Kevin Drum is asking an interesting question and coming to a conclusion that I think is completely wrong:
But here’s a more interesting question: after driverless cars become widely available, how long will it be until human-driven cars are made illegal? I say ten years. It will vary state to state, of course, and there will likely be exceptions of various kinds (specific types of commercial vehicles, ATVs meant for fun, etc.). Still, without a special license they’ll become broadly illegal on streets in fairly short order. The proximate cause will be a chart something like the one on the right.
I think this is an interesting question, but when I went to visit my in-laws last month, there were still horse and buggies on the road. And those have been technologically obsolete for a century now.
There are a few things that I think Kevin is getting wrong. First, there is a massive distributional issue. Driverless cars will by definition be new cars. The first wave of driverless cars won’t be 100% adapted. Some people will be technophobic, others will like driving sticks, others will be reluctant to put their life into the hands of a piece of software even if that software is statistically a much better driver than the average human (as we are all above average drivers in our own internal estimation it’s just those assholes who are honking at me that can’t drive). And others will decide that they don’t want to spend the money.
Even assuming that there is a fairly rapid shift in the share of proportion of driver controlled and driverless vehicles sold over a couple of years so that in five or ten years from the first good autopilot to 90% of new cars being sold are driverless or driver minimized vehicles, there will be millions of new vehicles that require drivers on the road.
No state government is going to tell tens of thousands of middle class or better voters that they need to junk their $20,000, $30,000 or $40,000 capital investments for safety reasons.
Furthermore, the used car market lags the new car market. My primary used car in high school was made several months before I could walk. Factory fresh, it had sixty six horsepower and by the time I bought it for $50 it could just hit 65 MPH going down hill with a good tail wind. It did not help my social life in high school but retrospectively that car kept me out of a lot of bad decisions simply because the car simply would not allow me to show off and be stupid.
The typical American car has at least a fifteen year lifepan.
There is no way any state government is going to successfully tell most of its working class voters that they need to scrap a $5,000 to $20,000 capital investment for a marginal safety improvement.
What is far more likely once there is good data on operational usage of driverless cars is that they will be treated like anti-lock brakes and skid-control features by the insurance companies. Driverless cars will receive a massive insurance discount because they’ll be far less risky as they remove the most common source of error (human error) from the equation. But driven cars will still be available and still be insurable but at a higher rate.
Sad that Car & Driver magazine will be called Car & Passenger in the future.
Gin & Tonic
Given how slowly GPS-tracked usage-based auto insurance pricing is catching on, I think anything in this area will move more slowly than even you imagine. Pricing models and rate filings, as you know, take time.
Having worked in the computer industry for 40 years I can assure you I am not rushing to buy an auto-auto. I hope the companies that do produce them and the ones that write the software have really good liability insurance. The software will fail and in doing so people will be hurt and killed. When cases are brought to court the jury will side with the human against the computer until proven the human did something so stupid another human would have been amazed.
Then there will be the issue of hackers.Not just the ones who would find great fun in limiting your car to 26 MPH or make your breaks not engage for laughs but serious people who want to tweak the software. I work with a guy who is custom coding the firmware that runs his Saab so that it increases the performance. He does not really care what the ramifications of the mods are as long as he gets a lot more torque (yes, he is also an ammosexual Xanist). What those folks will do is an exciting future I am not interested in but will be forced to share.
Well said Richard. I could see Drum’s scenario in like 50 years, but 10? That’s outrageous and would require some government handout a la digital television but obviously more expensive to even be considered.
Screw driverless cars. I’m sick of every car geegaw thing fetishized over like it was GM’s Futurama exhibition at the 1939 World’s Fair while non-driving–the actual idea of moving people and things for utiltity and need–goes begging.
The best thing this country can do for drivers is get drivers to pay people to not drive, to spend on systems that get cars off the road. Otherwise we’ve paved ourselves to a damned standstill.
First there will be lanes set aside for driverless vehicles, then specific roadways.
But never all of the latter. Both types of vehicles will have to co-exist for many a moon.
Likely will be an insurance incentive, prece-wise, for driverless vehicles.
Gin & Tonic
@Schlemazel: Commercial airliners mostly fly themselves without mysterious software failures.
I wonder what effect the inevitable crashes between driverless and driven cars will be on the rate of adoption? If it becomes clear that tens of thousands of deaths each year could be prevented if the people who caused the crashes didn’t insist on controlling their own vehicles, maybe driverless cars would be adopted more quickly? The insurance differentials might be massive if that proves to be the case…
This could be a looooooooooooong time. I, for one, would not trust a computer-driven car going 70 mph around my car filled with family. One computer glitch, one freeze up, one serious hacking, and I could be toast. Yes, people cause accidents too, but in general I trust the judgment of a human driver in a non-ideal weather condition than a computer. I think most humans feel this way. Or maybe I’m just old.
@Face: Or you haven’t looked at the data.
@Face: How could you not trust your new “Camry, Featuring the Windows Driving System”?
/warning: joke only applies to loyal Mac users
The trouble with the horse-drawn buggy analogy is that buggies are safer than cars… slower, yes, but the rates of serious accidents, injuries, etc. are just trivial.
Now, [email protected] makes a very good point, that juries aren’t going to find criminally negligent drivers criminally negligent when the other option is finding a car company criminally negligent. That’s just something the legal system is going to have to solve, possibly through highly pointed jury instructions. And possibly also by requiring disclosure of the AI piloting code.
@Gin & Tonic:
Have you read Patrick Smith’s Ask The Pilot blog? (Salon should never have let him, or his blog, go.) He keeps having to repeat that modern planes do not, in fact, fly themselves, and that human pilots are still very much needed. The software automates some of a pilot’s work, but nowhere near enough to fly the plane by itself.
I don’t think driving will be made illegal but I think insurance pricing might make it a moot point. Also is it possible old cars could be retrofitted with new technology. Driverless cars have such far reaching consequences,making driving illegal is not even in the top five issues. I wonder how many DINK couples with regular hour office jobs would give up one car.
Realistically, the earliest they can ban human-driven cars is ~15 years after they stop selling new human-driven cars in the U.S. Because many cars last that long. (I drive a 2000 Accord.)
And I think Kevin’s being incredibly optimistic about when self-driving cars are (a) introduced, and (b) get down to a price point where a lot of people buy them. I’d be surprised if (b) was within the next dozen years, say 2028 at the earliest.
And then there will surely be at least a decade where both human-driven and self-driven cars are sold in large quantities, which gets us to 2038.
And then another 5 years before human-driven car sales taper off to the point where the manufacturers are close enough to ready to stop making them anyway that they’re willing to go along with a government-imposed halt to new human-driven cars. So 2043 for that.
And then 15+ years for the last of the human-driven cars to get old enough so that the government can buy up the last of them at Kelly blue book prices: 2058-2060.
So I won’t live to see it. (And this is my optimistic scenario.) Too bad, really – I’d like to see the day.
This is not even a first world problem this is a first world problem in distant future. The statistics that driverless cars are safer is based on what sample size exactly? Doesn’t it also depend on traffic patterns other vehicles on the road etc. Has anyone other than Google looked at all their data?
@Gin & Tonic:
But, despite 6 degrees of freedom, the issues faced are much simpler. A 737 is not surrounded by 100 other 737s, some turning left, some pulling out from the curb and a lot of other issues. The number of lines of code is much less than it will be for a car I bet. Also, because of the margin in a plane cheap-screwing everything including the programmers is less likely. Add to all that, for a number of reasons are not attacked by the black hat crowd nor by sky-jockey pilots wanting better performance.
And get off my lawn! I know part of my attitude is old-man syndrome.
What if this driverless car needs to reboot when you are going at 70mph, after a firmware update.
Like I said in the last thread, there’s a lot of wave of the future bullshit in this stuff, and urban dwellers seem to have some basic inability to grasp the scale of the infrastructure build required to make it work out in the sticks.
The railroads haven’t done it, it’s cheaper to pay train crews over wiring up and maintaining a RC solution over a 500 mile mainline.
I’d also like to see the math on the efficiency and safety gains in the urban environment. Those are some complex and dynamic flows, and I kind of doubt the gets are all that. Slapping in some Roomba level software and avoiding curbs is one thing, improving average commute time is ah… bit more complicated.
Driverless freight vehicles make a lot more sense in the foreseeable future than do driverless passenger cars. The contents of the former are more forgiving of rapid maneuvers and crashes than the latter. I can easily see an Amazon van (e.g.) working from its local depot and making its optimized rounds (cf. the Travelling Salesman problem). As with rockets, human-rated vehicles are a lot more expensive than the other sort and thus harder to insure.
On the psychological side of it, who would spend a lot of money on a personal vehicle they can’t control?
peach flavored shampoo
I dont get this idea that computer driven cars are auto-matically safer. There’s a lot of body language involved with driving, at least in urban settings, that disappears when the driver does. Consider: hand waves that allow the pedestrian to know the driver is waiving his right to cross the intersection at that moment so he/she can cross. How will a computer know there’s a pedestrian in need of crossing the intersection? Or humans that can stop when they see a lost child or pet or road hazard on the side of the road and take care of them/it. What about subjective decisions, like slowing down or changing lanes to allow on-ramping cars to merge…how does a computer car know when and why to exercise this courtesy? More importantly, how does one program a car to drive in a blinding snowstorm?
@cokane: Not even in 50 years.
Everyone – including Drum – is confusing illegal with unaffordable.. What is far more likely to happen is that, as driverless tech becomes more available and reliable, the cost to own it will drop – and the cost to insure and operate a vehicle without it will begin to rise. Eventually insuring a manually-operated vehicle will be equivalent to current rates for flying a private aircraft, and a license to operate a manually-operated car will require similar – if slightly less stringent – training and certification to a pilot’s license. And as “driverless” and “electric/hybrid” are likely to reach comparable levels of acceptance in the market at relatively similar times, the manually driven car, which is less unlikely to be petrol- or diesel- powered, will have significantly higher costs to operate. There won’t need to be laws when only the 1% can afford the petrol, the training and the coverage; but there won’t ever be a law telling people what they can and cannot drive.
Drum’s “they’re coming for our cars” is as sensible and likely as Bundy’s “they’re coming for our guns”.
@Betty Cracker: google said every accident they had when they were test driving the driverless car was the fault of the other car driver.
@low-tech cyclist: the driver-less car wont start as a bonanza among regular consumers, it will probably take off first as a taxi service and/or sharing thing like Zipcar. I’m guessing the first majority users of driverless cars won’t be actual car owners themselves. So I think you have the wrong focus.
I would think a driverless car/truck would be such easy pickings for carjackings. I can see the Mob really making a killing off-loading jacked truckloads of Bose stereos and boxes of jewelry….
Thinking about driver-less cars, I realize that I would probably like to try one. I don’t drive and the reason I don’t drive is a problem with my eyes and sunlight, also a problem with attention span. If a driver-less car could take me short distances it would make my life easier. Taking buses means I often spend much more time commuting to a store than I spend in the store. It would be safer for me (and others on the road) than driving myself.
@BGinCHI: Wasn’t there some big hoopla not so many years ago about a USN CG that was fitted with an automated navigation and control system – which crashed when the ship stopped because the programmers had never thought about sitting still and never allowed for “speed=0” in the programming?
Sorry to interrupt this thread but I’m bored waiting for my tests to start & just cruising the tubes when I cam across this:
Chess forbidden in Islam, rules Saudi mufti
It is sad that the region of the world that introduced the West to the game, really the area of the world that saved knowledge while the West set itself aflame with ignorance and superstition and then returned knowledge to us without charge has chose to return itself to those blackened days of yesteryear and try to attain its own Dark Ages.
@peach flavored shampoo:
From Google, the makers of driverless cars is promoting this idea.
@Gin & Tonic:
Which is why commercial airliners have a trained pilot and a backup pilot. Because the plane flies itself.
The software on airliners is super-redundant technology that is there to assist the pilot not completely replace them.
And as I posted on Drum’s site – I think the same thing will happen with “driverless” cars. A lot of this tech will be put into place to help the driver, and there may be some situations where “super cruise control” can take over for a while (such as long-distance interstate drives, or well-mapped urban areas), but I find it doubtful that we’ll reach the point where all cars are driverless even in my son’s lifetime, let alone my own. The field of AI is full of 90% solutions where people are sure that last 10% is just around the corner – until it isn’t. The advances in driverless cars are real, but advancement in the AI field is rarely linear – it’s usually logarithmic. We had a big explosion in understanding that pushed us way up the curve, and now even an equivalent explosion will only push us just a tiny bit further up the curve.
We already have “driverless cars.” You know. The “drivers” that are talking/texting on the phone.
Why should insurance work any differently than it does now? You buy insurance, and if your vehicle was at fault in an accident, your insurance company pays out, regardless of whether you or the computer was piloting the car.
And it’s not like anyone’s found criminally negligent these days when they drive their car into a pedestrian or whatever. Why should that change? (The car companies will have good lawyers, and will bring up piles of comparable cases with human drivers where no charges were filed.)
So has the driver of every car that has ever been in an accident.
@Mai.naem.mobile: Why should we believe them? Aren’t they an interested party?
@BGinCHI: I’d subscribe to the edgier Autopilot Overriderz
@cokane: I know how often my computer needs to be restarted. I know how often it freezes up. I’d prefer my fellow highway travelers not require said restart at 110 km/h.
Besides, what data? Please share with me the peer-reviewed data (read: not shit published by Google) that shows a significant reduction in traffic accidents versus a human control group.
For some reason, this discussion reminds me of the famous ‘Slow Down’ sketch in Taxi. The premise here is that Reverend Jim has just been hired by the taxi company, and so has to take the written driver’s test. Reverend Jim reaches the question “What does the yellow light mean?”
Because we are not talking about 1 accident, we are talking about a Toyota automatic accelerating, Takata air bag level of accidents, deaths and dismemberments. Yeah, those companies all have some insurance but ask VW and its insurance company if they were fully prepared to cover the costs that are looming in their future for their software screw up.
@Punchy: Lock it down and use a encrypted key software to guard entry to the comupter, and it seems to me you have a nice mobile warehouse. YMMV, literally.
mike in dc
I don’t think that driverless cars will be 50-state legal until probably 2030(though they will be legal in some states several years earlier than that). I think they will become the majority of cars on the road within 10-15 years of that, and a super-majority by 2050. I think non-autonomous vehicles will require special permits after around 2060 or so. Just spitballing.
@Face: Data already existed before Google even began their project.
But, it seems like you already have some strongly held preconceived notions, so I suspect nothing will change your mind.
I know I’m a lousy driver. Can’t wait to turn that thankless job over.
I do a lot of driving in the Colorado Rockies. I wonder how driverless cars do in storms and continually changing adverse conditions. Modeling chaos theory is really difficult
That said, computers might be better than the worst of the drivers out there is crazy conditions
Speaking of reliable software, my dear, dear, friends in Redmond have been pimping their tablets on NFL games all year. How did they perform on Sundays big stage? Well The Patriots say their all crashed in the middle of the game. M$ has of course officially said they fault is not with their software.
Word is that instead of “OMAHA” Manning will be shouting “BLUE SCREEN OF DEATH!”
@NonyNony: The single best result I can see of implementing driverless auto techology is the increased pressure on drivers to be merely as good as their electronic counterparts. Assuming the driverless tech is more reliable and safer, the pressure to be as good a driver, not merely as your neighbors but as the massively researched and tested driverless tech you now have to share the road with, will be significant. That is nothing but good.
Add to that the first TXTnDRV citation that winds up in court, where the car’s logs are subpoenaed – and prove that Junior was in “manual control” mode even though his fingers and eyes were on the phone and not the road. Between those two influences, the people who remain in manually-driven cars will become significantly better drivers as well in very short order.
@Schlemazel: I work with computers too and can shout at clouds with the best of them. OF COURSE self driven cars will glitch, and be hacked, and cause crashes and kill people. It is the nature of technology.
The thing is, if progress could be made … if we could have self driven cars that kill 90 people a year (instead of the ~90/day currently killed by driver driven cars) we would be in a much better place.
I was recently T-boned in an intersection, totalling my 2015 van*. It wasn’t a close thing, it wasn’t someone trying to stretch a light. It was someone just completely ignoring the red and plowing into the intersection full speed. (an interesting note: When side curtain air bags go off it is really disorienting. One second you are driving normally, the next second you are in a tunnel that came out of nowhere)
But that wasn’t a software glitch, so y’know … whatever. Look the other way.
I want self driving cars, not because of me, but because of you. And yes, I _do_ expect that glitchy, hackable, limited computer drivers will do a better job than you currently do. They don’t need to be perfect. Just better.
* replacement value insurance – I don’t even remember why we went with it, but I am _so_ glad we did.
@Wag: Computers would probably slow the vehicle down in adverse conditions. That alone would make them better than 90% of the drivers out there.
Exactly. When states wanted to use red-light cameras, they passed laws saying that the owner of the car was responsible for the ticket no matter who was driving (because the cameras don’t identify the driver.)
I, for one, welcome our new automatic driving overlords.
My job has me driving all over western NY. I would love to be able to let go of the wheel and be productive. I would be fine with using the technology only on the thruway/expressway and major main roads.
I’m looking at getting a new car, and I’m looking at the kinds of safety features people are scared about because they don’t understand the technology. I think lane departure warnings, cruise control that automatically adjusts your speed, and the ability for the car to keep you in your lane if you fall asleep/get distracted, are great safety features and should be on more cars. Again, I drive all over the place, sometimes leaving early (6am) and getting back late (midnight).
I suspect the people who were scared of steam engines or the first automobiles used similar arguments against those vehicles: I know when my horse is getting tired. I can control my horse. I can feed my horse anywhere.
I do think that automatic driving vehicles are the future, but I also think we will always be able to override/take control of a vehicle and drive it “old school”
You forgot to mention the patriots who will never give up their FREEDUMB to manually operate machines of death just b/c some Islamofascicommiefurrinervegan plot.
That describes my mother to a T. She’ll be the first to demand self-driving cars be made mandatory… for everyone else.
@Schlemazel: But VW’s software didn’t screw up. It did exactly what it was supposed to.
Rob in CT
Glad you’re ok.
Re: insurance… shouldn’t the insurer of the person who hit you pay, given that it was blatantly their fault? Or were they uninsured?
Re: driverless cars – I too think it will come via slow feature-creep. We already have some of this stuff. It will keep getting added on and added on.
Eventually, probably 50 years from now, we might get to fully automated. One issue, though: you say all the computers have to do is be better than humans. In a rational world, yes. In a world populated by actual humans, no: they will have to be MUCH better. I think they can be, but it will require a lot of infrastructure we currently lack.
There is a possibility that insurance rates for driven cars will go down too — since as the percent of ai cars go up the percent of drunk drivers, incompetent drivers, drivers having a heart attack will go down and there will be less collisions generally (if you buy that the computers will generally be safer). Safer driver for many will mean safer driving for all.
Similar thing happened to me a few months ago, except the person decided to turn left in front of me (“I didn’t see Central Planning”). I was able to swerve so they clipped the back end of my car. Side airbags deployed, went off the road. $16k in damage to the car.
I wonder if a computer-aided driving system would have been able to calculate that accelerating would have been better than hitting the breaks. If the other car had been self driving or had other automatic driving enhancements, perhaps the car would not have been able to make the left turn.
@NotMax: This sounds right. Like the way E-Z Pass went from a time-saving convenience to a necessity in some areas — a lesson I learned the hard way when I mistakenly used an E-Z Pass-only exit on a Pennsylvania highway. I was billed $60 for 10 minutes of highway driving. Ain’t right — but it does force the issue.
We’ve had driverless cars for many many decades now.
They’re called trains.
Saw a NewYorker cartoon last week. They typical “pulled over by a cop” template. But the cop is asking the guy “Do you know why my car pulled your car over?”
I think the single best consequence of driverless cars will be that most people in urban areas won’t own cars any more. Instead, you’ll have a service where you can call a car to you when you need it. No more need for parking lots taking up space everywhere that is empty much of the time, no more maintenance. In short, the equivalent of mass transit that is personalized enough that people who can’t stand sharing a train or bus will use it. It won’t have all of the upsides of mass transit, but it will have a lot of them.
Just break out your tablet and phone and be productive. It seems to be what a lot of folks are doing when I go to work
One thing I expect to see is much tougher licensing standards. We let people drive today who are marginally capable just because their lives would be so difficult without the mobility a car gives them. Once driverless cars are readily available and clearly much safer, that attitude will change. The test to get a license in the first place will be made tougher, and it will be much easier to lose a license. We’ll stop looking the other way for people who struggle with the vision test or who have age-slowed reflexes, and many more driving infractions will result in losing your license, and for a longer time.
@Waldo: I got nailed by that shit in Orange County. There are no booths and they take a picture of your plate. The fine was insane but I think I got it reduced.
Anyone here have a kid who is learning to drive? I think driverless cars will follow trends with automatic vs manual transmission.
At some point manuals stopped being taught to new drivers and if you wanted to learn that you would seek out the extra training (from mom/dad whatever). But basically that level of driving is for enthusiasts or people who need the performance difference (snowy mountains etc).
At some point teenagers are going to say “hey it’s easier and cheaper to get certified to own a driver less car”. When that happens turn on Drums 10 year countdown.
but it is the scale of the issue I was referring too. Although now that you bring it up VW knew it was screwing up but still was not prepared for what followed. It will be interesting if they have insurance for such things how that company would react as I assume the could walk away on the grounds VW knew it was taking an unacceptable risk and didn’t share that info with the insurer.
@Schlemazel: A Saab-driving ammosexual? Now I’ve heard everything.
Seeing exactly how many people use their car as a moving party pit, complete with booze, the driverless car can’t come soon enough.
I know. It’s terrifying. I once texted my son during his drive time with the drivers ed teacher. His phone was in his pocket. It chimed and he left it there and didn’t try to fish it out and see why it was buzzing. He said the drivers ed teacher was impressed. That scares me too – that should be the default for everybody.
In fact, wasn’t there a lawsuit relatively recently (NJ maybe?) where the person who texted a driver was sued because they knew the textee was driving and that caused them to have an accident?
I drive a GM car, and there have been a sickening amount of recalls on it. A few months ago I was driving on a busy highway and the power steering died. I could barely turn the wheel. Turned out the electronics controlling the power steering were under recall. (I never got a letter about that, oddly.)
Also, my ignition switch went under recall. I DID get a letter about that one. People died because the ignition would turn off while they were driving 65mph + on the highway.
So sure, I’m completely comfortable getting into a fucking GM car that will drive me where I want to go.
And of course I’ve used GPS for driving in unfamiliar areas, and it’s made so many stupid, roundabout decisions while being completely clueless about detours due to roadwork.
So yes! GPS combined with those smart, honest people at GM! Sign me up for a driverless car.
Just One More Canuck
@cokane: That 90% of accidents are the result of human error doesn’t come as any surprise. What Face was asking was where is the data (not conjecture or PR puffery) that indicates that driverless cars will reduce accidents
However, I think that was a mechanical problem and the weight of they keys would pull the ignition into the off position. I see more and more cars with push-button start and RFID keys. Seems like that is the wave of the future too.
@Central Planning: A mechanical problem they ignored for many years.
I think it was too expensive for them to fix, or something… until enough people died that the gubmint got involved.
Americans have love affairs with two of our most deadly forms of technology — guns and cars/trucks.
When word goes out that the government is going to take away “our cars,” I’d expect a massive backlash.
With guns, it’s a minority of Americans, but a largely crazy, irresponsible minority. It remains to be seen how crazy the car and truck nuts are and just how many of them there will be. Unfortunately for these lunatics there is no mention of automobiles in the Constitution.
Personally, I think the real driverless car scenario is that people don’t buy cars – the driverless cars become an extension of ride-sharing services, initially used by people as a substitute for taxis/car services/Uber/Lyft, then leading to some people not buying cars at all who otherwise would have. In that scenario, Drum is sort of right, in that relatively few people will get driver’s licenses, so being allowed to drive will be more unusual.
Still, it seems likely that his estimate is off by decades, and it’s going to be gradual, in any event.
I love the idea of driverless cars, but this absurd notion that computers are infallible belies the fact that human beings are still far better at driving than computers are. Most of us will drive for decades without being in an accident. The idiot driver is the exception, not the rule. And humans are far, far better at dealing with unexpected conditions. So definitely if you’re tired, or drunk, or distracted by the latest text from your cousin, switch on the A.I. But if you want to make sure you get back home at the start of a blizzard, you’re better off keeping control of the vehicle yourself.
Not so much. Remember, the biggest thing driving the paranoia that the government is coming for our guns is the gun companies, who are worried that tighter limits on gun sales will hurt their profits. In contrast, the car companies will love the idea of forcing people to ditch their old-style cars and buy new, self-driving ones.
And the love affair with guns is mostly limited to our RW friends (with some exceptions) while the love affair with cars cuts across the wide political spectrum. The Romance Of The Road is a major theme of Rock music. Springsteen hasn’t written songs celebrating taking a train or greyhound bus, as far as I know…
@Rob in CT:
I really don’t understand how that all worked. My insurance company said that other driver’s insurance accepted 100% responsibility. But the settlement and check still came from my insurance. I’m assuming that the 2 insurance companies settled up in some manner, but I really don’t know.
One thing that really surprised me was that with 3 obviously totalled cars and a 4th damaged, the police had no interest whatsoever in taking a report. As soon as they determined that there were no injuries, they told us to exchange info and focused solely on clearing the intersection.
White suburban guy, the car is about the only thing about him you couldn’t pre-suppose after you talk to him for a while. Apparently there is a very active Saab modification community online.
Cite? It’s not just high speed/expensive crashes. It’s the little things too: parking lot dings, backing into vehicles, black ice detection, hitting parked cars, keeping cars on the road if the driver falls asleep, drunk drivers, distracted drivers, and there’s probably more that I’ve been a part of or witnessed that I haven’t thought of.
The part that’s really interesting/scary is the *economic* impact of driverless cars. One of my favorite youtubers, CGP Grey has a 15 minute video on the subject that I think is worth the time.
Like Uber, but without them pesky drivers?
I know it is illogical, but I will probably quit buying a car when and if this comes to pass. I can’t stand to have anyone but me driving. And there is no way I trust the technology and never will. Like I said, illogical but that’s just the way it is.
Plus, I hate that money is even being spent on this when we have little to no mass transit in this country. If I refuse to ride in the car I can’t control, I at least want another option. Hell, I would get rid of my current car in about a nanosecond if there was any sort of mass transit worth the name around here.
@Schlemazel: Exactly. Me too. I’ve been in software development for going on 30 years and no way will you find me riding in one of those by choice.
@Germy: Most trains still need drivers. When a subway train left Braintree without a driver last month, it was a very big scare
The problems I see for driverless cars is that purely legal operation is completely impossible in some places, and adopting the conventions that cover for this would be unprogrammable. When parked cars block the view, how do you pull out from the legal stop point? Street lights hanging wrong, or out of order & a cop is directing traffic? Can a driverless car even tell a traffic cop from a wino?
Super Dave Osborne did a bit on the back of a Grayhound It cracks me up every time I watch it.
@Redshift: It took 57 responses to get to this. Once you don’t need to find a parking space, using a car is really easy. Reducing requirements for parking is also a huge win for cities over
Also, because of this, no need to buy a car at all.
As to software hackers, my suspicion is that they are a very small number. How many people rootkit their android phone? How many rootkit their apple phone? Is that number rising or falling? If falling,
then the problem is going away.
@Gin & Tonic: Yes, and the FAA has now come out strongly against pilots utilizing autopilot all the time because the pilots lose proficiency and don’t have the necessary reaction time when a casualty happens.
@catclub: How many does it take to cause a major event during hi-traffic times of the day?
@Downpuppy: I wasn’t clear, sorry. I meant the passengers don’t need to drive the trains. They sit back and leave it to the (human) engineer.
They can the keys to my air cooled 911* when they pry them from my cold, dead hands. ;-) Back me up on this Mustang Bobby.
* I haven’t actually bought it yet but the funds are there and I just have to find the right one. Mark my words, one will be in the garage before the end of spring.
Internal memos from GM on the Corvair and Ford on the Pinto show that both companies knew their cars would kill people & that the cause of death was preventable. In both cases the Boards of Directors weighed the cost to fix the problem against the potential for lawsuits. In both cases it was cheaper to let people die. That is the mentality consumers are up against when it comes to driverless cars.
Also, this is a good reason to argue against liability limits. Courts have the power to punish offenders when they have reason to believe the offender made these sorts of choices and skate on earlier lawsuits. Limiting liability tips the scales even further in the ‘let people die, it is cheaper’ direction.
@Gin & Tonic: And yet we still have pilots and copilots in the cockpit!
And honestly, it’s a bit different when a close call with another vehicle means one came within 5 miles of you.
I just hope driverless cars are more considerate to bicyclists and pedestrians as my kids and I commute by feet or bike three seasons of the year.
it’s true, this happens with Teslas literally all the time. they crash left and right due to random firmware upgrades and computer crashes. serious death traps.
@Amir Khalid: More like Car2Go, except that the car drives itself to your front door at the time you want it, rather than that you have to locate the nearest one via the app and walk to it.
I don’t think you will need to buy one. on-the-fly rental. I think GM’s Maven? project is already on this.
A phone app to enable you to drive a given car.
@cokane: What data? People opining about robot cars state with absolute authority that they will definitely be safer, but this is merely wishing on their part until this is proven in the real world.
It simplifies the paper work to go through your own company. We did that when a workman set fore to our house. Our insurance paid & then collected from his. I have no idea if they got a discount but I did find out later that the claim counts against us with our insurance company! It has the same affect on rates as if I had been at fault. You might want to ask about that
There is a driverless train line which has been part of the commuter train system in Kuala Lumpur for 18 years now.
@C.V. Danes: How many will keep doing it after it is treated as terrorism?
@Schlemazel: All we need is driverless cars combined with a republican president and congress writing liability laws.
I’m glad I moved to a town where I can walk for most of my errands.
I -think- there are many more cars than trains, so the market for the software will be much bigger.
@Amir Khalid: Italy has some driverless trains.
I spent the first 14 years of my career in AI working on problems similar to the driverless car. This was a long time ago, but I periodically check out the state of AI research to see if they’ve solved any big problems yet with some wonderful new algorithm. The answer is no. A while back, out of curiosity I looked into how the driverless Google cars work and the answer is not what you would expect. These cars do not have anything approaching something like situational awareness. The idea that one of these cars is going to analyze what is happening as an accident is occurring and make a better decision than the average human is ludicrous. While this may happen in the future, the technology is nowhere near that state now.
For those interested: the reason it appears AI has made advances is that computer hardware is so much faster and has so much more memory that the same algorithms developed years ago that were pretty much impossible to run can now run on your smartphone.
There’s Downbound Train off the Born In The USA album, but that song is no celebration of anything.
@Central Planning: I remember Dave Osborne from the old Smothers Brothers show. He was the cop who’d show up at parties to arrest everyone. Funny man. He was on Seinfeld’s “comedians getting coffee” show last year.
@Schlemazel: a workman set fire to your house? Holy shit, my worst nightmare. I’ve witnessed some serious contractor blunders, but never anything like that.
You know, once we all buy in to the meme that humans need robots to get around, then what else will we be compelled to turn over to the robots? Spouse abuse is a big issue; maybe we need the robots to choose our spouses for us since our track record os pretty bad there, too. How about child raising? Perhaps we should turn that over to the robots, too, since that seems to be another area where humans need help. Where does it end?
Btw, I think that other countries will be way ahead of the US on this, probably by at least a decade. For electric cars as an example, 14% of new cars sold in Norway are electric compared to 0.7% in the US, and they are looking to phase out the sale of fossil fuel cars in the mid 2020s.
@catclub: I dunno. About the same as now. What would stop someone from launching the attack from Russia or China, assuming that these robot cars are also Internet connected?
@BGinCHI: When DED, jr. was still the bigwig there, calling Car and Driver Car & Driver (sic) would have gotten you banished. Similarly with C&D vs C/D. They had this rivalry with Road & Track, you see…
No idea whether they care about such things now.
(Who has a brain that is far too cluttered with trivia like that.)
The first place where we’ll see this is on freeways. They already have controlled access, which is a huge software simplification. I can pretty easily see these as a trial run, where there are initially machine-only lanes – similar to carpool lanes – that then can become machine-only roads.
After all, cruise control exists and it’s useful. I think that this is a logical extension of this, just like automatic transmissions are.
Gin & Tonic
That’s called “subrogation.” A very common and well-understood process.
@Marc: Except that many people still drive a stick because they feel in better control of the car, myself included.
Trains don’t even seem to compare, they’re only on tracks which really simplifies the context. Plus, like planes, they’ve still got oversight that’s not automated and keeping an eye on the context and really tricky elements, Air Traffic Controllers for planes and I don’t know, Rail Traffic Control for the other. The sort of people that could figure out how to safely halt that light-rail train that took off without its driver and the dead-man switch by-passed.
Actually, the FAA definition is 500 feet for what it calls a near mid-air collision. In general, I think commercial jets are considered too close when they’re nearer than a mile. At 600 mph, a mile is about 10 seconds flying time, and since planes are more like boats than cars in terms of how fast they can change course and speed (especially speed – cars can decelerate remarkably quickly), that’s pretty close.
One thing that you can do with autonomous cars is essentially end all rear-end collisions because they can be programmed to maintain a safe distance and check it with radar (which, because of quicker reaction time, will be closer than for human-driven cars).
I’ve been saying my 3 year old niece will be the last person in the U.S.A. to know how to drive a manual transmission car. Her father (my brother) and I both swear by them.
Gin & Tonic
40 or so years ago I read a short story in one of those pulp SF magazines that were popular at the time, and the premise stuck with me (although not the author’s name), and it was precisely that. Truck traffic was all auto-piloted on the highways, and at very high speed, in physically-segregated lanes. Climax of the story came with someone in a driver-controlled car got himself into the 200-mph truck lane to pursue a trucker.
My brother got t-boned by an idiot in an SUV a couple of years ago. Fortunately, he was driving an almost brand new Honda Civic with all the latest airbags and was wearing his seatbelt, so he walked away with a few bruises. Fifteen years ago, he probably would have been dead, or at least severely injured.
Gin & Tonic
@Eric S.: I made a point of teaching my children, now grown. One of my sons-in-law later said that one of the three factors that attracted him to my daughter was that she knew how to drive with a clutch.
He was using a torch to remove radiator pipes. Its an older house with an attached garage & the garage was built with 4 walls, so a double wall between the house & garage. He set fire to the inner wall and assumed he had it extinguished but it had crossed over to the outer wall. Because the bathroom above had been gutted the fire made it to the roof pretty quickly. The only thing that saved the house was our son had gotten home from school and was doing his homework in his room. He was annoyed that the smoke detector was not shutting off so he was going to ask the guy what he was doing. When he opened the door he saw the bathroom engulfed! He ran downstairs and the workman was attempting to bring the garden hose into the house to put the fire out! Son called 911, grabbed one cat & searched briefly for the second before being driven from the house by smoke.
In an odd way the fire reaching the roof so fast saved a lot of the house. Having a way out the fire did not spread as fast horizontally as it would have without the vent.
Second cat was OK, we found her hiding under a bookcase, terrified, after the firemen let us back in.
I suspect that humanity will be wiped out by an unfriendly ASI diligently working to optimize the conversion of the mass of the solar system to knitting needles well before human drivers are banned from the roads.
The assumption that I keep sensing is that the driverless option will come gratis with a driverless car. My guess is that it will be a subscription service. It will be sold as a safety/convenience feature. Insurance companies may give a discount for having a car with the feature, much in the same way there is a discount for having a car alarm. But this will not “pay for itself.”
The big savings will come for businesses being able to lay off drivers. For consumers who currently “pay” themselves to drive, driverless cars will be for middle class consumers and above for at least a decade until someone decides that government needs to subsidize the service so that the lower classes can afford to pay to sit in their vehicles. (HAHAHAHA. Like paying a subsidy to keep lower class drivers safe won’t be controversial.)
I know that disorientation. 40 years ago we got rear-ended by a guy doing 40 while we were stopped to make a left turn. He was literally under the dashboard trying to get his windshield wipers to work – hit the back of his head on the steering wheel.
The sensation I felt was that the steering wheel had leap out of my hands. He hit us so hard it tore the seat off its bolts and I ended up reclined looking up at the roof of the car. Very strange sensation.
@joes527: Computers that will be inside driverless cars can do better calculations faster than people so those risky turns you see people make would be a thing of the past. Plus auto-driving cars will probably be slower in accordance to safety rules unlike people with lead foot syndrome.
We used to think that computers were going to be the worstest thing for humanity, but really, we couldn’t function without them now. Driverless cars will be a tool and it will make car ownership a more democratic thing as cars can be bought/leased on a shared deal.
If there is a slow safe way to do something, and a fast way with fire, the workman will choose the fast way.
Sure, but the problem for trains is much simpler, especially because the railroad has a lot more control over its tracks than the car company does over the roads. Also, each individual train is a lot more valuable than each car, so the amount of money to be invested isn’t as much smaller as the raw numbers would lead you to believe.
@randy khan: Subaru has an “eyesight” option which is supposed to do this today. It uses twin cameras (not radar) to see ahead and will automatically activate the brakes before you hit anything.
At least thats the theory. The manual is quite specific that you shouldn’t test this feature. (the same system is used to maintain following distance for adaptive cruse control, and I _can_ verify that that works)
It also has lane assist which if it notices you drifting out of your lane it will nudge the steering wheel back into the lane. This is quite nice in that unlike driving aids that decrease the driver’s involvement and attention, having the car nudge the steering wheel instantly grabs your attention and focuses you on the driving task.
I’m enormously impressed by the image processing that the car has to do to figure out where the lane is, where the car ahead is, whether it is in my lane or the lane ahead, what is the relative speed of the car ahead … but it is a long way from a self driving car.
Rob in CT
Ah, yes, the claim was subrogated. So yeah, the other person’s insurance paid. Ergo, the details of your policy actually didn’t matter (but could have, if the other driver had been uninsured/underinsured).
J R in WV
I have retired from the serious software development industry after gaining a BS CS in 1984… I don’t use Windows-based computers now, although my final years in managing software development projects were using Windows systems both for development and implementation of complex systems in our user community.
I would never allow a Windows-based system (or anything produced by Microsoft) in anything that moves, even if I was driving the vehicle. Lots can go wrong with systems that are completely stationary, well tested, and not subject to wildly changing external conditions, like sleet, fog, deep snow, slicks of oil on a driving surface, etc.
Someone above remarked about using GPS systems to navigate in an area they were not familiar with. I have attempted to do this over the years since such systems became available. First in Kona, HI where I sought to drive our rental car to my first destination. It wanted to use a road that was mapped years ago, but never constructed.
There is a reason that every bit of documentation regarding driving with GPS mapping systems tells you to carefully watch current conditions lest your navigation system put you onto a non-existent bridge, into a lake from a recently dammed stream, an interstate lane that has been crushed and removed for reconstruction, etc.
Later in my driving history, we toured the Navajo Nation which covers parts of Arizona, New Mexico, and little bits of Utah and Colorado. We were on a US highway, and turned onto either an NM maintained road or a Navajo Nation road, which was invisible on our (big brand name here) navigation system. None of the Navajo Nation roads were on the GPS mapping navigation system. I’m talking well maintained, paved, marked roads that run for hundreds of miles, not invisible tracks across the desert at all.
While I have caused a couple of fender-bender level accidents in my nearly 50-year driving career, all the 2 serious accidents I have had were caused by other drivers. Once every automobile, delivery vehicle, bus, ambulance, and fire truck on the road is driven by computer, I will consider riding in an automated vehicle. I expect that won’t happen in my lifetime, which I expect to run for at least another 15 or 20 years.
The Space Shuttle was a computerized auto-pilot as far as launch to orbit went, and there were three complete systems doing the same task, with a fourth system monitoring those systems, and if the results were not identical among the three, the master system (that fourth system) would select the result of the two systems. Dunno what if all three were different, probably disaster.
Anyone hoping and praying for this to happen is dazed and confused, and pig-ignorant about software development.
One impediment to autonomous cars is their refusal to exceed speed limits. NOBODY will put up with that nonsense.
@ksmiami: In order for the computer to do these calculations, it must be tracking and projecting the intention of all the cars in its area at all times. The driverless Google cars do none of this – what they are doing is much closer to what joes527 talks about wrt to Subaru cars. The driverless Google cars are a long way from what is needed for truly safe driverless cars in all conditions on busy city roads with lots of traffic changing lanes, bicycles, cars parked on the side that may open a door, and pedestrians.
@Central Planning:This list of How to Sing the Blues has been around for the internet’s version of forever, and is a valuable resource.
The idea that the government would outlaw an entire class of privately owned transportation products, just because, is absurd. I own and regularly drive cars older than I am and they haven’t been outlawed (and won’t be).
Driverless cars will certainly have a place in the future (especially great for elderly and disabled folks, or kids too young to drive), but “old” or self driven cars will always have a place on the road.
Google cars still have trouble in the rain (i.e. they don’t let them self drive in it). If something that basic is still so problematic, they have a long way to go before realization of a true “driverless” car.
Also, too, waking up this morning prior to accessing the necessary blues transportation mode.
The question is: why do people speed? I don’t speed because I like the thrill of going fast. I was in Boston last week. The drive back was about 400 miles. 5 or 10mph over the speed limit can take a decent amount of time off your trip.
If there were driverless cars that didn’t speed, I could still get work done in the car. I win with that even if it takes me 15 more minutes to get somewhere. That drive time + 15 minutes is time I won’t have to spend on the computer at home once I get home from work.
@Gin & Tonic: “Traffic Problem” (author unknown)?
I remember a story like that where an antique (1979) Ford was spotted on the road and directed to an offramp in the sky which went nowhere–just a Thelma-and-Louise fall through the air–as a way of getting excess traffic the hell off the road.
What we need is real automatic weapons, triggerless ones that “know” who needs to be shot.
Dumbass at Malheur: Hey, why’s this thing pointing at m
The thing that scares me shitless about driverless cars is what if I hit one? I’m a human of probably-slightly-better-than-average driving ability, good judgment, and alertness, which is to say that I’m an idiot. I sometimes drift into other lanes. Sometimes when I’m changing lanes I get so focused on what’s going on behind and beside me that I lose track of what’s directly ahead for way too long. Sometimes I drive when I’m really sleepy. Or I just daydream a bit and my situational awareness suffers.
Now, if I cause a crash with some other human, then, whatever. The cars get dinged up a bit. The drivers tell their stories, and nobody pays much attention because humans can’t be counted on to see things clearly, understand what they’re seeing, or honestly repeat what they’ve imperfectly perceived and imperfectly understood.
But if I hit a machine-driven car? Ferchrissakes. That thing will have fucking telemetry on my ass. It’ll be able to describe my speed and trajectory with many-decimal-places accuracy in the entire time leading up to the collision. It’ll be able to show exactly how I fucked it up and exactly how a minimally competent human who was paying attention to what he was doing could have easily avoided the crash.
The main thing is that technology that people feel is useful is quickly adopted. And even if it is initially expensive, technology that people find to be useful quickly becomes cheaper and is quickly adopted.
Also, technology that is useful transforms society in surprising ways. How old do you have to be to get a learner’s permit? What is the minimum age to get a driver’s license?
Would you still need a driver’s license in order to ask for a driverless car? Or would there be a different parental control restriction, e.g. it would be impossible for a person under age 16 to drive more than 100 miles from home without parental authorization.
@BL: I think you’re correct, but it’s easy to envision an environment in the not too distant future with cheap sensors on practically everything generating data that can be processed instantly.
While I am skeptical that auto-driving cars will take over the roads any time soon, I do find it kind of amusing that people talk about how the terrible risks of having computers run cars. Modern cars depend on computers for many things already (just to pick three items – antilock brakes, automatic transmissions and emission control, as the VW saga reminds us). Car manufacturers actually are quite good at making software that simply works.
@Rob in CT:
I believe I would have gotten the NADA value with a standard policy, not the price of a new car (or very close to it) which is what I got.
Also? No, of course the government isn’t going to outlaw human-driven cars, and how is that an interesting question anyway? The government didn’t outlaw horse-drawn wagons either and I haven’t seen a single one on the road in months.
This’ll have everything to do with “do they work” and “are they cost-effective” and nothing at all to do with “will governments ban human-driven cars.”
If the technology becomes sufficiently advanced, I love the idea of driverless cars. (Although I would really miss driving. I genuinely enjoy it.) My biggest concern is that it will be expensive. I’m guessing it won’t be as simple as just buying – or renting – a driverless car and off you go. You’ll probably need some kind of “data plan” that covers as many hours or miles as you intend to cover in a given period of time.
Piling that cost on top of fuel, insurance and a car payment doesn’t sound awesome.
@trollhattan: There are various versions of that document, and I posted a version that I should have read through. I much prefer:
“I got a good woman” is a BAD was to begin the blues, unless you stick something nasty in the next line-“I got a good woman with the meanest dog in town.”
It’s called the Internet of Things (IoT). From techopedia:
You should check out the sci fi series, Humans. Only 8 episodes.
We have been turning over jobs to automation. Robots are just higher level automation.
@Central Planning: A friend of mine has an app on her phone that causes it not to buzz or beep in any way if it is going over 35 mph. Instead, the texter/caller gets a message saying that the person they are trying to reach in in a car and will be notified of the message when they slow down.
It should be a manditory feature on every cell phone
Somebody mentioned that the Horse and buggy was safer. I think that on a per passenger mile basis, horse drawn travel was just as deadly as automobile travel.
Plus, LGM has an occasional feature on dead horses in history.
I believe the best driverless car scenario would require the driverless cars in the same area to communicate constantly so that they can collectively determine the best continuing response to keep all the cars safe. This is what the traffic collision avoidance system for planes does. But this would require a set of standards to be developed and regulations put in place, etc, etc. Not going to happen in a long time, especially in today’s political climate.
Don’t think of it as owning a robot-driven car. Think instead about living in a world where it’s easy to quickly hail a taxi anywhere, any time, and the rides cost a quarter of what they do now.
I took a half-assed shot at this question a few years back: “What’s my real cost to own and operate a car? And what would it cost me instead to use taxis?” Real cost includes the cost of the car, maintenance and fuel, insurance, and parking. (Parking cost includes any permits I buy, any meters I feed, and wild guesses at what fraction of my rent represents the parking space out back, and also my pro-rata share of what my employer pays for parking.)
Short answer was, if taxis cost a third as much as they do today, I’d be better off not driving.
So when I think of a driverless-car future, I imagine subscribing to a car service that promises cars for my regular commutes, plus whatever I need for quick trips. Fee structures will vary from one club/service/company to the next. Some will give a break if you commute at off-peak times, or charge a premium if you want more luxurious cars.
I certainly don’t imagine owning one personally. Cars are running for a few percent of their service lives at most. At any given time, upwards of 90% of our car fleet is just sitting around taking up valuable space. We accept this because car-sharing logistics are incredibly hard. Once the cars can drive themselves around, that all changes.
Would everyone even “buy” a driverless car, or just the right to use a certain class of vehicle? Why not share a vehicle? Here, the vehicle is more than a taxi, but less than a private car.
So, for example, you call for a driverless vehicle to take you to the store, where it waits for you in the parking lot and then takes you back home. But if you have the vehicle take you to work, it could be available to someone else during the day, and the vehicle that takes you home in the evening would be the same type of vehicle, but not necessarily a single car that you own.
Telemetry is coming whether we get driverless cars or not. Dash cams are an obvious example of a primitive attempt to get telemetry for insurance purposes.
If statistics were all it took to change public policy, we’d have gun control.
@Central Planning: “Sorry boss. I was late because the robot wouldn’t go any faster.”
Sounds like a good excuse to me :-)
As it is in most human endeavors, the last 10% takes 90% of the time and effort. Getting it done is the easy part, making it work properly is not.
I work in machining. Have for decades. We owned “computer” controlled machines in the early 70s. Today’s machines do amazing things, things not even dreamed about 50 yrs ago. And yet they still require a human to program them, tell them what to do, within their parameters, every day. And while they are far more reliable, they still break down regularly. We have a regular commenter who travels the hemisphere fixing them. What happens when a self driving car trips a circuit breaker? Where’s the back up pilot? Sure commercial airplanes can fly themselves but there are 2 human backups to take over when the shit hits the now not spinning fan.
@Roger Moore: Telemetry is already here.
Most new (and many not-so-new) cars have a black box that can be downloaded after an accident. Any sensor that your car has is probably being recorded by your car’s black box today. (and your car probably already has sensors that it isn’t telling you about)
Between those two influences, the people who remain in manually-driven cars will become significantly better drivers as well in very short order.
What’s the incentive difference between now and then to make this miraculous change? Personality replacements? Cranial/anal removialectimies?
It ain’t happening. It’s not the incentive or the cost, it’s the basic operator competency.
Then we could all be rolling down the street, smoking indo, sipping on gin and juice.
Sorry, had to add that.
I’m fairly sure horse-drawn wagons are not allowed on the freeway.
@Central Planning: Uh, thanks, I’m aware of the term. Just didn’t want to get all jargony!
That’s what we do here!
There’s an Arthur C. Clarke novel from the 80s, “Imperial Earth,” it takes place in the 24th century and one of the plot points is that autonomous cars are the norm. There’s no law against driving them yourself, but there’s a big social stigma attached to it, almost like smoking, because people driving their own cars is uniformly recognized as statistically a hundred times more dangerous. That seems like a more likely outcome.
@Schlemazel: @schrodinger’s cat: I drive a lot for work and while this is anecdotal, the.majority of accidents I see are fender benders where it appears the rear driver didn’t leave enough time to stop or started in stop and go traffic when they shouldn’t have. While you guys are right,Google has an interest in the matter I tend to believe Google. Also I would love to know if Google did the driverless tests without a physical driver(i highly doubt it)at the wheel because just the sight of a driverless car would cause accidents.
@Laertes: Not sure I love the idea of living in a world where every time i want to run out for a gallon of milk I have to call – and wait for – a cab to pick me up.
if we share cars, smoking vrs non smoking is going to heat up. I guess we could program non smoker cars to report you and get themselves cleaned.
@Bill: Right now that taxi situation is pretty shitty. If I could reliably hail a cab within 5 minutes, though, that’d be good enough. I wouldn’t be surprised if that were possible once we’ve got AI-driven cars.
Whether human-driving vehicles get outlawed will depend entirely on how the infrastructure changes to adapt to self-driving cars. For instance, self-driving cars don’t need visible traffic lights. That can all be negotiated wirelessly. In all likelihood, by the time we get to that state of infrastructure the automation software will be so good that the incentive to banish human-driven cars will be gone. It will, however, be prohibitively expensive for most people – and that will come quickly. Self-driving cars represents a liability shift from driver to manufacturer that will save people money so quickly that they’ll choose not to drive. And that’s where the car sharing comes in – it’ll be an economic driver.
I suspect that the first stage will be most 2-car households will become 1-car households because they really had 1.5 car needs – one use-every-day vehicle, and one as-needed one. The other big shift in this first stage will be liberating millions of teenagers, senior citizens, and people with disabilities that have always been reliant on fixed mass transit. That will become a fairly massive economic boost to the regions that invest into this – car sharing services, infrastructure for automated vehicles, etc. Transportation is critical to economic growth. Any time you have a person that could be working but can’t because they can’t get to the work, you have lost GDP. Any fixed-cost infrastructure that you can buy almost always pays for itself pretty quickly. Automated cars also cover the marginal cost infrastructure as well by basically eliminating the need to pay the bus driver and replacing it with a higher productivity worker that periodically maintains/cleans these vehicles.
@Bill: In such a world it’ll be faster/cheaper/easier to just have the milk drive itself to you. And that’s already starting to happen in the Bay area.
I agree that Kevin’s far too optimistic on the changeover.
Toyota is saying that the future of transportation must be electric if we want to control CO2 emissions. Presently ~ 25% of CO2 emissions is from transportation.
Getting off gasoline and diesel is going to be a big change. But that change is likely coming before fully autonomous driving, for at least some of the reasons outlined above.
IIRC, Google’s cars are limited to 25 MPH. The calculations are easier for slower travel, obviously. Autonomous trucks may be coming before cars (they have the space for the computers, it’s easier to roll the, say, $10k cost for the system in a $100k truck than in a $20k car, etc.), but even there it’s hard to see them being used anytime soon except for interstate to warehouse uses (and the warehouses would be very close to the off-ramps of the interstates).
Most of the US lives in or near urban areas, and that trend shows little sign of changing. As fuel and infrastructure gets more expensive, it seems likely to accelerate. But cities can’t add more roads. So, to handle the needs to get more people in and out, and the stuff that more people need in and out, it seems likely that tolls, congestion pricing, etc., is going to become ever more common. If car traffic is restricted to, say, 4 people per auto, and the traffic is bumper-to-bumper, the it becomes very easy to effectively mandate that car’s have a smart “EZPass”-like feature that lets the system control everyone’s speed, spacing, etc. That seems far more likely to me to be implemented reasonably quickly than full autonomy.
I also think that transportation won’t get cheaper as robots take over. Computers get cheaper for a while but then quit – we still can’t buy a full PC or a Mac for say $5. (Arduino and OLPC are attempts toward that, but they’re not there and won’t ever be there, IMO, because there’s so much more involved in the price than the electronics.) It’s just rather than mostly paying for the hardware, gas, insurance, parking, we’ll mostly be paying for congestion fees, taxes, bandwidth, subscription fees, etc.
I don’t think that we need to solve the full autonomy problem to have substantial benefits. And even MS can contribute (they’ve got money to buy up people who know how to do it. ;-)
It’s a challenging problem and it will be interesting to see how it’s rolled out over the coming years.
@Ruckus: Financial. Far higher insurance rates; the end of “no-fault” coverage (with one party clearly in manual control, one under a reliable driverless control and logging on both, fault will be much clearer – and the big driverless operators won’t be at all interested in keeping no-fault coverage if it is more expensive and they can easily prove their systems weren’t to blame); and social expectation that, to be a driver with as good a record as the machine you need to be as good a driver as the machine – and not being such will cost a lot.
US citizens are no more or less intelligent or capable than those in other countries, yet many other nations have far more stringent driver education and licensing processes than the US does: any increase in driver proficiency will improve driving safety. And any lack will be addressed by the insurers far more quickly than any legal or social equivalent.
Consider SUVs. They’re monstrous, but the insurers are convinced that they’re safer (fewer serious claims, fewer injuries to pay for, etc) so they multiply on the roads. Marketing alone can’t explain why they are so popular: it takes TCO to account for that and insurance is a big TCO component. Similar benefits accruing to driverless vehicles will push people to those cars, and to driving in driverless mode more often – and facing the kinds of fines and higher insurance rates for manual driving will discourage bad driving.
It won’t make the crazy distracted driver go away – but it will make inattentive or distracted driving far more costly, even when there’s no accident and nobody gets hurt.
@srv: Sure, because cars with drivers used as IED’s made them take all of THOSE off the road, too. I miss driving.
Does the hypoxia interfere with your chawing habit?
@Raven: Thanks for the reminder. I was just down in OC on a trip and took the toll road a couple times, still haven’t paid. Hopefully not too late.
@srv: As opposed to the potential for millions of IEDs now on the streets?
The worst thing a driverless IED would do is muck up a commute or destroy a parking facility. One vaporised vehicle plus all the ones behind that screech to a stop; and one or two time-delay bombs bring down a garage. They can’t race into a building or compound without manual control, and they can’t eyeball their targets and select the greatest collateral damage: all they can do is go to their preassigned target and detonate as programmed. Given how many (two? three?) vehicles used as IEDs in the US so far, the death toll would be more than offset by the reduction in road fatalities currently suffered; while it’s highly unlikely that the cold hard math would sway public opinion, it’s still powerful evidence.
Vertical parking: A space-saving solution when early cars invaded cities
Kevin Drum has over-optimistic views about artificial intelligence, I think.
In the case of driverless cars, one of the things that a lot of writing in the popular press tends to miss is that the interaction between the driver and the car is not completely understood in all its details; add automation to the picture and we have even more open questions.
For example, last year Ford recalled more than 10,000 Lincoln MKC SUVs, to fix a dashboard design in which the push-button starter was too close to other controls, so that people were turning off their cars by accident. We have a century of work on automotive design and on human factors, and these things still happen.
On automation, it’ll be hard to get an automated system to do everything right, in every situation it might encounter. Know what will be even harder? Figuring out how to safely give control back to the people riding in the car so that they can figure it out. We have lots of cautionary stories about airplane crashes happening because of the challenge of pilots and co-pilots regaining situation awareness to take over from the autopilot when things go bad.
I can pretty much guarantee we will see an openly gay President before we see the roads filled with these futurist-dreaming, libertarian masturbation machines. According to the whiz-bang kids in the 1960s, I should be flying around in my own personal mini-helicopter at this point.
@RSA: Exactly. Either you trust the system to always do the right thing all the time so that you can twitter away on your smartphone, or you have to be situationally aware at all times in case you need to take over. But you also have to take proficiency into account. People lose proficiency with complex tasks relatively quickly. Airline pilots have gotten into trouble not just through situational awareness (oh shit the alarms are on I gotta do something real quick now what) but also in lack of proficiency (now that I got the stick, what the hell am I supposed to do in the 5 seconds I have to make a decision).
Putting a robot in charge is akin to riding in the passenger seat. You better damn well trust that the driver knows what it’s doing or be prepared to take over at any time, similar to riding with a teenager behind the wheel.
Hell, when self-driving cars are readily available, their parents are going to say, “if the computer is driving the car, I won’t have to worry about the kid doing something stupid behind the wheel.”
I’m the father of an 8 year old, and he’s a pretty flaky kid, and I don’t expect that to change. I don’t expect driverless cars to be readily available in 2023, but if they did, it would be an answer to prayers.
With all this said, I still want to know why driverless transport is such a whiz-bang thing when rail transit – which offers nearly all the benefits of driverless vehicles – is in such a sorry state. If driver/passenger/pedestrian safety were really such a big deal then rail travel would be far more popular and accepted. And if safety were so good then all the freight accidents that make (or don’t make) the news would be a thing of the past. Securing the existing rail infrastructure, and making it safer and more efficient, is orders of magnitude less complex and expensive than what driverless cars will require.
And we mustn’t forget that oversight of driverless systems is a high-level skill. Enabling an individual driverless vehicle is relatively simple and requires little management: managing a system full of driverless vehicles will require significant oversight, for popular acceptance alone if not for reliability and effectiveness. Oversight will require technical training, and a lot of people watching the systems. All it will take is a system as poorly managed as WMATA or MARTA for the entire proposal to fall into disfavor just because the people running it are underpaid, undervalued and so demoralized that the system fails from their inattention.
Yeppers. My wife and I could be a 1.2-car family if you could timeshare a car. When an extra car can come to you whenever you need one, you only have to own as much car as you need on a regular basis.
The other big thing I’d expect is that not only will you not own more cars than you need routinely, but you will also not own more car than you need routinely. We have 4+ passenger cars, not because we normally have a bunch of people with us in the car, but because every once in a while we need that extra room. Once we can summon up a multi-passenger car only when we need one, single-occupant electric self-driving cars will become the norm, because that’s as much car as most of us need on a regular basis.
@C.V. Danes: An issue that Sci-fi successfully foresaw: http://variety-sf.blogspot.com/2008/03/jack-williamson-with-folded-hands.html
Unlike, say, the Internet.
The answer to this is quite simple. Rail transit has a “last mile” problem. It can’t take you door to door.
Quick example. Here in Southern California, a chunk of people who lived in Orange County and worked in Norwalk would take the Metrolink rail line. Some people lived up to two hours away and it was a nasty commute. At the Norwalk station, shuttles would take a bunch of people who worked there to City Hall. Another city shuttle would take workers to their offices in industrial based Santa Fe Springs. And of course there would be pickups in the afternoon.
When the shuttles were eliminated because of budget problems, people abandoned the Metrolink in droves because no local buses ran conveniently to their jobs.
Rail requires density that few large cities in the US attain.
@Kylroy: Yeah, I always think of that story whenever I hear people talking about turning everything over to the robots because, you know, people are too stupid to take care of themselves.
@Brachiator: @catclub: I guess someone forgot to tell the Europeans and Japanese.
So, do you know how the Europeans and the Japanese solve the last mile problem?
@Brachiator: @catclub: No arguments here. The point I was trying to make is that transport without one’s own vehicle, with the presumed inherent safety and efficiency improvements driverless cars offer, is currently readily available to many – and still falls far short of what is needed simply because a certain segment of the electorate (cough cough Republicans cough) remain adamantly opposed to investing properly in it, both in terms of the infrastructure costs and in terms of the staffing. Implementing a larger passenger rail system is significantly less expensive than building an entirely new driverless automobile infrastructure, yet the US can’t be arsed to do it. Even the metro systems, where density is high enough to support continued rail service, are woefully underprovisioned, and the situation is not improving. Assuming that driverless vehicles are some miracle answer, when earlier, simpler and cheaper alternatives languish in public disinterest, is Pollyanna-esque.
“Last mile” is indeed a problem for rail – and Brachiator’s example of what happens when that last mile is abandoned due to costs is perfect – but it will be for driverless vehicles as well. What happens when your car drops off the monitored-and-managed thoroughfares? Does self-drive automatically kick in? Does the car drive a “best guess” route to your destination? And regardless of which option then takes effect, who owns the liability if you do get in a wreck? I haven’t heard anyone (least of all Drum) address any of this, yet it will all factor heavily into adoption – and into costs – for the technology.
@Brachiator: By cramming people into much, much more population-dense living situations with no American counterpart beyond NYC.
@Brachiator: Yes. It’s called walking.
Wrong. But thanks for playing.
@Brachiator: Buh? I think both my and C.V.’s answers correctly address the question. What non-walking means do Europeans and Japanese use to get from mass transit to their homes?
pseudonymous in nc
A friend of mine joked of a future where the people who wanted to drive cars would head out into the lawless desert, and the world of Mad Max exists while everybody else is in green, robot-powered cities.
The National Drivers Association motto will be *You will pry my steering wheel from my cold, dead hands.” Seriously, there will be well-organized, grass-roots resistance and conservative demagoguery.
pseudonymous in nc
@Brachiator: as others have noted, walking is not a ‘bzzt wrong’ answer unless you’re dealing with the kind of American who needs supplementary oxygen after five steps.
It helps when you build out suburbs with footpaths and cycle paths, not just sidewalks.
I don’t agree, and don’t agree that you can even blame the Republicans for this. Rail is a 19th century solution to 21st century problems.
In the past, when there was more rail, and trolleys, etc, you also had tons of taxis and trucks that would make deliveries of goods (milk, bread, etc) door-to-door. But people liked the control and the convenience of having their own vehicles, which allowed them to decide when to go, with much more relative comfort.
Also, some cities can probably make better use of rail than others.
And of course, there is local political BS (as in why the Los Angeles Green Line rail does not go directly to the airport, or to the beach)
J R in WV
This would mean that burglars working with hackers can pre-inventory your stuff and decide which to take with them, before they even break in. The Tax Man will know if you buy new custom kitchen cabinets, and how much they cost, before they write you a privilege tax bill for the month.
If you’re buying champagne and classy munchies, without declared income to support that life style, you will get a tax bill for the investigation into where you got that money.
I don’t need an Internet of Things. My things are none of your business. I don’t even need a new Samsung Fridge with cameras inside for inventory shrinkage reporting, not even for the low, low price of $5,500, which is current cost of video-monitored Fridge.
@pseudonymous in nc:
Actually, those who mentioned bicycles are closer to one of the actual solutions used.
There is no controversy here. Look up “last mile” in Wikipedia, or talk to people who deal with transportation issues.
@bjacques: long-haul driverless freight vehicles seem like they might actually be coming soon. As far as speculating about driverless cars, I really doubt it’s going to happen, ever. Google admitted sometime in 2015 that their cars can’t handle a large class of roadway changes. So either we have a ratcheting up of the misguided policy of the current transportation system where everybody better get out of the way of the cars because they can’t control themselves. Since public transport is still much more efficient and cheaper, I think we will still head that way instead
@J R in WV: I’m sure I’m not the first to note that you can’t spell idiot without IoT
@Brachiator: Just a reminder here: we have NO EFFING IDEA how much the driverless vehicle infrastructure will cost. GPS is already in place (yahoo). The computer systems required per vehicle are still in their infancy. Interaction systems for major throughways aren’t on the drawing board yet. And as many have pointed out the only thing the public sector seems good at for highways is photographing cars for assessing tolls and catching speeders, and that’s still a problematic system. Add to that the regulatory and indemnity systems which must be added or updated to handle autonomous vehicles without direct human control and the costs are currently unknowable, but predictably uknowably immense. And every ten years the systems are in development or testing adds nearly a significant digit to the final cost.
Rail improvement may not be “21st century” but its costs are easily knowable and calculable. “Old” isn’t always – or even often – ineffective, inadequate or less costly in direct or indirect calculations: if it were then everyone who just yesterday was making horrible noises about factory farming and other horrors of carnivorism who was touting small-scale sustainable farming would be irredeemable Luddites who could have been easily silenced. Hardcopy written material is an ancient concept, and we don’t see paperwork waning much even with the advent of the “paperless society” twenty years ago. Modern materials come with nasty aftereffects like PCBEs. The most modern solution is not always the most acceptable.
True, road transportation is often required between rail stop and end destination; but we don’t need to provide rail service to the hinterlands of Nebraska to make it effective, and we don’t need a train for every single person to make sure they get close enough to their work/shopping/home to make it worthwhile. We do need to do that for driverless vehicles – yet we see community after community unwilling to invest in the road infrastructure to make that worthwhile with the “dumb” roadways we have now, and we see instance after instance where privately-funded toll roads turn into cash cows for international investors rather than a benefit for the communities the roadways serve.
On YouTube I watched about three minutes of a 1940s pilot training film about a big two-engine plane. The film was about half an hour long, and it was nothing but the preflight checklist, demonstrated. The top comment: Yay for computers.
@joes527: There’s a new Mustang that has a display worthy of a videogame so you can watch how you did around a racetrack. Everything your dashboard tells you, plus GPS, a host of sensors, and a camera. Should make things easy for the insurance adjuster.
Plus, what’ll all the cops do? Bust the skulls of pedestrians? Write tickets out to Sundar Pichai?
J R in WV
They still do this in NYC, We saw a brand-new parking lot with cars stacked up 4 or 5 levels high.
ETA: We saw this 3 or 4 blocks south of Madison Square Garden.
Absolutely true. And the discussions here are mere speculation, about on the level of “when will all cities have sky trains and monorails” as promised early on by Disneyland. But again, I believe an observation I made earlier that any technology that is seen as beneficial rapidly becomes affordable and adopted. There is a cool visualization of this here.
I think there is potentially huge demand for driverless vehicles. Whether the promise can be fulfilled remains to be seen. There are at present tremendous hurdles.
That said, at least in California, I don’t see trains as a solution to much of anything. I make exception for the Metro system, but even this has to be supplemented by buses and other vehicles.
J R in WV
In France we were staying is a small hotel in a small village. They warned us, told us that they hoped the train wouldn’t bother us. We were a little concerned, as trains in West Virginia are diesel power unit trains carrying hundred of thousands of tons of coal and freight.
The regular trains in this village were electric, two cars, for people traveling from anywhere to anywhere in France. We saw larger but still light rail trains in cities, and rode the next to best (70 MPH) rather than 200 MPH super trains from Tolouse to Paris.
We need the trains the French have. And population density is not a factor in France!
pseudonymous in nc
Look up ‘extracting a stick from your arse’ on Instructables. Walking may not necessarily be ‘last mile’, but it’s certainly last half-mile or third-mile, and non-stupid planning can accommodate it very easily.
@C.V. Danes: Ever taken a taxi in your life? What you’re talking about is actually not a big deal.
@Brachiator: I actually see driverless mass transit long before I see driverless personal transport. Municipalities are far more likely to afford the technology (replacing expensive staff) than individuals forking over the cash (or auto loan) for the virtual chauffeur.
Trains in CA are moderately useful: three hours on an HST from SFO to LAX is far preferable to an hour to SFO, an hour getting through SFO security, an hour in flight, an hour collecting bags at LAX and an hour getting into the city; however hopping on the subway or Metrorail to do you shopping is not especially likely. NorCal does have four separate rail systems in place (BART, MUNI, CalTrain and Amtrak service), all of which have significant ridership, so if we consider rail as reduction multiplier for personal transport rather than complete replacement there’s still good reason to invest in it.
The Northeast and Chicago have the bones of good systems, but it seems only Boston, NYC, Philadelphia and Chicago make good use of the structure: DC in particular has a system that is absolutely horrible to rely on and unappealing for moving tourists around, yet of course totally uninteresting as a public infrastructure benefit so not well funded. THAT is my beef: that so many cities (where driverless vehicles are most likely to be initially deployed) have systems that with a little additional support could be worthwhile to drastically reduce the number of cars on the streets, yet thanks to the Taxed Enough Already™ and the just-below-the surface racism (“public trans is for Those People”) there’s not enough support to make anything worthwhile happen.
Acela is a joke: pretty rolling stock condemned to ride on freight trackways not designed for half the speed the consists can attain, and demanding double the fare to shave 30 minutes off a ride from DC to Philadelphia or 45 minutes from DC to NYC. Yet the various private and municipal authorities who own the trackways can’t be arsed to deploy the improvements necessary to make the run faster (and safer: the improvements required would make travel at any speed less dangerous, for both passenger and freight consists). Anyone traveling the NE Corridor who isn’t in full-blown white-rabbit “I’m Late” mode is at least as well served by the regular service.