Google Self-Driving Car in Crash: Like Always, Humans at Fault
Human drivers in other vehicles seem prone to rear-ending Google's robot cars.

A van t-boned a Google self-driving car as its driver ran a red light. The car was pretty well wrecked but apparently no one in either vehicle was greatly harmed. In a statement to 9to5Google, the company noted:
A Google vehicle was traveling northbound on Phyllis Ave. in Mountain View when a car heading westbound on El Camino Real ran a red light and collided with the right side of our vehicle. Our light was green for at least six seconds before our car entered the intersection. Thousands of crashes happen everyday on U.S. roads, and red-light running is the leading cause of urban crashes in the U.S. Human error plays a role in 94% of these crashes, which is why we're developing fully self-driving technology to make our roads safer.
Red-light running human drivers have hit Google vehicles before. For example, Google's latest montly report on its self-driving car project noted:
In that same report, the other four accidents involved human drivers in other cars rear-ending Google vehicles.
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
They need to stop driving so infuriatingly slow.
Do they always obey the speed limit?
That could be a problem on a lot of roads.
I Make up to $90 an hour working from my home. My story is that I quit working at Walmart to work online and with a little effort I easily bring in around $70h to $86h..Go to this website and click tech tab to start your work.Visit this web... http://tinyurl.com/hygs5jl
It's not that hard to understand; the necessary component that must be in place for self driving cards is for it to be illegal for humans to drive. They haven't trotted out that component yet because they know it will be unpopular, but lets face it who needs freedom when we can have an increase in safety?
Right? That is what we're saying here isn't it?
So far, I see no proof of this in any sphere of robotics or technology that these marvels can account for irrational human behavior en masse. Go ahead and use your voice to text option on your phone. Yes, it works 'well' but not 'perfectly'. When you're talking about billions of multi-ton vehicles, that imperfection translates directly into some percentage of death and property damage and massive liability for...you? The automaker? The software developer? Doesn't matter. That part can be resolved, but the technology can not. At least not in the way most people picture it.
It's essentially handing over your personal travel autonomy over to both the Government and a 3rd party intermediary. I don't see how that can end badly. Can you?
"Red-light running human drivers have hit Google vehicles before."
Makes one think the cars' situational awareness algorithm for anticipating a potential red light runner might need some tweaking.
Six seconds is too far away to guess reliably. I always look before pulling out, but any car six seconds away would be too far to see most of the time, and even if sight lines were perfect, I'd expect a car six seconds away to be planning on stopping. It doesn't take six seconds to stop in most cities.
I looked up the address on google maps. Looks like northbound Phyllis is a 30 mph zone, 45 fps, 300 feet away at typical speeds. 30-40 mph only takes 100 feet to stop, from decision to applying brake to actually stopped. Even if intending to slow down slowly, any car 300 feet away is not usually slowing down yet.
About the only thing an autonomous car could do is see the oncoming car and accelerate or brake or steer or otherwise try to avoid the collision, all dependent on being able to make any evasive maneuvers, and assuming the errant driver doesn't wake up and take his own evasive action.
Depending upon traffic, it's quite likely several cars had already gone through the intersection already. The car ran a light that had been red at least 6 seconds (more so if there was a left turn light in the same direction as the Google car).
"Our light was green for at least six seconds before our car entered the intersection."
"In that same report, the other four accidents involved human drivers in other cars rear-ending Google vehicles."
I'm going to go ahead and assume that the Google car was not the first in line at the light for the first quote. If it was, rear-ending seems justified.
Also, as has been noted before: Self-driving cars? Great. You first, Bailey.
You missed this one
"At the time of collision, our vehicle was travelling at 2mph"
Granny lowing through the intersection...perhaps it would have made it through if it used the accelerator a little instead of riding the brakes.
This seems like spin to me. These cars drive ridiculously slow to the point where most people would consider them awful drivers, like the old lady. But like the old lady, they are rarely going to be 'at-fault' by the way the laws. But their inability to keep up with the flow of things is the real issue.
It's not really human error when the vast majority of people don't drive like this car and people adjust to that.
I checked out that intersection on google maps.
In order to complete the turn it would need to travel at least 60ft total and would require around 20 seconds to make the turn if 2mph was its average speed.
If that is typical of how it makes turns and such, its a wonder they don't get rear ended more often.
From the Google's report quoted in the article:
It was being driven by a human.
So this whole story is completely irrelevant.
The 9to5Google story says the first accident (the 9 Aug 2016 accident is the second accident in the story ? there's a lot of mixing up of the accidents in the comments) is the worst accident a Google car has been involved in.
Previous accidents have involved only minor damage. This accident (there's a picture in the linked story) crushed one side of the Google car.
I find it kind of surprising that anyone is willing to come out and say that driving according to the traffic laws, or driving like a grandma, is "bad driving".
It seems presented as an acknowledgement that no one follows traffic laws unless they feel like it, and can be arsed to pay attention. It presents the reality of driving as a chaotic ballet of conflicting motivations requiring situational alertness as well as a base understanding of probability factoring non-compliance as a given.
Whereas the robot driving is clearly superior, efficient and obedient to the cocooning network of traffic and safety laws, and only works if everyone drives the same way.
As a metaphor, it ain't bad. Not to pick on you, Brochettaward. As you were.
and only works if everyone drives the same way.
Which will never happen as long as humans are allowed to drive. Basically, the only way these things will make us safer is if they take away our freedom to drive themselves. Bailey sees that as a positive and is happy to give up that freedom. I and hopefully a lot of other people don't. I am not trading my freedom and autonomy for some marginal increase in safety. It is not like driving is that dangerous to begin with.
This is just dumb. Should jetliners hold a raffle at the start of every flight to decide who gets to be the pilot? Every flight you've ever taken required that you give up autonomy for reasons of safety, and you'd happily do it if you're not a fool.
The data absolutely do not agree with this assertion. This is one of the reasons I'm still a minarchist instead of an anarchist. The reality is that most humans are atrociously bad drivers who do dangerous things routinely and have been lulled into a false sense of safety.
Driving is one of the most dangerous things you can do. If we treated bad driving like a preventable disease, then self-driving cars will be the inoculation that changes everything and you'll end up the anti-vaxxer in the next twenty years or so.
Yes, I am willing to give up my "autonomy" and yours if it has the effect of ending automotive accidents as a leading cause of death. It won't be "marginal", it will be dramatic. It is similar to rules requiring that firearms be operated in a responsible fashion (at a range, on private property, fired away from people and property so as not to endanger others). The only difference is that it's "normal" to launch two ton projectiles towards others recklessly.
Also, we put our lives in the hands of autopilots on planes extremely often. It's one of the reasons that air travel is far safer than ground travel. No amount of complaining is going to change the fact that human reaction time and decision-making is already trumped by algorithms. A computer can take into account a million factors and choose the best course of action before your brain has even begun to send the impulse through your nerves to move your foot from the accelerator to the brake.
Being a libertarian doesn't imply neo-luddism or some mistaken belief that we're all John Henry and can defeat those dastardly machines at every task through our rugged individualism.
Also, we put our lives in the hands of autopilots on planes extremely often. It's one of the reasons that air travel is far safer than ground travel. No amount of complaining is going to change the fact that human reaction time and decision-making is already trumped by algorithms. A computer can take into account a million factors and choose the best course of action before your brain has even begun to send the impulse through your nerves to move your foot from the accelerator to the brake.
While true, this is an argument for autopiloted cars in a vacuum, or when they achieve some certain amount of market saturation. This ignores that the technology will likely be phased into the market, not available to all cars in a short period of time. Simply put, the robots are going to have to share the road with humans for an undetermined period of time and while they do they are going to rigidly adhere to local laws meant to be loosely enforced by and against humans. This makes them a hazard unless and until they get the roads to themselves.
Absurd. We all had to learn the same rules of the road. We all had to pass a driving and written test that supposedly demonstrated competence. The complaint is "the cars drive like granny." Would you tell granny that she's not allowed to drive because she follows the rules? Well, are you a libertarian or aren't you? And you have to be able to handle driving with the expectation that you could encounter granny on the road at any time. If you can't do that, then you're an incompetent driver.
The vehicles have to follow the law for a very simple reason: everyone was taught that same law and we expect the machines to behave in extremely consistent ways. You're really making the case that they should be coded to drive like assholes because everyone else does, which means introducing a lot of inconsistency into the algorithms, the same kind of inconsistency that gets people killed right now. You're making the case that they should be made objectively worse than perfect drivers.
Would you tell granny that she's not allowed to drive because she follows the rules? Well, are you a libertarian or aren't you?
Not allowed to drive? No. Tell her she's a hazard? Would and have. As Acosmist points out below, "Driving below prevailing speed is worse than speeding and is objectively bad driving." Why is that? Well, to steal again, "driving [is] a chaotic ballet of conflicting motivations requiring situational alertness as well as a base understanding of probability factoring non-compliance as a given."
The only times anyone follows the letter of the law is when they are taking their road test or within view of a cop. That's the reality, and the manufacturers are pitching a product that flies in the face of it. Granted, this isn't entirely their fault: the vehicles will have to follow the law because regulators will demand it in order for them to greenlight the technology, as will the insurers that will be on the hook for any problems that arise.
You're making the case that they should be made objectively worse than perfect drivers.
No, you're making the case that traffic regulators at all levels of government know what constitutes perfect driving, instead of defining standards in service of other interests, and that programmers are performing a service by coding in support of those conclusions. As a libertarian you should find the stupidity in your attitude self evident.
@Zero Sum Game
One need not drive like granny to be following the rules of the road.
Jetliners are not cars. I don't own a personal jet sadly. Neither do many other people. The car is the greatest improvement in freedom and autonomy in human history. It allows me to go nearly anywhere whenever I want and not have to depend on someone else' schedule. Turn that over to a machine and I don't control that anymore. I think I do but really I don't. The programmer controls it or more likely the government controls it by not intervening and shutting my car down.
Driving is one of the most dangerous things you can do.
Most is a relative term. It is only the most dangerous because our lives are so safe. In absolute terms it is very safe. The chances of you being seriously injured or killed in an accident in your entire life is very small. What is dumb is that you can't understand relative risk versus absolute risk and thus think driving is dangerous when in fact it is not.
Yes, I am willing to give up my "autonomy" and yours if it has the effect of ending automotive accidents as a leading cause of deatj
This is one of those circumstances where people's emotions trump their reason, and it apparently affects purported libertarians.
The market is bringing a solution to a real problem and people's emotions overrule facts, reason, and evidence.
Free minds and free markets. It's the very first phrase we see.
The data is already in. Reason tells us that autonomous vehicles are safer than ones with human drivers. Yes, the cars have to obey the law. They'd never be allowed on the road otherwise, and for damned good reason.
Imagine a future where the law itself no longer makes sense, where there are no lights or traffic signs, and intersections are full of vehicles that splice between the gaps in ways that would cause your sphincter to clench if you were asked to perform that maneuver yourself.
What is "autonomy?" I like the idea of having the autonomy to read a book, nap on the way to work, to be free to do whatever I please while the vehicle does all the rest. And I still get to be in control of the car's destination. I can go into the store and tell the car to go find a parking spot on its own. I can go to a bar and when I'm ready to go home I don't even have to think about whether I'll be a safe operator (or anyone else leaving the bars are either). Beg or pay someone else to pick me up at the airport? No way, just call my own car to come and wait for me when I arrive. These things sound like liberation to me, not like tyranny.
"Imagine a future where the law itself no longer makes sense, where there are no lights or traffic signs, and intersections are full of vehicles that splice between the gaps in ways that would cause your sphincter to clench if you were asked to perform that maneuver yourself."
There are countries where the roads basically lack signs, lines, and signals of any kind. I've driven on them, and amazingly enough, it was not like this. Doesn't make these things bad or pointless, but amazingly enough, it's not the anarchy you describe. People are pretty good at coming up with their own rules, and the same thing happens now.
I don't agree with John entirely on self-driving cars, but the goal of these things should be to make them drive with humans. And if they can't do that, there's a legitimate issue and John's concerns are entirely valid. Free markets and free minds. Reason is writing to undercut preemptive regulations that would snuff out a technology. The opposite end of that is self-driving cars are adopted, and become mandatory so people no longer have a choice. And the technology is open to abuse by government. That is not a libertarian ideals. Nor is this a matter of anarchy versus minarchism.
I like the idea of self-driving cars. You would have to be blind to not see how our government can fuck it all up.
I'm not implying that it would be anarchy or bad. I'm implying that the cars would do a far better job managing it on their own and it would be a glorious symphony rather than the human-controlled clusterfuck we have now. The cars could do shit that would be insane for us to attempt. We'd have to tint all the windows so the humans wouldn't shit themselves.
Yep, I agree that the government could fuck it all up, and they'll certainly try. The point is that even governments can't stay in the way of technical progress forever. Automobiles are far more dangerous than horse-drawn carriages, but they won regardless.
As far as the "market share" argument is concerned, that will sort itself out more rapidly than anyone can imagine. Everyone is thinking about passenger cars, but the reality is going to be fleet vehicles for deliveries will be the first to completely switch. The insurance will be cheaper. A fleet of perfect drivers that won't be found at-fault for anything and don't do stupid shit that causes maintenance costs to be high, automatically drive with fuel economy in mind, and don't get tired, distracted, or collect a salary. That's market forces at work right there, man, and it will be overwhelming.
Zero Sum Game,
You don't understand that decisions are made on the margin. A new product must create more marginal benefit than its cost. Will these things have cheaper insurance? Maybe, but insurance isn't that expensive to begin with so the marginal reduction can only be so much. And "perfect" is a relative term. Most people would find riding in a car that never breaks the speed limit or makes any judgment calls to be miserable. Even if you can run your robot taxi service and charge lower prices than human ones, you still likely won't make it because people are unlikely to find the small savings worth the cost in time and aggravation of the way the things drive.
As Brooks says below, these things are the answer to a question very few people were asking. Driving is not a particularly onerous task and relieving people of doing it is a very small marginal benefit.
I don't think you know what the word "marginal" means. We'll just see won't we? In twenty years when most vehicles are autonomous, you'll be buying me the beer, yeah? Everyone invests all this money and effort into getting this right because they know it's going to be huge business. Even if some holdouts and nay-sayers are thinking with their emotions rather than their reason and business acumen, the sea change is coming and it will raise all ships whether they want it or not.
Everyone invests all this money and effort into getting this right because they know it's going to be huge business.
Really doubt there will be all that many takers for a feature that adds 5,000 to 10,000+ to the cost of the vehicle.
More likely they are counting on a government mandate that post a certain date all cars must be autonomous, and in the years before then, massive subsidies such as those available for EVs.
I know full well what marginal means. I just explained it. What does this technology get me over what I have now. And the answer is not much. The marginal benefit is how good these things are compared to just driving myself. And from the individual perspective, that is not a very big benefit, especially when you factor in how slow they will drive.
So what if the insurance is cheap. I only pay about $600 a year for insurance. Even if they made insurance free, it still would likely not be worth it.
If these things are in wide use in 20 years, it will be because the government forced them on people not because they are efficient or make any sense in most cases.
You broke it, John.
No shit. You've truly topped the Johnism by arguing that humans are capable by destroying the entire comment thread with a careless open ended italic.
Let see if this will end it.
Nope. Closing the italic bracket doesn't work.
I have destroyed the matrix.
I wonder if this will work.
Everything Italics is kinda damned annoying [/i]
If you're going to compare autopilot on an aircraft, you might want to ask yourself the question of do airplanes actually land with autopilot? No, no they do not. Nor do they taxi, or do any other ground movement with autopilot. Nor does autopilot fly the plane in any type of weather or complication. Airplanes operate in an entirely empty and obstructionless environment, and only need a heading and course corrections. If you can't see the difference between that and highway or downtown traffic, you're either an idiot or willfully blind to the obvious limitations of technology. (And don't even get me started on the implementation of rural or nationwide highway implementation, how much money are you willing to burn for virtually no benefit?)
That isn't to say that one day these might be possible, but that day is so far into the future I would be shocked to see it in my lifetime in any meaningful way. This is right up there with flying cars, solar panel roads, or A.I. as a futurists pipe dream. It is literally easier to get to Mars than to implement this idea.
Um...yes, yes they do.
And you do realize, that which you call impossible is taking place as you type?
Um...yes, yes they do.
The fact that they can land themselves does not mean it is typical.
Its really only used in cases of very limited visibility. Manual landing is the preferred and typical means of landing.
And the point he was making was that automated flying (maintaining heading, altitude, airspeed) is easier than automated navigation of a road.
On the ground you've got many other cars in fairly close proximity, potholes, signs to interpret, poorly striped roads, random objects entering the roadway. All manner of potential chaos.
In the air, particularly above 10,000 feet when such automated systems are typically in use, the only things to run into are other aircraft(at that height everything should have a transponder), storms, and terrain.
Even if the pilots are using these systems they are still monitoring whats going on and ready to take over if the terrain or aircraft collision detection systems signal a potential problem.
And that technology is 48 years old!
Unless the conditions are beyond the capability of the pilots. Then they use the more reliable system.
I don't know that I would say its more reliable. Its certainly more able when landing in very low visibility than a pilot would be, but I don't know about in clear air.
I rather suspect that autoland (as it currently exists) is far less sophisticated than what is necessary to navigate a car because the area around a car is very dynamic. The skies around an airport are strictly controlled so you don't have to worry (much) about running into small craft and such. There are fixed radio transmitters that define a sort of virtual road to the runway for landing, all the plane has to do is follow that road and not run into the ground. Without that strictly controlled airspace and the fixed transmitters, autolanding(as it exists today) does not work at all.
For a car, a road has paint stripes and they do indeed provide an excellent frame of reference for a car to follow. If driving a car on a controlled access highway following the paint stripes and not running into the back of other vehicles was all we expected an autonomous car to do then we could have had that fifteen+ years ago. That is easy. Whats not easy is navigating less controlled environments such as what you find on a city street.
Oh, I see.
Driving below prevailing speed is worse than speeding and is objectively bad driving. hth
Emergent/spontaneous order, as compared to planned order. It's a libertarian thang.
I drive 100 mph everywhere I go and change lanes like a schizophrenic maniac, and if you don't, FUCK YOU.
Gojira drives like he masturbates.
He needs Viagra for his car also?
As long as you use turn signals we're cool.
He leaves it on 100% of the time, just to make sure.
Who doesn't speed?
I'm a pretty conservative driver, I've never has a speeding ticket and have been pulled over 4 times in my life. And I am almost always going over the speed limit when I'm driving.
If these cars strictly stick to the speed limits, they are following the law, but not really conforming to the rules of the road as they exist in reality.
I don't think its surprising, people have been saying that for a long time. Pointing out excessively slow speed limits and things like that. In a lot of states you can get out of a ticket if you can show that a road's posted limit is too far out of whack compared to what the engineering assessment says it should be.
And local governments *love* that. After all, you can't rule law-abiding people - or at least you can't extract revenue from them at will. If 90% of the population is driving 'safely-but-unlawfully' then you can pick and choose who's going to get the ticket. Pop a cop out for a couple hours of aggressive enforcement and then have him go do something else for the rest of the shift. No need to even hire extra officers to make quota.
My little down does this - the main road through town drops from 50 mph to 25 through town (which is moderately understandable as that section is also heavily populated and full of businesses) and its the major through route for people traveling between San Luis and Yuma. So they park a cop at the appropriate end for rush hour traffic.
They would fit right in around here.
I've posted it before; there's a lot of that from the Google cars. The driver cohort that automated cars most closely resemble is the elderly. They don't technically cause them, but they get in a shitload of accidents.
I don't expect this to change, the incentives are such that it's in the company's best interests to do so - just look at Bailey's uncritical reporting. Add in the legal ramifications for making Google cars act like humans and there's just no incentive to have them be anything different.
I remain confident that I'm going to spend the majority of the next 25 years cursing their existence while I commute.
Here is the other thing about that, it is going to make them much less attractive to purchase or use as an uber or Taxi. You think it sucks being behind one driving like that, imagine being the passenger. Ron acts like driving is some deadly proposition instead of the low risk activity it is. Think about it, which would you prefer a human driven uber or cab that is going to strategically break the traffic laws to get you where you need to be as quickly as possible or one of these things that is going to obey every law to the letter?
I don't think the robot Uber is going to be as popular as Ron thinks it will be.
That car was t-boned, not rear-ended.
I understand that. I was just suggesting that if a car sits at a green light for six seconds it deserves to be rear-ended.
Okay, I googled "Self-Driving Car in Crash: Like Always Humans at Fault" like the title said to. Now what?
Appropos.
NSFW: Language.
SJW/Soavetarian Trigger Warning: Stereotypes.
"At the time of the collision, our vehicle was traveling at 2mph"
So self-driving cars will be right at home here in Seattle. A little fast for the local drivers, but right at home.
Until the cars begin assembling and programming themselves, won't humans always be at fault?
Also, from the sound of this and the Google report, it sounds like both accidents were human driver vs. human driver and thus normal car accidents with little-to-no relevance to anything except that the car happened to be owned by Google.
Lets see, the last time my car and another car came into contact, It was almost seven years ago. I was backing out of a parking space and some guy going 30-40 through the parking lot ended up behind me in the moment when I was checking the clearance with the parked car beside me. I dented his door and cracked my tail light. I might have been going less than two miles an hour at the time.
Unlike the Google car, I was trying to get out of a parking spot in a not so wide lot.
The parking lot at my work is terrible. Everyone goes 30.
You know who else thought their idea would work if other humans were just more cooperative
Niccolo Machiavelli?
Bitch, bitch, bitch! Fucking luddites.
Self driving cars are fucking awesome!
It's a good idea. I don't expect to be implemented properly.
Yeah, because private companies don't ever improve upon their products.
Some companies have a bad habit of shovelling out unfinished product and relying on patches.
Microsoft Auto?
I would hope so. You can't respond to the customer unless the customer is allowed to give feedback.
What operating system is iPhone on?
It's the unfinished initial release that's the problem. As in, essentially broken until later patched.
But it's not broken. It's been in what, 5 accidents, all caused by humans? They are safer than humans.
And I'm fairly certain they've run a cost/benefit wrt liability. If it's so dangerous that they thought they'd get the shit sued out of them, the cars wouldn't be on the road.
Yeah, but you should never download version X.0 and always wait for version X.1.
If your car won't drive until you update you're kinda fucked.
Fair nuff. That's prudent. But it's gonna be better than the internet. And I'll be able to get plastered at bars and wake up in my garage.
Realistically, you should always wait for version 'n+1'.
Release early, release often.
When the incentives are in place to do so. I predict that instead we will see the autobot makers pressing the government to steadily increase the cost of manual driving instead of adapting their cars to conditions as they actually exist.
This will be done in the name of safety, and they'll use the carrot of more efficient transportation in the future, once all those dastardly clingers are removed, to distract from the stick they are beating us with in the present.
But it isn't being implemented that way. The design is such that they interact with human drivers.
You're saying we shouldn't implement this great benefit to mankind because statist fucks might (stressed might) use it to restrict our freedom?
That's like saying we shouldn't have nuclear power because someone might make a bomb. Or the internet was bad because of how the government is using it to spy on me.
No one is saying these should be banned. The danger is they will try and ban human driving.
Unless they manage to do that, these things are not going to be nearly as big of a deal as their backers claim. If there were no cars and these things came along, they would be revolutionary. We already have cars. So the marginal benefit they give is not having to drive and them being safer. When you realize that driving isn't difficult nor particularly dangerous, you understand how overblown this is. The other thing is that you are stuck in the car anyway. If nothing else driving gives you something else to do.
And as is pointed out elsewhere, they will inevitably be programed to abide by all the traffic laws at all times. That annoyance of that will outweigh any marginal advantage they give.
I will respond to you, just this once.
Marginal benefit? No big deal? Overblown?
The average American wastes 37,935 hours over his lifetime driving his car. That's an average of 611 hours a year. If you make $15/hr and you committed that time to doing productive work it'd be an extra $9100/year or $568,000 per lifetime. Multiplied by 310,000,000 people.
Long haul truckers...3.5 M in the US. They make (average) $50K/year. That's a savings of $175B per year in the trucking industry alone. What does that do th the price of goods? More money in everyone's pocket to be spent on things that increase their quality of life.
Then we can get rid of pilots, who make $200K each per year. What's that do to the cost of your airline ticket.
And I'm sure I'm just scratching the surface.
This is huge!
And they will still be spending 37,935 hours a year in the cars. Only a nerd of the highest order thinks that people would want to spend all of that time working or whatever. Do people do much work on planes? Some do but most don't'. Most people fuck off And that is what they would do in robot cars. And if all the robot car buys you is the ability to fuck off and watch porn or play Angry Birds, it isn't that revolutionary or really worth paying any money for. You still stuck in the car and driving is for most people as good a way to pass the time as anything.
And the long hall trucking thing sounds great until you realize how the liability laws work. No way in hell are they going to be able to put trucks on the road without a human backup. And if you have to pay the human anyway, the robot truck hasn't gotten you much. And that doesn't even address the larger issue of network security and hacking. How many trucking companies want to risk the liability and loss potential associated with making their entire fleet and cargo venerable to hacking and hijacking? Not many.
These things are going to be the next big technology for a very long time.
*sigh*
I was placing a value on time to prove a point. ANY activity is more productive/valuable than driving. Sleeping, eating, masturbating, playing games, commenting on H&R are all more valuable than the act of driving. I'd rather be doing almost anything but.
I know you were. The problem is that you assume that the extra value of fucking off or playing video games versus driving is worth the cost of having a self driving car. For most people it won't be. A lot of people like to drive and most others at least don't mind it. It is not like you can do that much in your car anyway. The bottom line is that the cost of the self driving car over a regular car must be less than however much you value the burden of driving. Since driving isn't much of a burden and is for some in fact a benefit, these things are going to have to get very cheap to be marketable.
More than that, however, is the fact that they are going to likely drive in a very conservative and slow manner. When you add in the cost of doing that versus driving at your own pace, these things make very little sense for all but people who can't drive or have a very unusual loathing of it.
You will never find me agitating for the prohibition of self-driving cars. I simply think they're going to be terribly annoying drivers that I will be forced to share the road with. Like, bottom-15 percentile, "merging is not that fucking difficult", get-the-fuck-out-of-my-face-already quality.
Since I live in a densely populated area I deal with a lot of this already, and I'm going to complain about them the way I already complain about everyone else.
If these damn things are actually slowing down for yellow lights then that might just be the problem. Who does that?
They slow down for green lights where I live. So quit complaining.
In Connecticut they slow down on freeway on-ramps prior to attempting to merge.
I have almost rear-ended soooo many cars when I was living there.
Great, so self-driving Google cars programmed with all of the rules of the road keep getting into accidents. What should the real takeaway from this be?
Google has great insurance?
Repeal the rules of the road. Also no driver licenses for anyone born past the year 2000.
It is always someone else fault. What Ron fails to mention is that avoiding accidents after other drivers have made an error is part of safe driving. If these cars can't do that, then they are not as safe as advertised.
Try explaining this to the millions of drivers that cause accidents while following the rules of the road.
Huh. It's almost like we should drive like other people suck.
Coast Guard 'Rules of the Road' - Rule '0': Departure from these rules may be required to avoid collision. Compliance with these regulations does not relieve you of your responsibility to avoid collision.
In one sense, yes it is the human's fault. In another sense, not so much. The reality is that conditions chance and it is not enough to drive by the rules for yourself, you have to drive the conditions and for other drivers. Sure, if you never speed up to get through a yellow light and always come to a stop even if it means braking hard, people are going to rear end you more often. Is that their fault? Technically yes it is. A rear end collision is the fault of the person who was in the rear. But that fact doesn't excuse the self driving car's responsibility for the increased risk of that happening. So unless you are going to ban human drivers, this is a problem for robot cars.
I will never understand how anyone who values freedom and self reliance could like these things. There is more to life than safety Ron.
In this particular case, however, running a red light 6 seconds after it has changed will cause problems for humans too.
Yeah six seconds isn't even close.
Ban humans.
I am hoping that just as I start slowing down, these kinds of cars will be available. I do not relish giving up my ability to up and go where I want, due to age and infirmity. OK, car developers - I give you 25-30 years!
Yeah, if you already are old and drive like an old person I don't see the harm in offloading the responsibility to the Google. But for someone who likes driving, like me, these will be the bane of my existence for the next 40 years.
What does it do at a stop sign trying to turn left onto a busy road with no stop sign and traffic going both ways? Pull out at 2 mph? Or dies it wait forever to go?
I was going to say it would have made a poor routing choice if stuck at that stop sign. But then I realized it probably uses google maps for that, so of course it's clueless.
If you can't drive defensively, then you can't drive.
Id love an old sports car like that. BMW's 2002 is growing on me
DO IT. I have had three; one of which I drove for nearly 20 years. I miss that car.
I want to test drive one to compare to the e30, but I really like the look.
I had an 83 3 series when I lived in Germany. I loved that car. BMWs are not the cars they once were.
did you see the new 2002?
I like it. That is pretty cool.
I don't have time to dig around at the moment...
...but does anyone have a short, simple answer why Reason has such a freaking hard-on for "Driverless Cars"?
There's nothing about it that seems at all self-explanatory from a libertarian pov.
in fact, lots of things surrounding the issue strike me as completely un-libertarian. Everything about them will require more govt intervention in consumer choice/behavior, demand more govt spending on adapting roadways, new forms of policing, and so on....
i get why its interesting from a purely "technology is good" POV, but everything else about it is just.... bleh.
Hmm, personal guess: Reason has pretty consistently argued that technology will be a primary means whereby the state is overthrown or drastically reduced. That is, statism will be overwhelmed by technological advances.
That is, statism will be overwhelmed by technological advances.
I have never understood that belief. States have harnessed and used every technology in the history of the world for their own purposes. I am not anti-technology I don't think technology is necessarily bad or good. What is bad or good is what we do with it. I don't' understand how someone could think that technology will necessarily make us more free. As you point out, however, a lot of Libertarians believe it will with an almost religious conviction.
Counter point, the NSA
A lot of Libertarians will buy into anything if it is wrapped in new technology. Also, not all nannies are leftists. Some of them are Libertarians ironically enough. Bailey is just a nanny on this issue. He doesn't like driving and thinks its unsafe and it bugs the living hell out of him that people out there doing unsafe and in his view irrational and inefficient things. He sees these cars as a way to cure that. You would think people who claimed to like freedom would also like the big messy dangerous and uncertain world that comes with that. But not all do. Ron and a lot of others most certainly don't. They like freedom but it needs to be done safely, rationally and efficiently.
Mostly, I think, its *cool technology* and partially its an interesting example of history being made. And its an instructive demo of how the private sector can innovate faster than the legislature can legislate.
For example, all this automated cars stuff has gone from 'maybe we'll need some special road markers' to the government pushing V2I (which, not coincidentally, requires massive spending, lots of rulemaking, and allows an incredible amount of monitoring and control over travel) to 'nah man, computers and sensors are now cheap and powerful enough that we can do everything onboard' - which eliminates the infrastructure building, the rulemaking, and the monitoring - and the government was left sputting 'but, but, but . . . *V2I*! while the companies moved on. And this is in 20 years.
It really, IMO, illustrates the principle that lawmakers *do not make law* (and should not) but *legislate* - as in codify already existing practice. Instead of trying to get out in front of this and looking like they're 'leading', they should be standing by watching and act as a referee of last resort when it comes to conflict.
I don't think it is going to be as revolutionary as some think. People look at these cars as if we didn't already have cars. No, they are judged on the marginal benefit they give over what we already have. And when you consider the ease and safety of driving, they don't really deliver that much, especially in comparison to their cost. The proof of that is in the fact that governments are having to mandate that cars have more and more safety technology. The latest is every car has to have a backup camera.
If backup cameras really made economic sense for most people, the government wouldn't need to mandate them. The government is only mandating them because driving is so safe and the marginal benefit of the cameras so small only people with a lot of money will pay for them.
I agree that they don't deliver a whole bunch. But they do deliver some nice pluses.
1. They increase the range of my mobility. This is the thing I'm most excited about with them. I live in the middle of nowhere and that comes with a cost - less of a cost in the internet age, I would have hated living here in the '80's - its 3 hours to Phoenix or San Diego, longer to Tucson. So anything I would want to do there involves a 7 hour trip just to get there and back.
Automated driving means I could faff about on the computer, read a book, whatever during that time.
2. They increase the *age* of mobility. Not something I'm worried about yet, but maybe I'll live long enough for the eyesight to completely go, the reflexes and attention span to go.
As it is not, #1 and #2 require extra costs to achieve and I can't afford a dedicated *human* driver.
There's other stuff - but it is predicated on a significant percentage (not necessarily a majority) of other cars being autonomous to take advantage of and, frankly, things like increasing road capacity is not an issue where I live anyway.
Some people will want them. But many and I think most won't unless the government forces them to have them.
The same principle is going to apply to robot cars only more so. Given a choice, how much more money would you really pay for a robot car over a regular one? Most people wouldn't pay that much I don't think. This is especially true when you remember they are going to be programed to drive like old ladies. If the price of letting your car drive is it never speeding or violating any traffic law no matter how minor, unless you are just blind drunk or physically unable to drive, when would it ever be worth not just doing it yourself? Same thing with taxis. Who the heck is going to want a taxi that drives like that? Even if were free, it would still probably be worth paying for the human driver.
The bottom line is the only way these things are going to be put into wide use is if the government mandates them. And that is exactly what they are going to try and do.
Like any automation, its going to be a cost-benefit trade-off. Like fast food ramping up automatin in response to wage hikes. You're replacing labor with capital.
But, the costs will come down - like they do on all (non-veblen) goods over time. Eventually someone in my income bracket will be able to afford this technology, and then sometime later it will make economic sense to spend a little extra to effectively get an on-call chauffeur.
Probably take another 20 years before a decent used-car with this tech will be affordable to me but it will happen eventually.
I think what you will see is cars that can virtually drive themselves but don't have to or always do so. Cars will increasingly step in to save us when we screw up or drive themselves when we get board. But they will not be the robot cars of Ron's dreams.
The way to get around the liability issues and the resulting driving like an old lady is to have just enough human interaction to make the human still responsible so the manufacturer can give them the ability to tell the car to drive faster or whatever.
I think that is the future. Fully autonomous cars in big fleets that make it so no one actually owns their own car is just a fevered dream of nerds who don't understand economics or human nature.
Who broke the italics?
John. It's always John.
...but does anyone have a short, simple answer why Reason has such a freaking hard-on for "Driverless Cars"?
Is it reason, or just Bailey? Other than as a masturbatory technocratic "solution" to the chaotic nightmare of freedom and independence, I don't get the appeal. It seems to be an extremely expensive and complex answer to a question very few people have really asked. I get the "I'm too lazy to drive myself; all that paying attention to my surroundings is so last century!" argument; pushing on pedals and turning a steering wheel is terribly draining. And, of course there's the precautionary principle. Think of the lives we'll save, blah blah blah.
We often loath each other Brooks, but you are absolutely right about this. That is exactly what this is; the answer to a question that no one other than a few tech geeks ever asked.
I seem to have botched my tag.
Not me!
Everyone
Also, DOOMco-
Check out some of the "updated" 2002s out there. I stumbled across a youtube of one with a small (2.5) six* in it. It sounded SWEET, and was way livelier than a stock 4-cyl 2 liter 2002.
*Apparently a comparatively straightforward swap.
Yeah, I've seen people put the 318i in, apparently all bmw is pretty much plug and play back then. A little more power would help! They do it with the 80's 3 series. Put in a 5 series engine or a mid 90's 3 series.
Most likely noted above, but since humans programmed the thing, pretty much a given that ultimately humans are at fault.
Why is everything in italics?
Looks like John left his italics bracket open way back at 1:50 and we just can't manage to close it.
I am excited for self-driving cars. As soon as they can interact well with human drivers and the marginal cost of one over a regular car gets small enough, I am all for it. I don't like driving and I don't like other drivers. I would rather do anything else than drive. For example, I could argue about things on an internet site while my car drives me somewhere.
People who want to keep driving shouldn't be stopped, but once an autopilot is good enough, I am going to be glad to not drive.
Why is everything still askew here???
This defense goes to 11
Here, take my fj40 around boulder. Or a highway. For a day.
A lot of the bad driving is because cars are so easy to drive today and have such good brakes and handling characteristics that you can be a bad driver and get away with it. The thing Ron never mentions is that driving is not a particularly risky thing to do. And it gets safer every year. These things are being sold to solve a problem that really isn't very significant. Sure, you can pull out the raw numbers of people killed every year and that sounds scary. But when you consider the hundreds of millions of drivers and if I am not mistaken trillions of miles driven every year, those numbers are miniscule in comparison.
It's more that the unfamiliar situation heightens awareness. I was never more focused on the road and what was going on around me than when I was driving a uHaul, because I am accustomed to driving a sedan. Thus the reduction in familiarity leads to increased focus and attentiveness. It's the reason why most changes to an intersection result in a near-term drop in accidents. The familiar has been stripped away, so people approach it more cautiously. Once it stops being unfamiliar, the numbers go back up. (up by less if the change was good, up by more if it was a bad idea)
I propose a giant spike protruding from the steering wheel of every vehicle pointing at the driver's face.
...dad?
Buying a motorcycle made me a much better driver and gave me a much better appreciation for how lousy most other drivers are.
I'm wondering if I'm the only person who regards themselves as a subpar driver and has avoided accidents by realizing I'm not that good at driving.
Okay, you guys had an H-n-R get-together at Delaney's a month ago, right? Now it's shutting down. What the hell?
Deal! Id love an old sports car like that. BMW's 2002 is growing on me, but I think I'm still an e30 man. My dad has an Austin Healey 3000. It is beautiful.
I saw articles on that - and a complete lack of information on long-term results, just the immediate-term, which can be explained by a spike in caution that will erode as familiarity returns. I would love follow-up data. My personal wager is that the number is going to be close to whatever it was before the change.
No, it fell through. I'd like to get one going though. When done the close?
If you realize your limits, you are not a sub par driver. You are a good driver. The bad drivers are the ones who don't know enough to realize that driving fast is more difficult than it looks. The thing about cars are everything feels great right up until the moment they are not great and it is likely too late to do anything about it. Everyone should have to take their car on a skid pad so they can realize how quickly cars can lose traction and how violent it is when they do.
I had enough small dingy accidents when I was younger to realize this. Its been 11 years since I caused one, which is the only single car accident I have ever been in - I lost it on ice and totaled my car. Going backwards down the interstate is not fun. Nor was hitting the guardrail.
I have still been hit twice as often as I have hit. And avoided 3 times that many.
I'm a guilty conscience driver. I still remember minor (like, nicking a curb with no bad consequences minor) driving mistakes I made ten years ago and beat myself up over them.
I think it elevates my driving game from "mediocre" to "mediocre skills but doesn't do stupid shit", aka good.
The newspaper said it closed last Sunday.
The 2002s are great. If you can afford one, go for it. You could do a lot worse.
Risk tolerance hasnt necessarily changed, so as cars get safer, you can drive faster/riskier with the same risk profile as before.
I've been promised that or the tvr vixen he inherited from his uncle a few years ago.
He rebuilt it when I was 5, bought another one and played a mix n match.
What a cool car tundra. I like the cream. Same color the 3000 is.
With the upside of reducing travel time.
Frick. What bad luck. The Irish place closed too, I tried going there recently.
Last coats on tonight after work. God, this has taken forever. But it's a lot less noise now. And I wired in some back up lights and a third tail light up higher. Hopefully to avoid a rear end.
English cars and wires don't mix. Lucas system I'd assume?
My 7 inch rounds are not led yet. I wanted to switch, but was afraid to pull the trigger at $500+ for a pair of JW speaker ones. I saw a jeep with some knockoffs and they seemed good, so maybe the 100 dollar range is OK.
All my other lights are. I used superbrightled.com for the turn signals and brakes. They have been great so far. Make the plunge.
Maybe they just got a bad rap early.