This is a cache of https://www.metafilter.com/146782/I-cant-have-you-smart-folk-mucking-up-progress-with-your-silly-dystopia. It is a snapshot of the page at 2019-07-08T02:39:46.949+0000.
I can’t have you smart folk mucking up progress with your silly dystopia | <strong>metafilter</strong>

I can’t have you smart folk mucking up progress with your silly dystopia
February 5, 2015 5:48 PM   Subscribe

I Find Your Lack Of Faith In Autonomous Cars Disturbing [SLMedium] (Previously on the blue.)
posted by joseph conrad is fully awesome (175 comments total) 20 users marked this as a favorite
 
TL;DR (YouTube)
posted by joseph conrad is fully awesome at 5:50 PM on February 5, 2015 [1 favorite]


You know what’d be better than building another bus system? Building an autonomous bus system that could self-optimize its routes and coordinate logistics with all of its customers, service providers and city planners simultaneously.

Sorry, no. Give me a fixed route with frequent service or give me death.
posted by wikipedia brown boy detective at 6:01 PM on February 5, 2015 [24 favorites]


Well, that seemed borderline reasonable for a short while, but then it swerved into an argument against public investment in mass transit (which seems more like par for the Medium course; scratch a techno-utopian and you usually find a libertarian just beneath the surface). So I wonder: does this guy have any background in planning or public policy at all? lol course not he works on web sites
posted by RogerB at 6:03 PM on February 5, 2015 [51 favorites]


I agree that we should support autonomous vehicles. I also think we should support programs to build a fully operational Death Star.

But I kinda feel like we should put more money into the Death Star thing, because that is more likely to happen in our lifetimes.
posted by koeselitz at 6:05 PM on February 5, 2015 [5 favorites]


Huh? google already has autonomous cars. They haven't worked out the manufacturing/distribution/legal stuff, is why you can't buy them yet.

Could've sworn we had a couple threads about them...
posted by LogicalDash at 6:08 PM on February 5, 2015 [1 favorite]


Thread from 2014
posted by LogicalDash at 6:10 PM on February 5, 2015


google Maps (and the GPS in every vehicle I have passengered in) STILL thinks my home address is 100 yards farther south in a vacant lot. Until they get that fixed, I'm not setting my butt down in anybody's autonomous anything.
posted by oneswellfoop at 6:12 PM on February 5, 2015 [13 favorites]


Sorry, no. Give me a fixed route with frequent service or give me death.

Why?
posted by layceepee at 6:13 PM on February 5, 2015 [2 favorites]


LogicalDash: “Huh? google already has autonomous cars. They haven't worked out the manufacturing/distribution/legal stuff, is why you can't buy them yet. Could've sworn we had a couple threads about them...”

google doesn't have autonomous cars. google has machines that can safely carry people in controlled environments if fed excruciatingly detailed information about those environments ahead of time.

They've apparently managed to make this work on a few specific highways – probably largely because highways have fewer changing variables than other roads. And they've managed to allow passengers to ride in these "autonomous cars" in (surprise!) controlled test situations on their own lots.

google will have invented an autonomous car when they get one to carry somebody from Los Angeles to midtown Manhattan on a road trip. Or even outside of controlled circumstances in California and on a few closed campuses elsewhere. Until then, this is a silly pipe dream. Roads are extraordinarily complicated: the variables on them change daily, new signs get put up, temporary signals are erected, etc. These are more things than any "autonomous car" has ever been able to manage.

My pet theory is that google's "autonomous car" program exists for two reasons: first, it's a good testing ground for next-generation mapping quality; second, it's great PR.
posted by koeselitz at 6:14 PM on February 5, 2015 [31 favorites]


Mycroft was similar. In fixed lines that is.
posted by clavdivs at 6:15 PM on February 5, 2015


google Maps (and the GPS in every vehicle I have passengered in) STILL thinks my home address is 100 yards farther south in a vacant lot. Until they get that fixed

You have used google Maps' Report a Problem link to tell them where your house really is, right?
posted by radwolf76 at 6:20 PM on February 5, 2015 [2 favorites]


Sorry, no. Give me a fixed route with frequent service or give me death.

Why?


Detours mess up my naps. My subconscious knows when to wake me up to get off the bus based on what turns we've made.

Also, with a fixed route, I know more or less where I'll be on the route at any point in my trip, so if I need to detour myself (say, I want to get off and meet someone for dinner or whatever) it's easier to plan.

Also, I like my transit buddy acquaintanceships that I probably wouldn't get to keep if every day were a totally different route. There's something pleasant (to me) about seeing the same faces while I'm out and about.

Also, and most of all, I'd rather invest the money in something we have the ability to do well Right Now with the tech, laws, and resources we have.
posted by asperity at 6:24 PM on February 5, 2015 [12 favorites]


~koselitz: Right. You could have autonomous cars, provided you don't mind google establishing a surveillance regime that would make the NSA blush. Which, given the Valley ethos of "freedom for me but not for thee", might be what google is after.
posted by Cash4Lead at 6:25 PM on February 5, 2015 [2 favorites]


Oh and yes, BTW, that's fraudulent for-profit university company EdMC on young Turck's resume. So he's been into destroying public infrastructure for profit for a while, but he's switching sectors from education to transportation because the cars have a better sci-fi sales pitch.
posted by RogerB at 6:27 PM on February 5, 2015 [10 favorites]


Here's a video from 10 months ago, showing a visual representation of the google car's awareness of the world. You can see it recognizing cars, bicyclists, and pedestrians, planning its path through the environment, etc.

The car may now be reliant on on pre-scanned map data, but there's no reason that it will not eventually be possible to deploy sensors that can read the road far better than a human driver.
posted by rustcrumb at 6:27 PM on February 5, 2015 [9 favorites]


After years of trying to commute in the Chicago area I would welcome autonomous cars with open arms if only because I'm so sick of driving in the insane traffic. This would allow me to read! Add to that the potentially reduced number of sleepy drivers on the road and I would welcome it with open arms. Add to that a potential increase in commute speed and I would be so ecstatic I would gladly kiss the inventors... on the lips... tongue optional.

I hope I see this in my lifetime.
posted by Jernau at 6:31 PM on February 5, 2015 [7 favorites]


Ctrl+F employment

Huh.
posted by phrontist at 6:32 PM on February 5, 2015 [4 favorites]


Why do techno-utopian articles always have to be written in such a smug tone? It's like I'm being lectured by the college libertarians I know. Criminy, it's like it's impossible for projections to be wrong, for costs to be underestimated, for there to be the sorts of unforeseen flaws and faults that show up in any system - and if they're acknowledged at all (which is unlikely), it's always met with the same "well, humans are pretty flawed" statements.

It's not that I am really that opposed to autonomous cars, it's that I'm creeped out by this techno-libertarian, Silicon Valley belief in the purity of markets and technology to solve every life problem - and if you don't agree, then it's literally laughable. This guy literally calls objections to his ideas "hilarious."
posted by teponaztli at 6:34 PM on February 5, 2015 [29 favorites]


After years of trying to commute in the Chicago area I would welcome autonomous cars with open arms if only because I'm so sick of driving in the insane traffic. This would allow me to read! Add to that the potentially reduced number of sleepy drivers on the road and I would welcome it with open arms. Add to that a potential increase in commute speed and I would be so ecstatic I would gladly kiss the inventors... on the lips... tongue optional.
I hope I see this in my lifetime.


I get this in Chicago already if only the cars would get out of the way. It's called the CTA.
posted by srboisvert at 6:35 PM on February 5, 2015 [8 favorites]


rustcrumb: “The car may now be reliant on on pre-scanned map data, but there's no reason that it will not eventually be possible to deploy sensors that can read the road far better than a human driver.”

google has openly admitted that mapping the whole country this way would be extraordinarily inefficient, and would involve a cost far beyond what they'd ever be willing to put into it. Since their design is almost entirely map-based, it seems unlikely that this project can ever go anywhere.

When I hear that somebody has an "autonomous car," but that it needs super-detailed information about an environment before driving safely, I question whether that car is really "autonomous." google's cars have some amazing stuff, sensor-wise – they do promise that they can manage even unmapped stop signs, and I believe them there – but they're so far from truly autonomous vehicles that I think it's silly for us to be hoping and dreaming that this is a thing that could happen any time soon. Moreover, I think we should be clear-headed in approaching google's bluster about this, seeing it for what it is: marketing.

I would love to have awesome self-driving cars on our highways. But it's not going to happen on this side of 2050. We really need to focus on the here and now, and look at the admittedly mundane but nevertheless utterly essential solution we already have in front of us right now: public transportation. It may not be sexy; it may not be shiny; but it will fix the problems we have now and allow us to get where we need to go.
posted by koeselitz at 6:36 PM on February 5, 2015 [17 favorites]


Well, carjacking would become a one-sided argument.
posted by clavdivs at 6:38 PM on February 5, 2015


Buses aren't disruptive though.
posted by Ferreous at 6:38 PM on February 5, 2015 [23 favorites]


NPR: The most common job in every state

Spoiler: it's "truck driver" in, like, 3/5ths of the country.
posted by Rhaomi at 6:39 PM on February 5, 2015 [9 favorites]


ALSO: Do we really want to be aiming our future at replicating the fuel inefficiency of single person transports? Sure you can handwave away that with "well we'll have XYZ new fuel sources" but maybe we wont.
posted by Ferreous at 6:40 PM on February 5, 2015 [4 favorites]


So you're telling me this essay isn't a Swiftian satire?
posted by figurant at 6:44 PM on February 5, 2015 [8 favorites]


"OPTIMUS PRIME CAME HOME AND GOT A DAY JOB".
posted by clavdivs at 6:45 PM on February 5, 2015 [1 favorite]


So I wonder: does this guy have any background in planning or public policy at all? lol course not he works on web sites.

I'll take an evidence-based argument against public transit over baseless snark in favor of it. If you think small/medium cities should invest more in train lines or other permanent fixtures, I'd like to hear actual reasons why. (If by "public transit" you meant buses, well, a fleet of buses fits perfectly into the paradigm of autonomous vehicles.)
posted by Rangi at 6:46 PM on February 5, 2015 [2 favorites]


Do we really want to be aiming our future at replicating the fuel inefficiency of single person transports?

Hopefully not. The ideal would be if most/all of the autonomous cars were also electrically powered.
posted by Rangi at 6:48 PM on February 5, 2015 [1 favorite]


NPR: The most common job in every state

Spoiler: it's "truck driver" in, like, 3/5ths of the country.


I thought of this immediately when I was reading through the FPP article.
posted by codacorolla at 6:49 PM on February 5, 2015 [1 favorite]


A fleet of autonomous buses because what's more utopian than higher unemployment?
posted by evilDoug at 6:50 PM on February 5, 2015 [3 favorites]


Automatic trains are a solved problem.

Just saying...
posted by schmod at 6:51 PM on February 5, 2015 [9 favorites]


ALSO: Do we really want to be aiming our future at replicating the fuel inefficiency of single person transports? Sure you can handwave away that with "well we'll have XYZ new fuel sources" but maybe we wont.

I generally disagree with the article, but he does sort of address this:

...hell, autonomous cars could tell us that autonomous cars are a bad idea. They could tell us the most efficient option for a particular area is one mass-transit vehicle running the same twenty-point route every day, and that cars are totally unnecessary there. The cars generate data, and data has no hidden agenda. As long as we can remember how to listen, autonomous cars will tell us how to change our reality for the better, in ways we’ve yet to imagine.

*shrug* you can poke holes at that but I don't think its the same thing as "aiming our future at replicating the fuel inefficiency of single person transports."
posted by juv3nal at 6:52 PM on February 5, 2015 [1 favorite]


Automatic trains are a solved problem.

Really? I assumed there was some need for a conductor, since the NYC subways still have them. Is that because automating the trains would be too expensive, or simply inertia?
posted by Rangi at 6:54 PM on February 5, 2015


A couple of moments in this article really rubbed me the wrong way. But I liked the argument about public transit. It really is the case that NYC got massively lucky with subways and no one else can match it. Living in DC makes that clear: we have pretty good metros, but they'll never expand to the size of NYC's.

So when we think about truly stupendous public transit projects like that, we should really ask ourselves if they're as cost effective as the boring stuff, like buses with dedicated lanes, or as sexy as the sci-fi stuff like autonomous vehicles.

You won't own a car that drives itself in the next two years, but you very well might in the next ten.
posted by anotherpanacea at 6:56 PM on February 5, 2015


Really? I assumed there was some need for a conductor, since the NYC subways still have them. Is that because automating the trains would be too expensive, or simply inertia?

There are some completely automated commuter trains in the world. There's one in London, I'd be surprised if there weren't more in Europe and Asia. I think there are probably lots of subway trains where the "driver" is literally only there to pull a lever in an emergency.
posted by RustyBrooks at 6:57 PM on February 5, 2015 [1 favorite]


data has no hidden agenda

How do we know what data is important? How do we know what data to expect? How do we know what algorithms to apply? How do we know a solution when we see it?

All these things are decisions made by people, and people are hidden agendas.
posted by the man of twists and turns at 6:59 PM on February 5, 2015 [13 favorites]


Related links from Human Transit: no, autonomous cars will not "abolish transit" in dense cities and Luca Guala: Why "personal rapid transit" evolves into fixed route transit

I'm a lot more excited about the prospect of autonomous cars reducing the rate of fatal road accidents and injuries, which it is my sense are a major public health hazard in the USA.
posted by en forme de poire at 6:59 PM on February 5, 2015 [5 favorites]


RustyBrooks: Cool, you're right. And apparently even in New York, the L train is computer-controlled, with plans for the 7 as well.
posted by Rangi at 7:03 PM on February 5, 2015


There are some completely automated commuter trains in the world.

Not that I've used it personally, but Vancouver's SkyTrain is another pretty famous example; automation lets the service run very frequently throughout the day and into the night for a lot less expense than a staffed train line with comparable service.
posted by en forme de poire at 7:03 PM on February 5, 2015


Rangi: “I'll take an evidence-based argument against public transit over baseless snark in favor of it. If you think small/medium cities should invest more in train lines or other permanent fixtures, I'd like to hear actual reasons why. (If by ‘public transit’ you meant buses, well, a fleet of buses fits perfectly into the paradigm of autonomous vehicles.)”

'Sure, you can say that I should give up trying to gain magical flying abilities and just get a car, but, well, that fits perfectly into my paradigm of magical flying cars.'

In other words: if buses fit into the "paradigm" of the as-yet-impossible autonomous vehicle, that doesn't mean we should pour all our money into something that still seems to be a pipe dream. That means we should pour all our money into buses.
posted by koeselitz at 7:05 PM on February 5, 2015 [1 favorite]


lol course not he works on web sites

From his resume: "Strong advocate of meritocracies and 360° reviews in the workplace. Arch-nemesis of inefficiency."

😒
posted by en forme de poire at 7:06 PM on February 5, 2015 [17 favorites]


google may not have truly automated cars yet, but I think it's a technology that is inevitable in the long term, and research like that done at google, even if their methods turn out to be a dead end, is part of the process of getting there.
posted by Pope Guilty at 7:08 PM on February 5, 2015 [2 favorites]


That means we should pour all our money into buses.

Sure, but since buses are automatable, part of the public money earmarked for "buses" can go towards R&D into automating them. There's no need to spend it on manufacturing as many buses as possible. And of course the private car manufacturers and tech companies can invest their own money in whatever they predict will work, which right now seems to be self-driving and battery-powered cars.
posted by Rangi at 7:12 PM on February 5, 2015


I bring this up in every automated car thread, but just because they're so awesome and yellow ... Morgantown PRT

* it is actually very fun to ride and I wasn't even drunk yet
posted by RobotVoodooPower at 7:17 PM on February 5, 2015 [6 favorites]


Be a lot easier and saner to build a city designed around people than to keep designing everything around cars.
posted by fifteen schnitzengruben is my limit at 7:19 PM on February 5, 2015 [17 favorites]


cities are already built though. Can't really tear it all down and start again.
posted by Ferreous at 7:21 PM on February 5, 2015 [1 favorite]


Of course self-driving cars will become a reality, because the world they will be operating in will have been tailored to make it so. Huge areas will become tightly controlled wastelands of metal and concrete, where it is forbidden to walk, where plants or animals are actively exterminated, and any disruption to the safe and efficient operation of the machines is severely penalized. And to our (grand)children, this state of affairs will seem the most natural and unremarkable thing in the world.
posted by dmh at 7:27 PM on February 5, 2015 [4 favorites]


Who are all these snarky people?
posted by adgl at 7:31 PM on February 5, 2015


huge areas will become tightly controlled wastelands of metal and concrete, where it is forbidden to walk, where plants or animals are actively exterminated, and any disruption to the safe and efficient operations of the machines is severely penalized.

We have this already. They're called highways.
posted by el io at 7:35 PM on February 5, 2015 [31 favorites]


google Maps (and the GPS in every vehicle I have passengered in) STILL thinks my home address is 100 yards farther south in a vacant lot. Until they get that fixed, I'm not setting my butt down in anybody's autonomous anything.

A friend's Garmin insisted he drive a 90 mile round-trip to another city to get to my house.

He was parked outside at the curb.
posted by wallabear at 7:37 PM on February 5, 2015 [3 favorites]


I did not read TFA. Those betting against autonomous vehicle are going to lose in the long run--ten to twenty years IMO. There are too many societal and financial forces favoring them.

All major automotive manufacturers have programs on autonomy. The reasons are demographic. On the upper end there are millions of baby boomers who reaching the age when they should not (cannot) drive. But if you have an autonomous vehicle, it is still worthwhile to buy a car. On the lower end, fewer and fewer young people want to learn to drive. It's expensive, it's dangerous, and it isn't the ticket to freedom that it once was--they can connect with their friends more easily through smart phones and computers. In fact, driving is a distraction--you are not supposed to text or play an online multi-player game while driving. These people will not buy a car that is not autonomous. And why not just put your ten year old in the vehicle to take him/her to school?

Autonomy will emerge gradually. Already there are smart cruise controls, automatic braking when collisions are imminent, self parking. Soon enough there will be sensors for automatically avoiding pedestrians, cyclists, and other vehicles and for not running stop signs and stop lights. You might not even really notice when cars become fully autonomous.

Don't want an autonomous vehicle? At first you won't be required to have one, but insurance companies will strongly disincentive not having autonomous safety features. No autonomy = much higher premiums. Autonomous vehicles will save many, many lives. Drunk? Just have the vehicle drive you home. No road rage, no distracted drivers, no speeding, no reckless driving, no tailgating, no sleepy drivers, and so on. Eventually, I expect that autonomy will be required, just like brake lights and turn signals.

The primary impediments are (1) cost of sensors and (2) the regulatory climate. Computing is cheap. Regarding the former, the cost of these sensors is decreasing day by day, and may just use vision in the future. (Cameras are very cheap.)

The regulatory climate is a different story. Right now when you kill someone it's your fault. Your pockets are finite. But when your autonomous GM vehicle kills someone, it GM's fault, and GM's pockets are very deep. The automotive companies need regulatory protection for them to deploy fully autonomous vehicles.

Deployment of autonomous vehicles will likely be a lot like air bags. When air bags were first deployed, they saved a lot of lives, but they also killed some people, especially children. Improvements in the technology and public information eventually eliminated most of those kinds of incidents. So it will be with autonomous vehicles. They will save a lot of lives, but they will fail sometimes in ways that may not be anticipated.

For myself, I cannot wait. I would much rather read than drive.
posted by haiku warrior at 7:38 PM on February 5, 2015 [18 favorites]


We have this already. They're called highways.

Exactly. Now imagine every road everywhere is like that.
posted by Sys Rq at 7:39 PM on February 5, 2015 [7 favorites]


Here's a basic rule of thumb: Anybody who starts out talking about autonomous vehicles by talking about every single vehicle on the road being autonomous is full of shit.

To start with, when you say "autonomous vehicle", people think "car that drives itself entirely". In fact, the NHTSA has a really nice framework of four levels of autonomous vehicle (AV for short), and those are level 4 vehicles. (Your car is probably level 1, if it has cruise or ABS. google cars are level 3, because they still require driver input some of the time - rarely, but not never.) Or possibly (my terminology now) hypothetical level 5 AV, where vehicles control themselves in real-time with communication with each other. The platoons of vehicles on the highway, the "get rid of traffic signals", and the robotaxi fantasies are all in this bin.

But anyway, so we're talking about a level of technology that doesn't exist, but is near, and I'll grant that. Heck, let's be ambitious and say that there already exists a level 4 AV somewhere in a skunkworks lab covered by a sheet. So what happens next? Well, it takes 3-5 years from concept car to production model, so the first AVs roll off the line with a 2020 production year. But they won't be bottom-end models; car companies put the top features in their premium vehicles first, because that's what makes them premium and that's where they make their profit margin. So it takes a while to filter down to all the cars.

The history of the airbag is instructive; it's similar in that it's a major, expensive safety system. The airbag was invented in 1953, and the first commercial vehicles with airbags were sold in 1970. They didn't work very well, so let's move forward to the first modern airbag, which - no surprise - was in a high-end Benz in 1980. Through the 80s, they became more common, and by 1998, they were standard equipment mandated by law. 45 years after invention, 28 years after the first commercial use, 18 years after the first modern use.

So the first AV is sold in 2020, and with a similar time frame for rollout, maybe the first year all mass-market cars are AVs is 2040 or so. But, of course, it's not like everybody sells their cars every year. And even though all this dude's tech-hole friends may have luxury sedans on 3 year leases, their previous cars are still circulating and circulating. I happen to have a big dataset of vehicles from a major US state, and there are a lot more old cars out there than you think. Maybe not in your part of town, but 2% of the fleet is over 30 years old. (7.5% is over 20 years old.) These aren't (for the most part) antiques; I don't think anyone has great love for the cars of the late 70s and early 80s.

And maybe once 98% of the fleet is AVs you can start talking like all the vehicles are autonomous, but 2% of the modern US fleet is 5 million cars. And of course, no surprise, these cars are disproportionately owned by lower-income households. But in any case, we're now talking about 30 years or so after the full-fleet rollout of AVs, so roughly the 2070 model year.

It's not that we don't need to plan for 2070 - I'm working on a 2076 transportation scenario right now for one client. It's that it's ridiculously too far in the future to make any sort of global pronouncement based on a horizon where everybody talking about it will be dead. It's like the space colony visions from the Sputnik era. Interesting to think about, cool to work on, but you're wasting everybody's time if you think that those moon colonies will save us from the problems of the present day so we shouldn't bother trying.
posted by Homeboy Trouble at 7:39 PM on February 5, 2015 [54 favorites]


Thanks for the airbag history lesson Homeboy Trouble. I learned something, and it's in line with my thinking.
posted by haiku warrior at 7:42 PM on February 5, 2015


The regulatory climate is a different story. Right now when you kill someone it's your fault. Your pockets are finite. But when your autonomous GM vehicle kills someone, it GM's fault, and GM's pockets are very deep. The automotive companies need regulatory protection for them to deploy fully autonomous vehicles.

That's probably not the case. The next step is level 3 autonomous vehicles like the google car, which return control to the driver during times of conflict. Except if you read the articles about google's cars, when they do this, they store all of the data from all of their sensors, and plot a course that shows what they would have done. Excellent engineering practice, but it also happens to produce a near-infinite amount of evidence that the accident did not need to occur - that their algorithm would have done the correct thing - and that it must be your fault. Good luck going to the judge with that, once you've deciphered the terabyte of LIDAR data that google sends your lawyers during discovery.
posted by Homeboy Trouble at 7:45 PM on February 5, 2015


google will have invented an autonomous car when they get one to carry somebody from Los Angeles to midtown Manhattan on a road trip.

Does this mean the non-autonomous car wasn't invented until they got one to carry somebody from Los Angeles to midtown Manhattan on a road trip? The first transcontinental automobile trip in North America was nearly 20 years after Karl Benz built the Benz Patent-Motorwagon.
posted by layceepee at 7:45 PM on February 5, 2015 [3 favorites]


cities are already built though. Can't really tear it all down and start again.

Unless we're talking about people's lives and careers, of course. Those can be torn down as often and as unexpectedly as the kind of people with time to dream about a world full of autonomous cars would like and the market will just take care of transforming all those out of work truck drivers and insurance reps into highly-skilled technical workers in the blink of an eye. Hell, they'll probably make more money as engineers anyway!
posted by saulgoodman at 7:47 PM on February 5, 2015 [2 favorites]


yes, i have been drinking.

So, when you have a car that can drive itself, you don't really need a driver, right?

So, when the lady in the google ad says that she can help her kids with homework, what she really means is "I can work more hours while the car goes to pick my kid up from school. I have to work those hours so I can afford one car for each of my gradeschool kids."

Kids in cars?!?! But that's dangerous!

Oh, obviously kids don't need a flashy car that's fun to drive. They're just going to school after all! Introducing the all new bulletproof, hackerproof kid coffin. It's a box you put your kid in and it drives it to school. It functions as a locker while in school, and a bed when the kid gets home.

Soon, you don't even have to look at your kids! Whcih is good, because you won't have time, what with all those extra jobs you have to work to keep on top of your car payments! anything for the safety of our kids, right?

But, yeah. I guess my point is: cars have been a long, slow slide into the death of our species.
posted by rebent at 7:47 PM on February 5, 2015 [5 favorites]


Airbags might not be the most instructive example, because they didn't represent a fundamentally different economic proposition than not-having-airbags. (Not dying aside).
posted by graphnerd at 7:48 PM on February 5, 2015


koeselitz, I think you are too pessimistic about this technology. Think about what it takes to drive a car safely. Basically, you need to be aware of the terrain, and of the other objects moving around you.

Right now, it looks like the car is doing a really good job of tracking the "dynamic" objects that are in its vicinity. In that video, you can see it reacting appropriately to bicyclists, pedestrians, other vehicles, and stop lights. It's doing this with the spining laser range-finder thing on top.

From the video, it looks like the car is using the pre-scanned map stuff to help it stay in its lane, avoid curbs, etc. If you did not have the maps, the car would have to use an array of cameras to read the lane markers and spot the curbs in real-time, just like a human driver does. The car seems to be doing a little bit of this already, reading road signs and traffic lights, which is not surprising because normal cars have been shipping with this capability since 2008.

So why doesn't the car work without the maps now? I think there's a few reasons. One, it's computationally easier to take an image from a camera plus your appoximate position from GPS, to derive an exact position from the map data. It would take more hardware to be able to assess the terrain in real-time. Two, you can use the map data to cheat on path and speed planning. A robot car with no memory would probably just have to drive a lot more cautiously if it could make no assumptions about things it cannot directly see. Three, using the maps gives the car a margin of safety, which lets the designers test more of the capabilities of the car without worrying that its going to cause a wreck. If you know the maps work 100%, then you can test the car on roads using the systems that will replace the maps, disregarding the conclusions of the untested systems until they finally achieve the level of reliability desired. Like, the camera system might say "it's OK to move ahead," the maps system says "nope, that's a curb," and then you just go back to tweaking the software.
posted by rustcrumb at 7:51 PM on February 5, 2015 [2 favorites]


Give me a fixed route with frequent service or give me death.

Thanks to corporate synergy, google's robots will give you both!
posted by a lungful of dragon at 7:54 PM on February 5, 2015 [2 favorites]


"I know most people are smarter than me." Seems legit. But I'm suspicious.
"I found out by running my mouth too much that these smarter people don't think everything "technology" does is great; and I identify myself with technology so they're really insulting ME!

Well they just have too much HUBRIS to GET IT, maaaaan. So now I will proceed to mansplain something I have a hobbyist background in to these people who I admit are smarter than me. While inoculating myself from criticism by pretending to be self-aware."

I admit I kind of want to find out how he thinks being cautious about the dangers of too much automation is hubris of all things but the opening paragraph is too rich. Also that resume was terrifying. It reads like the hi-def screenshot easter egg of an evil villain's backstory.
posted by bleep at 8:06 PM on February 5, 2015 [2 favorites]


I couldn't deal with this guy's shtick, but I do think autonomous cars are coming. This morning my car drove itself for 20+ miles of my commute. For that distance it braked, steered, and worked the gas while I listened to a book. It wasn't perfect weather, either. It was snowing.

Even the "faux" autonomy of lane assist and adaptive cruise control have dramatically improved the quality of my life. Real autonomy will be amazing, even if it has some limitations.
posted by grudgebgon at 8:10 PM on February 5, 2015 [2 favorites]


I like driving. When I bought my last car, I made a special point to get a manual transmission, and when trying cars I'd evaluate them based on how the clutch and gearbox felt (ultimately went with a Honda Fit, of all things) because it's important to me. It's important to me that I can hop into my car and drive on some curvy roads to clear my head. Driving simulators are fine, but they don't really beat real life. You can't feel g-forces in video games.

The author addressed this, but I think he did it poorly.

"So, that just leaves the most problematic “freedom” fighters: the people who actually, honest-to-goodness love to drive. People like me.

But the simplest answer is often the best, and answers don’t come simpler than this. If America takes away the ability to drive, what will you do? Why, you’ll invest in a chunk of land that can be made into a race track or off-road trail, and charge people to drive recreationally of course."


I don't think investing in private land in order to have spaces to drive is going to replace manual driving. I think if that's going to happen, it's going to gate it off to monied elite elites. It'll basically be like horseback riding. I'm hoping it becomes more like motorcycle ownership, where manufacturers still come out with new stuff and there's a thriving community. They're allowed on roads, and drivers assume risk.

Also, consider this part:

"The automobile is the ultimate American symbol of status and character. Taking away your car is like taking away a part of you.Okay, I buy that. Just do me a favor though: stop reading this, look out into the street, and tell me the car you own is the ultimate expression of who you are. Of all the choices car manufacturers offer so that you can showcase your personality, be entertained, be pampered, be envied, overcompensate for your shortcomings and so forth, your spirit vehicle is a silver 2006 VW Jetta? Doubt it, bro.

The reality is that most of us will never take the leap to own anything less practical than a 2-door, let alone an orange car or a convertible, let alone a Ferrari or Bentley. What your car says most about you is, “I’m nobody.” "


This is after he asserts himself as a "car guy" because he owns a Nissan 370z, and is an auto racing fan. He comes across as incredibly insulting in the whole article, and these points are probably the worst. I am a car guy, if your spirit car is a 2006 Jetta then I get you. Hell, if your spirit car is a frickin toyota camry then I get you.

My spirit vehicle may not be my 2008 Honda Fit but I like it very much. The previous owner took meticulous care of it and he liked it very much. It carried me and my stuff across country. A friend of a friend is taking their beat-up corolla on a 6 year anniversary drive because they like to drive and they appreciate their car. These are not bro-cars, we do not have to compensate for anything, we like to drive. If that's taken away, then we will lose a great experience that can't really be replaced or emulated.
posted by hellojed at 8:15 PM on February 5, 2015 [2 favorites]


hellojed: “I like driving... These are not bro-cars, we do not have to compensate for anything, we like to drive. If that's taken away, then we will lose a great experience that can't really be replaced or emulated.”

I agree that this guy is kind of awful that way. Note that he isn't just a "car guy." He is a guy arguing for the abolition of public transport, to be replaced entirely by a "fleet" of self-driving cars that nobody gets to turn the wheel of or rev the engine of. Basically, he's talking about taking away everything you (as a car person) and I (as a non car person) hold dear.

I have an alternate proposal:

You keep your car. Meanwhile, we invest heavily in public transport, which will both cost less than self-driving cars and be more accessible to people of all classes and incomes. Many people – who would rather not have to drive to work if they could avoid it and if it were convenient – would stop driving, making the roads freer and easier to drive. Easier to drive for you, that is.

And – that way, there would be much fewer accidents! Everybody wins.
posted by koeselitz at 8:23 PM on February 5, 2015 [26 favorites]


You can't feel g-forces in video games.

I had a chance to try out an Oculus Rift roller coaster demo. Going into it, I thought I would feel the same way, that it might look impressive but my inner ear would know I was just sitting in a chair. Not so! I was really surprised how much the simple visual immersion created a sense of actually moving around in a real space.
posted by rustcrumb at 8:24 PM on February 5, 2015 [1 favorite]


The next step is level 3 autonomous vehicles like the google car, which return control to the driver during times of conflict.

What a terrible idea! Aviation experience has shown that humans do an absolutely terrible job of monitoring autopilots that might fail once a year. It would be better to have the computer monitor the human and intervene when the human screws up.
posted by monotreme at 8:24 PM on February 5, 2015 [5 favorites]


well, a fleet of buses fits perfectly into the paradigm of autonomous vehicles

Have you ridden the bus lately? How would you automate stopping at an unmarked stop and not opening the door until there is no cross traffic for the special needs guy who always gets off there? How would you automate helping an old lady up the stairs? Waiting for the guy down the street who is running towards the stop waving his arms? Bus drivers have a lot of functions other than driving the bus, from calling the police when there's a safety issue to helping passengers who need assistance to answering questions from tourists.
posted by Dip Flash at 8:25 PM on February 5, 2015 [10 favorites]


This article's focus is characteristic of someone who doesn't have a lot of serious thoughts about privilege, planning, policy, or the public. It focuses almost exclusively on the technology itself and barely on the massive social and political questions inherent to operationalizing a new form of transportation. At various points he implies that these autonomous fleets are both personal vehicles and a form of transit. Let's be clear about something: the difference between personal transportation and transit is only incidentally about technology or physical form. It is determined, rather, by political/social economics.

How do the people in this fantasy all obtain autonomous cars? If there is no Magic Dad giving us all free cars, what happens to the people who can't afford them? Why are they not accounted for in this fantasy world?
posted by threeants at 8:26 PM on February 5, 2015 [22 favorites]


It's because guys like this don't think about anything outside their own navel.
posted by bleep at 8:30 PM on February 5, 2015 [2 favorites]


I've been driving cars with some degree of autonomy for a while now. I really don't care about solving the edge cases of pedestrians and dogs in the suburbs. Once cars have mastered interstates from on ramp to off-ramp, that will be autonomous enough for me. And we are really, really close on that one. Elon Musk is promising that my Tesla is going to do that pretty soon, and I've driven an Audi and an Infinity that were very very close. The emergency braking assist in most equipped cars is a lot better than the average driver and the steering to stay in your lane is way better than damn near everyone. Automated cruise control to stay a safe following distance has completely changed my attitude towards stop and go traffic.
posted by Lame_username at 8:30 PM on February 5, 2015 [3 favorites]


Have you ridden the bus lately?

Definitely. I think autonomous buses could probably work well in the context of Bus Rapid Transit setups, where they're basically like trains without rails, complete with limited-access fixed paths, defined stations, at-grade entrances to eliminate or reduce the need for passengers to navigate stairs, and a plethora of large system maps and wayfinder signs.

But for a standard local route dealing with the usual sorts of unpredictable circumstances that come with picking people up from places that aren't closed-access at all, human drivers are almost certainly necessary. (My morning drivers usually look down the street to see if I'm sprinting for my stop. Robots will not be set up to do that because it's inefficient, even if we do someday achieve bus driver AI that would allow them to predict human behavior to that degree.)
posted by asperity at 8:37 PM on February 5, 2015 [1 favorite]


I had a chance to try out an Oculus Rift roller coaster demo. Going into it, I thought I would feel the same way, that it might look impressive but my inner ear would know I was just sitting in a chair. Not so! I was really surprised how much the simple visual immersion created a sense of actually moving around in a real space.

If you get on Youtube and search for Oculus Rift rollercoaster videos, there's a meanly hilarious subgenre of videos of people trying to watch the rollercoaster video and remain standing.
posted by Pope Guilty at 8:46 PM on February 5, 2015 [1 favorite]


I kind of want to find out how he thinks being cautious about the dangers of too much automation is hubris

By questioning this thing I believe in wholeheartedly, you are claiming you might know more than I do: hubris!
posted by Steely-eyed Missile Man at 8:50 PM on February 5, 2015


Actually I was thinking about it more and I think it might actually be "You puny human, you think you can question the will of ALMIGHTY TECHNOLOGY*???


Produced and controlled mostly by me, a white male! Pay no attention to the man behind the curtain"
posted by bleep at 8:54 PM on February 5, 2015


Which is basically what you said, except I don't think he's that self-aware.
posted by bleep at 8:56 PM on February 5, 2015


koeselitz: I have an alternate proposal

This is one of the few times in my metafilter life when I really wish I could favorite the hell out of a comment.
posted by Greg_Ace at 9:05 PM on February 5, 2015 [2 favorites]


metafilter: It reads like the hi-def screenshot easter egg of an evil villain's backstory.
posted by mstokes650 at 9:06 PM on February 5, 2015


Yeah, autonomous cars will probably happen, starting with the keeping-up-with-the-very-rich-joneses roll-out discussed upthread and then descending into some sort of disruptive uber techtopian wet dream that conveniently ignores the digital divide. From a personal safety standpoint, this might even be a good thing. And from a "cool, robot cars" perspective I'm stoked.

From another perspective, this just means perpetuating/preserving the massive pollution, inefficiencies, inequalities, and ruptured commons that single-occupancy combustion-fueled vehicles have so reliably produced.

Also what koeselitz and threeants said.
posted by aspersioncast at 9:26 PM on February 5, 2015 [1 favorite]


How do the people in this fantasy all obtain autonomous cars? If there is no Magic Dad giving us all free cars, what happens to the people who can't afford them?

Just after the most recent autonomous car discussion here I was driving on the freeway and passed a guy in the right lane driving an early 1990s Ford Tempo and was struck by how inaccessible even modern car technology is for many people, never mind full automation. There are a lot of people out there driving crappy junker cars because that's all they can afford, and any proposal to automate the entire car fleet needs to think about how that guy is going to get around in that brave new world.
posted by Dip Flash at 9:45 PM on February 5, 2015 [4 favorites]


If they even think they want it.
posted by Greg_Ace at 9:53 PM on February 5, 2015


There are a remarkable number of people in this thread who make handcrafted buggy whips. Good luck with that.
posted by Awful Peice of Crap at 10:00 PM on February 5, 2015 [1 favorite]


Buggy whips, huh? Sorry, I'm not keen on arguments that delay having working public transportation right now because we're going to have automatic cars that will get us door to door in a few years! And oh, they'll totally be just as accessible to everyone as public transportation is now, and having the costs fall entirely on individuals in multi-thousand dollar chunks up front rather than a couple bucks at a time is a feature, not a bug!

Hey, I love hearing about the automatic safety features described by the people upthread with the fancy new cars as much as anybody, and those are undoubtedly a wonderful thing for all of us that will improve highway safety dramatically as they slowly become available in more cars on the road. But that doesn't help us right now, and tomorrow morning, I've got to get to work.
posted by asperity at 10:10 PM on February 5, 2015 [2 favorites]


It's possible that autonomous cars that drive themselves exclusively on highways (under good conditions--no snow, nice clear lines, your camera isn't obstructed by condensation, etc.) will become common in our lifetimes, but driving on city streets is AI-hard in my opinion. In fact, harder--most people can't do it all that well. It's not the sort of task that computers are good at, either, so throwing more CPU at it won't make a dent. You think solving Go is hard? Try modeling the world in 3d, inferring other people's intent to avoid an accident, inferring how an intersection is supposed to work when the signage is damaged, navigating in bad weather, avoiding floods, downed power lines, zombie hordes, etc etc etc.

I'm sure you could come up with a car that works well in 9x% of everyday situations, but handling the long tail of crazy shit that happens on the street is a harder problem than things computers still suck at (like computer vision, which is just one part of this problem). The author handwaves this objection away, but predictions about AI are notoriously optimistic, and this one takes the cake.
posted by jewzilla at 10:23 PM on February 5, 2015 [3 favorites]


Wow. I was going to make a comment earlier about how the interior of major cities shouldn't need (automobile) roads at all, but considered it too radical, and not worth it to defend. I didn't expect the thread to turn to this opinion earlier on.
posted by halifix at 10:30 PM on February 5, 2015 [2 favorites]


Autonomous cars, Elon Musk, and me are all INTJs, so this is a sure thing.
posted by benzenedream at 11:04 PM on February 5, 2015


My city invented autonomous vehicles in the 1980s, and they are amazing.

Cars though? A ridiculous and inefficient technology, optimizing only profits for auto, gas, and rubber companies.

His arguments against subways (which can be autonomous) are downright silly. I'll pick on just two. 1) "Subways are not profitable." Well, no. They are a public service, like the bridges and roads I suppose your autonomous cars plan to drive on. 2) "Subways only work in NYC." Again, no. Having lived in Beijing and Toronto, this is plain wrong. Vancouver has great subways, too. As does Moscow, and dozens of other cities that are smaller, less dense, or less rich than NYC (though those things do help).

It's nice that this guy is interested in better forms of transportation, but he doesn't even seem to be aware of ideas like new urbanism, or transit-oriented development that are really central to the field he is critiquing.
posted by the thing about it at 11:10 PM on February 5, 2015 [7 favorites]


LogicalDash: “Huh? google already has autonomous cars. They haven't worked out the manufacturing/distribution/legal stuff, is why you can't buy them yet. Could've sworn we had a couple threads about them...”

google doesn't have autonomous cars. google has machines that can safely carry people in controlled environments if fed excruciatingly detailed information about those environments ahead of time.

They've apparently managed to make this work on a few specific highways
-- koeselitz


This is not accurate information. google has been using self-driving cars for many of their street-view camera cars. They drive all over the place, including down crowded pedestrian-filled city streets. Here's an article from 2011.

Yes, they have a person behind the wheel, ready to take over at any moment, as required by present law.
posted by eye of newt at 11:18 PM on February 5, 2015 [3 favorites]


By the way, at least in the suburbs, many bus routes have buses that are nearly or often completely empty. How many miles per gallon per person (or, more likely gallons per mile per person) do you think these things consume? How much pollution? Cars are incredibly more efficient and less polluting than these buses on these routes.

But they exist, because otherwise many people who can't drive wouldn't be able to get to their job, to school, or to the doctor. But they definitely don't exist because of efficiency.

At least in the suburbs, buses are oversold as an efficient form of transportation. Certain key routes, certainly. But many routes, just as certainly not.
posted by eye of newt at 11:23 PM on February 5, 2015 [1 favorite]


We'll have proper autonomous cars in mass use at the same time we'll have flying cars.
posted by MartinWisse at 11:30 PM on February 5, 2015


A lot of people here are missing the main point of the article. Its not the car, its the fleet. I don't think you're going to buy autonomous cars, your going to be subscribing to an autonomous car service. You'll request a car, it will pick you up at your house, drop you off at your location and then head on out to the next request. It might send me a 2 person vehicle to take my wife and I to Ikea, and then give us a cargo van when we leave. The single occupancy commute we Americans seem to love isn't great, but I'd rather see a fleet of mini Cooper sized vehicles making the commute then the mix of cars and SUVs I see now. That special needs guy that needs some extra time getting on the bus? He won't need the bus, a car will pick him up right from his house, in a vehicle with extra wide doors or a rear loading wheelchair ramp or whatever his situation requires and it will take him anywhere, vastly opening up the transportation options available to him. How will the poor pay for it? Create a subsidised subscription benefit, just like we do with bus passes. And now lots of poor people who can't scrape together the cash to purchase a car will have access to all the benefits of one. In the end it will save fuel, money and so much time. No more circling for a parking spot, no more filling up the gas tank, no more taking it in for an oil change, no more worrying if your 16 year old has been drinking and wrapped the car around a light pole. Sure this will be a huge disruption to society, but I really like the upside.

And I don't know why there's the impression that these are only being driven on closed tracks in controlled environments. You can find google Cars driving autonomously all over Mountain View right now.
posted by BenNewman at 11:30 PM on February 5, 2015 [6 favorites]


Did anyone go through the videos in the article? There's one titled "testing the Q50" or something which is about lane guidance on the highway. So the guy records it - all very impressive and everything. But then he records it after he GETS OUT OF THE DRIVER's seat while travelling at highway speeds, which is some serious bullshit because while lane guidance and auto cruise control etc are very compelling technologies, these are not toys to played with in such a cavalier fashion. We are nowhere close to the stage where no driver attention is needed (esp in commercial systems that are not designed to be fully autonomous, which the Q50 falls into).
posted by modernnomad at 11:34 PM on February 5, 2015


“Autonomous cars will destroy tons of American jobs.” Boy, will they ever.

How many jobs would you like to bring back from the dead today my friend? Blacksmith? Town Crier? Alchemist? Powder Monkey? Technology creates complex jobs and destroys menial ones, pretty much as a rule. It’s kinda the point of technology, no? To make our lives easier so that we don’t have to struggle for survival? There must be a thousand different jobs that don’t exist today so that your life can be as comfortable and entitled as it is. Nobody seems to care until it’s their job on the line, and then all of a sudden it’s more about their struggle than a billion other people’s struggle.

I appreciate that the topic of “jobs” is a highly political one. I have my opinions, and you have yours. There are thousands of destinations on the internet where you can argue your little head off about it. Here is not that place; there’s nothing about autonomous cars that makes the argument unique. Your beef is with cultural progress in general, so get used to fighting that fight on a daily basis from here on out.
This is the thing that really worries me. A whole lot of jobs for taxi drivers, van drivers, couriers are at immediate risk. Other jobs are vulnerable to de-skilling: truck and bus drivers replaced by minimum wage casual labour.

Yes, in the past technology has created complex jobs to replace menial ones... very slowly. It took more than half a century for the displaced 19th century agricultural jobs to be substituted. So yeah, as long as history repeats itself, by 2065 or so things will be just fine.

Moreover, it's not clear that this process is still happening. Lots of semi-skilled and skilled jobs are being destroyed, but this time they seem to be being replaced by unskilled menial jobs, like warehouse pickers.

In England and Wales there are 231,000 registered taxis and minicabs. So yes, my little head is worried about what's going to happen to all those drivers. Mysteriously I care even though my job is not on the line (I'm a computer programmer currently working in the medical sector). I guess my beef is just with cultural progress in general.

I don't think banning autonomous cars is a good solutionn. But I think we need to face up to the issues of a society where the returns to labour are decreasing while the returns to capital increase, and look at ideas like taxes on capital and basic income schemes instead.posted by TheophileEscargot at 3:12 AM on February 6, 2015 [2 favorites]


Come on people, this one's a gift

Who are all these snarky people?: metafilter

One of the things worth considering about traffic congestion is not a uniquely American problem. Most traffic congestion is happening elsewhere. Whether or not autonomous cars take off in the US, which is debatable, there are problems all over the world due to car ownership climbing and shortsighted (and probably corrupt) transport policy development.
posted by asok at 3:17 AM on February 6, 2015


Nobody seems to care until it’s their job on the line, and then all of a sudden it’s more about their struggle than a billion other people’s struggle.

This guy seems to think that since he's primarily motivated by self-interest, everyone else is. Or maybe that people who try to act from compassion are idiots; I'm not really sure.
posted by sebastienbailard at 3:23 AM on February 6, 2015


cities are already built though. Can't really tear it all down and start again.

Maybe not, but gradual change is possible. Adapting a city for public transport and bicycle use is quite within reach in most cases.

I seem to remember that Park Avenue in New York used to be an actual park. I am sure I saw that on metafilter, but I can't find it now.
posted by asok at 3:31 AM on February 6, 2015


So, I work as a transport planner and modeller. I've been involved, directly in completely automating a subway line and in one of the first deployments of Automatic Train Operation on mainline rail (incidentally, ATO on metro rail, yeah that's solved. The victoria line has been automatic since 1968, mainline, that's a bit trickier).
My first degree is also in Robotics and Automated systems.

So...
Where to begin.
Firstly the line: "The upshot is a transportation model that exerts energy at the exact level of consumer demand" referring to cars is just comically and obviously wrong. Cars as a transport model will always operate at significantly below peak efficiency, simply because they are big hunks of metal which are almost always used to transport 1 person.

He spends some time decrying the new york subway. Particularly focussing on Infrastructure costs. But of course roads are free. Very few people take into what an enormous subsidy to the car industry government run road building and maintenance is. Not to mention that the sheer number of people moved by a subway is staggering.

But the annoying thing here is that he's not ultimately wrong. I mean, yeah okay, he is wrong in terms of the absolute view that all you need is autonomous cars. You need high speed rail, you need bikeshares and so on, because it's a scale problem. High Speed Rail get's you from New York to San Francisco, then the autonomous car gets you the rest of the way, because why the hell would you travel intercity in a car, trains can be roomier, more comfortable, faster, and have restaurants. The bike is for when you want to move just you a few miles.

Next up. Why would your autonomous vehicle fleet look anything like a traditional car fleet. He references this a few times, "turn up in a porche" for example.
It wouldn't. It would be a box with seats in, probably swively ones. You would have different vehicle types for different problems, large coach sized things for regular, popular routes (why clog up a road with 50 autonomous cars when you can have 1 autonomous bus), smaller taxi sized things for when you have a small group going out (lounge style seating, minibar, if you want). Cargo haulers for when you need to move stuff, which might have a single passenger seat, but might not. Tiny one person pods for individual journeys which no one else is going to.

Oh, another interesting point I think, which he skirts around but then misses. An autonomous fleet, with continuous connection connected to a large scale database would be incredibly valuable in terms of maintenance monitoring. You can see emerging patterns in the data that show road deterioration and prioritise repair. This is something we can (and should) be doing with rail, and we are getting there, but it is one of these solutions that sounds so wonderfully easy but is full of hundreds of barriers to actual implementation (both technical and political). Actually a lot of the value he's talking about here isn't derived from autonomy, it's derived from connectedness and good data mining. We could do that with current solutions, except of course when you own your car you would, I imagine, be opposed to it reporting back useful data.
posted by Just this guy, y'know at 4:04 AM on February 6, 2015 [14 favorites]


How would you automate stopping at an unmarked stop and not opening the door until there is no cross traffic for the special needs guy who always gets off there? [etc.]

Automate bus driving (keeping the bus running on the scheduled route unless interrupted) but put a human conductor/guide/guard/ombudsman on each bus. The bus would never be forgetful, tired, drunk, or distracted, so it would be safer for everyone under normal circumstances, and the conductor could manually stop the bus, open doors in the middle of a block, wait for the guy huffing and puffing after the bus, etc. The conductor could also answer questions about the route, make sure everyone pays, intervene in passenger disputes, kick troublemakers off the bus, and call the police when needed. You would be much safer and happier riding a bus like that than you are riding a bus driven by an overworked human.
posted by pracowity at 4:48 AM on February 6, 2015 [3 favorites]


I don't think banning autonomous cars is a good solutionn. But I think we need to face up to the issues of a society where the returns to labour are decreasing while the returns to capital increase, and look at ideas like taxes on capital and basic income schemes instead.

This absolutely.

Assuming we'll have an autonomous fleet of heterogenous taxi-like special purpose vehicles, we can't pass up on gains in efficiency from:
- Way less trucks/SUVs exist, on average vehicles in the fleet will be way smaller than they are now.
- Pizza delivery now done by fridge-sized box on wheels, not 2 ton sedan.
- The fleet can easily be all-electric because the "range-anxiety" problem becomes moot, vehicles will manage to keep themselves charged without getting anyone stuck anywhere.
- An all-electric fleet means economies of scale for energy (centralized power production), and as our power production shifts to renewable the fleet automatically transitions to being powered by renewables.
- With the "most vehicles spend 95% of their time parked" issue gone, we'll need way less vehicles. So, less stuff to be created and then somehow dumped/recycled.
- As extension of the last issue, most parking lots and garages can be reclaimed for better use.
- Way less accidents.

And these are just the major wins. I can't imagine passing up the benefit to the planet, to individual mental health, and societal mental health.

At the same time, we can't ignore the human costs - which is why I think it's incumbent on us to pave the way by pushing for ways to disconnect the vagaries of the unregulated market from human wellbeing/flourishing. (IE Basic income or similar.) If we manage this, I think we'll discover a long list of things in addition to autonomous cars that we as a society will be able to implement that will be net wins, which we'll actually be able to feel good about because they won't come with so much collateral damage.
posted by tempythethird at 4:51 AM on February 6, 2015 [1 favorite]


A lot of people here are missing the main point of the article. Its not the car, its the fleet. I don't think you're going to buy autonomous cars, your going to be subscribing to an autonomous car service. You'll request a car, it will pick you up at your house, drop you off at your location and then head on out to the next request.

That's how the taxis of the future will operate, but many people will still insist on commuting to and from work in their own personalized cars full of their own stuff and not waiting for a taxi with an empty commercial-grade interior to pull up out front. When people don't have to drive their cars, they will want to climb into their own rolling recliners with their own entertainment systems, well-stocked refrigerators and bars, games, books, toys. They will want to make the windows opaque and spend their commute time napping and fapping, not sitting where maybe some stranger has been napping and fapping ten minutes earlier.
posted by pracowity at 5:03 AM on February 6, 2015 [3 favorites]


And they've managed to allow passengers to ride in these "autonomous cars" in (surprise!) controlled test situations on their own lots.


I have taken a ride in a google self driving car. If you consider US-101 near rush hour "their own test lot" then this is accurate.

While we were driving cars cut us off, unless they paid stunt drivers, the vehicles response was impressive.

Sure, they might not have figured everything out, but I think you'd be surprised how much closer they are than you think.
posted by vorpal bunny at 5:04 AM on February 6, 2015 [9 favorites]


Actually I was thinking about it more and I think it might actually be "You puny human, you think you can question the will of ALMIGHTY TECHNOLOGY*???

That's more to the point than the way I put it, I think. Techno-utopians see the March of Progress as completely natural and inevitable (which is really quite Victorian of them, as I understand it), and regard any arguments to the contrary as nonsensical on their face. That's probably why so many of them are libertarians: both viewpoints require a pronounced historical myopia.
posted by Steely-eyed Missile Man at 5:28 AM on February 6, 2015 [2 favorites]


They think people who won't travel on buses with the great unwashed are going to share a car with them?
posted by fullerine at 5:33 AM on February 6, 2015 [1 favorite]


How many jobs would you like to bring back from the dead today my friend? Blacksmith?

Well...yes, actually.

It just occurred to me that this guy, and people who chant "buggy whips" and "whale oil," make an argument (regarding job creation, etc.) that goes something like, "Things are radically changing all the time and at a rapid pace, but also it will happen exactly like it did (a) hundred(s of) years ago! Also, I have zero details of how or even if people adjusted, but somehow we got from that world to this, so I'm sure everything will be fine for everyone, or at least for me, and really that's what matters."
posted by Steely-eyed Missile Man at 5:47 AM on February 6, 2015 [6 favorites]


No see anytime you question a techno futurists assumptions you're instantly a Luddite or some obsolete job.
posted by Ferreous at 5:50 AM on February 6, 2015 [2 favorites]


As a side note, if you want actual dystopian fiction involving self-driving cars, Gene Wolfe's "Forlesen" pops out.
posted by CBrachyrhynchos at 5:59 AM on February 6, 2015


I always have to roll my eyes when they give me that baseline figure of $10000 per year for a privately owned vehicle, because I bought an older full-sized pickup truck three years ago and even with 13-15 MPG, insurance, maintenance/repairs, and heavy use in the context of my work, my costs all hover around $4300 a year, so I'm still saving $2200 over the magical autonomous car. These things always assume you're upper middle class and am mandated to buy new, because SCARY UNCERTAINTY OMG.

Of course, if my Miata hadn't wrecked my back, I could still be chugging along at a hair about $3000 a year, thanks to 31 MPG, and I'd be cuter, too, at least until I grunted my carcass out of the car and walked in a permahunch.

Techno-utopians love this shit, though. Remember how the CFL bulb was going to save us twenty dollars in electricity over the lifetime of the bulb and last ten years? Yeah, everything looked like bluish yellow phlegmy jaundice, the fuckers lasted a year at best, and left you with a wad of electronics and mercury to dispose of. Progress!

We have almost driverless vehicles already—trains and trams and to some extent buses. Everything else is Robert Moses junk for rich white people to get all Futurama about.
posted by sonascope at 5:59 AM on February 6, 2015 [5 favorites]


Everything else is Robert Moses junk for rich white people to get all Futurama about.

This is some extreme baby and bath-water disposal. Is there any room for the possibility of legitimate widely-beneficial technological/scientific progress without belief in such automatically making one a libertarian techno-utopian or a useful idiot?

Also, while the early CFLs sucked, the current ones and LED bulbs are living up to pretty much all their promises, in my experience.
posted by tempythethird at 6:11 AM on February 6, 2015 [3 favorites]


This is not accurate information. google has been using self-driving cars for many of their street-view camera cars. They drive all over the place, including down crowded pedestrian-filled city streets. Here's an article from 2011.


It's amazing watching a whole crowd of people fall, hook line and sinker for marketing. That article does not say what you think it does.

"Statistical reasoning" has some fundamental limitations as an approach to AI. Building a better autopilot for a car is a vastly smaller problem than true autonomous driving... and a solution with distinctly smaller utility. Constrain the problem and your autonomous car starts to look like a train.

But facts don't matter because the whole discussion is fantasy driven.

metafilter: I learned something, and it's in line with my thinking.
posted by ennui.bz at 6:12 AM on February 6, 2015 [5 favorites]


This is some extreme baby and bath-water disposal. Is there any room for the possibility of legitimate widely-beneficial technological/scientific progress without belief in such automatically making one a libertarian techno-utopian or a useful idiot?

There's plenty of room for amazing things. I just played a damn fine live gig on the modern equivalent of the Hitchhiker's Guide To The Galaxy, which I got for a hundred bucks because apparently four year-old things are now worthlessly obsolete. I can call for help or look up videos on trendcadis mosaic on a computerized telephone from a lush green mountainside in rural West Virginia. Hell, I can ride on a forty year-old driverless transit system in West Virginia. We're living in the world of the future, but free-roaming autonomous transit is a bizarre fixation straight out of Motorrama and I contend that it's more a way of pushing transportation out of the hands of the working poor than a magical transformation of experience. With this article's $6500 per year cost for autonomous vehicles, I'd be out of a livelihood.

And CFLs sucked for twenty years and never stopped sucking before they were replaced. The makers of LED lighting learned a lesson from CFLs, which is why they arrived nearly fully-formed—don't sell us bullshit that doesn't work and tell us we're wrong when we know for a fact that it sucks.

There's plenty of stuff worth holding back in that draining bathwater, but driverless cars just solve a symptom, not a problem, and they do it in a way that leaves out a substantial fraction of the population. Of course, your mileage may well vary.
posted by sonascope at 6:30 AM on February 6, 2015 [3 favorites]


I have taken a ride in a google self driving car. If you consider US-101 near rush hour "their own test lot" then this is accurate.

While we were driving cars cut us off, unless they paid stunt drivers, the vehicles response was impressive.

Sure, they might not have figured everything out, but I think you'd be surprised how much closer they are than you think.


Thank you for some firsthand information. It seems that when I run across discussions regarding google self driving cars, there are far too many head-in-the-sand proclamations of NEVER EVER EVER that completely ignore its current progress.

Of course, we aren't going to have some overnight shift and suddenly next week or next year, we'll all be riding around in google cars. This is a technology that will gradually make its way into transportation, over time. It has already begun, with adaptive cruise control, lane departure assistance, and braking assistance available on current vehicles that effectively create a highway autopilot. This technology will slowly, bit by bit, take over more of the driving resposibilities.

Millions of Boomers aging will create a large demand for assisted and automated transportation, and that's going to accelerate its adoption and progress.
posted by Fleebnork at 6:32 AM on February 6, 2015 [8 favorites]


In related news, Uber basically just leased Carnegie Mellon's National Robotics Engineering Center, for an undisclosed sum that apparently includes funding something like 50 full-time research positions. At least according to TechCrunch, "Uber has 'cleaned out' the Robotics Institute".

Reportedly: "The partnership aims to develop new aspects of mapping, vehicle safety and technology with an eye toward autonomous taxi fleet development. [...] 'This is the way, largely, I roll and Uber rolls as a company,' [Uber Chief Product Officer Jeff] Holden said in a Monday evening conference call. 'I was at Amazon for a long time and learned a ton about investing in the future and the long term of a company and planting seeds for five to 10 years out.'"

I don't think they are planning for 2070. They seem to be betting that they can make this stuff pay dividends in the corporate version of "long term", which isn't that long. I'm not sure they are entirely right, but I don't think it is something to be casually dismissed either.

My guess is that you will see some sort of proof of concept of automated taxis or automated delivery trucks within the next 5 years. The problem from an engineering perspective is not, as some allege, actually AI hard; it's a fairly simple set of rules; the difficulty is mostly on the sensing/machine vision side. And if you're rolling out to a specific geographic area and have money to spend, you can increase the accuracy of the sensor systems by doing stuff like upgrading signage and lane markers (e.g. cut small slots in road signs so that they have unique radar signatures based on the type of sign; use strongly fluorescent lane-marking paint, etc.). And of course you do terrain mapping. I could see a company like Uber doing that in a particular city as a demonstrator. Uber can also do the mapping economically by using their existing vehicle fleet as they drive around, and they can keep it up to date; just because it doesn't work with google Maps' business model doesn't make it impossible.

Driverless PRTs operating in controlled environments (e.g. Disneyland, or those death's-waiting-room golf cart communities in Florida) are even more straightforward. There's no reason why you couldn't build an automated golf-cart system, running on a defined cart path using augmented GPS and minimal machine vision to avoid collisions, right now; the cost just has to come down to the point where someone can market and sell it correctly. And once there, people will begin to see driverless systems as something unremarkable and boring.

A 100% car replacement—one that can go from any point on the road network to any other, in any weather—is admittedly probably a long way off. (That said, there are a lot of drivers who aren't capable of driving a car from any point to any other point on the road network. Or at least a lot of people I wouldn't want to try that with.) But you don't need to get all the way to that point to be disruptive and have a day-to-day impact. I think we're going to start seeing bits and pieces quite soon.

It'll never be impressive the way the techno-utopians want it to be, though, because by the time technology comes into wide use, it is by definition almost always boring.
posted by Kadin2048 at 6:40 AM on February 6, 2015 [4 favorites]


I will start believing the autonomous car hype when they test them successfully in snowy winter city conditions when streets can be a uniform field of white. It is hard enough to find the lanes as a human driver, never mind berms of snow blocking side streets that you have to ram through and giant banks of snow narrowing the roads. And then there are the potholes everywhere during thaws that will destroy tires, rims, and suspensions. Throw in pedestrians walking in the streets because the sidewalks are not cleared, vehicles slipping and sliding on ice and snow, winter cyclists, snowplows, and all of the other crazy shit on winter roads, and it is pretty chaotic. Will they work in ideal road conditions in sunny California? Sure, maybe. But I don't see them becoming widespread in my lifetime anywhere in the north.
posted by fimbulvetr at 7:14 AM on February 6, 2015


How many jobs would you like to bring back from the dead today my friend? Blacksmith? Town Crier? Alchemist? Powder Monkey?

I definitely want blacksmiths and town criers back.

A blacksmith and horses and the ability to ride them in every town, which means more green space and a better connection with nature for everyone, and maybe slower paced roads. I used to love it when the horse-drawn potato cart would come walking up our street and the driver would call out "Potatoes! Potatoes!" to let you know he was coming if you wanted to buy a sack. Bring back the potato guy and his cart and horse and the blacksmith to keep the horse shoed and the roads that allowed for a horse-drawn cart to share space with cars.

And I want a town crier walking around the neighbourhood on real walkable sidewalks or right up the middle of walkable streets, letting people know what's up and keeping the neighbourhood together in a way you won't get from a local news blogger, though the crier could also be the local news blogger. Crier as news gatherer and reporter. If churches can ring their big fucking bells, let the crier ring his.

And maybe bring back alchemists. Why not? We're not talking about government positions paid for by all. If you can have homoeopaths, you can have alchemists. Just verify their claims or tell them to get out.

And I want automatic fucking cars. I want to see insurance companies price manually driven cars right out of existence. No more boy racers zooming around town. No more drunk drivers. No more road rage. And so on. Just people converting all those wasted commuting hours into something good, even if it's just getting a little extra sleep in the morning. Or studying how to be a lamplighter.
posted by pracowity at 7:15 AM on February 6, 2015 [6 favorites]


My guess is that you will see some sort of proof of concept of automated taxis or automated delivery trucks within the next 5 years.

I agree. I think that the prospect of cutting labor out of existing automobile-related industries is going to drive the first round of implementation with the technology.
posted by codacorolla at 7:24 AM on February 6, 2015 [1 favorite]


Not to mention, how well will autonomous cars work when they and all of their sensors are covered in a layer of snow, ice, slush, salt, and grime? You can't drive down a winter street without everything getting covered in slush. Even if you automate driving, the car still has to be cleared off by the driver. And when the temperatures dip to -30°C and lower, even current simple modern cars start to behave weirdly and get strange electronic gremlins when the temperatures drop low enough. How will the complicated electronics and sensors of an autonomous car behave?
posted by fimbulvetr at 7:25 AM on February 6, 2015 [1 favorite]


A fleet of autonomous buses because what's more utopian than higher unemployment?
A fleet of autonomous buses shipping the ever-dwindling number of commuters employed in the one still-viable industry—disruption—to their campuses every morning. Where they will work on the technology necessary to automate "disruption," hence techno-libertarianizing themselves out of their own jobs. The last category of employment left standing will be writing smug, anti-human screeds like this one—gushing over the shiny perfection of machines vs. people, sneering at your "hilarious" faith in democracy and your own skills, and rubbing your face in the fact that There Is No Alternative because CHINA. By the time someone automates this kind of literary production (and, I guess, invents a fully autonomous machine audience to read and believe it) the coming environmental collapse will have completely overwhelmed and destroyed everything simultaneously, because we no longer believe in redundant systems.
posted by Sonny Jim at 7:29 AM on February 6, 2015 [2 favorites]


The problem of fewer people needing to be meaningfully employed (and how to organise that) goes back to Bertrand Russell's Pin Factory (and probably earlier). Automation will not stop. The problems of distribution of labour won't go away. They desperately need to be addressed sooner rather than later.
The solution is not going to be "let's not automate stuff". We all know that it isn't. An ostensibly market capitalist system won't let you do anything but produce as much as possible for the smallest cost, which means automation, de-skilling and lower pay.
I don't know what the solution is, but I don't believe it can be "Smash all the robots"
posted by Just this guy, y'know at 8:01 AM on February 6, 2015 [5 favorites]


The hate for this developing technology is astounding in this thread.

I'm not going to bother to try to refute all the claims in this thread ("there won't be driverless cars anywhere until they can navigate the most snowfilled roads" or whatever), but I will mention this...

Right now people rightfully discuss how a single medical emergency can bring a middle-class family to bankruptcy (hopefully the affordable care act can help address this).

But for a poor family that owns a car, a single mechanical breakdown can financially cripple them. They can't commute to their low-paying job and can easily be out of a (difficult to get job) from a single car problem. If they miss a single car payment their car may be remotely disabled. A single fender bender can make their insurance skyrocket to the point of being financially fucked.

The middle class sees cars as freedom. The poor sees cars as a necessity that gets them to their crappy jobs, but could fail at any moment (since they can only afford cheap used cars), making their already tenuous financial position ruinous.

While self-driving cars wont' solve the problems anytime soon (first the upper class will get them, then the middle class), the prospect of change from this awful situation is something I greatly look forward to.

And to those that say 'fuck driverless cars, we need better public transport'... Yeah okay, in the city yes, but there are a *ton* of poor folks in the country, and public transport improvements aren't ever going to improve their quality of life (indeed you might end up taxing them more to pay for the city public transport improvements).

The idea that cars = freedom is a position of privilege.
posted by el io at 8:07 AM on February 6, 2015 [5 favorites]


there are a *ton* of poor folks in the country, and public transport improvements aren't ever going to improve their quality of life (indeed you might end up taxing them more to pay for the city public transport improvements).
Oh come on. Autonomous vehicles won't be for the poor. (Where's the money in that?) Instead, they'll be good for forming the heavily armoured convoys necessary for resupplying the Great and the Good, rumbling from one Green Zone to another. (Or, much more rarely, whisking some of the Great and Good between the zones when they regrettably have to spend some time in meatspace.) It'll be like a high-tech, drone version of the road from Baghdad airport, c. 2005, except on a global scale.
posted by Sonny Jim at 8:19 AM on February 6, 2015


Sonny: Air bags weren't for the poor either. And while they weren't the first (nor the second, nor the third) folks to get them, they have them today (in significant numbers).
posted by el io at 8:26 AM on February 6, 2015 [1 favorite]


"Hate" is a bit strong of a word. I see a lot of well-thought-out skepticism. Also, to nitpick a bit, I said that I don't see them being widespread in the north, not "anywhere". Living in a city that is snowbound for at least 4 months of the year, winter conditions are not a trivial problem.
posted by fimbulvetr at 8:29 AM on February 6, 2015 [1 favorite]


The idea that cars = freedom is a position of privilege.

It's an illusion unless you don't worry about money at all. Cars are a financial albatross around the neck of anyone who owns and operates them on a regular basis (unless, I guess, you really, really enjoy driving and do it as a leisure activity, then they're just hobbyist equipment), which is to say nothing of the absurd amount of time in my experience most drivers spend thinking and talking about routes, traffic, etc. All that mental space being taken up getting from point A to point B, where point A and point B are only as far away as they are because cars have made it possible.
posted by Steely-eyed Missile Man at 8:31 AM on February 6, 2015


Air bags weren't for the poor either. And […] they have them today

It's amazing how easily a bitterly contested regulatory battle, in retrospect, becomes just the good of the market, benevolently trickling down.
posted by RogerB at 8:36 AM on February 6, 2015 [8 favorites]


Could maybe highways have autonomous cars at first? All the car would have to know is highway driving and how to pull into a rest stop so you can switch back to manual for the rest of the way.
posted by yoHighness at 8:48 AM on February 6, 2015


Kadin2048: My guess is that you will see some sort of proof of concept of automated taxis or automated delivery trucks within the next 5 years.

codacorolla: I agree. I think that the prospect of cutting labor out of existing automobile-related industries is going to drive the first round of implementation with the technology.

Item 1: the trucking industry in the US is sorely in need of new drivers (Feb. 2015)
Jerry Klabacka, president and an owner of Diesel Truck Driver Training, visible from Highway 151 near Sun Prairie, said there is a huge shortage of drivers and he doesn't see that changing anytime soon.

"The number of truck driving students is down by 50 percent from 10 years ago," he said.

Klabacka said industry pay, in general, hasn't kept up with the demands of a tough job with lifestyle sacrifices that are part of an irregular work schedule.

He said emphasis on education is stronger than ever for young people and truck driving "is not on their radar."
Item 2: Dutch plan for self-driving trucks could boost introduction of driverless cars (June 2014)
Under a plan by a group of logistics and technology companies unveiled on Monday, trucks without drivers could begin delivering goods from Rotterdam, Europe's largest port, to other Dutch cities within five years.

The Netherlands is reviewing traffic laws to make large-scale testing of the technology possible on public roads, Infrastructure and Environment Minister Melanie Schultz van Haegen wrote to parliament in a letter outlining the plan.

Initial testing would start on computer simulations and the trucks will be tested on a closed track before ultimately driving out on public roads, her letter said.
Leading to Item 3: Dutch approve driverless cars for public, large-scale testing (Saturday, January 24, 2015!)
The Dutch government on Friday approved large-scale testing of self-driving cars and trucks on public roads, saying the technology could greatly reduce traffic jams and improve road safety.

"The cabinet has agreed to adjust road rules to enable large-scale testing of self-driving trucks and cars on public roads," the Dutch infrastructure and environment ministry said.
Trucking needs drivers TODAY, so automation isn't just to cut labor costs. But it sounds like Uber is heading there for taxi-type services, which is definitely an attempt to cut out the existing labor in taxis and their own Uber drivers.
posted by filthy light thief at 9:01 AM on February 6, 2015 [1 favorite]


The future is rather difficult to predict. What we think we'll get in 20 years can take 100. What we'll actually get in 20 years is something we haven't even imagined.

There are ways to fix our problems which can be done now. There are unarguably much better ways to fix our problems which can't be done quite yet (nuclear fusion, driverless cars). I'm in favour of going with the former.
posted by dickasso at 9:05 AM on February 6, 2015


koeselitz: we should pour all our money into buses.

The thing is that there are many sources of money. Federal funds are going into evaluating technology that is being developed by private companies who are working on the private benefit, whereas development of new bus rapid transit services (high frequency, separate and fixed routes, unimpeded by other traffic flows) are funded by public funds for the public benefit.

Also, more driver assistance will come first, before fully automated vehicles ever come into play. Connected Vehicle: The Future of Transportation is a 7 minute video from U.S. DOT, about ways that "apps" can inform drivers of surrounding issues, which in theory could be tied back to the vehicles for automatic response instead of relying on drivers to respond to a vibrating seat, an audio tone or a visual display, building on what already exists for automatic braking.
posted by filthy light thief at 9:09 AM on February 6, 2015


I will start believing the autonomous car hype when they test them successfully in snowy winter city conditions when streets can be a uniform field of white.

Fair enough, but "test successfully" should just mean "does better than the median human driver" or "does better than the 90th percentile human driver."

Not to mention, how well will autonomous cars work when they and all of their sensors are covered in a layer of snow, ice, slush, salt, and grime? You can't drive down a winter street without everything getting covered in slush.

(1) They could avail themselves of the same technology that human drivers use to keep their windshields clear.

(2) Probably also moving-film systems like they have on external cameras on racing cars that keep the camera's view clear after it's been blasted with a horrible oil-water-rubber mix.

And when the temperatures dip to -30°C and lower, even current simple modern cars start to behave weirdly and get strange electronic gremlins when the temperatures drop low enough. How will the complicated electronics and sensors of an autonomous car behave?

Well, only a miniscule proportion of the world's population lives anywhere that sees -30C more than once every 10 or 20 years anyway. But in any case they'll behave just fine when it's -30C out because putting the electronics in an insulated compartment, and putting a small heater inside that to keep things at 0C or whatever, is not rocket science. I mean, it is rocket science, because that's how they keep electronics working on Mars, but you know what I mean.
posted by ROU_Xenophobe at 9:09 AM on February 6, 2015 [3 favorites]


dickasso: There are ways to fix our problems which can be done now.

There is not one monolithic entity pushing forward on solutions to current problems. There are market-driven innovations, regulation driven responses by private entities, and public projects, to quickly break innovations and actions into three overly broad categories. For instance, "less cars, more buses" is a great goal, but unlikely unless more public funding is allocated for transit. Private companies aren't going to provide most forms of mass transit, because it's not profitable.

But that doesn't mean it shouldn't exist. After all, the same thing could be said about most roads - except for toll roads, governments don't "make money" off of road development - it's a public good and improvement, a benefit for all. (I'm mentally responding to comments heard elsewhere about the problem with the return on investments for public transit.)
posted by filthy light thief at 9:14 AM on February 6, 2015


"It's amazing how easily a bitterly contested regulatory battle, in retrospect, becomes just the good of the market, benevolently trickling down."

There will be another bitterly contested regulatory battle, I predict, within my lifetime. Mandating cars have self-driving capabilities. The car companies will declare that this will make their non-self driving cars less affordable (true) while the pro-regulatory forces will point to the deaths that are caused when humans are in control of the vehicles. If you thought the air-bag battle was a bitter one, just wait until this one comes up.

People weren't passionately anti-air bag (not that I remember anyway), so that battle will be a doozy.

As far as 'trickle-down', it sure doesn't work for tax structures, but it tends to work with technology (refrigerators - those were for the rich, computers, cell phones, power-steering, anti-lock breaks, flat-screen televisions, etc, etc). Hell, the only way the rich can show off with gadgets these days is by encrusting them in jewels.
posted by el io at 9:16 AM on February 6, 2015


Have the rich ever shown off with "gadgets"? Sincere question.
posted by Steely-eyed Missile Man at 10:00 AM on February 6, 2015


how well will autonomous cars work when they and all of their sensors are covered in a layer of snow, ice, slush, salt, and grime?

Carnegie Mellon is in Pittsburgh. I think they've got this one.

The potholes thing too.
posted by Kadin2048 at 10:07 AM on February 6, 2015 [1 favorite]


Well, only a miniscule proportion of the world's population lives anywhere that sees -30C more than once every 10 or 20 years anyway.

Granted, Canada only contains a small portion of the world's population. But we do have a giant landmass and drive quite a bit. I wish -30°C only happened once every 10 or 20 years, but that ain't the case. Below -15°C (normal conditions during winter here), road salt is pretty well useless and all the stuff frozen on your car is rock hard and extremely difficult to remove.

I'm not arguing that autonomous vehicles won't happen. I just don't see them as a near-future thing for personal automobiles. I expect that long-haul trucking will be the first to adopt it, but even then I would be surprised if it will be much more than an advanced autopilot that still requires a driver.

I think the bigger problem with true autonomous road vehicles is primarily an infrastructure issue -- road maintenance, repair, and conditions. Just installing the infrastructure for and convincing people to adopt electrical vehicles has been a daunting task. The secondary issue the fact that it would take a long time to be evenly adopted. It is one thing to automate a well-maintained rail line with defined stops and predictable traffic, quite another to put robots on the road.
posted by fimbulvetr at 10:09 AM on February 6, 2015


Have the rich ever shown off with "gadgets"? Sincere question.

Eyeglasses, watches, electricity, phones, plumbing. All originally gadgets for the rich.
posted by uandt at 10:12 AM on February 6, 2015


Have the rich ever shown off with "gadgets"? Sincere question.

I was around when the original StarTAC came out. So I can tell you: yes. Yes they do.

The damn thing cost a thousand dollars and, more than that, implied that you had Important Business that was worth $1+/minute to talk about, not just from your car, but anywhere. They sold like goddamn hotcakes and almost overnight became a mandatory accessory for the hedge fund douchebag on the go.
posted by Kadin2048 at 10:12 AM on February 6, 2015


Just installing the infrastructure for and convincing people to adopt electrical vehicles has been a daunting task.

The problem with electric vehicles is that the cost/benefit for the individual owner isn't especially clear.

Setting aside the range issue and looking just at TCO, you'd have to drive a Leaf for 8 years, or a Chevy Volt for 26, to achieve breakeven, and chances are the technology is going to improve and decrease in price over that time, so you'll be stuck in retrospect realizing you paid a steep early-adopter tax. Hell, even hybrids didn't make sense yet, for most drivers (and that's at $3.88/gal for gas, it's worse currently). There's no point in owning either, over a cheap 4-banger econobox, unless you either drive a lot of miles or get some sort of "intangible benefit" (smug factor, etc.) from it.

The rational economic argument for buying an electric vehicle, for most drivers, isn't compelling; if it seems like an uphill battle, it's because most people are better served by buying a Nissan Versa and driving the wheels off of it.

A self-driving car, on the other hand... that's a pretty obvious benefit. If you commute 45 minutes each way to work (unfortunately common, at least where I live), that's an hour and a half of "free time" that you suddenly get back in your day, to sleep or read or whack off or whatever you (hopefully) aren't doing when you're currently driving. So the value proposition becomes: what would you pay to have that extra time—basically an extra workday per week's worth—back?

That, I think, is something that people will be willing to pay cash money (or better yet borrowed money, this is America after all) for.
posted by Kadin2048 at 10:36 AM on February 6, 2015 [1 favorite]


me: “... we should pour all our money into buses.”

filthy light thief: “The thing is that there are many sources of money. Federal funds are going into evaluating technology that is being developed by private companies who are working on the private benefit, whereas development of new bus rapid transit services (high frequency, separate and fixed routes, unimpeded by other traffic flows) are funded by public funds for the public benefit.”

Exactly! This is what is so infuriating to me about this article.

There are many ways that public dollars – our dollars – can be spent. This fellow is arguing that we need to eliminate completely our attempts to build a viable public transportation system (except for a hare-brained "control experiment" in Los Angeles) and pour all public transport funds into providing specialized infrastructure, direct assistance, and tax breaks for the private companies that are going to build these so-called "automated" vehicles.

This is the same swindle private individuals and corporations have been pulling for centuries in the United States. It has never benefitted us before, and it will not benefit us now. Public funds cannot be diverted from utterly vital and essential projects simply to provide cash for the investments of the wealthy.

This is our ancient way in America! Before the nation was founded, schemers and swindlers teemed across the Appalachians to grab every last speck of wood, of fur, of river, of land, no matter who it belonged to. George Washington himself stood helpless against this tide of scoundrels which disregarded treaties and desecrated every natural human dignity; he strengthened his government against them but ultimately despaired that the thieves would be our undoing. And to a very great extent they were. They moved into Georgia, demanding that the government help them seize the land and properties of the native peoples they found there; when the government refused, they acted in open rebellion of it, driving the Cherokee over the mountains on a Trail of Tears, and didn't stop for three quarters of a century of rapine and bloodshed. Conscientious Americans sighed, but were ultimately powerless. Meanwhile, these same schemers and swindlers – the schemers and swindlers who always and everywhere seek to exploit the public commonweal for their personal enrichment – built in the American South a trade in actual living human bodies the likes of which the world has never before known, driving an entire people into seven generations of abject slavery. When this trade in bodies was threatened, they convinced the people of America to mount a war to defend it.

When they lost that bloody war, at the cost of half a million lives, they spent generations making sure that that trade in bodies continued in every way except in name for another century. During that century, they built up their industries, forcing new generations to destitute servitude in their factories starting at the age of eight or nine years old. They moved into the West, despoiling the land with foul mining camps and carrying the plunder back East to their mansions. They built oil rigs in Texas. They drove their cattle across the Missouri, turning that great river which the books describe as pure, clear, and emerald-green into a sad, dull, brown flow of near-mud. They fought minimum wage laws, and when minimum wage laws became a fact of life for us, they successfully managed to preserve them through thirty years of inflation, so that now federal minimum wage is lower relative to the cost of living than at any time in its history. They have continually made us poor and destitute for the sole purpose of lining their pockets and getting a payout.

So now – now, after all this, I will be god damned if I'll stand by while these modern-day robber barons of Silicon Valley coerce this nation into abandoning the project of building an adequate public transportation system so that they can take our money by spinning up a phony story about magical cars that – if they ever actually do exist – will surely be sold back to us at a high premium that ensures the investors another fine payday.

Fuck that. We pour our money into buses, into trains, into public transportation. Let private industry pay for its own damned experiments.
posted by koeselitz at 10:46 AM on February 6, 2015 [13 favorites]


You could use the same argument pro electric vehicles and con autonomous cars taken from another perspective. If you live in a dense city and do mostly short trips, electric vehicles make a lot of sense. I can't see any use for an automated car, personally. I live in the city. An autonomous vehicle wouldn't make any sense to own. We walk to work, and the car only gets used for short hops here and there to go shopping, with the rare occasional long run out of town. Our average driving speed according to the trip computer in my minivan is 25 km per hour. I bet it averages more mileage in parking lots than on the road. We rarely drive more than 5 km from our house. There would be no benefit or compelling reason to by an automated car in my situation, but I can see a lot of drawbacks and extra costs.

That is the basic problem with the idea of an automated personal car utopia. It seems to rely on everyone adopting it, but there is no compelling reason for a lot of people to bother. There is a reason why the base models of cars and standard transmission still sell, and why there are a lot of old junkers still on the road.
posted by fimbulvetr at 10:57 AM on February 6, 2015 [1 favorite]


koeselitz: We pour our money into buses, into trains, into public transportation. Let private industry pay for its own damned experiments.

One hang-up on that: connected technology, developed by private entities, is seen as a boon to public safety. I'm not sure how much public funding is going into connected technology, but there is an effort to set up standards in connection and communication, if nothing else.


fimbulvetr: That is the basic problem with the idea of an automated personal car utopia. It seems to rely on everyone adopting it, but there is no compelling reason for a lot of people to bother. There is a reason why the base models of cars and standard transmission still sell, and why there are a lot of old junkers still on the road.

Three words for a swift and dramatic shift to more intelligent vehicles: safety, and insurance rates. As covered above:

haiku warrior: Don't want an autonomous vehicle? At first you won't be required to have one, but insurance companies will strongly disincentive not having autonomous safety features. No autonomy = much higher premiums. Autonomous vehicles will save many, many lives. Drunk? Just have the vehicle drive you home. No road rage, no distracted drivers, no speeding, no reckless driving, no tailgating, no sleepy drivers, and so on. Eventually, I expect that autonomy will be required, just like brake lights and turn signals.
posted by filthy light thief at 11:05 AM on February 6, 2015


And CFLs sucked for twenty years and never stopped sucking before they were replaced. The makers of LED lighting learned a lesson from CFLs, which is why they arrived nearly fully-formed—don't sell us bullshit that doesn't work and tell us we're wrong when we know for a fact that it sucks.

I'm gonna second that CFLs have been fine for years, whereas LEDs were still really harsh last I checked - unless that's changed in the last year or two.
posted by atoxyl at 11:05 AM on February 6, 2015


I do agree that having an autopilot to take over if you have a long highway commute would be fantastic. But I have doubts about automation outside of the highway scenario anytime in the near future. The thought of automated cars in this half of the century in Montreal in January is laughable.
posted by fimbulvetr at 11:08 AM on February 6, 2015


I doubt the insurance companies will let a drunk be in control of an autonomous vehicle. You will still have to have someone competent to drive behind the wheel, even if the autopilot is doing the driving.
posted by fimbulvetr at 11:10 AM on February 6, 2015


I will be god damned if I'll stand by while these modern-day robber barons of Silicon Valley coerce this nation into abandoning the project of building an adequate public transportation system

I am not convinced we have such a project, though I wish we did.
posted by Steely-eyed Missile Man at 11:12 AM on February 6, 2015 [1 favorite]


We have almost driverless vehicles already—trains and trams and to some extent buses. Everything else is Robert Moses junk for rich white people to get all Futurama about.

The real problem is that a lot of the same wealthy technocrats who keep gushing over self-driving vaporware are working actively now to gut funding for public transportation systems or "transition" those systems to unaccountable privately-owned replacements that monetize public services. The utopian language is just window dressing for the rent-seeking behind the scenes — to privatize what was public (or publicly accountable) and collect profit without creating anything new. We don't have to wait for autonomous cars to show up for that to happen.
posted by a lungful of dragon at 12:05 PM on February 6, 2015 [7 favorites]


Light bulb derail: They're getting better. I just got my first one (recommended here) and have been really happy with the quality of light. No longevity test for me yet, but CFLs never lasted even a year for me (I think poor electrical system in my rental apartments).
posted by el io at 12:35 PM on February 6, 2015


wikipedia brown boy detective: "You know what’d be better than building another bus system? Building an autonomous bus system that could self-optimize its routes and coordinate logistics with all of its customers, service providers and city planners simultaneously.

Sorry, no. Give me a fixed route with frequent service or give me death.
"

Yeah if he still cared as much about web dev as he does about 360° meritocracies, he'd know sometimes people like stuff to be consistent and reliable.

Even if on average the result is slightly faster, I could see a lot of people being put out by the idea that maybe tomorrow the bus will decide to take a different route or the bus that you usually catch at 8:15 is going to be at 8:10 because of reasons.
posted by RobotHero at 12:35 PM on February 6, 2015


The big thing about buses and particularly subways is how much throughput they can handle. The roads here are packed to capacity, and while I don't doubt that, once in enough use, automated cars could make things a bit denser, I suspect it's not even remotely to the level that a current bus system handles. (And yes there's occasional empty buses, but almost never in high traffic routes or times.) And when you consider the throughput of a subway it's orders of magnitude higher than cars are. Now I could be wrong, but a route free bus system seems like a disaster. A main route lets you get people from point A to point B quickly, if you need to make little diversions off the path then it's going to be a LOT slower.

That's not a reason why automated cars won't change the transportation system. If people are used to transportation as service, not as something you own, they change their behavior. There's almost no reason to drive from coast to coast anymore. You fly, and if you need a car you rent one at your destination. Or you take taxis. With a real automated solution the same idea can be for local travel. Invest more, not less, into public transportation. Make most arterial/inter-metro traffic public (and reliable). Then you can solve the last mile problem with automated cars. Taking a taxi a half mile is annoying for everyone involved, but it should be easy with a good automated solution.

(Honestly, it seems like that's already how some people use lyft and uber around here, although it still seems silly to me.)
posted by aspo at 1:24 PM on February 6, 2015 [1 favorite]


Yeah, I'd actually expect autonomous cars to make traffic worse in urban areas, because the people who have them won't be as frustrated at how long it takes to get anywhere because they can be doing things on their computer or phone or what have you.
posted by Zalzidrax at 1:46 PM on February 6, 2015 [1 favorite]


If they allow more people to drive who might have previously been limited by say, parking, but really for any reason, they will absolutely make traffic worse. You can fit a lot more people in a bus - comfortably! - than you can if all of those people drove.
posted by en forme de poire at 2:35 PM on February 6, 2015


(in the same area of road I mean)
posted by en forme de poire at 2:36 PM on February 6, 2015


Glad to see this getting fisked hard.

He's "not even wrong" about LA ridership and public transit policy; LA ridership has been increasing and is increasingly multimodal, especially as inner core density increases. Subways work both in getting people where they want to go and in getting people to want to live near a subway line. Urban design around transit hubs makes sense in a way that autonomous cars don't.

And yeah, I got a rueful chuckle out of the airbags thing, since (IIRC) they were the costlier alternative to the mandatory seatbelts that consumers hated.
posted by klangklangston at 3:15 PM on February 6, 2015 [3 favorites]


It's true that google et al's autonomous cars currently cannot operate in snow, but it's not like that is an insurmountable problem dictated by the laws of physics. There's no reason to think that this technology will not ultimately become cheaper, widespread, and integrated into autonomous cars and make them safer in whiteout conditions than a human driver would ever be.
posted by modernnomad at 6:24 PM on February 6, 2015


It makes no sense to hope against private car ownership -- people are too spoiled to ever give up their safe cocoons -- so all you can do is make private cars safer and cleaner, which means automated and probably electric (or some other zero tailpipe emissions system). Automate them and eventually turn private cars into lots of little interchangeable modules that can putter around individually at low speed or be locked together and managed as part of a fast public transportation system. You'll ride the bus or train or boat or plane in your own cocoon.

And I'll continue to ride my bicycle to work.
posted by pracowity at 1:51 AM on February 7, 2015


It's amazing watching a whole crowd of people fall, hook line and sinker for marketing. That article does not say what you think it does.-- ennui.bz

What do you think the article is saying? I live right next to Mountain View and I see these streetview camera versions of the driverless cars driving around all the time. This isn't autopilot. The person behind the wheel is there by law, but isn't doing anything--no hands on the steering wheel, no foot on the brakes or accelerator. The person could just as well be asleep for the whole day's drive. And these cars are driving around town--crowded, sometimes pedestrian filled streets. Maybe this is hard to believe because it is so new to people. Here's an article/video explaining it a bit.

I will start believing the autonomous car hype when they test them successfully in snowy winter city conditions when streets can be a uniform field of white. --fimbulvetr

Now there you have a point. Driving around the almost always sunny, never ever snowing Silicon Valley makes the problem a heck of a lot easier to solve than in many other parts of the country. I wonder if they are doing any tests in snowier climes.
posted by eye of newt at 10:53 AM on February 7, 2015


By the way, you can tell a regular google streetview camera car from the driverless version by all the sensors hanging off the front/sides/back. When you see one, watch the driver and you'll see that his/her hands are not on the steering wheel. When an obnoxious friend of mind sees one when he's walking, he likes to play tricks like pretend to cross the street, then stop, then start up again. He hasn't fooled it yet, but he says it is pretty conservative in its decision making
posted by eye of newt at 10:59 AM on February 7, 2015


Oh come on. Autonomous vehicles won't be for the poor.

This is incredibly shortsighted. Take a taxi, remove the cost of paying a driver, and amortize maintenance costs over a larger user base, because the radically decreased cost will meet a greatly increased demand. Subscription to an autonomous car service will almost certainly end up being much cheaper than owning a personal vehicle today.

But why take my word for it? Let's get out an envelope and see where we get.

From 2014: "AAA released the results of its annual 'Your Driving Costs' study today, revealing a 2.7 percent decrease in the cost to own and operate a sedan in the U.S. The average cost fell 1.64 cents to 59.2 cents per mile, or $8,876 per year, based upon 15,000 miles of annual driving." According to consumer reports, a decent portion of the cost of ownership is depreciation costs, so maybe we say $4500 per year in 'real' costs, if you're planning to just buy a used car and run it til it's dead. Again, for a car sitting in place 95% of the time. It's a lot of money, especially for the poor.

Now let's think about the taxi side of things. A bit of internet searching suggests that an Uber driver's mainenance costs - as a full time+, 60-ish hour a week driver - are about $15,000 per year (including the depreciation costs, and probably higher mainenance costs due to extemely increased usage). These same drivers, in NY and SF, make over $50,000 a year after expenses (which include payments to the Uber overlords), and more like $65,000 a year if they're working 'extended hours,' which is probably more appropriate for thinking about a self-driving car service. So a current Uber taxi costs $80,000 a year to run, with a human involved. This means that actual car maintenance is about 19% of the total cost of running the car.

Now, suppose we have a worker who travels 15 miles and 25 minutes on their commute. In San Francisco's Uber rates, it would cost this person something like $10,500 per year to commute by Uber. (Their pricing formula seems a little sketchy; I took $2.20+MAX(mile_rate*miles, time_rate*minutes) ). Now multiply that by 19%, and you get $1979, or less than half of the cost of personal vehicle ownership, even with the cost of vehicle depreciation removed entirely.
posted by kaibutsu at 1:16 PM on February 7, 2015


Even if that math is correct, you are now spending half of the cost of a vehicle for less than half of the use -- no grocery shopping, no trips to grandma, no dentist appointments, no kids play dates...
posted by Dip Flash at 1:21 PM on February 7, 2015


Honestly, I don't, and have never, owned a car. I'm currently using Bart to get to work, and my area in the East Bay has plenty of options for groceries, dentists, and kids, which I'm personally more than happy to bike to. So forgive me for focusing on the car as a method mainly to get to a place where one makes money to offset the car...

There's also plenty of room for tuning the taxi scenario to reduce costs further, like allowing a window of transit with a couple-few other people picked up going to approximately the same place. I imagine this would be a key feature of 'low rent' car shares.
posted by kaibutsu at 1:39 PM on February 7, 2015


"Honestly, I don't, and have never, owned a car. I'm currently using Bart to get to work, and my area in the East Bay has plenty of options for groceries, dentists, and kids, which I'm personally more than happy to bike to. So forgive me for focusing on the car as a method mainly to get to a place where one makes money to offset the car..."

Yes, living carless in SF does lead to some tech utopian assumptions about car usage.

"There's also plenty of room for tuning the taxi scenario to reduce costs further, like allowing a window of transit with a couple-few other people picked up going to approximately the same place. I imagine this would be a key feature of 'low rent' car shares."

As per the blog post upthread about an experiment running autonomous car service, by the time you have critical mass to support this you end up having too long of a delay for people to actually be able to get into the cars compared to demand and it ends up being more efficient to run buses, and from buses it ends up being more efficient to run them on fixed routes on a schedule because people will shift their schedules to meet the buses and they're more likely to take it if they have a predictable route.

Honestly, one of the biggest boons to an increased adoption of driverless cars is that it would likely also optimize traffic for free-riders, i.e. make buses more appealing versus trains because the big knock against buses is that traffic congestion is hard to predict, meaning you have to take a much earlier bus than you would a train even if the train ride is longer.
posted by klangklangston at 2:27 PM on February 7, 2015


pracowity: "You'll ride the bus or train or boat or plane in your own cocoon."

Ooh, like shipping containers. All these automated cars can just be forklifts.
posted by RobotHero at 2:39 PM on February 7, 2015 [1 favorite]


Yes, living carless in SF does lead to some tech utopian assumptions about car usage.

To be sure, I've also lived car-free in Boston, Eugene, Toronto and (most easily of all) Davis, California, after growing up in St Louis. Living without a car is a conscious economic choice on my behalf, and moving to places where it's more feasible has, I believe, greatly enriched my life. Some of these places have good public transit, and others were simply small enough that a bike is more than enough. And others - like the east bay - manage a bit of public transit, but are laid out on a human scale where the need for a car is pretty much removed. When I built my cargo bike some years ago, all of the difficulties of groceries and so on were eliminated.

So for my part, the promise of autonomous cars is to make a couple-few days a year a bit easier, when I need to ship a pile of boxes too big for the cargo bike, or move a major piece of furniture. I think There's a lot of space for them to work well in mid-size cities, like Eugene, where no one is going to build a subway anytime soon, and the transit options are actually fairly restrictive.

from buses it ends up being more efficient to run them on fixed routes on a schedule

Sure, commutes may reduce to having better autonomous bus systems, but making side trips cheap and easy goes a long way towards eliminating the need for car ownership. Meanwhile, there's a lot of non-bus routes in the world that could benefit from a car-share system: there's a lot to be gained just from injecting a greater degree of flexibility into the transit system.

This thread, and a couple others like it, have been agonizing to read. Yeah, the article was dumb, but so many of the knee-jerk responses here are... let's say myopic, uninformed of where the technology is currently at, and/or lacking imagination. What does it look like when you mix a 50-80% drop in the cost of overland transportation with the 'sharing' economy, or cloud coordination? Probably uniformly terrible, right?
posted by kaibutsu at 4:35 PM on February 7, 2015


I wish people would stop posting anecdata about how they manage to own a car for the price of scrap metal and run it on oily rags. Median means that half the group is above and half below that line. You're not disproving the statistic, you're merely telling us where you sit in that statistical population and humble-bragging about it.

You want anecdata? I paid about $800 in commuter costs last year. If I wanted to commute every single day this year, it would cost me $1500. This in a major, and expensive, first-world city. That's the real figure this guy has to beat to get me on board, not that $10k median car ownership.

Also, it took me 35 minutes each way, to and from work, time I spent reading or otherwise pleasantly engaged. So I have that now, but he wants to sell it to me again at five or six times the price. Nuh-uh.

Autonomous, electric vehicles will transform the world, but they fall far short of replacing public transport. We need them, but they cannot replace other things we also need. It's the broken window paradox in a new suit. If we spend the money kn this, the money is no longer available to spend on that,
posted by Autumn Leaf at 5:46 PM on February 7, 2015 [1 favorite]


You know, I would settle for a traffic light to send (via IR or RF or whatever) information as to how long the light would remain red or green. With that, then, my car could calculate, given present speed and distance to the light, how much and whether I should speed up a little to catch the green, or slow down a little while waiting for the red to turn green and so to avoid having to stop entirely.
posted by one weird trick at 6:30 PM on February 7, 2015


I'm in the 'human level driving is hard AI' camp, but many of the current problems could be overcome with smarter roads and traffic signals that remove the need for cameras reading stop lights with glare. Turn roads into railways. Magnetic paint could help relatively cheaply to remove the GPS dependence. This assumes a functional forward thinking highway department however.
posted by benzenedream at 8:20 PM on February 7, 2015


Ooh, like shipping containers.

Yeah, sure. Like shipping containers. Or like sleeping compartments on a very nice train that takes you door to door.
posted by pracowity at 5:48 AM on February 8, 2015


Sonny Jim: “Oh come on. Autonomous vehicles won't be for the poor.”

kaibutsu: “This is incredibly shortsighted. Take a taxi, remove the cost of paying a driver, and amortize maintenance costs over a larger user base, because the radically decreased cost will meet a greatly increased demand. Subscription to an autonomous car service will almost certainly end up being much cheaper than owning a personal vehicle today.”

This kind of ignores the realities of taxi travel, I think. You can't "amortize maintenance costs over a larger user base." As far as I can tell – though I welcome data on this – taxis in major US cities don't sit around on street corners for a large proportion of their time; they're running, with fares, almost as much as they can be. There is no expanding their userbase without an increase in maintenance cost. That's why the service that's been able to "disrupt" taxi services and compete successfully with them is one which totally ducks this question entirely: Uber, which doesn't pay for maintenance of its "fleet" at all.

For this to work, Uber – or someone like them – would have to decide that they are willing to purchase a fleet of millions of self-driving cars and take the hit of paying for maintenance of them – instead of following the current course, which is allowing the drivers themselves to cover those costs and saving all that cash. Yes, Uber has made noises about self-driving taxis, but they're the same ones who've pushed their drivers into buying cars, so it seems more likely to me that they're not planning on ever having to make the upfront investment.

What is more likely, I think: people will buy their own self-driving cars, and Uber will be a service that pays for the privilege of people using them as taxis.

But all this is totally theoretical, and frankly unnecessary to go over.

kaibutsu: “ Now multiply that by 19%, and you get $1979, or less than half of the cost of personal vehicle ownership, even with the cost of vehicle depreciation removed entirely.”

Yes – $1979 – a very optimistic figure, by the way, for the reasons I outline above. Which is $164 every month to spend on taxi trips. This is more than double the cost of a full-service monthly pass for the public transportation in most major cities.

Autonomous vehicles won't be for the poor.
posted by koeselitz at 9:02 AM on February 8, 2015 [2 favorites]


This assumes a functional forward thinking highway department however.

And one that is funded. Transportation budgets are probably being starved on purpose, though, to help drive privatization of these public services, which will dovetail nicely with networks of self-driving cars.
posted by a lungful of dragon at 9:48 PM on February 8, 2015


Jut one question : Do you plan to protect users privacy? Or do authorities learn whenever anyone goes anyplace? Wouldn't that chill Freedom of assembly?
posted by jeffburdges at 7:52 AM on February 9, 2015 [1 favorite]


As far as I can tell – though I welcome data on this – taxis in major US cities don't sit around on street corners for a large proportion of their time; they're running, with fares, almost as much as they can be.

I don't know about that; I live near a taxi dispatch center / garage and there are a lot of taxis parked there, pretty much all the time.

They clearly aim for 100% utilization of the taxis that they have on the road at any given time, but the vehicles themselves don't seem to be the limiting factor.

The BLS claims that the average salary for a taxi driver is about $10.97/hr and $22,820/year, probably much higher in big, high-cost cities (one source says the median in NYC is $38k/year). But even using the lower number, the biggest cost component is the driver, not the depreciation or maintenance of the car.

The average NYC taxi travels about 70,000 miles per year. If you bought a brand-new Ford Transit van and drove it that much, the DOE says that the cost-per-mile is about 22 cents, or about $15,800/year. (And I think that includes the big first-year hit to the depreciation, but I'm not sure of the exact curve they're using.) Even at the national average, that's less than the driver salary.

In NYC, where taxis are typically leased, the end cost to the driver for the car is "$9.58 to $11.58 per hour" and gross revenue is $26-44. Net revenue to the driver varies between $14-31/hr, which is basically in line with the BLS salary estimate. So again, the vehicle operating cost is much less than the driver's cut.

I don't see any way of cutting those numbers that doesn't make the driver the biggest variable cost. Exactly what the fares would be from a robot-driven taxi is an interesting question though. I think they'd definitely be lower than a human-driven one, although you'd have to do a CAPEX calculation to see exactly what the upper bound would be on a robot cab. My guess is that $100,000 per unit—given that each one could replace more than one driver—would be easy to justify.
posted by Kadin2048 at 10:37 AM on February 9, 2015


a lungful of dragon: "Transportation budgets are probably being starved on purpose, though, to help drive privatization of these public services, which will dovetail nicely with networks of self-driving cars."

Oh man, yeah, they start switching everything over to private toll roads, automated with GPS tracking and stuff. After that transition is complete, the road companies start lobbying against "road neutrality" regulations.
posted by RobotHero at 11:02 AM on February 9, 2015 [1 favorite]


Jut one question : Do you plan to protect users privacy?

How do you plan to do it now?
posted by pracowity at 11:23 AM on February 9, 2015 [1 favorite]


The UK Just Made Itself a Fantastic Place to Test Self-Driving Cars (Wired, 12 Feb. 2015)
The United Kingdom just became the best place on earth to develop self-driving cars. Yes, the land of extortionist gas taxes and intense emissions regulations, where you have to pay extra money just to get your car into central London, is taking what amounts to an anything-goes approach to real-world testing of autonomous driving technology. It’s placing no geographical limitations on tests, not requiring special licenses or permits, and even opting not to require additional insurance.
...
Under the “code of practice” the UK expects to finalize within months, companies like Nissan, Audi, Mercedes-Benz, google, and Delphi will be allowed to do just that. Those that have finished preliminary testing on closed tracks and want to begin real-world testing need only have a trained driver at the wheel and an on-board data recorder gathering information about the car’s speed, location, steering and braking inputs, and whether the car’s in autonomous mode.
...
“Driverless cars are the future. I want Britain to be at the forefront of this exciting new development, to embrace a technology that could transform our roads and open up a brand new route for global investment,” says Transport Minister Claire Perry.

Along with the report, the Department for Transport announced $29 million in funding for three driverless car trials: a shuttle in Greenwich, a “pod” in Milton Keynes and Coventry, and a Wildcat developed by BAE Systems in Bristol.
And I find it a bit hard to see nefarious plots behind cutting transportation budgets. It's one thing to conspire against public transit to force people to buy (self-driving) cars, but it's another to underfund the basic infrastructure that vehicles need to get from point A to B. Everyone needs reliable roads and bridges, even autonomous cars.
posted by filthy light thief at 12:29 PM on February 12, 2015 [1 favorite]


Taxi medallions aka rent seeking is the major cost, Kadin2048. Yes, the driver would come second, just because people are usually more expensive than machines, but basically the medallion system skims "what the market will bear" from the customers for the benefit of investors.
posted by jeffburdges at 1:46 AM on February 13, 2015


When you look at the places most invested in autonomous vehicles, it's small, contained nations like Israel or now Britain. And that's no accident: their small size means they have a different set of needs. But I can well imagine that NYC taxi medallion owners (who aren't actually driving the cars most of the time, after all) would actually set the stage for a similar takeover in Manhattan. What better way to maximize their investment? (Taxi medallion have gone for a million bucks.) Now, I see a certain value in taxi medallions for reducing congestion (the cartel costs are a side benefit) but the way they encourage fleet conversion is perhaps more important. The Taxi of the Day After Tomorrow could easily be autonomous (perhaps with a driver still required to "monitor.")

In fifteen years some of the people in this thread are going to have some serious "told you so" action available, if they want it. I can't wait to find out which!
posted by anotherpanacea at 4:33 AM on February 13, 2015


« Older "To a scientist, that’s just incredible.”   |   UNFINISHED Newer »


This thread has been archived and is closed to new comments