where I work and riding my bike everywhere. We need more dense cities, a housing market that's flexible enough to let people live near their work, and a good transportation system to let them get other places that they need to go on a less-than-daily basis.

We don't need a technology-driven solution to validate/perpetuate California's failed and unsustainable model of suburbia. We need better cities, and we need to solve some social problems in order to get there. Driverless cars are not going to solve the social and land-use issues that are creating all of this traffic in the first place.
posted by schmod at 1:30 PM on January 15, 2016 [52 favorites]


won't be able to afford

Neither will I, and I wouldn't want to--I hate owning a car. I'm eager for a possible future where some car subscription service just debits $reasonable_amount from my bank every month and I summon cars when I need them.

I know that won't be a panacea but it will beat the heck out of the current way of doing things.
posted by sandettie light vessel automatic at 1:33 PM on January 15, 2016 [2 favorites]


There is never, ever, ever going to be a flat-fee carsharing service. There's no business model where that would ever make sense.
posted by schmod at 1:34 PM on January 15, 2016 [2 favorites]


What's it all mean for transportation planning professionals?

Hours and hours of exciting "what if ...? Then what ...?" discussions. Some of us love this kind of thing - I like to imagine a future where roads without traffic signs and lights are the norm, not some "only in Europe" oddity.


Can we stop with the driverless cars hype please. It just won't be happening at any scale.

As Floydd pointed out, driverless trucks are already a near-reality (real world example of current investments in this technology: How Canada’s oilsands are paving the way for driverless trucks — and the threat of big layoffs), and as of July 2015, Uber wants to buy Tesla’s entire first batch of self-driving cars. If Uber gets people out of the equation, they'll have a lot more control over their system, and possibly side-step the whole "are we or aren't we a taxi service" mess, and likely move them towards the tipping point where it's cheaper to user Uber than own your own car.

Remember, before 1923, cars were not dominant on roads in the United States, and the first freeway wasn't open until 1940, which was the first time there was a road that was strictly Cars Only. Car culture is relatively young, and still quite malleable. Car ownership is in decline (ignoring last year's bump that is in part thanks to low oil prices). More reliable, ever-present on-demand vehicles would make that transition to not owning a car all the easier.

The one thing no one talks about in regards to the major coverage and discussions of autonomous vehicles: what about rural areas? Unless autonomous car renting companies get some incentive to extend their coverage to exurbs and beyond, I think those will be the last bastions of car ownership and/or "old style" vehicles.


Basically, what I'm saying is that something that looks very similar to "driverless car revolution" has already been modeled, and it doesn't appear like it would bring about the benefits that are currently being touted. However, it does appear to be a very convenient excuse to neglect funding for public transit today.

Yes, this always irks me - public transit is great, except it's less direct (and thus generally slower) than people currently enjoy with personal vehicles. But people aren't throwing stupid sums of VC money at making better public transit, because where is the chance for ROI there? (Now that is an interesting theoretical exercise - how could VC money and ROI be infused into improving public transit, while addressing issues of environmental justice and equal distribution of services and support?)
posted by filthy light thief at 1:35 PM on January 15, 2016 [9 favorites]


We already have had problems with Toyotas accelerating out of control, due to faulty algorithms. But the security aspect that no one talks about bothers me a bit more. If some governmental or corporate entity in local control of the state can remotely override an autonomous vehicle, in order to make it stop or drive elsewhere, there is little that can stop anyone else with ulterior motives from doing the same. Not sure I would be comfortable handing over that much control over my life (and death) to some amalgamation of Google and local authorities, let alone a hacker who wants to take me or my loved ones for a Baltimore joy ride for laughs.
posted by a lungful of dragon at 1:35 PM on January 15, 2016 [9 favorites]


What I always wonder is, what are our cities going to look like with all the newly unemployed?

Nobody planning the kind of future where driverless cars are a central feature gives a rat's ass, so it won't be addressed.
posted by ryanshepard at 1:37 PM on January 15, 2016 [14 favorites]


Can we stop with the driverless cars hype please. It just won't be happening at any scale.

[emphasis mine] I'll take the next step for you: "I believe in driverless cars, just not in anthropogenic driverless cars."

Driverless buses on fixed routes?

These appear to be the buses mentioned in the first OP link, being tested in Greece at the moment.
posted by XMLicious at 1:41 PM on January 15, 2016


There is never, ever, ever going to be a flat-fee carsharing service.

No auto-driving Zip cars?

There's an interesting question about who will own these shared cars. Taxi companies? That seems pretty obvious. But what about coops? Might a condo board own a dedicated fleet? I can see that making sense too.
posted by bonehead at 1:42 PM on January 15, 2016 [2 favorites]


Basically, what I'm saying is that something that looks very similar to "driverless car revolution" has already been modeled

I would argue they're not all that similar. The difference between a frequent but fixed service along a set of routes and a completely customised door to door service is fundamental. The change in the convenience level is a large part of what makes owning a car worth it today --- you can be anywhere within say 50 miles in a hour, instead of 3 on foot or maybe 15 on a bike. Even a frequently served route means wait times and transfers and walking a portion of the final mile --- don't what exactly the average range would be, but I wouldn't be surprised if it's much closer to biking than driving.

Let's say you have a lousy job and live far away from it; there's no public transit so you drive a $2000 car that's fifteen years old, skimp on insurance and get your repairs from Joe down the street. Are you really going to be able to afford a driverless cab back and forth every day?

The keyword there is insurance. More and more I think that's the flex point, the fulcrum on which the thing turns. It may well be that the tech can't really work at scale, which is one kind of problem. But assuming the cars are smart enough to drive themselves, the insurance dilemma is fascinating. Driverless cars will reduce accidents, but not eliminate them. Who will be liable when one occurs? If driverless cars are in fact way less risky, then will it be that manual cars are made unaffordable for segment of society because they can't afford the insurance? It's already illegal to drive without it, people just get away with it because they're hard to catch. The fewer manual cars there are out there, the easier that would become, and with far fewer speeding tickets, cops may well be incentives to stop people just for that.
posted by Diablevert at 1:46 PM on January 15, 2016 [5 favorites]




Lot of places (in Canada at least) already have no fault driver insurance. I can see that model applying easily to autonomous cars.
posted by bonehead at 1:49 PM on January 15, 2016 [2 favorites]


The real question about insurance is that if automated cars get so much better than humans will insurance companies still allow you to drive your own car? Or will they charge such huge rates for the privilege that only the rich will be able to drive themselves?
posted by octothorpe at 1:56 PM on January 15, 2016 [11 favorites]


bonehead: "There's an interesting question about who will own these shared cars. Taxi companies? That seems pretty obvious. But what about coops? Might a condo board own a dedicated fleet? I can see that making sense too."

See also: Snow Crash.
posted by schmod at 2:02 PM on January 15, 2016 [3 favorites]


1) A significant number of people use their car not just as a point-to-point people mover, but as a kind of mobile locker/office/station. So, sure, I see self-driving cars maybe eventually making up a big portion of work commuter traffic, but series-of-errands or out-with-the-family traffic? Still a skeptic at this point.

It's possible to rent a car for a day. ZipCar and Turo already do this. You can keep the meter running on Car2Go (though it's usually not economical). You can't keep all your stuff in it all the time, but you can certainly run a series of errands (and I have).

2) What happens when I want a barely itinerated trip to the desert? The mountains? Up the coast? Self-driving is nice, but unless the rental economics change significantly, can't see the appeal of owning for this reason going away, and my guess is this is more true the lower the population mass of a given area is.

Whether or not renting a car makes sense depends on the relative costs. We (my husband and I) are currently running an experiment of going car free. We are in Seattle and are able to bus to work and back regularly. However, we do use Uber/Lyft/Taxi from time to time to visit friends who are outside the city or just because we're running late to work.

We had a beater car, but when we owned it, it still cost us $400-$500/month when all the gas/maintenance/depreciation/insurance was factored in. That's a lot of Uber rides. So far, it looks like we'll spend around less than that on an average month.

Without a car, you get a ton of time savings. You don't ever have to fill up. You don't have to park it (or pay for parking). You don't have to bring it in for maintenance. If you don't have money, you can decide to bus (trade time for money) whereas with the car, you're stuck with whatever monthly payments.

Furthermore, if you rent/share cars, you can choose which type of vehicle you need. For example, most times, I want a comfortable sedan. But sometimes I need an SUV or a van to haul stuff. Sometimes, I want to rent a sports car for fun, or a convertible to enjoy the sun. If I rent a car instead of owning, all of these cars are at my disposal, rather than a single one.

I imagine you can tell the car to go wherever you want, if you want to just wander around. That exit looks interesting? Reroute your destination. Done. Plus, with the new mapping technologies, I doubt there will that many "hidden gems" in the future.
posted by ethidda at 2:08 PM on January 15, 2016 [9 favorites]


filthy light thief: "The one thing no one talks about in regards to the major coverage and discussions of autonomous vehicles: what about rural areas?"

Well, in a post-industrial city of 120,000 surrounded by corn with basically no transit (and taxis take an hour, literally an hour, they are for pre-arranged rides, not on-demand rides), my husband and I imagine driverless cars on demand would enable us to become a one-car family. We're always thiiiiiiiiis close to it working out, living close to work and carpooling and him being a bike fanatic, but with no transit options and so many things located "out of town," we just too frequently each need a car for work/driving kids places/etc. But if I could call up a driverless car to hop me over to the doctor's office on the day my husband needs to drive to a rural courthouse, we could jettison that second car.

It's harder to imagine NOT owning a personal car because of the spread-out nature of the area and the frequent need to go long distances (plus the total lack of transit), but I can think of lots of families who could get by with only one car if a reliable car service were available on-demand ... and many of those families (not all by a long shot, but many) could probably pay a slight premium to by a single personal driverless car instead of two regular cars. (Think of the convenience in sending the kids to extracurricular activities without a parent having to drive! Think of all the work you could get done on all the boring drives to the middle of nowhere!) But yeah, I can imagine as part of the initial transition, owning one personal driverless car and using a carshare for the second car.

The other use case that gets me kind-of excited is school buses, especially "short bus" school buses, which sometimes transport as few as ONE child with a disability from his home to a particular special program. In any district you'll have a bunch of these routes with between one and three students with special needs, and typically at least one attendant on the bus in addition to the driver. Imagine if you could use little self-driving shuttles and hire just the attendant to get these kids back and forth cheaper, quicker, with more personalized routes! The limiting factor is quite often the ability to hire enough bus drivers (trucking pays better and has fewer job requirements, it's hard to pay a market-clearing wage), and having to have two adults for one child is of course very expensive. Similarly, think about splitting a bus of rural kids up into a few little self-driving pods for a half-hour bus ride instead of a two hour rural route, without having to hire additional drivers (even harder in rural areas).

It is also not hard to imagine self-driving farm equipment, that stuff has gotten pretty high-tech with the GPS and planting computers and whatnot.
posted by Eyebrows McGee at 2:11 PM on January 15, 2016 [11 favorites]


What I always wonder is, what are our cities going to look like with all the newly unemployed? No more bus drivers, no more Uber drivers, no more taxi drivers, far fewer cops (which, hooray in most respects), a dramatically different car sales and repair system, dramatically different gas station system - there's going to be a huge surge in unemployment, which means that competition for other jobs will get worse and wages will fall.

Do you blame your grandparents for not preventing the tractor from taking jobs from farm workers? Of course not. There will always be jobs, and yes wages will fall, and the concentration of wealth will be exacerbated. However, productivity will go up, and total wealth will go up, and with progressive redistribution, everyone will have some share in that productivity — just like we all have a share in the cheap food made possible by modern farming.
posted by esprit de l'escalier at 2:12 PM on January 15, 2016 [12 favorites]


What happens when I want a barely itinerated trip to the desert? The mountains? Up the coast?

It's already possible to rent cars for these things; I don't think that will change. I don't think that driving will become illegal or prohibitively expensive without the very core of our culture changing dramatically.
posted by tofu_crouton at 2:12 PM on January 15, 2016


Renting a driverlesss car will be dirt cheap. Close to, or under the cost of current public mass transit.

Why?
1) cars will be built to last, fleet vehicles with easily repaired interchangeable parts.
2)most expensive part of a taxi is the driver, then the insurance, then the car payments.
posted by Freen at 2:13 PM on January 15, 2016 [4 favorites]


It will make sense in dense urban environments and not make sense in rural areas and non-wealthy suburbs. It also pushes us further towards a total rental economy.
posted by grumpybear69 at 2:13 PM on January 15, 2016 [1 favorite]


What I always wonder is, what are our cities going to look like with all the newly unemployed? No more bus drivers, no more Uber drivers, no more taxi drivers, far fewer cops (which, hooray in most respects), a dramatically different car sales and repair system, dramatically different gas station system - there's going to be a huge surge in unemployment, which means that competition for other jobs will get worse and wages will fall.

I posed a question on AskMe about six years back about the apparent surfeit of former service stations. (indeed, when my dad was in town last fall for my wedding, I pointed out that the corner of the blovck where his hotel was now a park where once it had been a service station, to which he replied "Every corner was.") I don't know if the presumptive massive rise in unemployment among mechanics has affected culture that strongly.
posted by ricochet biscuit at 2:14 PM on January 15, 2016 [4 favorites]


The one thing no one talks about in regards to the major coverage and discussions of autonomous vehicles: what about rural areas? Unless autonomous car renting companies get some incentive to extend their coverage to exurbs and beyond, I think those will be the last bastions of car ownership and/or "old style" vehicles.

That was my immediate question, too. My home base is a bit under twenty miles from the nearest city of any import, and we have a small population, although the population grows substantially when the college is in session. Just about all of the nearby taxi companies are located in the city, although I think the village has one ride service (no hailing) based here. Would we have to buy our own cars, or would someone think we were worth the investment? Perhaps the college could buy its own fleet (they currently have an arrangement with Zipcar, so there's that). Alternately, the city could probably make a case for subsidizing driverless cars for the surrounding villages on the grounds that it would promote local business--I for one would go into the city a darned sight more if I didn't have to deal with the terrible parking arrangements and endless run of one-way/badly laid-out streets. But I suspect that, in the end, we'd be left either driving old-fashioned cars or having to rent our own driverless cars.
posted by thomas j wise at 2:17 PM on January 15, 2016


Can we also stop with the 'heavier than air flight' straw man.

Flight makes sense and people wanted it. Driverless cars are a tech wet dream solution looking for a problem that doesn't exist.
posted by GallonOfAlan at 2:19 PM on January 15, 2016 [5 favorites]


Am I the only one who predicts that driverless cars will result in... more or less the status quo? That people who own cars now will prefer to own their own driverless cars, and will now just use their driving time to read/watch tv/screw around on their phone?

I can see changes on the margins: maybe cars will park a quarter mile away after dropping a passenger off, and of course the labor effects of laid-off Uber and truck drivers.

But I don't predict massive changes in car ownership and use, beyond the already ongoing changes arising from the preferences of the young for urban living and reduced car ownership.
posted by crazy with stars at 2:19 PM on January 15, 2016 [4 favorites]


Insurance companies will jump on this like vultures and raise rates like crazy in urban environments to disincentivize ownership and human drivers. It will become a privilege of the rich to own and drive their own cars and poor people will be forced into yet another rental scenario. That much seems inevitable. The biggest impact will be in suburban areas.
posted by grumpybear69 at 2:25 PM on January 15, 2016 [5 favorites]


Driverless cars are a tech wet dream solution looking for a problem that doesn't exist.

What? I mean, some people like driving, but I know at least an equal number of people who absolutely hate driving. And even people who like driving don't usually enjoy sitting in rush hour traffic. Wouldn't it be nice to nap/read/goof off on your daily commute?
posted by ethidda at 2:25 PM on January 15, 2016 [15 favorites]


far fewer cops (which, hooray in most respects)

Call me cynical, but I think it's far more likely that all those cops will keep their jobs and focus on other ways to fill their (non-existent, of course) quotas. Which, boo in most respects.
posted by Greg_Ace at 2:30 PM on January 15, 2016


Driverless cars are a tech wet dream solution looking for a problem that doesn't exist.

What? No.

Truck driving is a shit job. No one wants to do it. It's the equivalent of cotton-picking. I mean, I guess some people like it as a job since it's simple, but it's a terrible use of a person.

There are close to 40,000 deaths on US roads every year. And yet the US spends billions fighting terrorists?

And there are millions of disabled people in US and around the world who lack the basic ability to go where they want, when they want in a car.

I don't think self-driving cars are going to be perfect any time soon, but there are a lot of reasons to pursue the technology.
posted by GuyZero at 2:31 PM on January 15, 2016 [27 favorites]


Fuck cars. I don't want driverless cars. I want cities without cars. I mean, yes, there will always be the need for some cars, some cabs, transportation for people who require special transportation. But seriously, fuck cars. I would much rather all this investment was going into meaningful public transportation and pedestrian or transit friendly design.

I have yet to read an argument in favor of driverless cars that convinces me they will solve most of the problems with car culture.
posted by crush-onastick at 2:31 PM on January 15, 2016 [23 favorites]


I have yet to read an argument in favor of driverless cars that convinces me they will solve most of the problems with car culture.

Not everyone lives in a high-density city.
posted by GuyZero at 2:32 PM on January 15, 2016 [9 favorites]


> What I always wonder is, what are our cities going to look like with all the newly unemployed?

> Nobody planning the kind of future where driverless cars are a central feature gives a rat's ass, so it won't be addressed.

Except, we do. "We" being transportation planners, who already are concerned about environmental justice issues (as in: impacts of the environment/place on the people who live there). Car makers? Not so much.


> Driverless cars are a tech wet dream solution looking for a problem that doesn't exist.

Counterpoints: Uber, who would probably be happy to not deal with real drivers, and trucking companies, which are currently facing a trucker shortage plus increased hours of service scrutiny and oversight.
posted by filthy light thief at 2:32 PM on January 15, 2016 [3 favorites]


But seriously, fuck cars. I would much rather all this investment was going into meaningful public transportation and pedestrian or transit friendly design.

Any form of public transit which relies on fixed routes is necessarily unequal when it comes to how much it benefits any particular citizen. There are going to be people who live a half-mile from the nearest transit stop. Driverless cars remove the fixed route problem and allow anyone to be shunted from point A to point B. Driverless buses could even optimize routes if there are clusters of people in certain areas with similar schedules and destinations.

Subways and walking are great but not a complete solution.
posted by grumpybear69 at 2:37 PM on January 15, 2016 [9 favorites]


Self-Driving Car Doomsday Scenario 1: To save money on parking, people let their cars circle the block downtown all day.

They posted 39 more ways people could make cities worse with self-driving cars .
posted by anthill at 2:39 PM on January 15, 2016 [3 favorites]


Screw driverless cars, I want Boston Dynamics to build me the robot horse from The Diamond Age.

It would have some practical advantage too—there's a bus stop fairly near to my house that connects to the transit network of the nearest city but inconveniently there's a river in the way which requires an extremely roundabout trip to get around by road. But the river's not much more than waist-deep during most of the year, so my trusty horse-robot could get me to the bus stop without having to change pants and/or have my legs amputated due to hypothermia. Then I would slap it on the rump and shout "Home, Bucephalus!" and get on the bus.

Maybe some day when you post to MeTa, an actual pony might show up on your doorstep.
posted by XMLicious at 2:41 PM on January 15, 2016 [7 favorites]


I love cars, I've always loved cars, I think car people of the future are going to be like horse people are today. You love cars? You're going to need to buy a place in the country with space to use your cars.

The only reason I drive in the city is because our public transit is shit. If we don't have to hire bus drivers to drive the buses we can afford to double the number of buses. Plenty of people drive themselves instead of taking Uber/Lyft/Taxis but if there aren't drivers to pay the cost of the ride will be a fraction of what it is today. If public transit actually worked or hiring a car cost very little money, nobody would drive themselves and parking wouldn't be an issue. We'd also finally have room for proper bike lanes.
posted by foodgeek at 2:47 PM on January 15, 2016 [2 favorites]


Just heard a talk by the head of the Goggle driverless car program at MIT.

Can we stop with the driverless cars hype please. It just won't be happening at any scale.

Smart pragmatic guy with a great sense of humor. He made a comment about his kids, specifically that he hoped his son would never have a drivers licence. The kid was 13.

Perhaps the guy was optimistic, but there are driverless cars on the road now, every day in all kinds of weather in situations that would be sure fire accidents. One example was a messy intersection in the dark with terrible drivers crisscrossing and a bike with no lights zooms in front of the google car. Biker probably never knew it the algorithm saved his life.
posted by sammyo at 2:47 PM on January 15, 2016 [4 favorites]


It's worth noting that not everyone has the funds to rent a taxi, or the physical and mental ability to walk everywhere or bicycle. Solutions that require everyone to be physically and mentally fit and wealthy in order to participate in the city have at least some unexamined problems.
posted by corb at 2:48 PM on January 15, 2016 [15 favorites]


Until someone invites an inflatable, portable carseat, anyone with kids under the age of 8 or so is going to need their own car.
You mean like this one? My six year old uses one of these when we need to take a cab. (Some of us are car-less urban dewellers.)
posted by fings at 2:51 PM on January 15, 2016 [3 favorites]


I think the self driving car will converge with the Amazon delivery drone. I mean, if you deliver the pizza with a self-driving car, does that car actually need the ability to carry a human at all? Heck, make the self-driving car into a self driving pizza oven, and the pie can bake on the way there.

It might make congestion worse if they circle the block while the pie finishes...
posted by elizilla at 2:53 PM on January 15, 2016 [4 favorites]


What makes you all think that the companies building the driverless cars are actually going to sell them to consumers or independent operators?

There are few enough companies who have the resources to produce such a vehicle, and it'll be much easier to rake in profits if they rent the vehicles or operate on-demand vehicles that charge by usage. Odds are, the network effect will leave us with a duopoly (at most), similar to what we have with phone companies, ISPs, and cellular providers.

Hell. These companies will probably be able to successfully lobby to make private ownership illegal or prohibitively expensive.

Remember how AT&T used to charge exorbitant fees to rent telephones, because that was the only way you could actually get a phone in your home? Why is this going to be any different?
posted by schmod at 2:55 PM on January 15, 2016 [6 favorites]


It's worth noting that not everyone has the funds to rent a taxi, or the physical and mental ability to walk everywhere or bicycle. Solutions that require everyone to be physically and mentally fit and wealthy in order to participate in the city have at least some unexamined problems.

It's worth noting that not everyone has the funds to buy a car, or the physical and mental ability to drive a car safely. Solutions that require everyone to be physically and mentally fit and wealthy in order to participate in the city already have at least some unexamined problems.
posted by elizilla at 2:57 PM on January 15, 2016 [17 favorites]


Driverless cars will only work out if you can come up with some sort of premium driverless car that the lets the owner somehow be a bigger asshole than everyone else.
posted by srboisvert at 2:59 PM on January 15, 2016 [6 favorites]


It's worth noting that not everyone has the funds to rent a taxi

How about a cheaper ride than a bus?

Many times on off hours I've been essentially chauffeured in a gigantic multi seat diesel limo all alone in the far back seat. For a buck. The driver on that run was making a lot more than a buck for that 20 minute drive and the buck probably barely covered fuel anyway. Perhaps subsidized but still a waste for that specific trip.

Add up the various savings noted in previous posts, make it electric with smarts to self charge and it may be a lot cheaper and "greener" than the bus.
posted by sammyo at 3:01 PM on January 15, 2016 [1 favorite]


A relevant BikePortland article was posted today about the possible downside to dirt cheap SOV transportation.
posted by twjordan at 3:02 PM on January 15, 2016 [1 favorite]


Perhaps subsidized

I don't know of an unsubsidized public transit system anywhere in the world.

For TriMet, 23% of their revenue comes from passenger fares. So your $1 trip required about $3 from other sources.
posted by GuyZero at 3:05 PM on January 15, 2016 [3 favorites]


problem that doesn't exist

The act of driving is a terrible use of time, unless you find it pleasurable--I know there are people who dig it. Otherwise, I could be doing anything else than the continual-heightened-awareness-stop-go bullshit that consumes absurd chunks of time getting from this place to that place. And yeah, bikes and feet are great in a lot of situations, of course.
posted by sandettie light vessel automatic at 3:08 PM on January 15, 2016 [3 favorites]


I loathe driving (despite coming from Los Angeles) and, as I mentioned, would spend a lot more time in the city, giving businesses my $$$, if I did not have to drive there. (Or, for that matter, going to more far-flung destinations in my neck of the woods.) My parents, who are in their seventies and still in CA, would not mind at all if they could swap driving their own cars for self-driving cars, and I imagine that many others in their position would say the same.
posted by thomas j wise at 3:12 PM on January 15, 2016 [1 favorite]


Until someone invites an inflatable, portable carseat, anyone with kids under the age of 8 or so is going to need their own car.


When you pull up the driverless car app, you check the box for child seat,
and the car that shows up has one.

Problem solved.
posted by ocschwar at 3:15 PM on January 15, 2016 [4 favorites]


Self driving cars: every question is handwaved away with "tech will solve that"
posted by Ferreous at 3:39 PM on January 15, 2016 [4 favorites]


I kind of think that self-driving cars will be the new fusion energy: a promising technology always 5-10 years away.
posted by GenjiandProust at 3:42 PM on January 15, 2016 [6 favorites]


The real problem with driverless cars is people. People in their forms as drivers, bikers or pedestrians get drunk, eratic or stupid, and cannot wirelessly handshake every one on a potentially intersecting vector to negotiate right of way.

The amount of AI / algorithms and lidar / radar / GPS perspicacity required for driverless cars to interact with people is an order of magnitude greater than that needed for them to operate on roads limited to themselves.

But it is not going to be politically easy to start to exclude driver-navigated cars or cyclists from the roads, and impossible to exclude pedestrians from surface streets.
posted by MattD at 3:42 PM on January 15, 2016


I so wish I could link to the video I saw at the MIT presentation of one of the google cars dealing with a woman in a powered wheel chair chasing a goose in the middle of the street.
posted by sammyo at 3:46 PM on January 15, 2016 [7 favorites]


and liability issues. Who is at fault if a driverless car hits a person/another car? The car maker, the occupant?
posted by Ferreous at 3:47 PM on January 15, 2016


Note: no geese or women harmed.
posted by sammyo at 3:47 PM on January 15, 2016 [1 favorite]


Autonomous cars are going to be a huge leveller for many (but certainly not all) disabled folks. There are a few blind folks in my extended family whose lives would be transformed by summonable cars that are cheaper than taxis. OTOH, If they're the same price as taxis they won't be very interested.
posted by bonehead at 3:48 PM on January 15, 2016 [12 favorites]


ladies and gentlemen, i can report experience that i already have with cars that are a public resource, and i can assure you, it is gross when your rental car is sticky and reeks of chemical cleaner. but of course nobody would choose owning when they could have that!
posted by indubitable at 3:49 PM on January 15, 2016


and liability issues.

yes, huge, huge, huge --- for drivers... when there are zero accidents caused by the software, even idiots leaping in front. Which is the case for over a million goggle-car miles in traffic.
posted by sammyo at 3:49 PM on January 15, 2016


gross when your rental car is sticky

I think the Uber model of feedback will resolve a lot of that. When a car messed up by a slob is regularly rejected by the next fare the "slob" will have trouble getting a ride until his habits change.
posted by sammyo at 3:53 PM on January 15, 2016 [1 favorite]


The mysterious stains and smells issues are one reason why I think limited coops might well take off for collective car ownership.
posted by bonehead at 3:55 PM on January 15, 2016


The amount of AI / algorithms and lidar / radar / GPS perspicacity required for driverless cars to interact with people is an order of magnitude greater than that needed for them to operate on roads limited to themselves.

You know that Google has like tens of thousands of miles logged driving public california roads in their cars in autonomous mode right?
posted by GuyZero at 3:57 PM on January 15, 2016 [5 favorites]


And the driverless buses being tested in Greece are shown on the road with people and other vehicles.
posted by XMLicious at 4:01 PM on January 15, 2016 [1 favorite]


Aren't those built on a system that heavily scans and documents the roads they are going on? Not something that's entirely reactive without mapping.
posted by Ferreous at 4:01 PM on January 15, 2016


Every public road in the US is mapped. So I'm not sure what the issue is?

Yes, we have not yet perfected a vehicle that can drive across Iraq to do supply runs from ports to the green zones. DARPA continues to work on it.
posted by GuyZero at 4:03 PM on January 15, 2016 [1 favorite]


Is this a good time for me to begin studying urban planning?
posted by gucci mane at 4:13 PM on January 15, 2016


Truck driving is a shit job. No one wants to do it. It's the equivalent of cotton-picking. I mean, I guess some people like it as a job since it's simple, but it's a terrible use of a person.


What? Lots and lots of people want to do it, they pay thousands for training and certifications to get a chance to do it. In the blue collar world, it's a pretty good, well paying job, while other jobs have been getting shittier and paying less.
posted by rodlymight at 4:13 PM on January 15, 2016 [8 favorites]


Google, for its part, is ready to accept responsibility in case some of its cars commit a traffic violation, and says that it should get the ticket, instead of the individual in the driver's seat. “What we’ve been saying to the folks in the DMV, even in public session, for unmanned vehicles, we think the ticket should go to the company. Because the decisions are not being made by the individual,” safety director for Google’s self-driving car program Ron Medford
Who or What Gets a Ticket in a Self-Driving Car?

Google wants to accept liability for tickets. They know that their cars will outperform human drivers, and that freedom from liability will be a major selling point for their technology.
posted by rustcrumb at 4:13 PM on January 15, 2016 [4 favorites]


Or is there something else I should go to college for? Note: I'm 26 and have no degree and I want some sort of job when I get out at 31.
posted by gucci mane at 4:13 PM on January 15, 2016


Every public road in the US is mapped. So I'm not sure what the issue is?

Not to the degree that self driving tech as it stands requires.
posted by Ferreous at 4:18 PM on January 15, 2016


yes, huge, huge, huge --- for drivers... when there are zero accidents caused by the software

CAR IS NOT... TO BLAME!!! MEAT JUMP-ED... IN THE WAY!!! CAR EN-ACT-ED... JUS-TICE ON MEAT!!! MEAT FAILS! MEAT FAILS!

JUS-TICE!

JUS-TICE!!!

(wet sounds and the smell of burning. a saw whirs. soft sobbing. more wet sounds.)
posted by GenjiandProust at 4:30 PM on January 15, 2016 [7 favorites]


At least once a week I see self-driving cars on suburban streets in Google's hometown Mountain View. All the big car companies have research vehicles driving themselves. Some even have models you can buy, which drive themselves on the freeway. Buses are already driving themselves in parts of Greece and France. The technology has the potential to save huge amounts of money for the taxi and trucking industries. Computers keep getting smarter and faster. Of course we will have self-driving cars.

The trick is to figure out what will happen. As humans we tend to over-estimate short-term changes ("my son won't have to get a driver's license") and under-estimate long-term changes ("commuting will look the same in 30 years as today").
posted by Triplanetary at 4:34 PM on January 15, 2016 [1 favorite]


Google wants to accept liability for tickets. They know that their cars will outperform human drivers, and that freedom from liability will be a major selling point for their technology.

Or they can just negotiate a better deal for liability issues with the local government than individuals can: "We have the tickets you wrote our cars. You have a non-zero number of people who live in your district and work for us, or have some economic relationship with us, who represent X number of dollars of tax revenue that is valuable to you. So let's work out a deal that is mutually beneficial." The negative impact on public policy and livability where companies have control over public transportation options is an observed problem.
posted by a lungful of dragon at 4:36 PM on January 15, 2016 [1 favorite]


> What I always wonder is, what are our cities going to look like with all the newly unemployed? No more bus drivers, no more Uber drivers, no more taxi drivers, far fewer cops (which, hooray in most respects), a dramatically different car sales and repair system, dramatically different gas station system - there's going to be a huge surge in unemployment, which means that competition for other jobs will get worse and wages will fall.

Huey P. Newton had the answer to this one way back in 1970:
As the ruling circle continue to build their technocracy, more and more of the proletariat will become unemployable, become lumpen, until they have become the popular class, the revolutionary class.
posted by You Can't Tip a Buick at 4:46 PM on January 15, 2016 [4 favorites]


Every public road in the US is mapped. So I'm not sure what the issue is?

Whenever one of these stories about Google's bright future as a car company comes up, I like to go on Google Maps and see how they're doing with Street View coverage. If you zoom in on random Midwest American small towns, you'll find many areas where most of the streets are still not included, never mind the more detailed mapping apparently required by their current approach to driving software. Those of us who don't live in cities will be the last to get self-driving cars that can reach our homes, just like we're the last to get DSL or fibre.

Recently I drove across a bit of southern Ontario (the one in Canada) to visit family for Christmas. Sticking to reasonably well-maintained highways as much as possible without making a big detour, it's down to only two parts of that trip not visited by google's cameras, an improvement over last time I checked. But it also included a substantial distance where there was six inches of loose snow on the road. Human driver top speed was about 30km/h. I'm guessing that the Google driver would've given up there, although that may be because it's smarter than me.
posted by sfenders at 4:46 PM on January 15, 2016 [4 favorites]


I read here that Google and Ford are working on the difficult problem of getting the self driving cars to work in snow. Which I would be thrilled if they manage to find a solution. Since even people who have lived their whole lives in the snow belt turn into complete idiots the first time the white stuff flies. Seems to me that the computer would know better than humans to slow the fuck down?

Or on preview, are there certain conditions were the car would just refuse to drive itself?
posted by weathergal at 4:50 PM on January 15, 2016


So in the public mind Google seems to employ two types of programmers:

1 - trans-genius programmers who create omniscient systems that see all, know all and sell all to anyone.

2 - idiots who develop self-driving cars and who don't know what snow is

Man, if only the ads developers could go work on cars for a while.
posted by GuyZero at 4:50 PM on January 15, 2016 [3 favorites]


so, presuming that google/apple/uber are correct about autonomous cars, the liability costs for driving your own beat up shitty car are going to skyrocket, not even accounting whatever else google/apple/uber do to disrupt personal transportation. which means i will be forced to pay google/apple/uber to taxi me around.

all of you guys imagine this future where you will own your own autonomous car, which runs on batteries you own, which are charged up by a solar power system you also own. you don't think you are declaring war on all of the rest of us who will never have access to the credit to own these things.

it's just so jarring to see people excited about what for me will be a dystopia, maybe jarring isn't the right word...

although, i actually think the edge cases (ie snow) are way bigger than a bunch of software engineers immersed in suburban california realize...
posted by ennui.bz at 4:57 PM on January 15, 2016 [7 favorites]


Aren't those built on a system that heavily scans and documents the roads they are going on? Not something that's entirely reactive without mapping

Every car is capable of scanning and mapping as it travels, so the shared communal map is constantly up to date and increasing in detail, and areas not in the map are added to the map as soon as anyone goes there.

Perhaps the cars drive slightly slower and more cautiously when mapping an unmapped area, but I don't think they're unable to proceed.
posted by anonymisc at 4:57 PM on January 15, 2016


are there certain conditions were the car would just refuse to drive itself?

Snow-related conditions? I would certainly hope so. Unless they have some kind of really amazing snow-penetrating radar, roads where they often get whiteouts can be effectively impassible to humans and robots alike.

Incidentally it was one of those genius programmer types who told me that snow was a difficult unsolved problem for self-driving cars. If they've made progress I'd love to hear about it.
posted by sfenders at 5:00 PM on January 15, 2016


If you live in an American city or metropolitan area, do the following exercise:

Total up the amount of time you spend driving in a week. Add the time you spend walking from your parking spot to your intended destination.

Total up the number of miles you drive. Subtract the miles you drive to get fuel. Subtract the miles you drive just to find parking, rather than your destination. Subtract driving to the dealership to maintain the car.

Take the ratio. You likely to be averaging under 18 MPH.

Now, compare with a driverless car:

A driverless car does not need you on board when it goes to park.
A driverless car does not need you on board when it gets fuel.
A driverless car does not need you on board when it goes for maintenance.

With those constraints in mind, you should realize that a driverless car can drive at 20MPH and get you around a lot faster than your own car.

And at 20 MPH, a driverless car can avoid accidents much better, including things involving erratically behaving pedestrians. And when they do happen, they happen at lower speed, with less injury to all involved. Designing a driverless car for safety is an incredibly simple thing:

Just program the fucking thing to err with a brake.

Camera getting fuzzy? Slow down. Computer getting less confident at classifying objects near the travel path? Slow down. Nearby cars slowing down for reason unknown? Slow down.
posted by ocschwar at 5:04 PM on January 15, 2016 [6 favorites]


are there certain conditions were the car would just refuse to drive itself?


Asking the wrong question. A car should default to insisting on staying at rest,
and only proceed when it knows with a high level of confidence that it is safe to move.

Same thing your driver's ed teacher told you, except made explicit in the car's programming.

Add a requirement that the source code be signed by a licensed professional engineer, to prevent any shenanigans.
posted by ocschwar at 5:09 PM on January 15, 2016 [5 favorites]


I'm not sure it's true that insurance rates will skyrocket. Self-driving cars will have to be extremely safe to gain wide-spread acceptance. They will not just not cause accidents, but will also have to avoid instances where the other driver might cause an accident.

So there will be a free rider effect that makes the roads safer for drivers as well. Less payouts equals lower premiums*.

*Perhaps not immediately. Nobody wants to charge less. But even if the market doesn't create a competitor who gobbles up market share by charging cut-throat prices, there's enough animosity against insurance that all the lobbying dollars in the world won't protect them from regulation.
posted by politikitty at 5:09 PM on January 15, 2016 [1 favorite]


Not everyone lives in a high-density city.

Considering there are about 3 real cities in America, yes absolutely.

Of course, there used to be a lot more before automotive interests and the government went on a 75-year car subsidizing binge.
posted by Automocar at 5:12 PM on January 15, 2016 [1 favorite]


Let's pretend that class issues, and edge cases are not problematic (a tall order, I think):

the Forbes piece neglected a few things.

1. Houses will end up looking very different, thankfully. There won't be a need to have streets full of garage doors and front doors. Home architecture can become interesting.

2. Short-term, there will be value in having parking areas in strategic places. This might hold for long-term as well, if up-time per vehicle doesn't increase.

3. With fewer at-use parking lots required, there will be repurposing at unprecedented levels. Property valuations could be volatile.

There's more, of course. But I thought Forbes would go further.
posted by yesster at 5:14 PM on January 15, 2016 [1 favorite]


4. If 90% + of current vehicles' lives are spent idle, then in the optimized self-driving future, we will need about 90% fewer cars. There are many implications for this that the Forbes piece didn't explore. If I were a long-term investor, I would SHORT: car dealerships; car manufacturers; large-footprint retailers. I don't know yet what I would go LONG on. (moot point, I'm broke)
posted by yesster at 5:19 PM on January 15, 2016 [3 favorites]


However, productivity will go up, and total wealth will go up, and with progressive redistribution, everyone will have some share in that productivity

What progressive wealth redistribution?
posted by Iridic at 5:20 PM on January 15, 2016 [6 favorites]


The law of unintended consequences may be very interesting. As noted above until public policy resolves the control of empty cars circling waiting for their passenger parts of cities could become significantly more congested. The effect on insurance could be crazy for a while. Idiots trying to see how close they can get to a running car will likely have some jackass style tragedies. Google may not be evil but Ford certainly is somewhat and Uber rather certainly. Sudden job loss in certain segments... well have not heard an opinion piece by AFL/CIO. Hmm, like the music business where musicians are hired but sit back stage, perhaps the trucks will be required to have a sleeping driver? A new term for Luddite is created by workers slowing delivery trucks by walking in the road?
posted by sammyo at 5:28 PM on January 15, 2016 [1 favorite]


Stop lights go away.

Children are sent out to run anywhere, the streets are safe.

Jet travel under 4 hours is vastly reduced as a "sleeper/lounge" car is more comfortable/cheaper than a drive to the airport/1 hour wait/1 hour flight/taxi to gate/1 hour drive into the city.

Electric becomes practical/cheapest... some polluted cities are more livable.
posted by sammyo at 5:35 PM on January 15, 2016 [3 favorites]


the liability costs for driving your own beat up shitty car are going to skyrocket,

Why? Your cost of insurance is based on your probability of having an accident. Your cost should not change just because there are driverless cars on the road. In fact it may go down if driverless cars improve the chances of your avoiding an accident.
posted by JackFlash at 5:46 PM on January 15, 2016


Your cost of insurance is based on your probability of having an accident.

Not exactly, I think. Your cost of insurance is based on your probability of being at fault in an accident. As the percentage of autonomous vehicles goes up, the chance that the non-autonomous driver is at fault goes up.
posted by yesster at 5:51 PM on January 15, 2016 [1 favorite]


As the percentage of autonomous vehicles goes up, the chance that the non-autonomous driver is at fault goes up.

No it doesn't. As the driver of a car, you either have an accident or you don't. But you don't have more accidents just because driverless cars don't.
posted by JackFlash at 5:53 PM on January 15, 2016


The promise of autonomous cars is that there will be fewer accidents. In the few accidents that do occur, the default position will be to assume that the non-autonomous vehicle is at fault.
posted by yesster at 5:54 PM on January 15, 2016 [1 favorite]


One thing I think hasn't been noticed by anyone else - if all cars are controlled by companies regulated by the state, how would that affect protests? If you don't own your own vehicle and can only choose to go where other people choose to transport you, that makes it hard to go places the state doesn't want you to go.
posted by corb at 6:09 PM on January 15, 2016 [9 favorites]


That's pretty astute, really. Much like the encryption topic that is currently relevant.
posted by yesster at 6:11 PM on January 15, 2016 [1 favorite]


Unless they have some kind of really amazing snow-penetrating radar, roads where they often get whiteouts can be effectively impassible to humans and robots alike.

I don't think they've completed solved snow yet. But as to amazing radar, they use LIDAR which records 700,000 points per second. They're much more advanced than the google maps cars. At least three different major players --- a consortium of automakers + Nokia, Google, and Uber --- are currently competing to map the country with this more advanced systems. This TED talk has a bit of a visual representation. You can skip in to about minute 8, it's all self-justifying intro bullshit before that.

I don't disagree that rural areas will probably be the last that will be fully mapped. But while door to door driverless might not be possible to go from Ottowa to some random lake in the middle of Ontario, you might well be able to get from Ottowa to Regina or Toronto by sticking to major highways, etc.
posted by Diablevert at 6:15 PM on January 15, 2016


We had a beater car, but when we owned it, it still cost us $400-$500/month when all the gas/maintenance/depreciation/insurance was factored in. That's a lot of Uber rides. So far, it looks like we'll spend around less than that on an average month.

I try and make this argument to people all the time. I had a similarly beater-y car and it was actually sort of a money pit. I did a lot of maintenance myself... but between insurance, gas, repairs/parts, and stupid ugh car shit like tickets, of which i got several illegitimate ones i had to fight to even reduce AND dumb stuff like the car getting towed from in front of my friends house more than once(stupid 48/72 hour parking rule, on a side street no one goes on, because of bratty neighbors) it never stopped being a hassle. I sold the car, and started just using car2go or the not-taxi services whenever i felt like it. It's unusual for me to even hit the cost of INSURANCE in rides every month. And i have to keep telling myself that, because "oh, a $10 ride? that seems like a lot!" is always the thought but... no. You can take 15 of those a month and probably still be ahead.


And that said, I'm STOKED for self driving cars, but i have utterly no interest in owning one. No, i want the service that's cheaper than car2go(of which the most expensive part, and the part that just caused a "safety surcharge" rate bump is INSURANCE). Car2go can already be half the price of an uber/lyft/taxi sometimes. Sometimes it's even better than that. But what if there was no liability risk of me crashing the car or otherwise causing an accident?

Car2go is already the closest thing to the service we're imagining here. It's just a tiny car that you pick up whenever and wherever you want, and leave it in any legal parking space. It's already cheap. Without that liability, and likely with electric drivetrains... how much cheaper could it get?

The comparisons to public transit in cost are apt here. A $6 car2go trip can be a $2.50-75 transit trip, but what if it was $3?

I'm an enormous supporter of public transit, whose voted for it and gotten out in person to support it since i turned 18. But there's trips that just don't make sense, where a 15 minute drive is an hour long bus journey. Or in areas like magnolia where the damn buses stop running at 10pm. We can fix some of that, but some trips just do not make economical sense to have a direct bus route connecting the two places.

This, if market forces push it down to as cheap as they should, will be a massive gamechanger. Even for low income people, or people on fixed income. And that's not even factoring in ideas like an even cheaper mode that auto-pools rides that are close enough to the same route.

all of you guys imagine this future where you will own your own autonomous car, which runs on batteries you own, which are charged up by a solar power system you also own. you don't think you are declaring war on all of the rest of us who will never have access to the credit to own these things.

I can't speak for anyone else obviously, but i'd like to think there's at least a few of us who have utterly no interest in owning any of those things. the 90% of the time i'm not in a car, i don't want to deal with it. I don't care what the car looks like, i don't care who owns it. I don't care where it's charged, or who fixes it, or how. I just want to pay someone else an agreed upon amount for the actual time i take up using a random car. Literally all i care about is that it's readily available in some reasonable amount of time, and that it gets me there.

Everything about owning a car seems like a bullshit hassle to me, and it's a hassle so easily fixed by economies of scale. One company repairing 500 cars can have their own repair shop. Now every repair is just a tiny fraction of the price of each ride. They can negotiate cheaper power rates(or the purchase of solar equipment) at much better prices than you ever could. Not to mention the battery replacements/refurbishments and even the cars themselves.

I just can't think of a way in which a self driving car doesn't make more sense as a service. They'll absolutely be for sale, and people will buy them. But unless you live somewhere super rural(and then, some kind of co-op STILL seems to make more sense).

Pretty much though, i think if you're approaching this from the side of owning one you're missing the point. What will make these dirt fucking cheap to use and change peoples lives is the division of fixed cost by the lack of downtime. Especially with electric cars that are inherently a lot more reliable.
posted by emptythought at 6:42 PM on January 15, 2016 [9 favorites]


If you don't own your own vehicle and can only choose to go where other people choose to transport you, that makes it hard to go places the state doesn't want you to go.

The state can already tell people where to go with police roadblocks and checkpoints, which are a much more direct and precise instrument than regulatory control. There's no reason to believe that simply regulating companies that make driverless cars would have any impact on protests.
posted by tonycpsu at 6:47 PM on January 15, 2016 [4 favorites]


Honestly, I still find it hard to believe that we're going to have 100% safe and everywhere driverless cars like, REALLY SOON (i.e. before 20 years are up). I mean, hell, my job's been trying to automate the same systems for over ten years and they still haven't managed to work out every bug. I think of my old boss who refused to let me do such-and-such because "oh, it'll be automated in a year" and uh....I forget how many years it's been, but it still hasn't been successfully 100% perfectly automated. I figure we're still going to need to have actual drivers behind the wheel in event of emergency/something going wrong for quite some time, so it doesn't quite seem like it'll be as perfect as everyone thinks right off the bat. Plus yeah, what about the hackers who hack your driverless car--that'll turn up in a TV show or two any second now.
posted by jenfullmoon at 6:49 PM on January 15, 2016 [2 favorites]


Diablevert: "But while door to door driverless might not be possible to go from Ottowa to some random lake in the middle of Ontario, you might well be able to get from Ottowa to Regina or Toronto by sticking to major highways, etc."

I can easily imagine -- and would be grateful for -- a system where I drive the 45 minutes to the freeway, and then my self-driving car takes over the freeway driving until I get to my parents' town, and then I drive the 12 minutes or so from the freeway to their house. I'd happily drive to a sort of "rest stop" next to the freeway where my car spends a couple minutes talking to the system before joining in; even if it had that sort of parking-lot entry and exit points for the freeway system, it'd be a big benefit.

Even if I drove my regular car to a regional transit center (bus/train/self-driving cars), and rented a self-driving car there, had it drive me on the freeway, and had my parents come pick me up at the regional transit center on their end, it'd be super-useful. I imagine the system would be madly overbooked for Thanksgiving and Christmas, but MAN would it be nice for a summer weekend visit. (And, indeed, I already sometimes take an airport shuttle from a regional transit center near me up to an airport near my parents' house and have my parents come get me at the airport! The downside is it mostly runs when the local college is in session and isn't very useful when it isn't.)

We haven't had passenger train service in 30 years, and intercity bus service is really limited to just a couple of cities. For those of us in more "rural" areas, it could really fill an intercity transit gap, EVEN IF it only operated on the freeways and you had to get a taxi or a ride or just walk when you got to your destination. Even if I had to schedule it 24 hours in advance. All these little cities in downstate Illinois -- Rockford, Springfield, Peoria, Bloomington/Normal, the Quad Cities, Chambana, Decatur, Alton -- all used to be served by intercity rail networks that are now gone. An on-demand public transit system that could make a profit carrying 1 to 4 people could reinvigorate some of those decayed links and provide a real benefit to college students, low-income residents (who use the infrequent shuttle buses now), the disabled, and people who just don't feel like driving. The Peoria-to-Chicago shuttle bus is $40 one way, so it just has to be about that affordable and slightly more convenient, and people will use it.
posted by Eyebrows McGee at 6:50 PM on January 15, 2016 [6 favorites]


("rural" in quotes because I live in a darn city, a WALKABLE darn city, but surrounded by cornfields and not a "metro area" like a big city. I don't know how self-driving cars would work in really, really rural areas that are three hours from a city of any sort, but in the Midwest mostly there are these legacy regional small cities that provide services to the surrounding rural area, and I am picturing self-driving car services similarly initially basing their services that would be accessible to rural areas out of these small regional cities. The typical pattern is, Chicago gets something, the suburbs get it, and then they spread it into the small downstate regional cities and along the connecting interstate network, and then into rural areas. Like cell service, or high-speed internet, or highways themselves. So I have a sense of some of the rural uses that would be welcome and useful in my sort of nearby rural area, but I'm not technically rural myself nor is my rural area like all rural areas in the US!)

Now I'm just gonna say "rural juror" over and over after that.
posted by Eyebrows McGee at 7:01 PM on January 15, 2016


The fact that driverless cars and other technology will eliminate jobs isn't a reason to oppose driverless cars and technology, it's a reason to lobby for a universal basic income.
posted by Solon and Thanks at 7:21 PM on January 15, 2016 [16 favorites]


I don't know of an unsubsidized public transit system anywhere in the world. For TriMet, 23% of their revenue comes from passenger fares. So your $1 trip required about $3 from other sources.

General reminder: the vast majority of public roads are paid for by public moneys, and there is no real mechanism to directly recoup those costs. You know, just like that bus, metro, rail or other transit ride that actually costs more than whatever the ticket costs.

The transportation systems of the world are subsidized systems, by and large. Tollways only work with heavily populated areas where there are enough people are willing and able to pay for faster access.
posted by filthy light thief at 7:43 PM on January 15, 2016 [7 favorites]


120k people is a decently big city in much of the US. Five states—Delaware, Maine, Vermont, West Virginia and Wyoming—do not have cities with populations of 100,000 or more. (This definition is limited to populations within the confines of the city limits.)

Here in New Mexico, there's one major city, and the entire Albuquerque metro area has less than a million people. The next biggest town? Las Cruces, which is under 100k, and is the little sibling of El Paso, Texas. The state capitol, Santa Fe, is less than 70k.

Still, we planner types like to look to the future possibilities, otherwise we're trying to catch up.
posted by filthy light thief at 7:50 PM on January 15, 2016 [1 favorite]


The scifi novelist Daniel Keys Moran wrote a series of books which I quite enjoyed. One of the, "The Long Run," includes a small sub-plot built on a society of people who refuse to give up their cars. (OK, they are flying cars.) This group, the Speed Freaks, stages a round-the-world road trip to protest the final outlawing of personal aircars…and the nefarious U.N. one-world government -- dominated by French cyborgs, of course -- steers a hurricane into their path and kills most of them. Early in the book there is a chase scene with an old, stashed aircar that is funny because the guy fleeing the cops barely knows how to control the thing, since no one drives anymore.

(Jesus, this sounds terrible as I type it. I really did enjoy these books, especially "The Long Run." Here, read it yourself for free.)
posted by wenestvedt at 8:41 PM on January 15, 2016 [2 favorites]


On MUNI busses in SF, the operator must both drive the bus and maintain order on the bus. I hope that self driving busses are still staffed with a conductor who can devote their full attention to helping passengers.
posted by rustcrumb at 8:56 PM on January 15, 2016 [2 favorites]


No it doesn't. As the driver of a car, you either have an accident or you don't. But you don't have more accidents just because driverless cars don't.

You responded to the idea that as the number of autonomous cars goes up the human driver is more likely to be at fault with the objection that the number of accidents the human is in doesn't go up. But that's a non-sequitur. Human drivers could simultaneously have a lower average number of accidents per year and be much more likely to be at fault in those accidents. Which is exactly what is most likely to happen.

It's like vaccination. As the number of people vaccinated for mumps goes up the total number of cases of mumps goes down and the likelihood that any individual who gets mumps is unvaccinated goes up.

Autonomous driving cars; a vaccine for human driving stupidity.
posted by Justinian at 9:22 PM on January 15, 2016 [2 favorites]


The scifi novelist Daniel Keys Moran wrote a series of books which I quite enjoyed.

Why did you have to mention him? I had almost... almost... forgotten how sad I am that we'll never see the rest of the Continuing Time. I need to know why Trent the Uncatchable is thought of as a god! I need to know what Camber Tremodian's story is! And Daniel Lord November (insane even by human standards). And Ola, Lady Blue, who was death itself and sorrow!

Why did you mention "The Long Run"? It's going to take me another 15 years to put it out of my mind!

includes a small sub-plot built on a society of people who refuse to give up their cars.

Note that California was exempt from the ban on personal internal combustion engines because cars were considered an integral cultural component of the state!

I wouldn't say that the world government was nefarious per se. Remember that the attempted second American Revolution was portrayed as a fool's errand and a mistake.
posted by Justinian at 9:29 PM on January 15, 2016 [2 favorites]


It's like vaccination.

This is nonsense. The fact that a person who gets mumps is likely to be unvaccinated doesn't increase the probability that they will get mumps. In fact it decreases the probability that they will get mumps, even if unvaccinated, if everyone else get vaccinated. If you are a health insurer, the cost of insuring you for mumps goes down because you are less likely to get mumps, even if you don't get vaccinated. Health insurance costs for everyone goes down if most people get vaccinated, even those who aren't vaccinated.

The insurance company only cares about your probability of having an accident (accidents per year). You don't have a greater number of accidents per year just because there are driverless cars. Your rates don't go up because your rate of accidents doesn't go up. (It more likely to go down).
posted by JackFlash at 9:37 PM on January 15, 2016


I don't think they've completed solved snow yet. But as to amazing radar, they use LIDAR which records 700,000 points per second. They're much more advanced than the google maps cars. At least three different major players --- a consortium of automakers + Nokia, Google, and Uber --- are currently competing to map the country with this more advanced systems. This TED talk has a bit of a visual representation. You can skip in to about minute 8, it's all self-justifying intro bullshit before that.

Chip scale atomic clocks are going to bring the GPS accuracy down to inches in much cheaper and smaller packages. That will allow cars to transport themselves using a combination of environment sensing (camera, ultrasound, LIDAR) and mapping.
posted by Talez at 9:48 PM on January 15, 2016


This is nonsense.

You seem to be completely misunderstanding the original comment (and my followup). The original comment was talking about the probability of the human driver being at fault, not the probability of an accident. That's it, so far as I can care.
posted by Justinian at 10:22 PM on January 15, 2016


That said, there is a difference between "given you are in an accident, what is the probability of you being at fault" and "what is the probability of you being at fault in an accident". The original comment was unclear which it was talking about.
posted by Justinian at 10:23 PM on January 15, 2016


The original statement was "the liability costs for driving your own beat up shitty car are going to skyrocket".

That is false. Using your analogy of vaccination, if more people are vaccinated, cost of insurance goes down, even if you aren't vaccinated. Who's found at fault has no bearing on the argument.

There is simply no reason to believe that insurance rates will go up if there are more driverless cars. If there are fewer accidents, particularly serious injury and fatal accidents, I would expect insurance rates to go down.
posted by JackFlash at 10:37 PM on January 15, 2016


If there are fewer accidents, particularly serious injury and fatal accidents, I would expect insurance rates to go down.

Rates for autonomous vehicles, yes. Rates for non-autonomous vehicles, no.

As the reliability of autonomous vehicles rises, their insurance costs will approach zero, since the risk part of the insurance equation practically disappears.

In the world consisting of reliably safe autonomous vehicles, the few non-autonomous vehicles are the biggest risk to road safety. Quite appropriately, their insurance costs should rise to the point that such behavior is strongly discouraged.
posted by yesster at 11:10 PM on January 15, 2016 [4 favorites]


Let me rephrase that.

In a world of autonomous vehicles, and non-autonomous vehicles, there will be few accidents.

However, the cause of those accidents will increasingly be due to the non-autonomous vehicles.

There is no insurance perspective on this scenario that doesn't push all financial responsibility on to the non-autonomous party.
posted by yesster at 11:16 PM on January 15, 2016


There is no insurance perspective on this scenario that doesn't push all financial responsibility on to the non-autonomous party.

Okay assume driverless cars cause no accidents and all accidents are caused by drivers. But that doesn't raise the price of insurance for drivers. Drivers cause the same number of accidents as before, no more and no less. Let's say the average driver causes one accident per year. Just because driverless accidents go to zero, the average driver still causes one accident per year. Therefore their insurance is exactly the same as before, the cost of one accident per year.

Yes, drivers now bear 100% of responsibility for accidents, but that is because driverless cars cause 0% of accidents. But the average drivers still causes one accident per year, same as before. No change in insurance rate.

Now under your assumption, the insurance cost for driverless cars theoretically goes to zero, but that is because they cause zero accidents. In this scenario driverless insurance goes down. But that doesn't mean driver's insurance goes up. It stays the same because they cause exactly the same rate of accidents as they did before.

I'm trying, but perhaps failing to explain that just because drivers bear all the responsibility for accidents, it doesn't mean that their rates go up. Before the introduction of driverless cars, drivers bear 100% of the responsibility for accidents. After the introduction of driverless cars, assuming they are perfect, drivers bear 100% of the responsibility for accidents. Nothing has changed at all. Drivers are responsible for the accidents they cause, just as always.
posted by JackFlash at 11:34 PM on January 15, 2016 [1 favorite]


Driverless cars already exist. They're called busses.

Stop wasting money on pointless crap when we clearly don't care about transporting people safely and cheaply.
posted by koeselitz at 11:38 PM on January 15, 2016 [3 favorites]


How many years before my uncle has a Red Barchetta?
posted by notyou at 11:39 PM on January 15, 2016 [1 favorite]


Insurance doesn't work that way.

Regardless of what consumers/drivers experience on their side, there's a whole bunch of behind-the-scenes stuff going on.

"Rights of subrogation."
That's a minefield.

Auto insurance companies have never been about paying for accidents. They've always been about finding the right people to pay for those accidents. If they're lucky, it's other insurance companies.

Autonomous vehicles present insurance companies with a trump card.
posted by yesster at 11:46 PM on January 15, 2016 [1 favorite]


One thing I think hasn't been noticed by anyone else - if all cars are controlled by companies regulated by the state, how would that affect protests? If you don't own your own vehicle and can only choose to go where other people choose to transport you, that makes it hard to go places the state doesn't want you to go.

This is a good point. In combination with the increasing ubiquity of facial recognition systems, perhaps the No Fly List will become the No Drive List. At least you might find out sooner that you're on the list, I guess.
posted by XMLicious at 11:49 PM on January 15, 2016 [1 favorite]


Really, a lot could be written about what a hideous idea driverless cars are, what an enemy of the future they are. They're a bit like artisanal twinkies - sure, what a great deal of effort went into making them, what a fine pile of money you'll make selling them, but don't go lying to yourself that you're doing a fine and noble thing in feeding people with that junk. It's just an improved version of something that was never meant to feed people in the first place. Similarly, individual transport is a stupid, dead idea that only continues to exist in order to line pockets with cash. If people have found ingenious ways of prolonging its existence, that's only because they're interested in the money they might make at it.

Anyone genuinely and honestly interested in city planning and good living for humans should be shouting down self-driving cars at each and every opportunity.
posted by koeselitz at 11:51 PM on January 15, 2016 [2 favorites]


the No Fly List will become the No Drive List. At least you might find out sooner that you're on the list, I guess.

Oh how fine! Just what we need! Another blacklist for political opportunists to exploit to marginalize labor activists and other undesirables.
posted by saulgoodman at 12:06 AM on January 16, 2016


...individual transport is a stupid, dead idea that only continues to exist in order to line pockets with cash. If people have found ingenious ways of prolonging its existence, that's only because they're interested in the money they might make at it.

Anyone genuinely and honestly interested in city planning and good living for humans should be shouting down self-driving cars at each and every opportunity.


You know that there are people who live outside of cities, right? People—lots of people, in fact—have to live out here if only to carry out all of the resource extraction and waste disposal that's necessary to keep your cities lurching along.

Call us after you've forced everyone in your city to ride only tandem bicycles because individual transport is such a stupid, dead idea.
posted by XMLicious at 12:28 AM on January 16, 2016 [12 favorites]


What I always wonder is, what are our cities going to look like with all the newly unemployed?

Replace the people, seize the proceeds, come on now.

The only one of these arguments against the idea of autonomous cars that I really buy is that cars as a primary means of transit are a bad idea to begin with. Which, to be fair, is a biggie.
posted by atoxyl at 1:10 AM on January 16, 2016 [1 favorite]


Replace the people, seize the proceeds, come on now.

I mean that was glib but the technology is not the fundamental problem in your scenario. It's not by any means the solution either but in this case it makes some headway on some other very real problems.
posted by atoxyl at 1:12 AM on January 16, 2016 [2 favorites]


My vision of the future definitely does not involve owning one either but certainly one should not count on the future delivering things in the form one prefers.
posted by atoxyl at 1:17 AM on January 16, 2016


The only one of these arguments against the idea of autonomous cars that I really buy is that cars as a primary means of transit are a bad idea to begin with. Which, to be fair, is a biggie.

Yeah. I live in a (non-US) city with good public transportation, in the city itself and connecting the city to a lot of the surrounding towns/villages. About 500 people work in my building, of which fewer than 20% commute by car. Some walk or cycle; most get the bus, the train or the tram.

I don't drive (due in large part to a visual impairment), I have a kid and kid stuff to cart about, and I would much rather live somewhere less urban (we're planning to move anyway in a few years). So in a lot of ways, driverless cars would be great for me! But if everyone else was using one too, the city would grind to a standstill. Even putting aside the energy and pollution costs of 20-60 individual cars vs. one bus, the roads would be unusable with all the extra space required for those vehicles. At that point, it doesn't really matter if you're driving or the car's AI is - you're still not getting anywhere.

It would make it easier to live outside the city, I suppose? But again, it depends on what public transit links are in place already. Not everywhere has a train station. Plenty do, though, and because trains can go faster and aren't dealing with traffic, it would often be be quicker for me to get to work by train than car. Travelling to more rural places, or the islands, then they'd be handier.

So it's not that driverless cars wouldn't be useful, for me personally as well as on a bigger structural level, but I see their greatest use as being to fill in the gaps that a decent public transit system doesn't cover. They're still cars, with all the structural problems that involves. They aren't going to solve the bigger problems that a car-centric culture has. Only effective mass transit is going to do that.

(And yes, I know most of the US does not have a decent public transit system. But, come on: if you can dream an intelligent driverless accident-avoiding schedule-optimised arriving-on-demand car, you can dream a better bus service.)
posted by Catseye at 3:54 AM on January 16, 2016 [1 favorite]


At that point, it doesn't really matter if you're driving or the car's AI is - you're still not getting anywhere.

If there were a coordinating traffic control system I think it might matter - in situations that would be stop-and-go driving for humans or independent computers, I would expect that cooperating or centrally-controlled computers could keep all the vehicles moving continuously but slowly.

But you make a good point about a better bus service; if you had a centralized computer system aware of all available public transit links as well as hireable cars, I'm sure you could offer travelers an option to transfer to a bus or train whenever possible and incentivize doing so. Certainly for me, most of my destinations usually are near or on public transit routes, it's just that the nearest endpoint in the system to my house is a good five miles away by road through hilly terrain. Maybe the key is requiring companies renting out (or Uber-facilitating) anything more than a handful of vehicles to offer this kind of integration.
posted by XMLicious at 5:29 AM on January 16, 2016


What happens when a shared self-driving car gets plowed in? This happens regularly to me in the winter. Whose responsibility is it to get it out?
posted by desjardins at 7:03 AM on January 16, 2016 [1 favorite]


But if everyone else was using one too, the city would grind to a standstill. Even putting aside the energy and pollution costs of 20-60 individual cars vs. one bus, the roads would be unusable with all the extra space required for those vehicles.

I think it's highly unlikely that that's how it would work. Why do you have a car? To get you from point A to point B in the most convenient way possible. Do you keep a car because you enjoy parking it? Buying gas for it? Building a garage to keep it in? No. These things are just the built-in costs of owning one and being able to travel in it. In a self-driving car world it is possible to keep the good part of owning a car --- being able to go wherever you want whenever you want --- without any number of built in costs. You don't have to have a garage, or stop at the gas station, or park it. It does those things itself. In a case like your office, probably a lot of people would keep commuting whichever way they do now. But for the people who do drive, instead of owning a car they keep at their house, they'd own a membership in a car service. The call up a car when they want it, when they're done the car goes off and serves other members, and if no one uses it it goes and parks itself in the service's garage. The amount of cars on the road would be fewer, not greater, even if more people choose to commute via car.

What happens when a shared self-driving car gets plowed in? This happens regularly to me in the winter. Whose responsibility is it to get it out?

If you own it, you. If a car service provider owns it, them. That's how Zipcars and such work right now.
posted by Diablevert at 7:19 AM on January 16, 2016


I didn't manage to go through all the links, but one thing that struck me the other day is that driverless cars would also mean that we could heavily streamline highway interchanges. Our current interchanges are designed to keep traffic moving smoothly, due to the way humans drive; in a driverless scenario, all but the most massive highways could probably just have a single 4-way crossing as the interchange, and rely on the vehicles coordinating with each other to move smoothly without any accidents. That is going to free up a ton of real estate.
posted by destrius at 7:24 AM on January 16, 2016 [1 favorite]


You know what I think is crazy? All these people who say that everyone will watch television and movies on a 4 inch screen. And then they say people will get rid of their record players and listen to music on their phone..? Don't get me started on how much I hated typing class in high school - they say that people will spend all day typing messages to each other on a teeny-tiny screen... even if they are in the same room! Look, a customer just came in and I've got to help them pick out a movie to rent...

Sorry to be so snarky, but this is what the anti-driverless-car argument sounds like to me. It feels like people aren't taking the time to consider how much things will radically change when driverless cars come into being. A majority of people would prefer to never drive - why do you think the Honda Civic is sometimes the best-selling car in the US? I have spent time restoring a 1967 Mustang; I've driven old Rx-7s and old Mustangs for a significant portion of my life. I currently have an 05 Mini. I love cars. I look at Bring A Trailer and Japanese Classic Car every day. BUT, driverless cars are clearly the future and they will so totally transform the car experience that having your own car will be like having a record player or having horses as someone suggested upthread. Yes, people will still do it. My mother has horses, in fact. But it will be an unusual hobby, because driverless cars will be orders of magnitude simpler. I drive around 100 miles total a week and I still spend at least $80 a month on insurance and gas. I can't imagine a constant, on-demand car rental service would cost more than that. That's not even addressing the fact that driving in a car is the most dangerous thing the average person does, because people make idiotic mistakes while driving.
posted by Slothrop at 7:42 AM on January 16, 2016 [7 favorites]


Boy, I can't wait to check back on this thread ten years from now to see how crazy wrong so many of you are.
posted by maxsparber at 9:25 AM on January 16, 2016 [5 favorites]


in a driverless scenario, all but the most massive highways could probably just have a single 4-way crossing as the interchange, and rely on the vehicles coordinating with each other to move smoothly without any accidents

Instead of building bridges there could just be a ramp and you'd jump the river or ravine like the Dukes of Hazzard! If necessary, drones would swoop down at the high point of the arc and catch you and bring the car safely to the other side, like Gwaihir bearing Gandalf to Lothlórien.

In the car chase scenes in movies, the cars could carefully maneuver around fruit vendors' carts rather than smashing straight through them.
posted by XMLicious at 9:29 AM on January 16, 2016 [4 favorites]


Jesus, this sounds terrible as I type it.

oh my god i love those terrible books
posted by poffin boffin at 11:32 AM on January 16, 2016 [1 favorite]


It seems like there are a number of really complicated technical problems that have yet to be solved around all this. I think it's misleading easy to label skeptics of self-driving cars as uninformed or anti-tech, but that doesn't seem to be accurate to me. For example, take David Mindell, MIT professor:
If robotics in extreme environments are any guide, Mindell says, self-driving cars should not be fully self-driving. That idea, he notes, is belied by decades of examples involving spacecraft, underwater exploration, air travel, and more. In each of those spheres, fully automated vehicles have frequently been promised, yet the most state-of-the-art products still have a driver or pilot somewhere in the network. This is one reason Mindell thinks cars are not on the road to complete automation.

“That’s just proven to be a loser of an approach in a lot of other domains,” Mindell says. “I’m not arguing this from first principles. There are 40 years’ worth of examples.”

...as Mindell also observes, there are many challenges to the Google model: Its cars must identify all nearby objects correctly, need perfectly updated mapping systems, and must avoid all software glitches.

Ultimately, Mindell writes, “Google’s utopian autonomy is a more brittle, less functional solution than a rich, human-centered automation.” He predicts that the fully driverless model will not be the most successful, both for technical and social reasons.

“The notion of ceding control of something as fundamental to life as driving to a big, opaque corporation — people are not comfortable with that,” he says. Additionally, other companies and research groups looking at automating cars are “very clearly not going for the Google approach to fully driverless cars.”
If full automation hasn't happened in domains--like undersea submersibles--where there are likely to be a lot less complications and difficulties than in urban driving, why would it happen for cars in cities?

And about those difficulties with mapping... Lee Gomes further describes them:
Google admitted to me that the process it currently uses to make the maps are too inefficient to work in the country as a whole.

To create them, a dedicated vehicle outfitted with a bank of sensors first makes repeated passes scanning the roadway to be mapped. The data is then downloaded, with every square foot of the landscape pored over by both humans and computers to make sure that all-important real-world objects have been captured. This complete map gets loaded into the car's memory before a journey, and because it knows from the map about the location of many stationary objects, its computer—essentially a generic PC running Ubuntu Linux—can devote more of its energies to tracking moving objects, like other cars.

But the maps have problems, starting with the fact that the car can’t travel a single inch without one. Since maps are one of the engineering foundations of the Google car, before the company's vision for ubiquitous self-driving cars can be realized, all 4 million miles of U.S. public roads will be need to be mapped, plus driveways, off-road trails, and everywhere else you'd ever want to take the car. So far, only a few thousand miles of road have gotten the treatment, most of them around the company's headquarters in Mountain View, California. The company frequently says that its car has driven more than 700,000 miles safely, but those are the same few thousand mapped miles, driven over and over again.
That was written in 2014, so it's possible that some technological progress or advancement has changed the picture. I'm super interested in hearing about it if that's so.

However, as recently as December 2015, Charlie Custer pointed out ongoing obstacles to fully self-driving cars, such as the way that even light rain causes serious difficulties. Earlier this week, Google released a report on the number of disengagements of the automated driving system in the past year, which included "341 significant disengagements, 272 of which were related to failure of the car technology and 69 instances of human test drivers choosing to take control."

So, yeah, as best as I can understand, it's pretty clear that cars are going to continue to become more and more automated. But what isn't clear is if all the technical problems standing in the way of full automation are solvable, how long doing so would take, and if full automation is even the best solution, either technologically or socially/politically/economically.
posted by overglow at 11:33 AM on January 16, 2016 [9 favorites]


The notion of ceding control of something as fundamental to life as driving

Does anyone know if it's possible to sprain your eyes from rolling them too hard
posted by Automocar at 11:47 AM on January 16, 2016 [3 favorites]


I've also seen several people in this thread state that since cars sit idle 90% of the time, we'd need 90% less cars. That math doesn't really work given that often the 10% of the time those cars are needed is often overlapping.
posted by Ferreous at 11:59 AM on January 16, 2016 [4 favorites]


For the record: as a cyclist who lives in a city and proudly doesn't own a car, I like the idea of driverless car technology. Makes it way less likely for pedestrians and cyclists to get mowed down by inattentive drivers.
posted by Solon and Thanks at 12:08 PM on January 16, 2016 [1 favorite]


before the company's vision for ubiquitous self-driving cars can be realized, all 4 million miles of U.S. public roads will be need to be mapped, plus driveways, off-road trails, and everywhere else you'd ever want to take the car.

For most people, "everywhere you'd ever want to take the car" is from home to work and back again. There is no reason a solution has to cover 100% of car travel. That is the same argument people make about electric cars and people are buying hundreds of thousands of them, even though you probably wouldn't want to drive one to Yellowstone Park.

People were skeptical when Google started to create street views and they managed to cover 5 million miles of roads in just a few years. They started with big cities like San Francisco, Seattle and LA. You don't need to do all 5 million miles at once.
posted by JackFlash at 12:10 PM on January 16, 2016 [1 favorite]


But you do need to keep them much more updated than street view. Like, any time something traffic based in the road changes updated.
posted by Ferreous at 12:13 PM on January 16, 2016 [1 favorite]


The idea that taxis are always cheaper than car ownership is just flatly not true for many of us who use our cars more than fifteen times a month, even including gas and insurance., especially those of us with kids.

My morning: I drive my daughter to school, then my husband to work. It is an hour and a half commute. It would cost me about 90$, per Uber's current pricing. Assuming I had no chores to do during the day - a bad assumption - I would then need to take an Uber to my daughters school and back -40$, I've done it when my car was in the shop - and then order an uber for my husband. We'd be looking at 150-175$ a day during weekdays. My insurance is 70$. A month. In gas, that trip is about 5-8$, depending on prices.

The idea that this would save everyone money only works with people who live, walk, and educate within the same 5 mile trafficless radius.
posted by corb at 12:50 PM on January 16, 2016 [1 favorite]


"The idea that taxis are always cheaper than car ownership is just flatly not true for many of us who use our cars more than fifteen times a month"

I don't think anyone in this thread has suggested that; in fact, there's been quite a bit of discussion about private self-driving cars, co-ops, and how these would work in rural, suburban, or small urban areas where the density and transit to support a car-free lifestyle is not there. I think you are reading an attack on your lifestyle into the thread that simply isn't there and that, on the contrary, people have been addressing thoughtfully throughout the thread.
posted by Eyebrows McGee at 1:00 PM on January 16, 2016 [3 favorites]


The idea that this would save everyone money only works with people who live, walk, and educate within the same 5 mile trafficless radius.

As hard is this is to believe, Corb, the world does not revolve around you and your particular needs.
posted by JackFlash at 1:01 PM on January 16, 2016 [1 favorite]


JackFlash, the same is pretty universally true: it doesn't revolve around any one person. No need to bring the hammer on Corb.

But there is a continuum of lives that range from "downtown hi-rise dweller" to "remote fire-tower crew," and a lot of us have needs that can't be met by mass transit simly because it doesn't exist in our area. *shrug*
posted by wenestvedt at 1:08 PM on January 16, 2016 [4 favorites]


A solution to a societal need doesn't have to be better or more affordable for outliers in order to be desirable, it merely needs to make things better on average for the population as a whole. We can always provide subsidies to meet the needs of the outliers, as we do with nearly all government services.
posted by tonycpsu at 1:12 PM on January 16, 2016 [2 favorites]


I guess my problem with the "skeptics", if you're identifying yourself as a member of such a group overglow, is that all that's happened is we've gone from absurd declarations about the impossibility of driverless cars as a technology, ones which could frequently be immediately disproved with video evidence, to various little quibbles about details of the technology or its implementation that seem intended to imply that no future can exist in which there are any material differences as a result of it.

Which seems more like a lack of imagination to me and an unwillingness to try to envision what the future might be like, rather than any substantial form of skepticism. And it does bear a striking resemblance to Creationists continuously pointing out supposed missing links in the fossil record and disputing minor details concerning evolution, as a proxy or totem for disproving the existence of the overall phenomenon; though I don't think that sort of behavior is uniquely anti-science.

If, on the other hand, the skepticism is simply a matter of pointing out potential problems and consequences created by self-driving cars, "skeptics" are kind of behind the curve on that one—if you were to look through past threads concerning self-driving cars I think you'd find that the people who acknowledged their existence were the ones bringing up many of the possible negative implications being discussed here.
posted by XMLicious at 1:18 PM on January 16, 2016 [1 favorite]


and yet you're the one making conjectures about what the future will look like based on tech that doesn't exist in any scalable fashion. Thinking the future might not have totally automated car fleets for single passenger use isn't really lack of imagination, it's just not the future you want to imagine.
posted by Ferreous at 2:05 PM on January 16, 2016 [3 favorites]


Nope, I'm just fine imagining that. Go ahead and describe these not-totally-automated car fleets of the future that aren't for single passenger use, but I'm gonna make a little prediction that they'll sound a teensy bit like a scaled version of the current technologies we've been discussing in this thread.
posted by XMLicious at 2:18 PM on January 16, 2016 [1 favorite]


If they aren't totally automated, then all the logic of "shared fleets that will come pick you up as needed!" doesn't work. Especially since last mile movements will be the hardest to design. What it sounds more like is a future that we are heading towards already on production scale, drive assist features on cars that require a human operator as a backup for liability and areas where the drive assist doesn't do everything. That plus underfunded but still workhorse public transport.
posted by Ferreous at 2:28 PM on January 16, 2016 [2 favorites]


I've been a google fan boy in this thread but I certainly have skepticism on a range of issues, but now that we have a legitimate "man will never fly" style proclamation from a legit MIT professor:

The notion of ceding control of something as fundamental to life as driving
David Mindell (thanks overglow)

I'm now quite certain that the adoption of robot cars will be really quite rapid, as in ubiquitous just a couple years or less.

Another point from the MIT talk I was at recently was that the google cars were not going to be released until they handled all problems. They were totally aware that a single robot death to a pedestrian would impact vastly more than the many per day that occur now.

The point that a top driver will handle the extreme cases (off road) better than a robot is spot on. But 80/20 rule rules. I'm a good driver but 99.9something percent of my driving is mindless and several highly tuned LIDAR sensors will beat my perceptions in many edge cases, certainly when I'm tired.

I was just took a walk through a monster car oriented mall to stretch my legs and unless you're driving from one side to the other to get to another store it was just unpleasant, restructuring so there's rational drop off points and good walking routes will be good for people in so many ways. It'll take some social policy, and with our wacky economic pressures it's going to be done wrong a bunch of times in the robot car age.

There are likely to be any number of "why didn't we think of that" but perhaps it's worth it to prevent 2-3 child fatalities per day.

Right, gotta finish up a thread like this with "think of the children" :-)
posted by sammyo at 2:29 PM on January 16, 2016


The notion of ceding control of something as fundamental to life as driving

lol if driving is fundamental to life, give me death
posted by entropicamericana at 2:39 PM on January 16, 2016


lol if driving is fundamental to life, give me death

Yeah I am *not* the sort of person to want to minimize the remaining unsolved technical challenges. It just seems like many of the people ridiculing the aspiration toward fully autonomous vehicles are out of touch with why the rest of us would like it to happen.
posted by atoxyl at 3:42 PM on January 16, 2016


But you do need to keep them much more updated than street view. Like, any time something traffic based in the road changes updated.

What exactly do self-driving cars do about construction? Google maps continually tries to route me down roads that are temporarily closed for construction, and the only reason I know those routes are bad is because I've tried them before or I've read the news (there's been a huge interchange project going on for years). There are detours all over the place. Sometimes one lane is closed. Sometimes there are flaggers, where only one direction can go at a time. I'm not quibbling over technicalities because these are not edge cases; they happen on a weekly basis to me, and I've lived in places with more frequent construction.
posted by desjardins at 4:00 PM on January 16, 2016 [4 favorites]


From what I've seen they mostly seem to stop and wait for things to go back to the expected situation when something unusual appears, so presumably after a long enough wait they give up, reroute, and tell the rest of the system.

When and if the threshold of needing a passenger that can drive is passed will be really interesting though, I can imagine a car version of air traffic control to deal with that sort of situation. It'd really smooth out property values in terms of commuting location and commercial hot spots. No more designated drivers and running for the last tube either.
posted by lucidium at 4:40 PM on January 16, 2016


drive assist features on cars that require a human operator as a backup for liability and areas where the drive assist doesn't do everything

So you're saying we'll have automated-to-the-point-that-the-human-does-nothing-but-sit-there-for-liability-reasons cars for single passenger use, except there will be places where no high-resolution mapping will be done for some reason and in those locations they'll... not suddenly veer off the road or anything, right? Obviously the other cars, trucks, motorcycles, bicyclists, and pedestrians on the road constitute a completely novel set of obstacles every time the car travels down the same route and the present technology can deal with that, so I'm pretty curious to hear what capabilities you think will be impossible under lower-resolution-map circumstances.

But even with whatever those limitations are in the areas that go unmapped for some reason, all of the above still doesn't count as a scaled version of current technology?

It seems to me that this vision is quite compatible with many of the predictions made in this thread and elsewhere; and that whatever this is skepticism of, it isn't skepticism of self-driving cars, because you have self-driving car technology there and it's for legal rather than technological reasons that a human ends up needing their own set of controls.

I will admit to disagreeing with you on the liability thing, though IANAL; having a human driving a car is already a hell of alot of liability, it seems to me, and if indeed it turns out that a self-driving car on its own exposes a company to higher liability than a human driver, I think it's probably a bad bet that the rental car company a few miles down the road from me is going to be unable to find any way to avoid paying an hourly employee to stop what they're doing and sit behind the wheel while the car I rent drives over to my house. After which I'll sit behind the wheel while the car drives him back to the dealership.

(Not to mention, it seems a bad bet that they would also completely give up on the massive new marketing demographic of customers who don't have drivers' licenses whom they'd otherwise be able to rent to.)

I mean liability isn't an obstacle that makes it impossible for small mom-and-pop pizza joints to have humans deliver pizzas; I'd think that the risk exposure from a self-driving car would have to be an order of magnitude higher before it would prevent multinational corporations from figuring out how to eliminate the paid employees from the equation. But there's no evidence whatsoever to suggest such a lopsided disparity in risk, is there?

So tl;dr it sounds like you envision extremely similar, material and substantial changes to society as a result of this technology—similar to what non-skeptics envision—so I think what you're expressing isn't skepticism of self-driving cars themselves, at most it's skepticism of the ability of corporate risk managers and actuaries to accommodate their operations successfully enough to eliminate all the jobs and thereby the human-sits-there-doing-nothing-while-the-car-drives requirement.
posted by XMLicious at 4:41 PM on January 16, 2016


What exactly do self-driving cars do about construction?

Identify the current lane and gestures of any flagpersons and follow the flow of traffic slowly and precisely.
posted by sammyo at 4:41 PM on January 16, 2016


a lot of us have needs that can't be met by mass transit simly because it doesn't exist in our area. *shrug*

I guess this conversation would generally be a lot less frustrating to have if more people understood that people that have "needs that can't be met by mass transit" live in built environments formed by powerful corporate interests with the help of local, state, and federal governments hell-bent on making money and keeping white supremacy alive as long as possible.

It's no coincidence that all of this (very minor) resurgence of American mass transit (usually in the form of BRT or light rail) has come about because white people are moving back to center cities.
posted by Automocar at 4:42 PM on January 16, 2016 [5 favorites]


I'm just as excited for driverless cars as anybody, but given all the breathless excitement, anticipation, and serious financial preparation we've had for them, I think it would be really funny if driverless cars just... didn't happen. Like we just never figure out how to make them work on a large enough scale to displace driverful cars. Somebody would definitely assassinate Elon Musk if he hadn't offed himself in despair already.
posted by J.K. Seazer at 5:52 PM on January 16, 2016 [1 favorite]


There is simply no reason to believe that insurance rates will go up if there are more driverless cars. If there are fewer accidents, particularly serious injury and fatal accidents, I would expect insurance rates to go down.

Wut.

So lets say that a driverless car gets in like, one at fault accident per 100,000 trips. Or even better. What if it's one in a million, or ten million trips. Insurance for them could plausibly be so cheap that it was a dealership addon like a warranty that you just pay up front. I would even be willing to bet something like 99/100 or even 999/1000(or better) instances of that occurring were due to improper maintenance or tapping yes on some kind of "weather conditions are making the path unclear" sort of warning screen on the control panel, and would likely involve inclement weather or something.

The point i'm getting at is that basically all accidents on the road would be human-operated car>self driving car or human operated car>human operated car. And i'd be willing to bet the majority of accidents would be in the second category. A self driving car can move in any direction to avoid getting hit. It knows the exact speed it can accelerate, brake, etc and could cut someone off for example without any risk of an accident to avoid one.(assuming it's programmed to drive aggressively to avoid a collision, which seems like a logical choice).

Assuming all, or even just some of this is true, who would want to insure a human driver in a car? It's like getting flood insurance on a flood plain. I bet some insurance companies will stop even carrying some drivers. It seems plausible that if you say, get a DUI then you can just never get manual car insurance again, or can only purchase it as something like an sr-22 but at ten times that price. The only people who get some kind of major moving violation or DUI and still get insurance will be fucking rockstars and tech millionaires.

I honestly can't think of any scenario in which manual car insurance isn't like, actually a luxury item. The insurance company carrying you would know that you were basically guaranteed to be declared at fault except in the, statistically dwindling, possibility that another human-operated vehicle hit you.

Additionally, i bet getting in even one at-fault accident would like quadruple your insurance if not more. Even like a really minor, or single vehicle one.

This has nothing to do with driving being safer in general, it has to do with being declared at fault. And barring like, a national law about no-fault insurance, that's how shit would be.

You know what would be awesome though, and still probably cheap? Riding a motorcycle. If almost all cars were self driving, riding a motorcycle would be really safe assuming you didn't do anything stupid and get in a single vehicle accident.

I seriously can't imagine a future involving self driving cars that isn't structured like manual cars being the equivalent of cars with "collector" or "historic vehicle" plates now, though. Seeing one 10 or so years after self driving cars are common would be a "woah, look at that" thing and would probably also be a weird or interesting car like an old ferrari or some shit. No one is going to be paying $500 a month for insurance to drive a beat up 90s buick.
posted by emptythought at 6:11 PM on January 16, 2016 [2 favorites]


I guess this conversation would generally be a lot less frustrating to have if more people understood that people that have "needs that can't be met by mass transit" live in built environments formed by powerful corporate interests with the help of local, state, and federal governments hell-bent on making money and keeping white supremacy alive as long as possible.

I can't help but notice that all you mention there are the causes, without denying that the needs exist. Assuming that there were actually a lack of understanding of these influences, how would mentioning them and dropping in white supremacy make the conversation less frustrating?

Doing that isn't going to make the vague "where you live should be more like virtuously-planned cities" suggestions for replacing cars sound any less obtuse to someone living on a farm in a town in the Midwest or to wenestvedt's remote fire-tower crew, or to anyone else living in a place with a population sparse enough that even coordinating school busing where all the kids are going to the same places at the same time can be a problem.
posted by XMLicious at 6:22 PM on January 16, 2016 [2 favorites]


I wouldn't say that I'm a "driverless car skeptic"... more like a driverless car agnostic. I don't know what's going to happen and I'm interested in learning more and conversing about the possibilities. It seems to me like there's a lot of unquestioning acceptance of the hype around driverless cars which, let's face it, is marketing done by corporations that don't always have honesty as their highest priority.

I posted the critiques of driverless cars partially because I want to see if anyone has rebuttals that are compelling. I'm open-minded about this. If someone said, yes, currently, rain and snow interfere with the radar and laser sensors of the cars, but here's [proposed solution that is in the works] or [informed opinion by independent expert], I would likely be convinced. So, far, I'm hearing more of the refrain, but it's totally inevitable!

But why is it inevitable?

I mean, lots of people have seized on the The notion of ceding control of something as fundamental to life as driving line and that totally makes sense. It's a pretty blinkered perspective and leaves out the historical newness of driving and the many, many people in the world (and US) who don't have access to cars and is totally unaware of the privileges involved in car ownership. It deserves to be mocked.

But what about the main thrust of Mindell's argument? That in other industries, people haven't pursued full automation but rather chosen to go with highly-automated systems that still incorporate human agents at some point?

I mean, it seems to me (and I'm aware that I don't know all the things about this and am happy to be informed/corrected) that it would be technologically a lot easier to create fully automated airplanes than fully automated cars. Because there's a lot less to navigate in the environment in the sky and a lot less unexpected things that a vehicle might need to quickly respond to--i.e., there's no pedestrians in the sky. But we don't have fully automated airplanes. Why would cars be different?

Again, there could totally be reasons why cars are different. Please tell us some!

Similarly, along the lines of what folks have been discussing, here's another bit from that Gomes articles, about the difficulties in dealing with construction:
Another problem with maps is that once you make them, you have to keep them up to date, a challenge Google says it hasn't yet started working on. Considering all the traffic signals, stop signs, lane markings, and crosswalks that get added or removed every day throughout the country, keeping a gigantic database of maps current is vastly difficult. Safety is at stake here; Chris Urmson, director of the Google car team, told me that if the car came across a traffic signal not on its map, it could potentially run a red light, simply because it wouldn't know to look for the signal. Urmson added, however, that an unmapped traffic signal would be "very unlikely," because during the "time and construction" needed to build a traffic signal, there would be adequate opportunity to add it to the map.

But not always. Scott Heydt, director of marketing at Horizon Signal Technologies, says his company routinely sets up its portable traffic signals at road construction sites. Frequently, they are simply towed to a site and turned on. "We just set one up like that in New Jersey," said Heydt. "You can be driving to work and everything is normal, but on your way home, discover a new traffic light." (Of this possibility, a Google spokesperson said, “We will have to be ready for that.”)

Noting that the Google car might not be able to handle an unmapped traffic light might sound like a cynical game of "gotcha." But MIT roboticist John Leonard says it goes to the heart of why the Google car project is so daunting. "While the probability of a single driver encountering a newly installed traffic light is very low, the probability of at least one driver encountering one on a given day is very high," Leonard says. The list of these "rare" events is practically endless, said Leonard, who does not expect a full self-driving car in his lifetime (he’s 49).
It seems pretty significant to me that Google admits that they don't yet know how they would keep their maps updated. That is a huge piece of the puzzle, and seems likely to be pretty daunting to fix.

So, yeah, I am authentically curious about the answer to these questions--how might these problems be solved? At the risk of repeating myself, it seems strange to me that people are simply assuming that it's inevitable that they will be solved. Not every predicted or hoped for technological advance turns out to happen. If people have evidence--either updates from Google or information about potential solutions--I would love to hear them.

I don't have enough engineering knowledge to know if the skeptics are likely to be right or not. I do think, based on my current understanding and knowledge, that it's possible that they are less like the people who said humanity couldn't fly and more like the people who are saying, yeah, colonies on other planets are probably not going to happen.
posted by overglow at 6:23 PM on January 16, 2016 [2 favorites]


I can't help but notice that all you mention there are the causes, without denying that the needs exist. Assuming that there were actually a lack of understanding of these influences, how would mentioning them and dropping in white supremacy make the conversation less frustrating?

Because at least then the people saying things like that would show some glimmer of understanding that the American suburbs are not some ideal state of human civilization and that maybe it would be a good idea to work to minimize their spread as much as possible.
posted by Automocar at 6:28 PM on January 16, 2016 [1 favorite]


I agree with emptythought's opinion that insurance costs would skyrocket for manual cars once self-driving cars become widely available. I'm curious what the tipping point would be for people. I don't really see the benefit of self-driving cars for me personally - I've never had an at-fault accident in 24 years of driving,* I like driving, and I want the convenience, flexibility and control of my own car. Even when self-driving cars become commonplace, you will always have to wait, you will always have restrictions, there will always be costs. I don't think increased public safety is going to be much of an incentive for the general public to give up personal control and convenience - we can't even pass any form of gun control in the US!

*not that I couldn't, but it would affect my decision making process
posted by desjardins at 6:41 PM on January 16, 2016


Okay, insurance rates aren't MAGIC. For the purposes of this example we're going to use a single-state insurance company. They're a combination of the claims the insurance company pays out plus the overhead costs of the company plus around a 5% profit margin. In Illinois in 2014, there were 269,014 car crashes, of which 845 involved one or more fatalities. (15,000 of these crashes involved deer; 85,000 were rear-endings. This is not relevant, I just thought it was interesting.) The total cost of these crashes was about $5.8 billion. There are 8.2 million licensed drivers in Illinois. If all of us are equally likely to have an accident, and we're only paying for claims paid, our auto insurance rates in Illinois would be about $707/year. (Which is, in fact, not that far off of my safe-middle-aged-lady-driver insurance rate, so once you add in overhead and profit, and then rate me back down as UNlikely to have an accident and driving a very boring car, that seems just about right.)

Car insurance is a highly competitive industry. So competitive that my insurer (State Farm) actually sends me a rebate check in December of any year that their premiums outweighed their payouts because they don't want me jumping to Allstate or Geico because they did a better job with their actuarial guessing that year. (Allstate doesn't want you to take the safe driver classes it's always on about in commercials out of the goodness of its heart -- it wants you to have less accidents so that it can pay out less in claims.)

Let's say that tomorrow we magically have 50% self-driving cars in the state of Illinois, and they are 99.9% perfect, so our accident rate immediately halves. That means there are suddenly 135,000 crashes costing around $2.9 billion a year. And suddenly 4.1 million drivers who are likely to have accidents (with the other 4.1 million having given over their cars to Google). YOUR INSURANCE RATES STAY THE SAME because while there is now a smaller pool of accident-prone drivers compared to cars, the number of accidents and their costs have similarly dropped, and while there will be some statistical noise in the insurance rates, car insurers cannot afford to stay in business if their profit margin is too high because it is a highly competitive industry. (In fact what you will probably see is layoffs as they can cut costs as the central office that processes claims because there are so many fewer claims.) But the total number of accidents and the total cost of those accidents will probably drop by MORE than half, because you're no longer a bunch of 80% good human drivers surrounded by other 80% good human drivers; you're a smaller number of 80% good human drivers surrounded by 99.9% good robot drivers. Even when the robots can't manage to avoid your swerving, they're going to be MUCH BETTER at mitigating the severity of the accident, so that accidents are less serious and thus less expensive. So chances are good that your insurance rates will actually DROP as more robot cars come on the road, because you will have fewer and fewer opportunities to get into serious accidents.

Now let's imagine that 90% of cars have been replaced by self-driving robot cars (which still leaves us with 820,000 human drivers in Illinois which is still a pretty significant market but let's pretend). It's not inconceivable at that point that, especially downstate, you may end up with a situation where it's not financially feasible for an insurer to operate in a particular (probably rural) county, because there isn't enough business for there to be any economies of scale. The first thing that will happen is that small local insurance companies -- they still exist! We have lots downstate! -- that do a lot of their underwriting by hand still will step in to that gap, as well insurance co-ops that already serve rural property holders who can't get adequate service or reasonable rates from big companies. This will probably cost a little more because you're essentially buying artisinal hand-crafted insurance, but you're getting it either from a two-man shop with very low overhead or a non-profit, so not a whole lot more.

But let's say ALL the insurers in a particular county shut down. We actually have examples and experience with this in other insurance sectors, and the typical response by the state (and remember, insurance companies require state charters to operate and are highly regulated) is that an insurer who is willing to take on the unprofitable and unserved counties gets a special state subsidy to cover the unusual costs of operating in the market. Sometimes, they also get a limited monopoly deal (like, they may be guaranteed to be the only chartered insurer in that county for the next three years, if they put a profit ceilings of 3% on their rates). Politicians are willing to let rural Americans pay more than their urban counterparts for services like internet and cable and phone, but they AREN'T willing to let that grow to a level of "obviously and unjustifiably being screwed, it just does not cost that much to run fibers, they are taking advantage of there being no competition there" because agriculture is a very important economic sector and farmers are politically sympathetic. There is no one in the statehouse SO STUPID that they'd let a downstate farmer pay ten times as much for his pickup truck insurance as a Chicago fatcat banker because the banker has Google Roads available and the farmer WHO BRINGS YOU YOUR FOOD has to drive on county roads that haven't been self-driving-car mapped and some of which aren't paved. (And I mean, there are some pretty fucking stupid dudes in the Illinois statehouse. But nobody THAT stupid.) They'll create exemptions for rural drivers, or they'll tack a $5 "accident risk pool subsidy" on to the self-driving cars' yearly insurance, or roll it into the vehicle registration fee you pay every year.

If it does, for some bizarre unmathematical reason, get to the point where driving your own car makes you an insane insurance liability, the state does step in -- as states have to ensure that low-income people have access to car insurance so they can drive so they can work -- California, New Jersey, and Hawaii, for example, all provide subsidies for car insurance for low-income residents. Illinois has a special risk pool for expensive-to-insure residents. States don't just let insurance costs float randomly.

Data on accident rates, costs, drivers, etc., all from IDOT's 2014 traffic safety reports.
posted by Eyebrows McGee at 7:29 PM on January 16, 2016 [4 favorites]


guess this conversation would generally be a lot less frustrating to have if more people understood that people that have "needs that can't be met by mass transit" live in built environments formed by powerful corporate interests with the help of local, state, and federal governments hell-bent on making money and keeping white supremacy alive as long as possible.

This conversation would be a lot less frustrating if people weren't assuming that everywhere in the United States outside of NYC and maybe three other cities was a corporate white supremacist stronghold. There isn't particularly good mass transit in most places. It is the rule, rather than the exception.
posted by corb at 7:36 PM on January 16, 2016 [6 favorites]


Because at least then the people saying things like that would show some glimmer of understanding that the American suburbs are not some ideal state of human civilization and that maybe it would be a good idea to work to minimize their spread as much as possible.

Bravo, I could not have come up with a better demonstration of the sort of thing I'm talking about than you flagrantly ignoring both me and wenestvedt specifically referring to places that have nothing to do with suburbs or cities so that you could righteously tell us peasants how inadequate our concern for cities and the ways they spread is. How dare anyone else be focused on getting to doctor's appointments in y'know, the community they actually live in, or getting to work at a construction site over in the next state, when you've got important things to do like preen and put your special and virtuous enlightenment on display?

Maybe your suburb residents not giving a shit about the rest of the people in your conurbation is a little of that karma coming back at you.
posted by XMLicious at 7:38 PM on January 16, 2016 [2 favorites]


overglow, while they do have experience and highly reliable software systems, the airline and sea-faring industries are not pioneers in the domain of software engineering.

They don't have the expertise, risk-tolerance, and funding to build software that can successfully navigate terrestrial roadways. The limited size of their problem domain (air and sea) and the dearth of potential customers are partial reasons these industries cannot produce the software and hardware required for more complex problems such as realtime terrestrial navigation.

To my mind Mindel, notwithstanding his expertise (or maybe precisely because of his expertise), can be compared to Ed Colligan (Palm CEO) infamously opining that "PC guys are not going to just figure [smartphones] out. They’re not going to just walk in." But walk in is exactly what they did.

So now Google, Tesla, Apple, Uber, GM, you-name-it, are working on this problem of navigating roadways with autonomous vehicles.

It not a question of "if" or "when" because driverless cars exist and are operating on a very limited scale. With every day that goes by, that scale will increase.

I'm a bad predictor of these things, but I'm pretty certain inside of 5 years it will not be uncommon to see driverless fleets in many major US cities. I think internationally more than a few countries will take to driverless cars like bees to honey (Japan comes to mind). I think 5 years after cars that have an autonomous option are available to the general public is when they will take off and the graph will go hockey stick.

Relatively wealthy people will buy them and will not lose a moment of sleep because they don't always manually control their cars. The benefits will be obvious and incontrovertible, just as the benefits of having (sorta) ubiquitous access to data and voice networks in the form of, first, cellphones and, later, smartphones, was.

Also, upthread someone dissed California car culture as the main contributor to unsustainable suburban life but Corbusier's vision of the Radiant City (failed in its specifics) was actually the model for city centers like Chicago and bedroom communities in New Jersey, Connecticut, and the Greater DC area.

In other words, don't you go to blaming only California for our unsustainable and poorly-conceived suburban developments. The blame for these problems goes almost everywhere you can look from municipalities to state legislatures to Federal transportation policy, to oil monies, to your middle-age crisis-fueled hotrod dreams.
posted by mistersquid at 8:03 PM on January 16, 2016 [3 favorites]


This conversation would be a lot less frustrating if people weren't assuming that everywhere in the United States outside of NYC and maybe three other cities was a corporate white supremacist stronghold.

corb, I know your comment counters a misperception of the homogeneity of the sociopolitical landscape of US cities outside the usual cosmopolitan centers, but let's be real here.

Every. Single. City in the United States is a corporate white supremacist stronghold.
posted by mistersquid at 8:09 PM on January 16, 2016 [7 favorites]


Except for wherever Automocar lives, of course, that's why they had to drop some knowledge. Just to make the conversation less frustrating, you understand.
posted by XMLicious at 8:41 PM on January 16, 2016 [1 favorite]


For things like construction, presumably signs of some sort could be created that would send out signals to self driving cars, that would allow them to more easily understand what is going on.

The roads will have to change in certain ways. I'm guessing down to things like the kind of paint used as boundary edges - there are things the government will be able to do with road building and maintenance that will help self driving cars function even better. It'll be a fraction of the cost to society as car accidents.
posted by imabanana at 8:50 PM on January 16, 2016 [2 favorites]


overglow - perhaps I am just as agnostic as you are. I don't see dealing with rain or snow as inevitable, I just don't see failing to do that as erasing the effect of the technology or otherwise as a reason to dismiss it. Even if we were to only get self-driving vehicles that just operate indoors or outside when there's no precipitation, I think they'll still have a profound effect on many aspects of our civilization.

Are you sure about the "we don't have fully automated airplanes" thing? I notice that in the article you linked to Mindell specifically talks about commercial airplanes.

Also, I'm not a drone hobbyist myself, but an acquaintance of mine had a quadcopter which was controlled over WiFi and claimed that if it lost the network connection it automatically returned to the place it was launched from and landed.

The stop light thing is certainly interesting, if they can determine the signal displayed by a stop light marked on their map but not recognize one independent of the map. (Also kind of ironic that Google's cars are shown braking and stopping in response to a variety of obstacles and potential problems, but a stop light is the thing they might not get.)

I'm a software engineer, but I haven't done any work in computer vision.
posted by XMLicious at 9:36 PM on January 16, 2016 [1 favorite]


Driverless cars will help you if you're
  • Old enough to take care of yourself but too young to drive
  • Young enough to take care of yourself but too old to drive
  • Disabled in any of various ways
  • Otherwise without a license for any of various possible reasons
  • Constantly spending your time and energy driving other family members to and from things
  • Commuting your life away instead of doing something productive when you're in the car
So they will come as soon as they are affordable, and poor people will benefit from driverless mass transit.

But don't expect people to voluntarily give up owning cars. A private car is an essential storage room, staging area, and safe room for a lot of people. Personally owned passenger vehicles may sit idle much of the time, but so do most of your clothes and probably most of the rooms in your home and most of the belongings in them. They are all in reserve for you.
posted by pracowity at 7:57 AM on January 17, 2016 [5 favorites]


Stop subsidizing car-based lifestyles and the problem of car ownership and car-based lifestyles will solve itself.
posted by entropicamericana at 9:42 AM on January 17, 2016


I agree with emptythought's opinion that insurance costs would skyrocket for manual cars once self-driving cars become widely available.

This simply makes no sense. Assume that diverless cars are perfect and never have an accident. Then drivers will be responsible for 100% of accidents -- exactly as they are now. Nothing has changed.

Let's say that the average driver causes $1000 worth of damage claims each year -- exactly as they do now. Nothing has changed.

Let's say that insurance companies charge a premium of $1200 per year to cover the damage claims and provide a profit -- exactly as now. Nothing has changed.

Insurance companies will charge the same premiums as before and have the same profits as before -- exactly as now. They have the same incentive to sell insurance and make the same profits. Nothing has changed.

So what change will occur to cause insurance rates to skyrocket?

There may be lots of pertinent arguments about the feasibility or desirability of driverless cars, but this red herring about soaring insurance rates is not one of them. If anything, you would expect insurance rates for drivers to go down because there are fewer error-prone drivers on the road to run into.
posted by JackFlash at 12:15 PM on January 17, 2016 [1 favorite]


Insurance companies will charge the same premiums as before and have the same profits as before -- exactly as now

How? If 50% of people switch to self-driving cars, insurance companies will make half as much as they used to, unless they increase premiums. (Although Eyebrows' point about competition is a good one.)
posted by desjardins at 12:38 PM on January 17, 2016


If 50% of people switch to self-driving cars, insurance companies will make half as much as they used to.

Ah, I see your confusion here. You are confusing revenues and profits. Whether a business is viable or not depends primarily on profit rate, not revenues. Yes, revenues will go down but that doesn't mean they need to raise premiums. What that means is some insurance companies may exit the market.

In the example I gave above, insurance companies will still make a 20% profit. Even if some insurance companies go away, other insurance companies will still be happy to sell insurance at the same premium if they still get the same 20% profit rate. If a 20% profit rate is attractive now, it will be the same in the future. The auto insurance industry will be smaller, which is a good thing, but the remaining companies will still be happy to sell insurance for $1200 same as before.
posted by JackFlash at 12:56 PM on January 17, 2016


I would expect the driverless cars to expand the range of their maps, by being partially manual at first. Maybe there will be three ways they can navigate:

1. Completely automated.
2. Completely manually driven.
3. Remotely driven by a human in a call center somewhere, via cameras and things.

So say you want to get an automated car to come to your house, and stop in your driveway to pick you up. The first time, you have to drive it there yourself. Maybe several times if your driveway is particularly challenging. But as it goes, it maps to the central repository, and eventually the maps are good enough that it can be automated most of the time, with occasional input from the manual driver at the call center, at which point you don't need to be a licensed driver to call it. Perhaps at first only small cars can get there automatically, but if you order a bigger vehicle it still needs some human help.

And this is how the maps get built out - they're built by use. Because people go to these places, and with repeated visits, the database is expanded. But until the system has enough info in the maps, to have confidence that a given vehicle can get in and out of whatever the spot is, you can't order one to that spot unless you have a driver's license.

Occasionally a car will go somewhere it can't get out of and the company will have to send someone to retrieve it. Once again so what? The retrieval might be a company driver, or it could just be a tow truck from the existing AAA dispatch system. Waiting for a professional to get you out of the ditch is not a new experience unique to driverless cars.
posted by elizilla at 1:16 PM on January 17, 2016 [2 favorites]


If 50% of people switch to self-driving cars, insurance companies will make half as much as they used to.

If 50% of people switch to self-driving cars, insurance companies will charge the other people (the self-segregated dangerous half) more.

But, yeah, eventually very few people would have the money to drive manually, and the practice would just be abandoned and outlawed as an unnecessarily dangerous exception to common sense. Just as you can't drive a horse and buggy down the highway, you shouldn't be able to zoom around in a manually driven car, endangering the rest of us as we patiently wait for our safe driverless cars to get us to our destinations.

Unless money, of course. Still, with very few actual drivers on public roads, the rest of us would be much safer.
posted by pracowity at 2:37 PM on January 17, 2016


If 50% of people switch to self-driving cars, insurance companies will charge the other people (the self-segregated dangerous half) more.

That is not true. Insurance companies in a competitive market cannot raise rates arbitrarily. If an insurance company made a profit selling a policy for $1200 before, they will do the same after. Anyone who tries to raise prices will simply lose customers to those companies that don't raise prices.

Look, if a car dealer complains that revenues are going down because they aren't selling enough cars -- they don't raise their prices, they lower them. If they were to raise their prices they would lose even more sales to the car lot down the street.

the self-segregated dangerous half

They are no more dangerous than they were before, therefore no need to raise rates. Don't confuse the possibility that driverless cars may have lower rates because they are less risky with the fact that drivers will be no more risky than before and therefore have the same rates as always.
posted by JackFlash at 3:20 PM on January 17, 2016 [2 favorites]


This house proposes that we nationalise Uber!

i'd say, maybe not nationalize, but why not have municipal level services -- like expanding TriMet's ride sharing services to everyone and providing an app for fleet vehicle access and scheduling -- whether driverless* in the future or not? that is, actually making the library an economic model for a true sharing economy in the public interest: Putting powerful platforms under cooperative control :P
At the Platform Cooperativism summit... several hundred people gathered to talk about the problems of an online economy reliant on monopoly, extraction, and surveillance—and discuss how to build a "cooperative Internet, built of platforms owned and governed by the people who rely on them."
re: simulations/models, not sure what the state of it is now exactly, but TRANSIMS, which came out of LANL, appears to be open source and i think is still being used by DOT:
The TRansportation ANalysis and SIMulation System (TRANSIMS) is a set of travel modeling procedures designed to meet the State DOTs' and MPOs' need for more accurate and more sensitive travel forecasts for transportation planning and emissions analysis.

TRANSIMS outputs detailed data on travel, congestion, and emissions; information that is increasingly important to investment decisions and policy setting. Because TRANSIMS simulates and tracks travel by individuals, the benefits to and impacts on different geographies and travel markets can be evaluated as well. Furthermore, TRANSIMS has the capability to evaluate highly congested scenarios and operational changes on highways and transit systems.

TRANSIMS is based on four primary modules: population synthesizer, activity generator, route planner, and traffic microsimulator. Using these components, TRANSIMS estimates activities for individuals and households, plans trips satisfying those activities, assigns trips to routes, and creates a microsimulation of all vehicles, transportation systems, and resulting traffic in a given study area.

TRANSIMS differs from previous travel demand forecasting methods in its underlying concepts and structure. These differences include a consistent and continuous representation of time; a detailed representation of persons and households; time-dependent routing; and a person-based microsimulator. These advances are producing significant changes in the travel forecasting process. To date the TRANSIMS models have been tested with data from Dallas, Texas, and Portland, Oregon.
i'm guessing urban engines (or google, IBM, etc.) are combining/training this, or something like it, with 'big data' from people's cell phones, traffic cams, etc. for the consumption of city planners, urban designers and the like?

also btw...
-Cars and the Future
-Ways to think about cars
-Anil Dash: Tech is failing us
-Obama plans $4bn boost for driverless cars
-Why Detroit is moving to Silicon Valley (map)
-The transformative potential of self-driving electric cars

---
*@elonmusk: "In ~2 years, summon should work anywhere connected by land & not blocked by borders, eg you're in LA and the car is in NY"
posted by kliuless at 4:37 PM on January 17, 2016 [2 favorites]


They are no more dangerous than they were before, therefore no need to raise rates.

independent of if/when driverless cars take the road en masse,* DOT is getting closer to mandating 'V2V' communications technology: "This is what connected vehicle technology can do. It promises to eliminate 80 percent of accidents in which drivers were not impaired."
U.S. Department of Transportation Secretary Anthony Foxx said the agency, along with the National Highway Transportation Safety Administration, will fast-track a proposed rule that would require all new cars, even those that need a driver, to have vehicle-to-vehicle communication, known as V2V...

The DOT will also speed up testing on the 5.9 GHz spectrum reserved for V2V to ensure wireless communications from connected cars aren’t obstructed by radio interference. Testing will be completed within one year of receiving production-ready devices, Foxx said in a blog post released at the time of the event.

Combined, these commitments will accelerate the introduction of V2V and vehicle-to-infrastructure systems, Foxx said. Vehicle-to-infrastructure, or V2I systems will allow automobiles to wirelessly communicate with roads and traffic signals. The regulatory framework will be designed to encourage innovations that increase traffic safety, Foxx added.
more from DOT fwiw...
-Communities Helping Communities: Bike-Ped case studies from FHWA
-Smart City Challenge seeks the best in integrated transportation technology for mid-sized cities

And this is how the maps get built out - they're built by use. Because people go to these places, and with repeated visits, the database is expanded.

here's how musk describes tesla's 'fleet learning network': "The whole Tesla fleet operates as a network. When one car learns something, they all learn it... [each driver using the autopilot system essentially becomes an] expert trainer for how the autopilot should work."

---
*Google steps up hunt for self-driving car partner - "Krafcik gave a non-committal answer when asked whether he had recently been in meetings in Detroit's Renaissance Center, General Motors' headquarters... Krafcik, a veteran of GM, acknowledged the significant gaps in automotive know-how that Google would have to fill to build cars... It was too early to say whether Google would work with multiple carmakers — or OEMs — on several vehicles or with a single one on one model, Mr Krafcik said."
posted by kliuless at 6:19 PM on January 17, 2016 [1 favorite]


"The whole Tesla fleet operates as a network. When one car learns something, they all learn it...

Elon Musk = Bolivar Trask, it's all so clear to me now!
posted by entropicamericana at 7:02 PM on January 17, 2016


pracowity: "If 50% of people switch to self-driving cars, insurance companies will charge the other people (the self-segregated dangerous half) more.

Except that insurance is an extremely competitive (I tried to count car insurers in Illinois but gave up when I got to 80) market, and insurance is a highly regulated industry, in which states already impose limits on profits for some types of insurance. If the market ceases to regulate costs adequately, the states WILL step in and impose limits on the profits car insurers can take.

Just as you can't drive a horse and buggy down the highway, you shouldn't be able to zoom around in a manually driven car, endangering the rest of us as we patiently wait for our safe driverless cars to get us to our destinations."

You can TOTALLY drive a horse and buggy on the highway, just not on the interstate. If you happen to visit the 24-hour Meijer in South Bend at 4 a.m. because possibly you're still drunk (YOU DON'T KNOW MY LIFE), you'll get to see all the Amish horse-drawn buggies parked there so the Amish can shop before the roads get busy and the farmwork for the day starts. They come up on the state highways in their buggies. It's moderately terrifying if you're the car because the buggies are black and the local Amish don't believe in reflective triangles, but they're all over. Horses and buggies, combines, and other odd vehicles on the highways are particularly common in rural areas, where people are probably most likely to continue needing to drive regular cars. There will be similar accommodations and social norms. These are highways posted 55 or 65 mph, not slow little county roads.
posted by Eyebrows McGee at 7:11 PM on January 17, 2016


I did that exact thing in that exact meijer not more than a few weeks ago. Only I wasn't drunk, just insomnia ridden and needing some god damn carpet cleaner. Decent amount of amish, guess that explains why there's more horse shit in that parking lot than one would expect.
posted by Ferreous at 8:12 PM on January 17, 2016 [1 favorite]


American suburbs are not some ideal state of human civilization

WHAT someone needs to make a movie about this
posted by grumpybear69 at 8:22 PM on January 17, 2016


laptop shows the world as the car sees it, a wireframe representation of the area that depicts all the objects around the car: pedestrians, trees, road signs, other cars, motorcycles—basically everything picked up by the car’s radar and laser sensor

The cars are still dependent on an ultra-high resolution onboard map which defines all road areas, traffic sign location and types, correct? They are not really reading all the information needed to form x_view in real-time, which is what all these articles imply.
posted by benzenedream at 12:11 AM on January 18, 2016 [1 favorite]


So, driverless cars might save lives, but won't that mean fewer healthy viable organs get donated to those in need? I wonder whether that might lead to an opt out rather than opt in policy for organ donation.
posted by peppermind at 3:46 AM on January 18, 2016 [2 favorites]


That is not true. Insurance companies in a competitive market cannot raise rates arbitrarily. If an insurance company made a profit selling a policy for $1200 before, they will do the same after. Anyone who tries to raise prices will simply lose customers to those companies that don't raise prices.

I'm not so sure. An insurance company make make a profit on a $1,200 policy when 1/3 or so of the nation's population has to buy $1,200 policies. If the demand for insurance drops drastically, you may get considerable industry consolidation until the market is cut down to just a few strong and nimble players who may be able to establish oligopolistic market power.

Insurance is unlike other industries, also in that what is being priced is shared risk. Say for the sake of argument in a driverless car world cars are much safer and the majority of people buy memberships in a car service and a covered under the blanket policy for that service. So what we have today is say, 100 million people buying insurance, at an average rate of $1,200 a year, for $120 billion in industry-wide revenue. From this you have to deduct the cost of accidents and overhead and then you're left with your profit. In driverless world, you might have only 50 million people needing insurance. The number of accidents overall might be more than halved --- but it's likely that in the vast majority of them your customer will be at fault, so "cost of accidents" doesn't necessarily get halved for the insurance industry. Not sure what happens to cost of overhead. So maybe that just means that car insurance is a less profitable business; few would weep, if so. (Warren Buffet, maybe.) But I can think of ratios where yeah, insurance rates would definitely go up for the remaining manual drivers.
posted by Diablevert at 4:52 AM on January 18, 2016


The number of accidents overall might be more than halved --- but it's likely that in the vast majority of them your customer will be at fault, so "cost of accidents" doesn't necessarily get halved for the insurance industry.

Yes, cost of accidents does get cut in half by simple math. "In the vast majority of them your customer will be at fault." People seem to be hung up on this thing. But in fact, in the vast majority of them your customer is at fault now, in fact 100% of them.

The math is, no matter the number of drivers, each average driver still has the same number of accidents each year as before (all else the same). If today the average driver's claim costs are $1000 per year, they will still be $1000 per year with half as many drivers. So they charge each driver $1200 per year now and likewise $1200 per year with half as many drivers to cover overhead and profit. Nothing has changed.

This discussion about insurance seems to be going like the Monte Hall problem. The idea that driver's will be found at fault if driverless cars are never at fault is leading people astray.
posted by JackFlash at 9:48 AM on January 18, 2016 [2 favorites]


The cars are still dependent on an ultra-high resolution onboard map which defines all road areas, traffic sign location and types, correct? They are not really reading all the information needed to form x_view in real-time, which is what all these articles imply.

Pretty sure they are reading it from the environment in realtime. The map reduces the overhead and uncertainty substantially - it lets the car know that the apartment block over there is a stationary fixture that will not enter traffic, but the map cannot and does not substitute for scanning and tracking pedestrians, traffic, reading traffic lights, etc. The cars are scanning everything (and AFAIK they are updating the shared map and/or mapping previously-unmapped areas) as they drive.
posted by anonymisc at 2:30 PM on January 18, 2016


Department of Transportation going full speed ahead on self-driving cars
Secretary Foxx wants to develop consistent autonomous car policies across the US.
At this week's North American International Auto Show in Detroit, Transportation Secretary Anthony Foxx announced that "in 2016, we are going to do everything we can to promote safe, smart, and sustainable vehicles. We are bullish on automated vehicles."
posted by XMLicious at 10:14 PM on January 18, 2016 [2 favorites]


"Bullish on automated vehicles" - so cutting funding for proven solutions like public freaking transportation and giving the money to Silicon Valley twerps. Well, that's just awesome.
posted by koeselitz at 1:09 AM on January 20, 2016 [1 favorite]


benzenedream: "The cars are still dependent on an ultra-high resolution onboard map which defines all road areas, traffic sign location and types, correct? They are not really reading all the information needed to form x_view in real-time, which is what all these articles imply."

Exactly - which is why this terrible idea will never work, and only exists to suck public money into the pet projects of billionaires.
posted by koeselitz at 1:12 AM on January 20, 2016


If today the average driver's claim costs are $1000 per year, they will still be $1000 per year with half as many drivers.

Took me a couple days, but I figured out what was bugging me about this. I don't think that's right.

Because:

Average claim cost= Total claim costs/number of drivers.

If there are fewer accidents, the numerator decreases. But the average claim cost still increases if the denominator shrinks as well. 100 billion in costs spread over 100 million divers=$1000 average claim cost. 50 billion in costs spread out over 25 million drivers =$2,000 average claim cost.

koeselitz -- exactly what public money do you think is at stake here? Google's spending its own money, Ford and GM and Uber are spending their own money. Tesla ditto. They all think it's a viable commercial tech, they're not begging for handouts to fund the research.
posted by Diablevert at 2:03 PM on January 20, 2016


If there are fewer accidents, the numerator decreases. But the average claim cost still increases if the denominator shrinks as well. 100 billion in costs spread over 100 million divers=$1000 average claim cost. 50 billion in costs spread out over 25 million drivers =$2,000 average claim cost.

This implies the rate of change is equal among both.

But you're likely to see more licensed drivers who don't drive much. Some will want to keep their old car, even if it's something of a luxury or back-up plan.

Even non-drivers will want to be insured. Either they own a smart vehicle that can be damaged or stolen, or they want to be covered in case one is damaged or stolen while in their care. Their insurance will be lower to reflect their lower risk profile, but it won't be zero.

Plus, you'll likely see a decrease both in damage per mile driven *and* cumulative miles driven. This also makes it more likely that numerator will decrease at a larger rate than the denominator.
posted by politikitty at 3:26 PM on January 20, 2016


koeselitz: "Bullish on automated vehicles" - so cutting funding for proven solutions like public freaking transportation and giving the money to Silicon Valley twerps. Well, that's just awesome.

I first read an article that mentioned $4 million and I thought "that's not so much. But that was a typo -- Secretary Foxx Unveils President Obama’s FY17 Budget Proposal of Nearly $4 Billion [over a decade] for Automated Vehicles and Announces DOT Initiatives to Accelerate Vehicle Safety Innovations.

While that's a serious chunk of change, the new five-year federal surface transportation authorization law, Fixing America's Surface Transportation (FAST) Act provides $305 billion for highway, transit and railway programs. Of that, $233 billion is for highways, $49 billion is for transit and $10 billion is dedicated to federal passenger rail. The allocations are split so that the amounts increase each year, and the linked article provides some more details.

People are lazy and selfish, so to make transit really viable in the US would require a significant shift in policy, which takes bold leadership (hah) or a significant change in public demand (ha ha) to push policy.

Unfortunately for public transit, it's on-demand services like Uber and Lyft that are changing public demands with some populations, as seen with people wanting car rentals to be more like app-driven services instead of rental-lot focused. (The discussion of client transactions for car rentals vs Uber bothers me, because car rentals are typically multi-day or at least multi-stop uses, whereas Uber is more akin to taxi services, providing origin-to-destination only, with some exceptions where people ask the driver to wait while they do something quick and then go to a 3rd location, but I digress.)
posted by filthy light thief at 7:27 AM on January 21, 2016 [1 favorite]




filthy light thief: "People are lazy and selfish, so to make transit really viable in the US would require a significant shift in policy, which takes bold leadership (hah) or a significant change in public demand (ha ha) to push policy."

Yeah – and if I have an air of fatalism about this, that's why. I still believe that using tiny boxes-for-hire to transport individual people around is a waste, even if those tiny boxes-for-hire drive themselves and don't require humans at the helm.

Another part of my cynicism is this: as far as I can tell, actually making automated vehicles viable across all roads in the US (not just highways) would require massive infrastructure spending and some paradigm shifts concerning how we build and maintain roads. To make road construction areas navigable by automated vehicles, we need policies governing road construction areas to be sharpened and standardized; as it is, road construction zones in different states vary wildly, and there's no way for an automated system to predict how they're going to work. That's just one area where automated vehicles are going to need a lot of help before they're viable – others are things like road closures and manually-directed intersections, neither of which automated vehicles can handle, and neither of which they'll be likely to handle until they become predictable in a standard way.

So while $4 billion seems like a healthy start (even though it thankfully is dwarfed by those planned transit budget numbers) it's not really enough. Furthermore, I'm not sure any money would be enough. As I said, this stuff is going to take paradigm shifts – making roads truly automated-vehicle friendly means changing actual roadway policies at the local and state levels. That's a problem you can't really throw money at, except maybe if you throw it at lobbyists; legislatures across the US are going to have to take it up if we want to make this work.
posted by koeselitz at 1:19 PM on January 21, 2016


Makers of driverless cars want pedestrians and cyclists off the roads

Well no, it's the CEO of Renault talking out of his ass.
posted by GuyZero at 1:20 PM on January 21, 2016


anthill: Makers of driverless cars want pedestrians and cyclists off the roads

Sadly, they're not alone - I've seen federal and state transportation guides and references talk about bicyclists as "hazards" for drivers. How quickly car culture has become the norm - streets are for everyone (for the most parts - some freeways are exclusively for motorized vehicles).


koeselitz: I still believe that using tiny boxes-for-hire to transport individual people around is a waste, even if those tiny boxes-for-hire drive themselves and don't require humans at the helm.

Oh, I agree, too, but some people value even small amounts of time "saved" by driving their own vehicle, even when the cost to take public transit at a similar speed is a fraction of the cost. And there's not as much money to be made in public transit as compared to selling new cars that will be driven (or ridden) by a single person for a fraction of the day, then left idle the rest of the time.
posted by filthy light thief at 1:39 PM on January 21, 2016 [1 favorite]


even when the cost to take public transit at a similar speed is a fraction of the cost.

When is this ever true outside of really big cities with train/subway? It would take at least 3 times as long for me to take public transit to work and I live less than 5 miles away.
posted by desjardins at 6:30 PM on January 21, 2016 [4 favorites]


Yeah, I'm pretty sure that driving your own car is almost always going to be faster than public transit. The seven mile trip to my office takes me about 25 minutes to drive in traffic but the two buses I'd need to ride would take almost an hour.
posted by octothorpe at 8:00 PM on January 21, 2016


The entire point of subways is that they don't have to deal with traffic or stoplights and even surface streetcars are no worse than car speeds through heavy traffic.

But those both require very expensive infrastructure investment which doesn't make sense outside of extremely dense urban areas which comprise a really tiny fraction of most countries by land area.
posted by GuyZero at 9:47 PM on January 21, 2016


The slowness with bus travel for for me is that I have to transfer which means a ten or fifteen minute wait in the rain for the second bus and that I have to walk three blocks to my office from the stop. My car takes me from my garage to my office's garage.

They've been talking about a subway to the East End for almost a hundred years now but I doubt that it will ever get funded.
posted by octothorpe at 3:59 AM on January 22, 2016


Yeah, I'm pretty sure that driving your own car is almost always going to be faster than public transit. The seven mile trip to my office takes me about 25 minutes to drive in traffic but the two buses I'd need to ride would take almost an hour.

Nope, that's a result of policy decisions.
posted by entropicamericana at 8:11 AM on January 23, 2016


Nope, that's a result of policy decisions.

The US could stand to have much better public transit. But it's also a young country, and the vast majority of its infrastructure was built post-car. There are huge, huge areas of the country that simply don't have the population density necessary to support a public transit system that would be faster and more efficient than individual car ownership.
posted by Diablevert at 10:09 AM on January 23, 2016 [1 favorite]


It is a result of policy decisions - we could have decided to build things differently to not prioritize cars - but the problem is that now the sprawl is already there. It's hard to undo that damage without significantly uprooting people. What are we going to do, just raze entire suburbs and force people into cities? And the problems are intertwined with others - people moved out to the suburbs in part to escape failing schools and crime and because taxes were lower (I'm not even going to get into racism right now), so school funding, poverty, drug laws, law enforcement, and property taxes (plus a dozen other things) all have to be taken into account in order to begin to address density and transit issues.
posted by desjardins at 12:56 PM on January 23, 2016 [1 favorite]


Is it really just the result of policy decisions though? Aren't there still usually lots of destinations it's quicker even to walk to than take public transportation to in a dense urban area, much less reach by bicycle or by cab or something like that? When I visit friends who live in cities that seems to be the case. I would think that this is a fundamental mathematical network topology result having to do with a large number of possible destinations rather than anything to do with humans.
posted by XMLicious at 4:39 PM on January 23, 2016


What are we going to do, just raze entire suburbs and force people into cities?

Being Americans, we are going to hold off doing the right thing until we've exhausted all the other possibilities. Whether we do this is before we turn the planet is a roasting hellhole is dubious.

If we wanted to act like adults instead of a nation of Verucas Salt, we could stop subsidizing sprawl, stop building sprawl, and let the market solve it, but like I said...

Is it really just the result of policy decisions though?

What else would it be from, magic? We made a decision starting in the 1920s and doubling down in the 1950s to prioritize the car above all other forms of traffic. If a car had to navigate a city where people had the right to the street they once had, there is no way a car would be quicker than other forms of traffic. Cars used to have speed limits of 10-12 hours on city streets. Look at Market Street in 1906— people walk and stand in the street, bicycles and cable cars all mingling. (Note that the same cars appear over and over again—it was an attempt to make SF look more cosmopolitan. Cars will still quite rare.

Aren't there still usually lots of destinations it's quicker even to walk to than take public transportation to in a dense urban area, much less reach by bicycle or by cab or something like that?

Yep! Just like there are lots of destinations where it is quicker to walk than take a car. Multi-modal, baby!
posted by entropicamericana at 7:24 PM on January 23, 2016 [1 favorite]


Cars used to have speed limits of 10-12 hours on city streets.

This is exactly my point: if a car (or a horse-drawn private vehicle, which you seem to be pointedly ignoring in that video) going 10 to 12 miles per hour, more than a century ago, or a bicycle or Segway or motorized wheelchair going the same speed today can reach a destination faster than public transit in even the most densely populated areas, it pretty clearly is not true at all that there is some possible configuration of mass transportation where anything but the tiniest fraction of all destinations can be reached by a succession of public transit methods in an amount of time similar to going directly via individual transportation. Not unless you live somewhere that paramedics can get people to the hospital faster through public transportation than by ambulance.

The town I live in and grew up in was incorporated before the country was foundeded and there isn't any possible sequence of differing "policy decisions" that would have let the people who live in it today reach their destinations by public transit in anything remotely similar to the speeds afforded by at least some kind of private vehicle, unless as desjardins points out you count "doing the right thing" by eliminating all communities like mine and resettling the populations as tenants within more worthy communities like yours as a "policy decision". And even then you'd probably have to build walls or other barriers all over the place to actually keep public transit speeds similar to those of individual transportation methods.

This is true no matter how insistently you refer to any means of travel anywhere that's faster than public transit as "magic". I don't think that lying about basic physics here is going to accomplish much more than climate change deniers lying about basic science does, particularly if it's all in the service of pretending that no one needs to live outside of cities and that people who live like you personally do won't need to make any major changes to deal with climate change, and plus it's all just some unspecified people in the past's fault too. But hey, go ahead and pat yourself on the back for being so adult about it.
posted by XMLicious at 1:17 AM on January 24, 2016 [1 favorite]


It's 7am and I'm on my first cup of coffee and I don't follow you. If you live in an east coast city, you must know it is completely unfeasible for everyone to navigate by a private automobile and that only a series of deliberate policy decisions in the 20th century made it feasible for cars to exist in the city.

You know who needs to live outside of cities? Farmers. Miners. Loggers. How many of those do you know? Non-rural dwellers were 80.7% of the population in 2010. Sadly, a lot of those people live in sprawl instead of environmentally and economically sustainable communities. The fact is, people that don't own cars are subsidizing people who do, so they can pretend to be ranchers in their F-250 or have their minimansion surrounded by our nation's largest crop.

Nobody I know of is suggesting marching your from your crackerbox tract home into the city at gunpoint, we're just suggesting we stop building sprawl and that sprawl dwellers start paying their fair share.
posted by entropicamericana at 7:12 AM on January 24, 2016 [1 favorite]


Since I guess it was so unclear: I have pointed out that your claim that policy decisions are the reason reaching somewhere via individual transportation is so much faster than traveling by a public transit system, is bullshit. It's pretty obvious that unless you simply don't build roads and streets at all, not even the sort of ones that horses would have traveled on a hundred years ago, or actively obstruct other forms of transportation, there's no way that a public transit system is going to be faster between most pairs of locations.

Favoring public transportation has tons of advantages on its own and is worth accepting slower median transit times, as does limiting urban sprawl; there is no need to make things up here to cause them to appear preferable. Yeah, I have always agreed that we should have much higher gas taxes like the rest of the world does as that Atlantic article suggests, but higher gas prices don't make driving a car anywhere take any extra time nor do they make public transit systems faster. The feigned incomprehension of this just gets a bit old, especially while accompanied by broadcasting sneering contempt for people who live differently than you do while congratulating yourself for supposedly being so adult.

It's not the relatively small number of rural people who aren't farmers, miners, loggers, (as if the voracious consumption of cities is actually provisioned by only those three resource extraction activities) workers in the tourism industry staffing places where city-dwellers want to go on vacation and/or own villas in the countryside, and all of the others who make it possible for those kinds of workers to live in actual communities despite their sinful shortcoming of not living in a city, who are the ones who'll be responsible for the planet turning into a roasting hellhole. Far more of the blame is going to fall on the members of that 80.7% who say "welp, I live in a city, did my part" and convince themselves it's entirely the fault of some nebulous Other who can be sneered at for supposedly having a lawn and being the sole beneficiary of fossil fuel use.

It's like the conservative penchant for blaming government expenditures entirely on some nebulous person who is quite like them but receives particular sorts of benefits, while ignoring all of the ways that oneself and one's social circle and one's favored institutions all benefit from government expenditures, and dismissing any single thing pointed out as being necessary unlike the profligate consumption of the Other that's the real cause of all the problems.

tl;dr But apart from serving as a psychological device in that way or as a justification for eliding basic facts about reality, yes the complicity of car-and-truck-transport-related policy in urban sprawl (most noticeable in areas that were primarily built up during the 20th century, as you say, but it certainly exists elsewhere as well) and other issues is worthy of condemnation.
posted by XMLicious at 9:44 PM on January 24, 2016 [1 favorite]


I still don't see how you make transit faster than driving for the majority of commuters. The two things that slow down any trip by transit are that the bus or train has to make stops along the way to pick up other travelers and that unless you're very lucky, you'll never get a point to point trip. You're almost always going to have to walk to the station and from the station at either end while a car goes exactly where you want it. And that's if you're lucky enough not to have to transfer. I have to walk a few blocks from one bus stop to another in the middle of my trip to transfer to the second leg.

And slowing cars down won't hell, I average only 15 MPH on my commute but I still beat the bus.
posted by octothorpe at 4:48 AM on January 25, 2016 [1 favorite]


I use public transit as my primary form of transportation and I would champion it loud and proud any day... but there's no way it can ever be faster or more convenient across the board. In some select cases it can be, if a commuter rail is given a direct route to where it needs to go, but that's only convenient and faster for those people travelling that exact path.

tldr: Playing the game Mini Metro has disabused me of any utopian notions I have about public transit.
posted by tofu_crouton at 7:28 AM on January 25, 2016 [2 favorites]


The idea that we should judge methods of transit strictly by the question "how quickly does this get me to work in the morning?" is a good example of what's wrong with roadway planning in America today.
posted by koeselitz at 8:35 AM on January 25, 2016 [1 favorite]


(Another problem: the idea that the question "how quickly does this get me to work in the morning?" is identical to the question "how quickly does this get everyone to work in the morning?" We only need to look at this thread to see that confusion consistently. Every time someone says "most commuters are going to have X experience," they follow it up with "why, just look at me, I have X experience!" No data here, apparently – just anecdotes.)
posted by koeselitz at 8:38 AM on January 25, 2016


Reality has a known urban bias.

Comparing US transit as it exists, having been woefully underfunded for decades, to commuting via private auto (which has been overfunded for decades) is idiotic.

I grew up in the country. I love the country. That's why I live in the city.
posted by entropicamericana at 9:29 AM on January 25, 2016


If you live in an east coast city...

This, right here, is the problem: bourgeois urbanites assuming that their experience either is universal or should be universal.
posted by corb at 9:30 AM on January 25, 2016


I live in Omaha and manage just fine without a car.
posted by maxsparber at 9:56 AM on January 25, 2016 [1 favorite]


Statements that one assumes are universal usually aren't prefixed with an "if", nor do they usually include statistics that explicitly point out that the thing they're talking about isn't universal.
posted by tonycpsu at 10:37 AM on January 25, 2016


If you're in or around San Francisco this summer and want to geek out about automated vehicles, Transportation Research Board is hosting Automated Vehicles Symposium 2016, a Workshop on the Future of Road Vehicle Automation on July 19-21, 2016.
posted by filthy light thief at 7:58 AM on February 2, 2016


This, right here, is the problem: bourgeois urbanites assuming that their experience either is universal or should be universal.

One could just as easily (and more accurately) say the problem is bourgeois suburbanites thinking everybody else should subsidize their economically and environmentally unsustainable lifestyle.
posted by entropicamericana at 9:51 AM on February 2, 2016


« Older “Randolph Tech,” Mr. Newburger said, making up...   |   brb getting a tattoo that says "The Hammock... Newer »


This thread has been archived and is closed to new comments