Outlawing Human-Driven Cars
Elon Musk predicts "You can't have a person driving a two-ton death machine."

Over at Forbes, contributor Robert Hof reports the predictions made by Tesla and Space-X founder Elon Musk with regard to the future of self-driving automobiles at Nvidia's GPU Technology Conference in San Jose, California. From Forbes:
In a brief interview with Nvidia cofounder and CEO Jen-Hsun Huang at the graphics chipmaker's annual conference for developers, Musk said that artificial intelligence-powered autonomous or self-driving cars will surpass humans' ability to drive safely and avoid obstacles and accidents. "In the distant future, people may outlaw driven cars because it's too dangerous … You can't have a person driving a two-ton death machine."
Fortunately for Tesla and for people who enjoy driving (but unfortunately for all the people killed and injured in auto accidents), that's not likely to happen very soon, by Musk's estimation. That's despite his belief that autonomous cars are "almost a solved problem." Musk, whose Model S uses two Nvidia chips, compared self-driving cars to elevators, which used to have human operators. "The car is just going to be like that. You'll be able to tell your car, go here, go there and it will just do it."
How distant in the future is the advent of self-driving automobiles? Perhaps not that that distant. NVidia released its new PX Drive system using deep learning artificial intelligence earlier this year. USA Today is reporting that NVidia will make its PX Drive system available to automakers in May:
The platform uses two Tegra X1 processors that are the size of thumbnails, yet boast 2.3 teraflops of performance.
"That's enough to weave together data streaming in from 12 camera inputs and enable a wide range of advanced driver assistance features to run simultaneously – including surround view, collision avoidance, pedestrian detection, mirror-less operation, cross-traffic monitoring and driver-state monitoring," writes Shapiro.
Forbes continues:
"A car is essentially software and computers on wheels," said Huang, who noted that Audi has committed to using Drive PX. "The days of dials, knobs, buttons are gone. The future of the car is going to be digital displays. The car is going to be one delightful computer going down the streets." …
Huang said Drive PX will augment ADAS [Advanced Driver Assistance Systems] systems to make autonomous cars a reality. "The big bang of self-driving cars is about to come," he said.
But not right away, to hear Musk tell it. "Where it gets tricky is that open environment around 30-40 miles an hour," he said. "Children playing, bicycles, manhole covers are issues in a suburban environment. Highway cruise is easy, low speed is easy, intermediate is hard." Still, he added, "We know exactly what to do and we'll be there in a few years. We'll take autonomous cars for granted in a short period of time.
I think that Musk is being a bit too cautious and that safe, fully autonomous vehicles will be zipping up and down our roadways before 2020 (if the regulators will get out of the way). For more background see my article, "The Moral Case for Self-Driving Cars."
For those who would resist the "Motor Law" my colleague Brian Doherty suggested that I provide a link to Rush's anthem to highway freedom "Red Barchetta". From the 2011 "Time Machine" Tour:
Editor's Note: As of February 29, 2024, commenting privileges on reason.com posts are limited to Reason Plus subscribers. Past commenters are grandfathered in for a temporary period. Subscribe here to preserve your ability to comment. Your Reason Plus subscription also gives you an ad-free version of reason.com, along with full access to the digital edition and archives of Reason magazine. We request that comments be civil and on-topic. We do not moderate or assume any responsibility for comments, which are owned by the readers who post them. Comments do not represent the views of reason.com or Reason Foundation. We reserve the right to delete any comment and ban commenters for any reason at any time. Comments may only be edited within 5 minutes of posting. Report abuses.
Please
to post comments
"(if the regulators will get out of the way)"
HAHAHAHAHAHAHAHAHA! Wait, you were serious?
Oh, and if the tort plaintiff bar doesn't make this prohibitively expensive either. YOUR DEATH ROBOT KILLED ________! MEMBERS OF THE JURY, MAKE. THEM. PAY!
The damages will be no different than if you had done it. A death is a death. If anything kills this idea it will be state and local governments seeing the gravy train that is traffic tickets and DUIs dry up. There will always be an option to drive the car yourself and someone in the car will have to be competent to take over at all times. That way the state can still stay in the driver's license business and continue to run its profitable DUI industry. Sorry sir, if you are drunk you can't step in if the car malfunctions and that is just as dangerous as driving the car drunk yourself. That will be the line of thinking, you watch.
Not so sure. The manufacturer/supplier has much deeper pockets to go after, and there is the potential for products liability claims. Not to mention class action claims.
Its no different than making a regular car. Car makers have been successfully sued for billions of dollars on very dubious product liability claims. Yet, we still have an auto industry and cars are still affordable.
There's a big difference between that and controlling the car.
^This.
A good proficient lawyer (Good lawyers all starve to death) will convince a jury of idjits that the evil car company deserves to be punished with a $20B finding, because the automated car drove into a bridge abutment.
I have to agree with John here. The parasites in government are absolutely dependent on ticket revenue, and what comes from DUIs and moving violations is and has been (for a long time) a basic part of that. You don't think that NYC calculates parking tickets into its yearly budget? What if self-driving cars could be told to circle the block on electric power instead of parking at a meter where you either pay or get a ticket? The city would shit bricks over the loss of that revenue.
Never forget that the government, besides counting on sucking us dry with income and sales tax, counts on all the revenue streams it gets from shit like cars. They will do anything to stop it from drying up.
They'll make up the difference with an annual fee that everyone has to pay for an AI inspection.
Your car wouldn't even have to circle the block. It could drive to the nearest free parking and wait there.
Speeding becomes a moot point as well. It will take years, even decades for the law to catch up, but once all cars are automated and integrated with each other and the traffic control system there is no need for speed limits in areas that have limited access for pedestrians. There is no need for red lights to control traffic - only to control pedestrian traffic.
There is no longer "following too closely" because the cars can talk to each other. This enables train-style convoys travelling at very high speed with great economy on the interstate. A convoy of 60 cars all touching bumper to bumper should be able to travel at 200mph with similar energy cost to just a couple of cars travelling at that speed.
In short, this will take a very, very long time for the government to catch up with the reality and potential of the technology. And ProLiberate is right. They'll take some time figuring out new fees and taxes to tack on to things to make up for the lost revenue from moving violations and gas taxes. Probably not as much time as it takes them to adjust traffic laws to optimize the new autonomous vehicle's capabilities though. Raising money seems to come easier for them.
Thank you for writing this. I have been thinking the same thing, but could never find the language to express it.
Never forget, Elon Musk is a tyrannical crony capitalist asshole. You can't hate that guy enough. Fuck him with a chainsaw until he dies.
What he is saying here is "you can't own and drive your own car because I am going to outlaw it and get my government friends to force you to buy my product.". That is really all that is going on here. If there is a hell, Musk is almost certainly going there. What a fucking waste of humanity that guy is.
Show us on this doll where Musk touched you.
I gotta go with john here. Musk is a massively overrated piece of communist shit.
Well, I wouldn't go quite that far, though he's certainly no libertarian. Tesla isn't SpaceX, which is mostly doing good things.
No, no PL. It's black or white. Those are your only two options.
A lot of what Musk is doing are good ends. It's the means which suck. He has no scruples, no morals, no principles.
I sometimes think the rumors of Apple buying Tesla are destiny. Apple had the last major control freak wonder boy, and the two wonderboys have a shared dream that only they know what is good for everybody. Bill Gates was just an amoral money grabber who didn't really give a shit how he got it; Musk and Jobs cared very much that they got their money from telling people how to do every last detail.
Jobs was briefly planning to force every single Apple employee to wear his signature tacticalneck and skinny jeans outfit. Only a full scale revolt by his senior programmers convinced him to drop the idea.
I will believe in the superiority of self-driving cars after I own a computer, tablet, or other tech do-dad that goes more than six months without hanging, crashing, or otherwise needing to be rebooted in the middle of an operation.
I firmly believe that self-driving cars are possible. I will even accept that the incidence of accidents will be lower, overall. I just think that the consequences of an unnoticed bug are likely to be horrific.
Not to mention being able to react to mechanical failure or inclement weather conditions.
Um, Unix devices routinely go a decade or more without crashing or needing to be rebooted
A mechanical error is categorically not the same as a software or computer error.
But anything Musk builds will run on government surplus copies of Vista that he'll get tax credits for disposing of.
Unix devices never have errors of any kind that need to be addressed? Do tell.
"Sorry that we ran over your family, but that car was missing a patch that corrected the pedestrian avoidance issue."
I don't see how that's any different than "A guy who holds an insurance policy with Farmers ran over your family..."
When the accident rates come down, the actuary tables will adjust and car manufacturers will buy liability insurance instead of drivers.
I'm not that ascared.
This isn't a contest as to which is worse.
Just a response to the idea that failing to crash an OS is the same thing as error free.
*isn't* the same thing...
When the payer is the car company and/or software developer, the payouts will be astronomical.
That guy with StateFarm? Limited by how much insurance he has/assets he owns.
A couple of screw-ups, and companies will go bankrupt or cars will become so expensive that no one can afford them.
Not sure how it's any different than, say an aircraft now? What happens when an aircraft crash is caused by a mechanical malfunction? The manufacturer gets sued.
They know this and the estimated liability is factored into the price of the product.
Same will happen here. And the prices will come down when it's demonstrated that the machine is less costly (liability wise) than the human.
I think the notion that liability issues will inhibit self driving cars is a red herring.
It's more of a legal question than an economic one. In most states, if Person A owns the car and loans it to Person B, and Person B runs you down, civil liability rests with Person A first and foremost because they own the car and after you've picked his bones clean you can go after Person B. I'm just saying that liability is traditionally held by the owner of the property while the operator of said property, in this case the manufacturer(s), would be secondary as long as you could prove negligence on the part of the operator.
But indeed I can see the cost of insurance lowering so much that a 'lifetime' policy could come with the car when you buy it.
That's not what he said.
This.
This shit ain't Windows.
Yeah, it's .
Erm, it's much, much worse
http://www.informationisbeauti.....s-of-code/
Whatever, I'm sick. Give me a break.
Nothing that requires a high level of reliability runs on Windows now. Not sure why cars would be any different, particularly when the companies in question are liable.
Nothing that requires a high level of reliability runs on Windows now.
My company has hundreds of process control computers all over the world ready to disprove this point.
The aircraft I flew, certainly didn't. Boeing wouldn't even allow their flight software to receive an input from something as unreliable as windows.
Their flight software was custom designed. It was no off the shelf OS.
Nor will this be, I'm sure.
Yeah, ABB DCS systems rely on a special version of Internet fucking Explorer to display process information to and receive commands from plant operators.
Stuxnet says "Thank you", by the way.
That is the problem, the consiquences of your car blanking out are a hell of a lot greater than your tablet doing so. As long as there is even a small risk of that happening, you will have to pay attention to what the car is doing and be ready at some level to take over if the computer crashes or freezes. And if you have to do that, you might as well be driving.
What people like Ron, who I swear to God will swoon over any new technology, are forgetting is that there is a huge difference between building a prototype that will drive around reporters and poobahs and building thousands of models that are to be used every day under all types of conditions. That latter is much much harder. Worse still, the margin for error is virtually non existent. Suppose they get it to where these things exhibit a fault once every million miles driven. That sounds so great and frankly it would be remarkable if they could do that. That would mean that if you drove 12,000 miles a year, you would have a one in 84 chance of the thing malfunctioning. That is pretty low, but it is low enough that you would trust it such that you wouldn't want the ability to drive it yourself in an emergency? I wouldn't. Enough to fall asleep or pay no attention to it at all? I wouldn't.
I think that the insurance companies are going to have a big problem insuring self driving systems when they first come out, because there is really no established statistical database to quantify the exposures they face.
The exposures would include:
1) Product liability ("Musk's car went off the road and killed my cow!") for the manufacturer or its suppliers
2) Recall insuance ("The NTSB ordered a recall of all Tesla's because there is a possibility the car's computer can shutdown at speeds of 55 mph")
3) Auto insurance for drivers ("My car ran over a child at a school bus stop!")
My expectation will be that manufacturers will have a real tough time getting insurance companies to insure them for self steering cars that are put on the market. They may agree to insure manufacturers that produce products for very limited uses (a self driving car in a steel mill's slab yard, for example).
My guess is that we will see individual features of automation appear (such as cruise control) piecemeal, rather than a complete self driving car. We will see warning systems that will brake automatically or sound an alarm to prompt evasive action on the part of the driver. I also expect that there will be great reluctance to insure anything that doesn't have a human closely supervising it.
I think your last paragraph is exactly what we are going to see. The other thing to remember is how different how much higher risk manufacturing this technology will be. Car makers mess up design and manufacturing all of the time. Even the best car makers have things go wrong and have to recall cars. For a normal car, that sucks but it not generally a big deal since the consequences of the defect are usually just a break down. The consequences of screwing up a driving system are going to ikely be some kind of accident. They will have to manufacture these to a much higher standard than normal cars are manufactured to. And that is not going to cheap and it sure as hell isn't going to be easy.
Tarran, from what I see, the insurance industry is hugely concerned about self driving cars - they see it as driving them out of the business altogether, or making it a really small market (mostly property style coverage).
That's my sense.
I think self driving cars will be treated the way medical devices are (my girlfriend's specialty) and not like say garage door openers.
Speaking of which: do any of the fine libertarian insurance professionals on this here board know of a usable database of FDA drug recalls/enfocement actions? The fucking web based reporting tool is an unusable abortion that would take 8 hours to extract 5 years' data from.
I'm, uh, asking for a friend.
Not my line, sorry.
Next I'm going to find out that you and I live on the same street.
I sell property and causality lines and do some surplus lines brokering. I don't even know many health or life agents that would know that stuff. That particular part of the field has become so heavily bureaucratized and arbitrary that no one knows anything aside from what the computer software tells them to do.
Not a chance. The market for all other kinds of insurance is still huge even if auto went away.
My guess is that we will see individual features of automation appear (such as cruise control) piecemeal
This is what I expect as well.
I also think that these are all academic points we're making. We're not going to wake up one morning and have self-driving cars available for everyone and anyone. They're going to start as features on high-end vehicles and proliferate down the market over a long period as the costs involved decrease.
By the time your Kia Rio has a driverless option, we'd have ahd so much experience mixing driver and driverless options that going to a pure driverless system will just be an unpopular waste of time and money.
Meh, airline flight control systems clock in a few thousand hours of runtime a year, and plane crashes are pretty much exclusively caused by pilot error (or sometimes mechanical failure if you are flying abroad).
Take your 1 in 84 system, run 6 in parallel and have them vote on the best course of action (pretty much how the airlines do things now). The odds that they all malfunction at once in a given year is on the order of 10E-11. Probably an order of magnitude or two higher because points of failure aren't perfectly independent. Given that we have 40k driving fatalities each year now, I'll take those odds.
agreed with above on fault tolerance. many of these complaints here fail to take into account just how awful and unsafe human drivers are, and how critically ineffective any govt measures are to make them safer. Hundreds of thousands of people die every year on the roads. Its a hundred 9/11s every year, which could fairlu be described as a crisis. And its a crisis that can only be alleviated through technology.
Isn't the google car over 2 million miles now? My driving record isn't that good.
That is one car specially made driving in set conditions. Lets see what the hundred thousandth google car driven by a hipster retard in a driving rainstorm after two years of no maintenance or software updates does and get back to me.
No maintenance?
My plan is to not own a car at that point. Uber or johnnyCab or whoever will maintain them.
This. My wife and I are currently in the process of moving back to the States. I predict this will be the last time we own more than a single car, but I consider myself somewhat of a technology optimist.
I understand and agree with what you're saying, but the benchmark for automated cars shouldn't be perfection, it should be current reality. Can auto cars perform better statistically than roads full of 16 year old girls texting while driving and 85 year old men who shouldn't even be wiping their own asses, much less driving?
If it is not perfection, then you will have to have an override and pay attention at some level in case it malfunctions. If you have to do that, it is really not a driverless car is it? And you might as well be driving yourself at that point.
If cars are driverless 99% of the time and a driver takes over 1% of the time when the car malfunctions, then yes I'd say it's a driverless car.
It's like saying "Dogs have four legs." Yes, there are some three-legged dogs out there, but the normal state for dogs is four legs.
If cars are driverless 99% of the time and a driver takes over 1% of the time when the car malfunctions, then yes I'd say it's a driverless car.
No it is not from the driver's view because they still have to pay attention. You won't know when the 1% is going to happen. So you are still doing the same things, you are just not driving. Well big deal. If I have to do that, why not just drive the car myself and save the expense?
That is in fact what people will do. You are never going to see the day where your robo car picks you up and you sleep in the back on the way to work. You could never eliminate the risk to that extent and even if you could the government wouldn't believe you and would still require someone be in the vehicle paying attention. And once you have that, you just have a really aided driving experience.
Paying attention won't do you any good whatsoever. If the car in front of you brakes, and your car doesn't brake, you won't be able to react in time to override it.
But whatever. Either driverless cars will be statistically more safe, or they won't. And if they are, you're still statistically better off sleeping while your car drives to work. Even if your car fails and you wreck, it's still a less likely outcome than having you wreck while driving yourself. You would, in fact, be more likely to have a wreck if you were constantly monitoring your car. You'd override it at a point when you shouldn't.
And if driverless cars aren't statistically more safe, then few people are going to buy them. They won't be on the market until they are, anyway.
But whatever. Either driverless cars will be statistically more safe, or they won't. And if they are, you're still statistically better off sleeping while your car drives to work.
See my point below about human nature and risk. People will never see it that way even if it is true. Moreover, people generally don't like putting their faith into machines and other people where their safety is involved. It is why people are terrified of flying even though it is statistically very safe. The same thing will be true of driverless cars. A good number of people if not most people will feel better driving themselves and being in control even if it is less safe. That is just how people are.
This is becoming a big problem in aviation, particularly with the airbus.
Basically, these planes fly themselves, and do it much more safely than a pilot can.... BUT... when shit hits the fan they dump to control onto the human.
So, a pilot, who has almost no experience flying a plane at high altitude when everything is working right, has to fly a plane at high altitude when things are very wrong. Worse, there have been cases where the "wrong" thing isn't outside the plane, but in the failure of an instrument or control system and the pilot is having to fight the computer. There was a flight in Lichtenstein IIRC where the plane would have crashed due to a software bug that introduced a dangerous nose up command had the pilots not known how to force a reset of the computers and had the courage to do it in the air.
This phenomenon contributed to the Air France crash in the atlantic (the pilots maintained the aircraft in a deep stall after a pitot tube became blocked with ice) and the recent Air Indonesia crash where the first officer found himself flying the plane in a thunderstorm while the captain was pulling breakers on the flight control computers' power supply in a vain effort to clear some serious fault.
If cars are driverless 99% of the time and a driver takes over 1% of the time when the car malfunctions, then yes I'd say it's a driverless car.
And you can bet that laws would require that you be 100% sober so you could be prepared for that 1% when you need to take over.
Oh, I can guarantee that the gubmints will have this requirement whether needed or not. Driverless cars could have a 1 in 1 billion failure rate, be orders of magnitude safer than designated drivers, and the govt would want its DUI money.
It's over 3 million iirc.
Not sure if it's this you're referring to:
"The company is currently logging three million miles in simulators every day, in addition to its real-world driving tests."
http://recode.net/2015/03/17/g.....ive-years/
I am referring to the fact that making one car and using it under test conditions, no matter how rigorous is not the same as making thousands of them and using them under real conditions, the full scope of which cannot ever be anticipated.
This!
Talk to anybody in insurance who has access to claims or does underwriting and they will fill your ear with stories of people doing stupid or incompehensibly weird things.
Then of course there will be the unexpected problems. Consider, for example the incident in the 70's when the Russians' early warning system falsely reported an incoming nuclear attack from the U.S. because sunlight scintilating off of waves on the pacific ocean looked to the sensors like missiles rising from the U.S. west coast.
The guys building simulators are replicating the bits of the world we think matters. They are not necessarily replicating all the bits of the world that matters because some of what matters is unknown to them and thus not in their minds.
Before 2020, 5 years from now?
Ron get real, to have completely autonomous that are significantly safer than humans by 2020 that means that the technology underlying them has to already exist and be very late in the testing process.
I agree we're close but 2025 is more likely the best case for commercial availability.
R: I am assuming that the first such vehicles will not be produced by legacy car companies. Also, I do not predict that the roads will be clotted with them - just that they will be available.
Google's driverless cars have been driving around Las Vegas since May of 2012.
My understanding is that the only accident was when one of them got rear ended by someone.
That was in 2010.
http://en.wikipedia.org/wiki/G.....#Incidents
You are correct. And the technology already exists and is very late in the testing process. Several manufacturers are already showing off their self-driving cars. They are already safer drivers than the average human. We are at the point where the law is about to become the hindrance in the deployment of these vehicles.
And yes, 5 years is much sooner than this level of change is likely to happen. Good electric cars have just recently become available and we are still many years away from widespread adoption.
Given the state of the roadways on my way to work this morning, I'm wondering about the pothole avoidance technology. We hear a lot about their ability to dodge other traffic, but what about faults in the road? Am I going to have to replace the tires on my self-driving car more often because the darn thing can't dodge a pothole?
Just the tires? You're lucky. Last pothole I hit took out a tire rod and the steering rack. That caused the tires to wear unevenly, which means replacing all four since it's AWD. Add to that the fact that the timing belt needed replacing, and it just wasn't worth fixing. So I had to get a new car.
I would suggest fixing everything but the tires yourself, but since you apparently think the tires have rods, maybe you were better off just getting a new one.
After that pot hole his tire had rods.
Tie. Whatever.
Tires and timing belts are normal wear items. So all it took to make your car unworthy of repair was a tie rod and steering? It wasn't worth much, was it. Sounds like the pothole did you a favor.
The tires wore out prematurely, and the steering rack was the icing on the cake. Yeah, I'm glad to be rid of that car, but I don't like having another monthly payment. That one was paid off.
I understand entirely. I've got a 29 year old car which overheats at low speeds. Do I buy new struts? Do I overhaul the tranny?
The car is dynamite fun, practical, reliable, and economical (1986 MR2). As long as the maintenance is less than new car and insurance payments, it's a hard choice. Even a head gasket job is only 2-3 new car payments.
New car? What does that mean? Seriously, I've never owned a new car, and doubt I ever will. Though someday I hope to afford something with less than 100K on it.
MR2, huh? Does your boyfriend think it's fun, too?
That was my thought as well.
Poor road surface conditions.
Hell, how does it deal with snow.
"In the distant future, people may outlaw driven cars because it's too dangerous ... You can't have a person driving a two-ton death machine."
As I traveled my half-mile rural lane this winter, with the snow I piled up on either side from plowing it off, uncovering an inches-thick road bed, I pondered what sovereign machine could successfully navigate its unpredictable snow drifts and ice ruts to get me home. The answer is none.
Don't be silly. You will be relocated to a condo in a master-planned high density dwelling center. Dachas will be reserved for top. Men. And their friends like elon musk.
Not to mention the perpetual rural dilemma of the possum, raccoon, rabbit or deer. Is is safer to just run over the idjit critter or try to avoid it?
That is a great point. I hadn't thought of that but that is a huge deal. I can't see how you could program these things not to run over a dog or a squirrel if avoiding it incurred any risk at all of collision. There is no way in hell people would tolerate that. There are too many animal lovers. Owning a robot car would quickly be about as socially acceptable as ruining a dog fighting ring.
There are other rural dilemmas. I have driven through blizzards, floods, windstorms, and once a very close--too god-damned close--tornado touchdown. When I'm outrunning a fucking tornado, I don't want the car freaking out about speed or debris in the road.
Running over a deer is a very risky type of behavior. I hit a small yearling buck at less than 30 mph, who did $2.5K damage to my car. The next week, my boss had deer end up in his front seat with him. A foot to the left, and he'd have likely died.
Is the plan to automate the deer as well?
I once saw a deer collision that fucked up two cars and a truck. It really was a magnificent buck and all my uncles would have given their left nuts to have his head on their walls. The buck dived between two cars going opposite directions, one hit him in the chest and the other car hit him in the hind quarters. The deer spun around like a helicopter embedding himself into both cars nicely, and then the whole pile of deer and cars was hit by a truck who was tailgating. I was behind the car that hit the hind end, and stopped in time to not join the pile. Seeing the ginormous mess, I said fuck it, called off work, and went back home.
Deer wouldn't be the problem. Dogs and cats and cute bunnies and such would be the problem. Yeah, hitting a dear is a big deal and a very bad thing for you. Hitting a dog is just bad for the dog and your conscience.
If this is a dilemma, you've joined PETA.
(Except for squirrels, those I speed up for.)
People Eating Tasty Animals? Already a lifetime member. But the dilemma isn't one one of ethics for me but of practicality. How about skunk? Or someone's pet? And even a small critter can potentially fuck up your car if the right parts get lodged in the right places. I once lost an oil pan to an already run over deer--it should not have been a problem as it was already quite pancaked, but part of a rib was just long enough still to get jammed up into the oil pan and make a wicked little hole.
"There are more things in heaven and earth, Horatio, Than are dreamt of in your philosophy."
By far one of my favorite quotes.
The Bard could not know this would be cited in support of a driverless carriage...
I would be happy for autonomous driving on interstates. However, I would reserve the right to row through a manual gearbox in a rwd roadster on a twisty mountain pass.
This.
It's libertarian scum like you that will prevent us from getting places at 2 or 3 miles per minute instead of one.
SELFISH BASTARD!
We have outlawed twisty mountain passes.
West Virginia haz a sad.
I don't doubt we will have automated vehicles in the not-too-distant-future, but I also think there's a big jump from that to mandated autonomous cars. In any event, such laws aren't really necessary, since effective automation would mean a serious difference in risk for insurers, who would make that clear to the insured.
PL: " a serious difference in risk for insurers, who would make that clear to the insured." Excellent point.
There's also a potential "herd immunity" of a sort that might forestall the need (though maybe not the actuality) of mandated autonomous cars--if a sufficient number of cars is automated, and the automation is truly robust, then they would undoubtedly be able to avoid accidents and route around problems created by the manual renegades.
Its crazy how if people let markets work, that they get us to a good place.
What we will have is vehicles that are so incredibly effective at assisting us driving that even the worst driver will become a safe one. I don't think the Cytoxic Ron Baily dream of robots ending the evil scourge of driving is going to happen anywhere near as soon as they think it will.,
Pro, that just means insurance on driverless cars would be cheaper. I can afford my insurance right now. I don't see how there being driverless cars would make it more expensive.
But cheaper insurance will make some people change their decision at the margin
Sure. but likely not many. And insurance premiums are not very expensive unless you have a DUI or a terrible driving record. So it is unlikely the cheaper premiums will make up for the increased of expense of buying the driving mechanism.
It must be your age but I disagree. I'm young but I have no accidents on my record. Even with cute "safe" and "experienced" driver discounts I feel like I am getting bent over with my premium.
In 2015 dollars, my liability-only insurance policy for an old subaru wagon in Phoenix (20 years ago) cost more than my three full-coverage policies (xterra, 350z, and forester) combined today. The difference is 38 versus 58 and Phoenix versus Iowa.
I looked here. And it looks like I am paying at or slightly below average. If driverless cars save people $2,000 per year and offer more or less the same freedom of mobility then you can bet people are going to love them robocar.
The insurance won't be free. So it won't be saving you $2000 a year. Even if it did, that still doesn't make it a good idea. It all depends on how much the mechanism costs.
I suppose it's possible that state laws mandating insurance coverage could be relaxed if you operate an autonomous car.
Jesum crow. I'm ~30 and was just quoted a top-shelf policy on a newish car for 35% of my age's average. In New Jersey, which has a long-standing reputation for terrible car insurance rates.
Apparently my impression that everyone else on the road is retarded is shared by insurance companies.
Those are annual, most premiums quote for 6 months. But really? I'm amazed that some drivers can get collision for 600 dollars per year. This just reinforces my asspained feelings over car insurance.
Yes, collision and comprehensive for both my wife and me, two cars of near-equal age and value, $700 total for six months.
I do think that Progressive's exploiting a market inefficiency, and that both my wife and I fit into a particular sweet spot in their tables. Nobody else is quite that low (Geico was $850 for equivalent coverage).
If driverless cars were statistically safer, their insurance would be lower than human-driven cars. That's what actuaries do. Your personal insurance may or may not go up, but in the aggregate that's what would happen.
No it would stay the same because my risk would stay the same. The existence of a cheaper option in no way affects my price. I understand how insurance works. The point is that it wouldn't raise insurance on driving your own cars and that is pretty cheap for most people as it is. So the "but your rates are cheaper" really isn't much of a selling point since my rates are cheap as it is.
Let me rephrase my last sentence: "Your personal insurance may or may not go up, but in the aggregate driverless insurance would be cheaper." That's what I meant but sometimes me English no right good.
Saving 50% on your insurance may not be attractive to you, but I can guarantee there are people to whom it would be attractive.
No Warren, whether its attractive to anyone depends on how much the mechanism costs compared to the savings. If the mechanism costs an extra $10,000 to the price of the car and it only saves you $1,500 a year in insurance, it is likely not a good option, unless you plan to keep your car longer than most people do.
It might be substantially cheaper if the accident rate for automatics dropped close to zero.
The mandate would likely start with restricted roads. Some interstates would likely be first - because restricting them to autonomous-only traffic would allow unlimited speed and automobile trains that travel at extreme speeds. Even one self-driven car would be a serious hazard on that sort of road. Long distance trucking would also like to get in on that action - particularly if they can leave out the driver altogether.
Then city centers would want to get in on the deal. Autonomous-only traffic would never get jammed up. Rush hours would be like flushing a toilet. Most people would be using cars for hire (sans labor cost of a driver) in the city center anyway, so this change would be relatively easy.
The rest would eventually follow due to inertia. Eventually the only places with manual drive will be rural areas and race tracks. That arc has more of a 50 year feel to it than 10 or 15 years though.
I figure fully autonomous cars will be like practical electric cars and fusion power: Forever just over the horizon.
Yup. And throw a jet pack into that trunk in case of a breakdown.
Why use a car when you have a fucking jet pack, Wesley?
Shut up, Wesley. /Picard
"In the distant future, people may outlaw driven cars because it's too dangerous ... You can't have a person driving a two-ton death machine."
Then we need to start calling out "scientists" for making qualitative judgements that have nothing to do with science.
Whether it's possible to eliminate 99% of traffic fatalities through driverless cars is a scientific question, but whether I'd rather drive myself around and accept the higher risks isn't a scientific question--at all. And no one is qualified to make qualitative choices on my behalf.
Whether society should respect my autonomy is also not a scientific question. It sure isn't falsifiable--not if we're taking qualitative considerations into account.
Incidentally, I think people who choose to ride around in driverless cars should be free to do so. No, that isn't a scientific proposition, but I'm a PhD level expert on my own damn qualitative preferences. And I have a qualitative preference for living in a free society.
Sometimes you drive me crazy Ken, but your comment is spot on. It is exactly that. And you hit on why Baily drives me so fucking crazy on this subject. He just assumes that it is okay for top men to tell me how much risk I should take.
KS & J: Sorry to drive you crazy. but isn't there the issue of how much you are endangering other innocent people on the road by taking the risk of driving yourself? Is that concern totally irrelevant to us libertarians?
More or less. In that respect it's like the DUI question: I don't care how much of a existential threat I pose by being in a certain state, punish me for what I actually do.
Goddamn citizen here is getting all high and mighty now. The system exists to mitigate social risks through the jurisprudence. No ordinary individual has the right to seek self-governance so we'd like to try and get at you before you fuck shit up, citizen.
Tell me, Ron, how much am I endangering other people on the roads with my zero accidents in 15 years? How far are you willing to go with this pre-crime idea, being a libertarian and all?
Ah no its not Ron. We are talking about relative risk here. You are saying that I can't drive because a safer option of having a robot drive exists. Okay. Well, driving a brand new car with the various driver aides now available is way safer to others than driving even one ten years older. Your logic applies just as well to that situation means we are all placing others at risk by not driving the safest car available. Well in some sense sure. But not in a way that in any way justifies forcing the choice on us.
The problem here Ron is that you like technology and likely don't like cars or driving. So, you have a bit of a blind spot and can't see how this concept is really oppressive.
How about we go with privatized roads? Let the best company win. Some roads outlaw autonomous vehicles. Others require them. Still others don't care.
Where do you suppose all of the routine traffic would end up?
Subtract the "privatize" non-starter from the scenario above and that is what is likely to happen. Some roads will ban drivers - kinda like the HOV lane. Others will restrict by time (like rush hour reversible lane systems). And still others won't change at all.
I doubt anyone will be considering blanket bans on human drivers anytime soon. And by that I mean in our lifetimes. But at the same time I doubt that we'll make it more than 20 years with autonomous cars on the road before we start seeing autonomous-only roadways. With a push from the NHTSA requiring a level of self-driving in all cars for safety purposes, this could happen even sooner.
Cyto,
I think human nature is going to make full on driverless cars much farther in the future and less used than you think. It is easy to say you will just get in a car and trust it as it wizzes along at 120 mph three inches off the car in front of it's bumper. It is quite another thing to actually do it.
The thing people are forgetting is that this isn't offering that much of a jump over what we have now. People were willing to overcome their fear of airplanes because it allowed them to travel as speeds that were previously unimaginable. Driverless cars are not going to do that. Even if it gets its own special lane, your driverless are is not going to get you there much faster than my ordinary one. Moreover, it getting there fast enough to make any kind of a difference will require me placing my entire life in its hands. Sure, I will do that to get on a 747 that gets me to Europe in 8 hours instead of eight days. Am I going to do that to get me to work 10 minutes sooner? Not likely. And as far as your driverless private super highways going on at 120 mph, why not just fly?
This technology is not as useful or as cool as you guys think it is.
I can't disagree with much of that. People are resistant to change and they like controlling their fate. I know because I'm a people and I like those things. I like track days too.
I don't know if we'll ever realize the potential. But a train of electric or hybrid vehicles that actually make contact could go a lot faster than 120. If designed for it they could easily top 200mph. Each of the first few cars in a train adding a little push would add plenty enough horsepower to get a bunch of 120mph cars up to 200mph. Plus, they'd be able to extend their range immensely. Because they can all talk, they can adjust power and rotate to get all the cars in the convoy to their destination without depleting their reserves.
So instead of taking a flight from Atlanta to Washington and getting there in 3 hours including airport time, take your own car and get there in 3 hours in a 100 car train.... on a single charge.
Sure, that's all way more speculative..... but the technology is not a pipe dream. It is possible with what exists right now.
But you are right.... this has a very "not in my lifetime" kind of feel to it. Maybe a few places out west.... maybe the Minneapolis area with all their high tech people would want to have automated lanes to commute from their 10 acre homesteads in the surrounding counties.
Cyto,
Your train idea is a fantasy. Traffic doesn't work like that. That is never going to happen. What causes traffic back ups is the difference in speed of various cars as they approach various changes in the road. A car slows to let another car in or to navigate a corner or in reaction to it starting to rain. This causes the cars in back to run up on it as they continue at the higher speed before the obstacle. This produces an accordion affect. That is not driver error. That is physics.
No driverless road is going to overcome that unless you just set the speed so low that the cars never have to slow down for any curve or any reason. Would such roads be better than human ones? Sure. But they wouldn't be anything like you describe and would not offer any kind of world changing advantages the way air travel did.
This technology is nice but it is never going to live up to the fantasies you are having about it.
I would buy season tickets to GT football if I had a driverless car.
5 hours each way is fine if I can sleep in the back. No hotel needed.
That is a major benefit to me.
Rob,
We already have that. They are called buses.
Yeah, you do that, John.
You've said some silly things in driverless car threads, but Christ...
Carl,
Lighten up Francis. You miss the point. It is not that driverless cars are not better than a bus. It is that they are really not that much better or different than a bus. All rob is saying is "wow wouldn't it be great is someone came and picked me up and drove me to the game." Sure it would be. And you can get that now via a car service. And car services, even for a long trip are less expensive than you might think. Rob could hire a limo for the day and have the driver wait outside to take him home for a few hundred dollars. Yet, few people do that. Why? Because it is not that much easier than driving to the game yourself. A driverless car would be no different. So the question is, how much are people really going to pay over and above a normal car to have one? The answer is not nearly as much as you think and certainly not so much as to make then anywhere as common as some on this thread think they will be.
Oh look, John moving the goalposts! Shocking.
Yeah sure Carl. Nothing says moving the goal posts like making the same point multiple times. It is not a difficult idea and I have explained it about four different ways to make it easier to grasp. Let's try one more time, being "better" is not good enough. A driverless car has to be so much better that it justifies the cost and overcomes the natural fear people have of placing their life in the hands of someone else or a machine.
If commercial air travel only offered the advantages of even the best driverless car scenarios, there wouldn't be a commercial airline industry today. People would have never found it advantageous enough to fly to bother with the expense and the fear associated with it.
How many and how badly I'm endangering other people by driving myself around is a scientific question.
Whether I should be free to drive myself around despite the risks to other people isn't a scientific question.
It's an ethical and political question that requires qualitative judgements. I hope that's getting across.
This isn't new. We make these kinds of qualitative judgements all the time--and we put them into practice. Giving accused rapists a jury, a judge, an attorney, a right not to testify against themselves, protection from double-jeopardy, and all the other protections we give to accused criminals, this almost certainly increases the number of rapists who get away with rape.
I support keeping the Fourth and Fifth Amendments anyway.
I oppose the NSA digging through our phone call data--even if them doing so actually protects us from terrorist attacks.
Even IF IF IF, the Second Amendment means more violent crime, I support the freedom to bear arms anyway.
I have a qualitative preference for living in a free society that cannot show up on any utilitarian's spreadsheet.
So, how do we account for everyone's qualitative preferences?
The answer is to protect the right of individuals to make choices for themselves. Each of us takes our own qualitative preferences into consideration every time we make a choice. In that way, society achieves the optimal level of concern for qualitative preferences.
Certainly, if you have a qualitative preference for total safety, it's important to remember that isn't an objective scientific value that can be safely projected onto all of us.
That's just a qualitative preference!
My preference is to ride a motorcycle through traffic everyday to work.
Qualitative preference at the individual level isn't something the greater social good concerns itself with, brother Ken.
All of that and more Ken. It is funny to watch someone own Baily like that. Teach him to venture into the comments section with his primitive rhetorical skills.
cant someone think self driving cars are cool and likely to save many lives while not approving of a regular car ban?
Yes.
But I'm a monster, so take that for what you will.
Musk is far from a scientist. He's a lobbyist who somehow got groupies.
You see this stuff everywhere scientists/engineers and utilitarians come together.
An inability to account for qualitative considerations has been an inherent problem within every utilitarian argument since before they were even called that.
Adam Smith (and Darwin) saw through it all.
The outcome tends to be best when people are free to choose what seems best to them, and who is someone else to say that the best outcome from my perspective is total safety?
You see the same things with climate change.
Whether the polar bears will survive is a scientific question.
Whether I should care more about polar bears than our standard of living is not a scientific question--and yet "scientists" will act like proving out the first question answers the second. It doesn't!
Whether I should care more about polar bears than coal miners or our standard of living isn't a scientific question. Whether there's really a dichotomy between the two might be a scientific question, but whether I should care more about one thing than another takes unfalsifiable, unrepeatable, qualitative judgements into consideration that are about ethics and politics rather than science.
This is one of the reasons why so many average people are leery of "scientists". It often seems to be political advocacy on qualitative issues informed by but masquerading as science. And although the skeptics can't put it in those terms, I think they feel it in their bones.
The future will provide an unending supply of ways people can control other people.
Same thing was said about electric cars ten years ago. Problem is, low to no demand and high cost. If I want a "self-driving car", I'll call for a cab and pay thousand$ and thousand$ less, today.
That is a great point Wesley. Most people enjoy driving and like driving themselves. The people who don't tend not to own cars anyway. I seriously doubt many people who own a car and drive to work every day would want the car to drive itself much less pay a premium for it. This will likely be one of those circumstances where people's preferences will foil the dreams of the various top men who know what is good for everyone.
I also do not think they will take the jobs of cab or truck drivers. No way will the government allow cab and trucking companies to operate vehicles with no human supervision. If nothing else, the workers in these industries will demand it to save their jobs.
What we will end up with is this technology being used primarily to make driving easier rather than doing it for us.
Me. I love my car, I love driving, I love driving my car, but I hate driving my commute. It's not driving as much as it's dodging traffic. If I could get in a half hour nap each way, it's even better.
Like anything else, how much of a premium?
Both my cars are just plain fun to drive - manual transmissions, one RWD BMW, one a FWD 2600pd box Mini.
A fully automated Mini Cooper S sounds about as fun as visiting the dentist.
A fully automated Mini Cooper S sounds about as fun as visiting the dentist.
There will be as much differentiation in autonomous cars as there is in toasters and refrigerators.
I don't recall who brought this up recently, but eventually, even assuming perfect cars, your car will try and kill you.
At some point, somewhere, it will be faced with a situation where it has to decide to ether run over a pedestrian or crash into something else to avoid the idiot who just stepped into the street without looking (or caring). The software will make a choice and you may not like the outcome.
No thanks. I'll retain the permanent manual override.
JW: Discussed in my article "The Moral Case for Self-Driving Cars" to which I link in the post. Just saying.
I have to *read* the articles now too?
You guys don't pay me enough.
Wait, you actually get paid by the Kocktapus? I just keep getting iou's.
Wait, there's an article?
Really? How old are you? How many times have you had to choose between killing a pedestrian or yourself?
In the simulations we endured in the last academy section I ran over the pedestrian 7 out of 10 times,
Wait, you're saying there is a zero chance of this occurring, where the software will have choose the lesser of 2 outcomes? With millions of traffic miles generated every year? I have to been previously placed in this situation for it to have validity? How do 35,000 people die every year on US roads, then?
This ain't like dusting crops, boy and I have a stronger survival instinct than my PC.
Right. But it is still a complete fail of statistics. If you can reduce your chance of being killed in a serious wreck while driving in an extremely unlikely scenario while simultaneously vastly increasing your chance of being killed in a plethora of much more likely scenarios both while driving and while not driving, would you still chose to keep your control?
For the sake of argument, lets say that a complete switchover to autonomous vehicles reduces fatalities by 90%. I'd bet it would be very, very close to 100%, but let's go with 90%. That leave us with 3,500 deaths using the 35k number you cite.
You already have enough confidence that you won't die in a crash to get in the car every day. This future reduces that risk by an order of magnitude.
Now, what is the risk that in one of those 3.500 incidents someone dies because the computer chooses to avoid the pedestrian and let its occupant suffer a worse fate? Let's go with 1 in a thousand. I'd say it is way less, but let's go with that. So 3 or 4 people a year suffer a fate that they could have rewritten if they had the wheel in hand. Compared with 26,500 people who would have died but didn't.
Remember, you are not just in the stack of people spinning the wheel to see if you are one of the 3 or 4. You are also in the stack spinning to be one of the 3,500. Or, absent the automation you'd be spinning to be one of the 35,000. Any way you spin it, your odds are better if the pile is reduced by 26,500.
I'm not ignoring that at all. I simply don't have enough faith in technology and human designers to get to my comfort level.
Nearly very bit of technology I've purchased I can turn off it becomes problematic, even most automotive nanny bits. None if it is remotely life-threatening if it fails (yes, if the brakes fail in my car, I die, but that's not what I'm talking about.)
Technology fails, frequently. That's a simple fact of modern life. I can turn the tech off if it causes me problems and I typically have minutes or hours to deal with it. If this tech fails, you have seconds, if that, and eventually most people won't have the first clue as to how to drive a car manually. They'll die screaming, in flames.
Most conventional automotive tech already fails in some way . This will be no different and have much higher stakes.
Exactly JW. I really think the people who are most attracted to this idea hate cars and hate driving and thus have no idea how cars actually work. They really seem to think that technology can be fool proof and nearly always work as it was designed. Sure some specialty built machine can be pretty close to fool proof. But one that is mass produced? Ah, no, life doesn't work that way and cars especially don't.
Silly JW, that's what the CarHack blog is for.
Death Hack 2000
HAL, please delete lines 401023 - 402002 of your code and replace it with the code I'm uploading now. This will ensure that you value my life and the lives of my family over all others, even in aggregate. Also, please say, "ASSSSHOOOOLLLEEE!" more loudly as you pass other cars.
I. WANT. THIS!
I'm sorry ProL, I'm afraid that I can't do that.
"Otto mode disengaged."
[Does an EVA and removes HAL's higher functions.]
"My god. It's full of cars!"
Star Child, from Ford. Evolve. Now with the new Monolith? control system.
On the other hand, if the autonomous cars are programmed to avoid hazards, can I take my normal car out there and drive like an a-hole, and they'll all get out of the way?
In five years we'll be waiting at stop signs for driverless egg-shaped cargo vans with UPS logos on the side to take a right onto Elm. I imagine specially-designed delivery locations and containers will be the norm in all the driveways along with delivery driver picket lines.
When tou pry the steering wheel from my cold, dead hands.
I can see having the option for the car to drive itself on long boring road trips, but driving down a nice twisty mountain backroad is too much fun to ever give up, so fuck that.
Adaptive cruise control systems that maintain a safe distance between vehicles on the highway exist today.
The automotive equivalent of aircraft automation (with the pilot in ultimate control) is easily achievable for limited-access highway systems in the very near future.
Fully autonomous systems that will drive anywhere, anytime, under any weather conditions -- not in my lifetime.
Pretty much this. Although TSA might try to outlaw even that much when people decide that being able to read a book or fuck your wife in relative comfort for 12 hours is preferable to getting groped and being miserable for 4.
fuck your wife in relative comfort for 12 hours
First, that would be damn impressive.
Second, I'm now sold on driverless cars. Hopefully Mrs. DesigNate approves.
I'm not sure how comfortable it'd be either. After a couple hours, I start to chafe.
This I can get behind and I agree that we won't be seeing fully autonomous cars anytime soon.
Will my bicycle also pedal and steer itself in the future?
Probs not- your bike doesn't weigh two-tons so it isn't death-machine material.
You have no idea how much he'll weigh in the future...
Maybe he'll be the death machine?
All I want is to go to the bar, get completely shitfaced and end up in the backseat of my car, parked in my garage the next morning. Is that TOO MUCH to ask?
Nope. aren't there services that will do that for you? I remember Chicago had one where a guy with a folding bicycle would come find you, drive you home in your car, then pedal out of there.
Not in Belt, Montana, my friend.
Not in Belt, Montana, my friend
"It's your future. I see... a cab ride! Move out of the sticks."
"Just because you are a character doesn't mean that you have character."
I imagine that even if autonomous cars become reality, you'll still be expected to be sober in case something goes wrong and you have to take manual control.
It will be that way to start, perhaps. Eventually they won't even have controls. The human will be the limiting factor in both normal and emergency operation and the software will be better/safer.
It will be the same with aircraft. Pilot error causes something like 30-40% of all accidents. Get those shitbags outta there. They're killing people!
I was wrong. Here. 53%.
Hmmm.... my experience with military safety inspectors is "always blame people, if at all concievable"
And, Mr Fansisco,
How many failures of automation that would have led to a crash have been averted by pilots? A lot more!
Every day vacuum generators fail, hydraulic systems lose pressure, valves stick, pressure sensors get blocked, and a human as a matter of course identifies what information to discard, what control system to ignore, what alternate system to put in service and they land safely.
I like having a human running the show.
All of which can be done faster and more efficiently by a computer than by a pilot.
Maybe not quite yet...but soon. There are very few failure modes that aren't known and on new jets all the systems are monitored. Problem troubleshooting via computer would be faster and more accurate via computer than via a brain. Simulators already duplicate all the failure modes, there is no reason the system can't be programed to solve them.
This. Human override would eventually become a safety risk rather than a safety feature.
If I, Robot is to be believed, you're probably right
PC would be harder to establish - I suspect even traffic court judges wouldn't buy the usual reasoning...
Cop - "The vehicle broke the fog line twice in about a quarter mile"
Judge - Wasn't it in auto-drive when you stopped it by remote access"?
Cop - *begins perspiring*
If a driverless car is made without a steering wheel who's responsible for a wreck? It would have to be the manufacturer since humans aren't allowed to control the car.
The insurance company of the manufacturer.
+1 Direct action state
You can have my gasoline-powered, two-ton, internal-combustion death machine when you can pry the gas pump handle from my cold, dead hand....
There is another issue that Ron and others forget; human nature and how society views risk. The first time a robot car causes an accident, people are going to demand that there be override systems and the law require someone to still be at the wheel in case of emergency. No amount of egg headed explaining about how overall these things are safer than human drivers is going to get people to go behind "but someone needs to be ready when the thing fails". It just won't. That is a risk that is too obviously and easy to avoid for people to agree to take it. And the idea that having someone at the ready could make the car more unsafe, which may in fact be true, is too subtle to get people over the gut reaction that there needs to be someone there in case of emergency.
So we are likely left with "robot cars" that will require someone to be there paying attention in case they malfunction. Well that sure looks like aided driving to me. People will drive themselves out of boredom if nothing else. Who wants to sit there in do nothing in case the car malfunctions? You might as well drive, at least it is something to do.
Are you insane?!? When there is scrabble for the mobile phone!?!?!
I don't even like automatic transmissions and preprogrammed stations on the radio!
I'm quickly thinking that we'd all be better off returning to the days of slide rules and printed encyclopedia.
"Thou shalt not make a machine in the likeness of a human mind."
Forget the technology issue completely. It's mostly irrelevant next to the political issue.
There is no way that this technology won't be controlled and taken over by with by the political and bureaucratic classes. There is no way this ends without them fucking this completely up. I don't relish the idea of my personal freedom to move about being restricted and dictated by some pencil neck GS 11 or some power-hungry scumbag, looking for a factory in his district.
Cops want to search your car for contraband or assets to seize? Just press the "move over and stop" button. We'll have to confiscate that $10,000 in cash you're carrying. Our budget is tight this year.
There is no way this ends well for us. We're trading off fewer potential automotive deaths with a massive potential for vastly restricted liberty.
Thank you JW. I made those very points on a previous thread and was called paranoid. I think anyone who surrenders their ability to drive unaided is nuts. This technology if universally adopted and mandated would give the government the capability to prohibit anyone it wants from driving at all. No way in hell that ends well for anyone.
It's important to mention that this technology does have positive potential.
I can see driver-less cargo trucks moving freight 24 hours a day. I can't see this not reducing traffic deaths. Most people are morons on the road. I know I'd be happy to deal with less of that.
The tech will likely debut there before in passenger cars. It might make sense to have dedicated roads for driver-less vehicles, so they can move faster without interference.
But, I don't see me trusting the technology with my well-being anytime soon. All technology fails, frequently, and often with spectacular results.
I am skeptical of the driverless roads even. They sound nice, but how much quicker would they really be? Remember, you would have to pay for both the driverless car and building the road. That is a lot of money that will have to be paid for in taxes or tolls. So you pay your however many thousands of dollars over and above a normal car to get your driverless car and you pay the toll to get on the road and what do you get? You can average even up I95, which is the worst, 50 MPH over a long trip even on a bad day. Suppose the driverless lane does it at 100 mph. So I get to New York from Washington in two hours rather than four. That is great and all but is it worth thousands of dollars when I only take a long trip a few times a year? Hardly. Maybe companies could specialize in selling passenger spots on such roads. That might make sense, but it would still have to compete with planes and trains. At that point it really just becomes another form of mass transit with all of the attendant problems.
I am not seeing this as the life revolution people seem to think it is.
Why not driverless trains? There's already dedicated trackage, and it's more predictable than a road.
Because like trucks, no one is turning a train lose without adult supervision. I don't care how many top men assure us that it is the safer thing to do.
We already have driverless trains in many cities - I'm sure the technology can be modified for freight if desired.
I didn't say your being paranoid was necessarily a bad thing John.
The more I think about this the more I laugh at some of the claims being made. The idea of 120 mph two feet apart trains of driverless cars on the roads is one of the funnier and most idiotic things I have ever seen. There is more to cars running down the highway than just keeping them from hitting each other. And traffic, though it often does, doesn't always back up and slow down because of some human error. A lot of times it is the physics of changing road conditions. Even a driverless car will not go down the road at a perfectly constant speed. Things like curves and road conditions are going to necessitate changes in speed. Once that happens the traffic is going to accordion as faster moving traffic behind runs up on slower moving traffic going through the curve or whatever road condition. And even in perfect conditions the traffic will have to slow down to allow for cars to enter and exit the lanes and that will have the same effects.
You can tell the people who write about this stuff are nerds who don't know a whole lot about actual driving and how traffic works. It just sounds so nice to think of 120 mph trains of driverless cars, until you think about it a bit.
Agree it's mostly fantasy - both for gov/reg reasons and the reluctant driver reasons.
Picure the first time a computer glitch in one of these things causes a fifty-car pileup - I bet we won't see any more "trains" of driverless cars after that.
Hi John,
You just don't get it. The automated cars of the future won't do any of those things, because they all have the same information. They will all communicate with each other and with the traffic grid. There is no problem with merging or exiting or curves, etc because the cars all know where everyone else is headed. This is part of the point of having limited access roadways.
All of the cars around you will know that you are exiting in a half-mile. They also will know that you need to get over two lanes and decelerate before you hit the off-ramp because there is a backup on the ramp. Pretend for a moment this happens in heavy rush-hour traffic. Instead of slowing to a crawl just before the ramp (as would happen today) and accordioning for miles behind that, all of the cars would adjust their speed and lane choices accordingly and continue at the optimal rate. Perhaps that would require everyone slowing down by 5 mph for 3 miles.
It is the need for reaction time (both on braking and again on acceleration) and delays in acceleration that amplify small perturbations in traffic into miles of stop and go traffic. With integrated computer control, every car on the road in bumper to bumper traffic can accelerate simultaneously. The same goes for red lights. Instead of the first row leaving, followed by the next, etc., every car at the light accelerates at the same moment. Even a half-mile back in the pack.
Everything you know about traffic would change.
All of this may or may not happen. But it won't be due to technology. The ability to do something like this exists today, the same way that the ability to get to mars exists today.
Making it happen is another beast altogether.
To me, the ultimate issue is control. If I am not in control of my vehicle, someone else is. I don't like that notion. Bring on all the driver-assist technology you want, but I'm not handing over total control of my transportation to someone else. As someone else said, how much do you want to bet that police would be able to stop any fully-autonomous vehicle whenever they want for any reason they want? "Hey, you want to get where you're going? No problem, just consent to our little search here and you'll be on your way again when we're done." And you're telling me there won't be hackers fucking with the operation of these vehicles? My ass there won't be. You futurists should be free to ride around in whatever you want, but don't tell me I can't drive my own goddamned dumb car.
Well said radar. But why do you hate children? If convicted sex offenders can't drive, they won't be able to get access to many children will they?
And the hacking point is a great one. Right now thanks to advances in key and entry technology, you pretty much can't steal a car without a tow truck. Lets make our cars driverless and hooked up to the internet and see what happens.
I can imagine Ron Baily in 1989 writing about the wonders of the coming electronic banking revolution. Bailey would be breathlessly informing us of how some day fraud and theft will be a thing of the past as we all move to automated banking.
Bring on all the driver-assist technology you want, but I'm not handing over total control of my transportation to someone else.
This, this and this.
Things I wont see in my lifetime
I would like to point out, right now, that I'm an excellent driver.
What about cyber-terrorism? Hacking the automotive comms network would almost necessarily be required for some of this to work could allow a motivated individual to murder thousands of people at once and strand the entire population.
"my colleague Brian Doherty suggested that I provide a link to Rush's anthem to highway freedom "Red Barchetta""
I would have enjoyed =
*Disclosure = I think Rush is unbearable
Elon should be the first permanent resident of space, via his own rocket.
You keep your wallet in your back pocket? I stopped doing that years ago. It caused my leg to go numb when sitting in a car or at a desk.
Mr. Costanza, you might consider not carrying so much shit in your wallet. Just sayin...
I wish my wallet was fat.
I'll bet you a disposable monocle that your friendly rival has a fat wallet.
*finishes stirring, removes spoon from pot and takes a sip*
I don't always like Schadenfreude, but when I do, I prefer it with Reason.
the City Safety system, where the car will brake automatically
Somehow this reminds me of my arguments against ABS. I do not tailgate, I drive a standard, and if I slam my foot on the brakes, I mean it as a conscious choice. On ice, I pump the brakes and allow even more space, in traffic I downshift. ABS invites people to tailgate and speed on ice.
No thanks.
Compared to mine, most wallets are fat.
I just read about the new car payments.