Forget self-driving car anxiety: In the early days human drivers were the fear
timeline.com
timeline.com
I'll bet they even ban hobbyist driving - like those who 'still have old cars' and want to drive on those 'old car roads'.
Statistically too dangerous. Far more dangerous than anything else we really do - certainly more than smoking. And when it's no longer a 'necessity' ...
Do you remember those movies before the 1960's when they get in cars and they don't even have seat-belts - and we 'gasp' ... 'OMG' ...
I think our grandchildren will gasp when we tell them we drove cars. They will think we were crazy and uncivilized, we'll call them wimps.
You'll want to teach your grandson to drive, your daughter in law will be upset and call you a crazy old man :)
I don't see that happening for a century. Unless you subsidize the cost of purchase of a new car for everyone who has a car now, and associated costs of owning a new vehicle.
Where I'm from(Poland) the average age of a car is 16 years. Buying a new car, even if it's something very cheap like a Fiat Panda, is still a huge deal.
Why would you care about what purchasing power is like in Poland you ask? Well, because at the moment I can drive my car literally around the world, without modifying it an inch. If some countries start banning manual cars, then suddenly you can't cross certain borders anymore - that's almost idiotic. If the ban of manual cars is going to happen, it will need to happen almost worldwide at the same time, and that's absolutely, definitely, completely not happening in the next 20 years.
I actually think your example with cars without seatbelts is great, because it illustrates what will happen with manually-driven cars - instead of banning them, they will eventually become so rare there will be no need to ban them, just like you can still drive your 1950 car without seat belts or airbags.
+ They will make them cheap. The AI is software, it's a commodity, and the parts (i.e. LIDAR) will be cheap by then.
+ In 20 years you won't likely be able to drive your old car in France, and no offence - they will probably mock the Polish for still having 'real cars' :).
But you're right - 20 years is too soon.
20 years to 'almost everyone in driverless cars'. 30 years to banning humans. 45 years to banning them even from driving hobby cars.
Add 15 years for East Europe and Russia.
Add 30 years for poor places.
Again, I don't think this will ever happen, just like we haven't banned horses on normal roads - there's just so few of them that it's not an issue. Likewise, if in 50 years 99% of cars are self-driving, it won't matter an inch if you want to drive yourself in your antique. If you find someone who will insure you in it, fine, go ahead, don't see why they would be banned - self driving cars will probably do a good enough job of avoiding manual cars anyway.
Just like horses which you cannot bring to a highway.
And cities autonomous only? So what about bicycles? Or tricycles? Or hand drawn carriages(rickshaws etc)? Would those roads have high walls around them ,stopping pedestrians from entering?
And finally - where would the money for that come from? Even very rich countries(I live in the UK) struggle to maintain their roads to a high standard, something simple like adding new pedestrian crossings can take years in planning and approvals, but somehow, we would modify all our roads to suit one specific type of vehicle?
I'm really sorry if I'm being cynical, but I really do have a feeling that this self-driving-car utopia is not going to happen.
I think however that at least for several years (decades) the autonomous vehicles must be able to work on the roads we currently have.
As a disclaimer, I would say that I am actually pro-ban of personal vehicles in city centers today. I really do not see a point of anything that is not an ambulance, fire truck, bus, a taxi or a delivery truck on a road. Why people wake up soon in the morning only to then be stuck in an over than one hour commute when a subway ride would take 20 minutes tops is really beyond me.
Banning all cars first and then progressively let autonomous vehicles (as part of a fleet) in would be a win situation in my opinion.
If you are talking about elevated distribution roads that people will pass under, then maybe. But I don't see it working on a city like the ones we have now. Maybe there's a major urban restructuring to happen soon...
Funny, the local SCCA group has autocross at an airstrip near an Amish community. The roads are swarming with horses and buggies the morning of the Sunday events. Interesting juxtaposition :)
There are some interesting statistics at http://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/pub....
For 2014 (latest year with all data available) there were 260 million vehicles on the road. In this year, there were about 7.6 million new vehicles sold. Of these, about 0.5 million were hybrids, having grown from 0 in 1998 and 0.25 million in 2006. If self-driving cars follow a similar track to hybrids, it's going to be a long time before the majority of those 260 million cars are self-driving.
Absolutely not. That will never ever happen. F1, DTM, MotoGP, MonsterTrucks, NASCAR , Rallye Dakar are all well too amazing to allow for that.
I agree, your commute form SF to Palo Alto and similar boring activities will be automated away, but there are way too many enthusiasts to allow hobbyist driving to ever disappear completely. Maybe, but only maybe, have them confined to racetracks or leisure roads, but I doubt it.
Source: I love cars and bikes and would go berserk if someone would tell me I am not allowed to drive anymore and I know hundreds of people that would feel the same.
I get it.
I know how you feel.
My grandparents grew up on farms - and there were always rifles around. My grandfather hunted.
They never, ever talked about 'guns' for self protection, I never saw a gun outside of it's locker (unless in the field).
My grandfather was an extremely normal and nice, down to earth person.
And yet often these people are treated as pariahs. 'City people' just think 'gun owners' are crazy red-neck killers. They would ban them in a hearbeat.
I feel you may be part of the 'NRA of the future' - ie a group that will be positioned by the press and pop-culture as 'angry old men out of touch' yada yada.
Everyone that dies in a car crash - it will make headlines like the random person who got shot in an accident or whatever - even though you car-enthusists are probably the safest and most responsible people imaginable.
Anyhow - I'm not saying I support the NRA or whatever - I'm just saying I think there will be a really weird social dynamic.
But you are right - drivers are pretty crazy and would fight pretty hard!
I understand hobbies and sometimes work (professional hunter, race car driver, whatever) and am all for people getting to do whatever they want to do (to themselves), but I think it gets interesting when put in a society, where not everyone can do whatever they want, and we need to come to a consensus (where consensus means the majority gets as close to what they want as possible, but a lot of people don't get what they want).
If human drivers in public means me and my family/people I care about/random people are at risk, and a safe option - driverless cars - exists and serves the same function, then I think human drivers should be banned (again, in public), and I'd vote that way. If there's a way for human drivers to keep off main roads and still get their jollies and whatever necessity they want out of it, then great.
Sometimes we trade things for safety. Often it's not worth it, but sometimes it is. In government surveillance I don't think it's worth it, even if the safety benefit would be measurable (which it's not). In both cases of gun regulation and future human driver regulation I think that trade is worth it, as long as the benefit is quantifiable (which may or may not be the case, I'm merely arguing on the premise that it is).
If you want to drive, you should be forced to do it on a closed course. Driving on public roads is NOT for your fun and enjoyment, it is for SAFE TRANSPORTATION. Your enjoyment of the activity comes second to everyone else's safety.
It WILL be illegal for you to drive on public roads some day.
Fortunately, I won't have to live there :)
As for me, I have no illusions that those tyrannies exercised for the good of their victims will ever eventually die out; self-righteousness provides a powerful high, and to get their fix those addicts will need someone (or something) to scold.
Humans do immensely dangerous things for fun all the time[1]. That isn't going to stop. People will continue to drive on private roads pretty much forever.
[1] For example https://www.youtube.com/watch?v=-C_jPcUkVrM
I think the reality here is that we don't have the fuzzy logic and AI to handle snow covering up lane markers or heavy rain and other conditions humans handle with relative ease. I also think, politically, its just going to be a tough sell and any edge case crash that leads to fatalities will quickly change people's minds about the future of autonomous driving.
>they don't even have seat-belts - and we 'gasp' ... 'OMG'
Yet seatbelts were mandatory in 1959 and the usage rate in the 1980s was around 50%, which led to mandatory usage legislation in the late 80's and 90's. So the "20 year" metric of everything changing is questionable as half the population ignored the seatbelts in their cars 20+ years later. Its only via expensive ticketing and enforcement that we convinced most people to even wear them, when their safety benefits were obvious!
I think this overly-rosy view of the future gets ad impressions for the tech press and gets readers excited, but the technical issues with night, snow, rain, etc visibility aren't solved. I imagine until we get better AI, they won't be.
Driving just has way too many difficult edge cases that lead to fatal outcomes and because of that, its not going to be something quickly replaced by machines. Augmentation and assistance for human drivers is the most likely outcome here.
Our roads and infrastructure evolved to accommodate cars, but that was at a time when governments were much better at encouraging and adapting to new technology. I wonder if governments will do anything to incorporate self driving technology into infrastructure.
True, but improved AI isn't the only way to address such problems. For instance, thermal cameras can see through light snow, so those may become standard. Also, something like RFID embedded an inch or two below the road surface could serve as marker enhancers. If all else fails, GPS and highly accurate maps can stand in for road markers in many areas. Besides, snow won't be as much of an issue if self-driving snowplows get started shortly after snowfall begins.
Cyclist are no angels, either. I always find it fascinating how they clamor for at least a meter space next to them when cars are passing, but when traffic is slow (not standing still – that's a different story), mere centimeters to my side mirrors are enough to pass cars themselves.
I think a drastic re-education of traffic participants would be necessary: de-emphasize small, inadvertant and temporary infractions and punish vindictive and deliberately dangerous behaviour much more aggressively.
If a car traveling 40 mph or faster hits a pedestrian or cyclist, the person hit will almost certainly die.
The two are really not comparable.
If we replaced all the drivers on the road with still-fallible but conscientious/polite human drivers who always followed traffic laws, always drove the speed limit, were effective at looking for and stopping for obstructions in the road (e.g. pedestrians crossing the street), followed at a safe distance, and adjusted their speed and following distance to match road conditions, the death rate from traffic collisions would drop dramatically.
If we replaced all the cyclists with perfect riders, the death rate from traffic collisions would change only marginally.
A cyclist that passes a slow car with, say, 30 km/h and brushes into the car may very well die.
It's exactly symmetrical to the situation of a car passing a slow cyclist with 30 km/h.
Furthermore, dying isn't everything that should be avoided. Injuries and material damage are also bad.
Further-furthermore, aggressive and careless cyclists can also kill a pedestrian. Not as easily as a car can, but still. As a pedestrian my number one danger are cyclists. Cars and pedestrians are quite predictable in their movements. Cars especially don't sneak up on you without any noise and then pass you on a random side, while you're walking around.
And on top of that - their brakes are useless, they in many cases drive as fast as they can, downhill or not. Yeah, for me as a driver, cyclists are the biggest threat in the cities and outside (effectively anywhere but highways). As a pedestrian, cyclists are again the biggest thread in the towns mostly on pedestrian crossings.
As a pedestrian stopping and going again is effortless. As a car driver the same. Only cyclists dread stopping.
You need very little kinetic energy to make a cyclist topple and fall. The fall itself can lead to injury and death, lying on a street in the midst of all the traffic is even more dangerous.
Any touching of car and cyclist is very, very dangerous. That's why I brought up the example of cyclists passing cars with practically no space between them. They feel it's safe, but it's incredibly dangerous.
Lying on a street - yes. Because cars might roll over you, cause of the kinetic energy.
Getting squished between cars - yes. Because of the huge mass of the car.
Banging your head against an opening car door - yes. This time it is because of your own kinetic energy that hits a small surface area which is immobile cause of the mass of the car.
So it is quite dangerous to get near to cars, yes. But almost always because of the huge mass of the cars, especially if they are moving. That was the point of ggp.
He dies – as you put correctly – because of all the cars around him. Of course kinetic energy is relevant there, but it doesn't matter to the argument, because the claim was that being touched by a car is much more dangerous than being touched by a cyclist. And that's wrong.
As soon as the cyclist falls he's in mortal danger.
A 1 ton car at 10 mph has 9 000 joules of kinetic energy, so about 30% more.
Many of these people where not going all that fast. https://en.wikipedia.org/wiki/List_of_cyclists_with_a_cyclin...
I have experienced the former.
Do you know how many careless car drivers I meet? At least one each time I bike more than 3km. Not crazy (those I find every 100km), just careless. But thats enough to threaten my life.
Either way, the cyclist is on the receiving end of the risk (we're not talking about bike vs pedestrians here). So no Symmetry. Mass matters. A lot.
Yes, it is. At 10km/h it's really not dangerous; at 30km/h an open door may kill the cyclist, and a hitting pedestrian may be severely hurt them. Speed is that a relevant factor.
Annual pedestrian deaths from collisions with motor vehicles in the US: a little under 5000.
Do you seriously believe 5000 pedestrians a year are being killed by bicycles? I would be surprised if there were as many as 500 pedestrian fatalities a year that had anything to do with bicycles.
Cyclists are not your number one danger. They are nowhere near to being your number one danger.
No casual cyclist can sustain 30kph on a level road. My average speed when I was in top shape and biking 20km every day was 15kph. I don't think you understand just how fucking terrifying 30kph is when you're on a bicycle. I don't want to ride at that speed, it's really way faster than I can react, and I feel it very well. When you're in a car, 50kph feels like nothing at all, but on a bicycle you're acutely aware how you're propelling your body sitting on a few thin tubes of metal and two flimsy wheels at a speed at which you have no control over. I've reached 50kph only once in my life on a downhill stretch of good asphalt road on a Sunday with no traffic as far as the eye could see and I don't ever want to do it again. 30kph is scary. 50kph is terrifying.
Right, this is because a car that's moving faster than a cyclist and traveling close can drag the cyclist off their bike. A cyclist that is moving faster than a car cannot drag a driver out of their car.
But if the cyclist touches the car, he can be dragged off. No matter who was passing the other and who was faster. Only the contact matters.
If I'm filtering through slow traffic at 10mph and clip the side of a car, I'll have a minor tumble - scuffed handlebars, grazed palms, dented pride.
If a motorist clips me at 40mph, I'm almost certain to suffer serious injuries. I might be hit on the back of the head by a wing mirror, I might be dragged under the wheels, I might be launched over the hood, I might be bounced into some street furniture.
No, this is completely backwards. Small, inadvertant and temporary infractions kill people just as dead as vindictive and deliberately dangerous behaviour. And whether it's the driver's fault or the cyclist's fault, either way it's the cyclist who dies. What we need to do is start holding drivers responsible for deaths that are their fault, "accident" or not. If you don't have the skill or concentration to avoid causing a collision, you shouldn't be controling a ton of metal in a public space.
I'll PM Elon immediately.
Meanwhile, in the United States alone, there are around 5 million vehicle crashes annually.[2] Although automobile safety has increased, I am curious as to whether our accidents per-capita has stayed the same.
Also, we have to consider the interaction effects between decisions taken by human drivers and self-driving algorithms. I have a feeling that they might be deleterious initially, but should be able to improve after some iteration.
[1] https://asirt.org/initiatives/informing-road-users/road-safe...
[2] pdf: https://crashstats.nhtsa.dot.gov/Api/Public/ViewPublication/...
I've been in a few situations where I had to work very hard to not kill/injure someone because a motor cyclist was being completely reckless. It is super annoying to have motorcyclist squeeze around you on both sides on a one lane ramp merging info traffic. Or a motorcycle sitting in my blind spot without even the decency to stay near the center of the lane.
Its not really different from any other driver, except the complete absurdity of driving among cars at car speed with literally no safety mechanism beside plastic wrapping your head should demand a heightened level of defensive driving.
There are terrible people in control of all vehicles.
BTW, there are good reasons why motorcyclists don't always drive in the center of the lane. One main one is to position the bike so you are visible in the left mirror of a driver you are passing on the left, meaning you ride in the 'right track' of the lane. Another is oil and gas tend to be spilt in the center of the road, and most bikes don't have ABS. Riding in the left track will help prevent a driver passing you on the left from cutting their pass short into your braking zone. It all depends on the surroundings and riding defensively.
The things that frustrate me most during driving are equally present when running on a trail, waiting for a subway train or observing cyclists, for instance. They are:
1. Not knowing where the heck you are and what you are doing. Many people seem utterly unaware of themselves and the effect that their actions have on others.
2. Refusing to either speed up or move out of the way (doesn’t really matter which of those you do but if you refuse to do both then you will become a huge obstacle, whether it’s a sidewalk or a road).
3. Believing you are better at multitasking than you really are.
If you’re blocking a sidewalk, you’re exhibiting poor “driving” in a walking scenario. If you’re blocking the entrance to the doors, you’re exhibiting poor “driving” in a train station. If you’re prioritizing your texting and interfering with me, you’re exhibiting poor “driving” in just about any situation (driving or walking or social for that matter).
Automation is not required to fix the biggest frustrations with driving: rather, human beings need to start paying a lot more attention EVERYWHERE in life and being a lot more reasonable to one another.
Now, "unifying theories" tend to be rather simplistic, and the parent did exactly publish a replicable study, but you didn't really rebut the thrust of the argument. Instead, you claimed the parent is the problem for expressing the argument in the first place. Even if the behaviors are not comparable in final effect (and I agree with you, they aren't), they are comparable in cause. On the face of it, there do appear to be strong, outward similarities in the frustrating obliviousness (or willful apathy) of people who inconvenience others while walking and people who inconvenience others while driving. One might result in death and the other might not, but that's really just a matter of nuance. The behaviors are fully comparable, it just happens that one takes place in an environment which is fundamentally more lethal.
As an aside, I find my reactions are comparable as well. One of the most bloody frustrating things I frequently encounter is someone who suddenly stops their shopping cart in such a way that it blocks traffic, then walks away from it to roam for produce.
That could, of course, be an incorrect interpretation of what was being said; I certainly agree that the immediate emotional response tends to more negatively weight things that are annoying and slowing me down more than things that are actually dangerous, which is part of why I think it's so important to say something about this sort of thing, and to make sure you change that weight.
Guns get a lot of press and a lot of emotion; really, a lot of emotion that I think would be better focused against cars. If there's not a gun in your house, you are dramatically more likely to be killed in an auto accident than by a gun. (or, if you want to put it another way, as an American, if you remove suicides, you are a lot more likely to be killed by a car than by a gun.)
It's not much different from that when a human is driving. The conditions do dictate some limits to how the driver can wander about (you can only mostly go straight on a highway) but everyone will stop to think something at times and maybe happens to slow down or slowly steer left/right until they refocus on driving again.
However, on the road, this behaviour is very irritating for those who are in their from-A-to-B mode because, unlike in a busy railway station where they can just snake through the crowd, with the worst case of gently bumping into bypassers while at that, they can not easily go around the slower drivers in busy traffic. Driving is basically monotonic robot work, or should be. It's not a natural way of being that lends to distracted behaviour as walking about does.
Having traveled over the world , I tend to chalk most human behavior up to cultural or personal quirks these days. If an issue is cultural in nature, the culture can be changed.
My point here isn't to start an argument (clearly the definition of "sucking at driving" is relative), but just pointing out that I'm fascinated that our roads work as well as they do, all things considered.
I guess a million deaths is also an acceptable number a year for you?
This is part of the problem. In some countries, they aim for 0 fatalities, but in most we accept the number of deaths we have now.
Not only do they suck at driving, but they're also reckless. You'll find them doing all sorts of crazy things while driving:
- Search for something in the glove box.
- Do their make up.
- Read novels.
- Play Pokemon Go.
I'm not making up any of the above. The last one has even made the headlines in Japan recently because of 3 deadly accidents caused by drivers playing Pokemon Go, and because of video of a tourist bus driver playing Pokemon Go on the job.
Ballistics is so complicated and yet, our brains and and muscles do it all naturally.
On the other hand, I'm sure a properly configured and programmed robot will throw the paper ball into the bin more reliably than me.
On the contrary it seems humans are very good at adjusting to varied and unpredictable traffic and road conditions without issue everyday.
I think ai proponents have got into the bad habit of underplaying the skills, instincts and adjustments involved or exaggerrating accidents out of context of total traffic. There is no data on how AI cars would do at this volume and level of varied conditions so the argument is dishonest.
An ai vehicle operating in constrained road and traffic conditions is a completely different propositon to the traffic patterns negotiated by human drivers everyday.
It is interesting to note that at least one of these laws seems to be returning. Reducing residential speed limits to 30 km/h is a trend now and is done to increase pedestrian safty. Driverless cars will of course also be so restricted. If it turns out that driverless cars are actually safer for pedestrians then pedestrians will return to the streets. Then "The motor-car is not to be made as useful as it should be.." will be an issue again.
To the tiny residential streets, maybe, but they never really left those. I walk my dog on the street in my neighborhood (no sidewalks). The bane of our walks are the stealthy electric things. With those around you are much safer at night when you can see their lights around corners. We will never see pedestrians taking back the main roads. The separation of pedestrians and cars is more efficient for both. Our cities would grind to a halt if every truck and ambulance was suddenly stuck to a walking pace due to pedestrians out for a stroll.
Motorcycles and bicycles also aren't going anywhere, not for decades. Telling those groups to give roads back to pedestrians, to put away their favorite two-wheeled toys and stick to a walking pace ... Bring a helmet when you tell them it's all over.
To the contrary, I haven't heard anyone complain about collision-avoidance systems; the paranoia comes in when the possibility of cars being driven by a single (artificial) intelligence is discussed.
There is no comparison, a horse is much smarter then any AI i seen [0].
Look at the traffic closely and observe all the edge cases and all the details that an AI couldn't handle properly. Like potholes, paint, animals and people coming from beyond cars parked on the street, works on the road, a dust cloud, a gravel road with big puddles, and idk what else. On top of that the thing is that going over 40kmh is not driving, it's ballistics.
So far i have seen that a car can be driven by an AI from A to B in good conditions, but i have not seen it drive in more difficult conditions (some are common in my city).
BUT - cars will eventually be better than horses.
Why?
Because they have a) the tech is very specifically geared to do things related to driving and driving only b) it's not inclined to 'emotions' or 'spooking' and finally c) it has tech the horse cannot have - it will be able to see further out, communicate with other cars 'real time', it will be able to make crazy probabilistic calculations in real-time with respect to 'what best to do in a negative scenario'.
I think it will take a long, long time to get through those corner cases, and probably a few disasters between now and then, but in 20 years maybe it will be crazy safe.
I think the concerns of these 19th century legislators were warranted and we made a massive mistake in promoting cars as the default form of transportation for everyone. We've unnecessarily exposed ourselves to an incredible amount of danger when there has always been other safer transportation solutions possible.
It's not too late to reign things in. Autonomous cars will help, but solutions exist today and there's no reason to not start working on making the streets safer right now. Cities need to limit speeds to 25 mph like NYC has so that crashes, when they do occur, are less likely to be fatal. Cities also need to provide more transportation alternatives such as public transit and protected bike lanes so that city staff can create smaller and safer roads without affecting people's ability to travel.
I've been in two wrecks. The last one left me with several broken bones and the inability to walk for a few months, although I did gain the amazing ability to set off metal detectors everywhere!
Human drivers are my worst fear.
Instead if you filter and split lanes and get to the front of the queue of traffic you're not only safer but also out of the way of that wave of traffic, since you're on two wheels and can accelerate away much faster than the rest of the traffic. This then allows you to have a buffer behind you.
To ensure computer drivers behave better than human drivers, re-legalize walking in the street. Also, ban human drivers.
[1] http://nomano.shiwaza.com/tnoma/blog/archives/20060601/m/dsc...
Though this does have the unfortunate consequence of lagging behind which can sometimes be severe.
Trains are engineered to absurdly high safety standards and every time there is a crash, it's widely publicized and an in-depth investigation is required along with suggestions for systemic changes to prevent repeat events, and the companies are required to pay large settlements:
http://www.npr.org/sections/thetwo-way/2016/10/27/499592760/...
At the same time, trains are vastly safer than personal cars. Yet when a driver kills someone, police can barely be bothered to attend the scene and rush to excuse the driver and let traffic flow again.
The safety standards on vehicles are such that billions are spent on engine development yet we tell people to watch out for the "blind spot" in that trucks mirror contraption. We're soon going to have autonomous driving AI yet trucks are roaming city streets that are unsafe by design and regularly claiming lifes. I don't see Volvo and MAN settling with anyone over that.
Most likely because trains essentially only travel in one dimension almost all the time, along a very well-defined path.
In car crashes, each case could be different and difficult to figure out what was the exact cause. Cars are becoming more complicated, as with farm equipment vs farmers or the "check engine light". Without transparency into the internals, I wouldn't know what broke or was it my fault or the car and I would probably have to rely on the honesty of the company to tell me or my family that.
For the same reason generally passengers don't sue airline companies when something goes wrong.
There is an unbelievable amount of energy and money that goes into safety ... so much so that flying would probably be 1/4 the cost if they simply didn't care that much about it, and so there's a degree of 'public goodwill' in the whole thing.
I think the 'moral' issues will come down to scenarios wherein the computer has to make a decision like: do I swerve and kill my 2 passengers, or hit the car in front of me head-on and kill it's 2 passengers: pick one of two outcomes. I think this is the moral dilemma that is already coming up.