The High-Stakes Race to Rid the World of Human Drivers

I don't want to live in a society where every car becomes a tool of law enforcement.

I think you underestimate how crazy and obnoxious people can be. Take San Francisco, as an example. There are people who are poor and disenfranchised, and mentally ill, who feel that it gives them some power to saunter out into traffic and make everybody stop. With an autonomous car, that deranged person can just sit down in front of the car, bringing everything to a halt. Take New York as another example. For years I only walked in NYC and I figured that all drivers were maniacs. Then I drove though Manhattan once and I realized the actual dynamic. If a driver doesn't put on a credible appearance of being willing to run down pedestrians, the pedestrians will never stop crossing the street. How does an autonomous car fit into that?

I think these cars are going to work brilliantly in very polite and orderly places like Zurich, where everybody takes their turn and neither pedestrians nor bicyclists nor drivers feel put upon by the system. I have no idea how they are going to work here in America where we dump our mentally ill on the streets, and then don't give them sidewalks and crosswalks or any sense of dignity. I have a feeling it's going to work quite poorly.

> Cars should yield to pedestrians and cyclists

The problem is if they don't and pedestrian gets killed there is not way to resurect the pedestrian, and start over. Therefore, pedestrian's goals is staying alive regardless of what laws are written in the books.

To put it another way, most countries out there have laws about "cars yield to pedestrians on crossings" but in a lot of places I wouldn't want to prove a point or rely on the drivers to obey the law vs just staying alive and in one piece.

> autonomous police cars who bringing reliable traffic enforcement to all road users.

Hmm. Now this is interesting. We are almost there -- red light cameras, speed camers, automated toll booths, cars are always recording and logging their speed, acceleration and hundres of other parameters, onstar system. Now we just need an incentive such as an insurance discount to have cars be remotely disabled and pulled over of they speed too much where it becomes unsafe, or if they run red light.

At first you'd think nobody would buy it, but maybe if car insurance goes from $500 a quarter to $150 there would be some takers. Then before you know it, it would become illegal to drive cars that don't have that feature.

You can't have controlled traffic flows with constant jaywalking. And traffic flows will be increasingly important in order to minimize power consumption by automobiles and keep our growing cities liveable.

I don't see why the cost would be any higher than it is today. If anything, driver-less cars would have lower insurance premiums. So if you can afford/justify your insurance today, you could still afford it in the future.

I could even see all insurance getting cheaper, since the risk of accidents for all cars should decrease as the mix of cars gets more and more automated.

>That is, the cost of insurance for human drivers will be exceedingly high.

right now, in california, you are only required to carry on the order of $35,000 in liability insurance.

Now, personally, I think this is completely crazy, that it defeats the whole point of liability laws, etc, etc, but the point is that unless you change that, even if the rates go way up for human drivers, if the insurance companies can continue to sell plans that are limited to only $35K in liability, people who don't fear bankruptcy will still very likely be able to buy that legally-minimal insurance for not all that much money.

> then what?

Whatever we want.

Peace and prosperity, I hope.

I imagine the same argument was made about the industrial revolution.

There is still a lot to work for once we remove the more mundane parts of our existence.

Then we distribute the free time to the whole population, accounting for the vast majority who never had the means to accumulate wealth

or

I don't know

Jaywalking is a crime for a reason, because it's potentially deadly to the pedestrian. It seems to be much worse in small/mid-size cities (probably because of less overall traffic). I've had people literally jog into my lane then walk the remaining 4 feet to the sidewalk while I slow from 35 to 10 to avoid hitting them. I think if half the cars on the road are autonomous and will just stop no matter what you will not see the decrease in pedestrian injuries and deaths some are expecting.

>> Pedestrians and cyclists may also learn that they can freely cut off an AI car at anytime they please. Why wait for a walk signal? Why walk all the way to the intersection? Just saunter across anywhere you like, so long as you cross in front of an AI car.

>That's probably a good thing. Car drivers need to get out of the mindset that they own the roads.

The problem with that is the time period before AI cars are ubiquitous. In my area I've already seen pedestrians act far more carelessly now that the pedestrian right of way laws have been enhanced and there's been an uptick in pedestrian fatalities. Sometimes giving people the impression that something is safe makes them behave so much more recklessly than they would have originally.

You do realize there are countries that have a significant chance of snow for 4+ months per year, right?

And that there are many places with stop signs that are almost completely obscured by foliage, or simply by the angle of roads on the crossroads.

Not everything is California. And not every city has a grid.

When autonomous cars can survive in a place like Rome or Paris[1], or even downtown Manhattan. Then we can talk about fully replacing humans.

[1] Rome and Paris are notorious for people having special city cars because they get banged up so much in the course of normal life

Here's an example of traffic in a modern European city: https://www.youtube.com/watch?v=HfJ-VokP3EQ

And another: https://www.youtube.com/watch?v=lay8aZlsbB0

How long until autonomous cars can do that? Hell, I'd shit my pants trying to do that.

PS: in the real world there are also many roads that flat out don't have lanes drawn on them.

There are just too many edge cases. I believe, that we'll get some kind of hybrid solution: computer drives on the highway during daytime otherwise drivers takes over.

In Switzerland we have those wonderful pass roads. On one side you have the mountain, on the other side there's nothing. Never in my life would I let a computer drive down that road with my family in it:

https://www.youtube.com/watch?v=HDilGBIiKxI

Sometimes roads are so narrow, that drivers naturally honk while going round corners. If there's a honk back you gotta slow down. Now add some snow fall and fog and drive down in the night. Good luck!

Google's self driving cars only drive on roads that have been comprehensively mapped by google at great expense. You couldn't just take it to a new country and expect it to work.

So yeah, there's a very long way to go until self-driving cars are ready to replace human drivers.

What's more, people are forgetting that normal traffic does not go through wilderness, and roads are not just slabs of concrete in a line. We draw markings on the road. We put signs next to them. It stands to reason that as self-driving cars will be slowly deployed (I don't see them rolling out simultaneously in sunny California and in snowy Switzerland), we'll be updating road infrastructure with signs and other objects designed for AI drivers - ones that could help them determine road conditions better.

I personally found that lane departure warning made me a much more precise driver. I like that a lot. It goes off in error sometimes too, but it's rare now that I've been trained to properly stay within the lines.

This is my first car with the ultrasonic parking sensors; they're nice to have for sure. I won't know the corners of the car as well, but, frankly, today's cars are getting harder than ever to see out of anyway, so likely I never would have been as comfortable in a new car.

Also, part of "knowing the corners" of your car implies at least some gentle contact at some point in time, which it's best to avoid entirely.

One of my favorite features so far is rear assist, which is able to detect oncoming cars as I'm backing out of a parking spot. The bigger the car the more useful this is. It doesn't tell you when you're 6" from a vehicle, it tells you when there's a vehicle traveling at speed further away that you're about to back towards.

Typically the car is not following the car in front, merely regulating speed to avoid ramming it. In my car, at least, if the car in front changes lanes or exits, then the cruise control system speeds back up to whatever speed I set it to.

This can make for some exciting speed differentials if, for example, everybody is moving 20MPH and they open up a big gap by changing lanes, and my set speed is 70MPH. But it doesn't accelerate aggressively, and this is a rare occasion anyhow.

I'm not sure if I've used it in the rain yet, but my understanding is that, yes, the sensor needs to be kept clean. It follows the SPEED of the car in front of me; I still steer. I assume it uses steering wheel sensors to detect that it should be following the car slightly to the left of me when, say, I'm going around a left-hand curve.

The lane keeping system uses a camera to detect lane markings. I'm not sure how it decides whether or not to follow exits, but I know from the Tesla autopilot descriptions that it's not perfect.

That depends on the sensors which are used. Most current ACC systems use radar, which should not be affected by weather but has other problems like high cost, narrow field of view and you only get speed and distance as data points which are not always easy to interpret correctly. Ideally it's combined with a camera to get more information about the environment like land markings, so that you know if the car in front of you is actually in your lane or just going around the corner in the lane next to you. The camera is obviously prone to glare, raindrops, etc.. When the car ahead changes lanes or exits the freeway, the system will try to find another car as a target vehicle and adapt the speed accordingly to keep a certain timegap, if it can't find a target it will behave like normal cruise control.

Didn't you hear? You won't be permitted to purchase an autonomous vehicle unless you can provide documentary evidence of your having been in favour of the technology even before it was proven.

On top of that, the cars can communicate more and faster than the drivers, and can be programmed to collaborate with the other vehicles. So you can have stuff like the cars in a packed highway opening up a diagonal for a car in the left lane to take an exit on the right.

My offhand anecdata suggests that the bulk of traffic jams are caused by people driving older cars. There is a little-understood cosmic force that causes cars to stop running in proportion to how critical the section of roadway is to regional transportation and in proportion to the age of the car. 1985 Nissan Sentra in a tunnel? Guaranteed to burst into flames.

More seriously, there are obvious reasons why older cars are more likely to just stop running, or to be involved in a fatal accident that snarls traffic. I think the tipping point is not going to come from the deployment of self-driving cars, but will arrive instead when the fleet of junkers nears eradication.

No, if you rear-end someone driving safely, you are at fault and reckless. This I can control the situation thinking are the famous last words of every idiot ever wrecking his car, or, sadly, killing others.

The truth is that people aren't all "idiots who should never, ever have been allowed behind the wheel". Driving is hard - ironically, harder than most drivers realize. It's been proven time and again that we absolutely suck at it. And the "normal driving standards and patterns" is a big part of the problem - our natural tendency to ignore the abstract rules and go with our gut is what kills people. It's hard to really blame a person for doing bad at a task that's not well-suited for their brain architecture. But it means we should leave driving to machines that are better at it, the very moment such machines become available.

Hrmm, I did find exactly one instance of the phrase on Google. I'm surprised, too.

As it was explained to me, a Chinese Soda Fallacy is a business plan wherein you stand to make a dollar selling a Coke to everybody in China. You will make a billion dollars! Except that the market-clearing price for Coke in China isn't a dollar, and not everybody wants one, and so on and so forth.

The idea that you will make a dime every time anyone drives a mile strikes me as perfectly congruent with that fallacy.

Edit: This article seems to address the topic without using the phrase.

http://www.wsj.com/articles/SB100014240529702048303045741335...

A COKE FOR EVERY KID IN CHINA This gambit rests its case on a plethora of secondary data to show how large and fast-growing a market is. The plan then makes a heroic leap and assumes that the new venture will grab X percent of that market it could be 1%, 10%, 30% or whatever. "Surely," the plan argues, "with the large number of customers in our market, we'll easily get enough. We only need a small fraction to have a very nice business."

Double clutching makes downshifting at high revs much smoother, and reduces wear&tear on the clutch & transmission. (I have a pretty grabby clutch, yes, it's aftermarket.)