Thoughts on Our Future With Self-Driving Cars

ztnedman1 wrote: You need evidence that thousands of cars entering a finite area will still cause traffic? Liquid by definition has perfect volume utilization, and does it autonomously! Keep pouring water into a finite area faster than it can release it and you got a problem.

Please read my statement again, maybe a little more slowly this time. I said I'm looking for evidence that there would be no appreciable improvement with self-driving cars, not that there would be no problem at all.

What you're talking about assumes there are enough cars on the road to all physically be locked together, bumper to bumper. The article I reference above indicates that this is NOT actually the primary problem. And you cant just look at one one ramp to solve the problem. You have to look at the previous and subsequent on AND off ramps. Cars get on. Cars get off. There would necessarily be faster and slower flows, but the total flow rate could be maximized by altering car density and speed based on variables that could be known by the cars communicating with each other in advance.

Lets say the car immediatly in front of you exits, but your car knows another will be merging at the subsequent on ramp, so it maintains the gap for that car to time it's acceleration up the on ramp to enter the flow of traffic into. No speeding up, followed by no slowing down. Traffic could also be diverted actively to minimize overcrowding on the few roads that people don't know any better than to do so.

Or maybe that your car needs to move over two lanes to exit. The immediately surrounding cars could compress the spaces between themselves temporarily, to open a large enough gap to let the exiting car through, and then redistribute the space back once the existing car has passed through. But as long as they're not physically touching, that would only have to affect a few cars forward and rearward, since they're all communicating with each other to know it's a temporary (and thus localized) disruption. Meanwhile human drivers would cause the entire line of cars behind to slow down (more and more the further back you look) in order to open the same 'temporary' gap.

Yes it gets far more complicated as the number of different possibilities are expanded upon, but that's where the advantage of computers talking to each other grows exponentially over human drivers reacting to each other based on no communication beyond a few lights, a horn, and a finger.

Computers can analyze so many possible scenarios that they can come up with solutions contrary to even the best human minds, that end up working better. Consider the computer that beat one of the best human players of the game Go, at one point making a move during it's path to victory that was so incomprehensible to the human player that he had to leave the room for 15 minutes to regain his composure.

Driven5 wrote: . In much the same way, one side will argue their freedom to drive their own car, while the other side will argue for their freedom to not get killed by somebody driving their own car...It's not a question of simply losing freedoms, but rather is a question of which freedom is more important than which other freedom to society as a whole?

Freedom is an illusion created by laws. The fact that laws exist is proof that freedom doesn't.

As far as this example I quoted though, that goes back to actions having consequences. You made the choice to get onto a roadway, surrounded by 2 ton objects traveling at great speed. You're choosing to put your "freedom to live" in everyone else's hands at that point, regardless of who is in control of what.

Having so much trust and faith in machinery, or more specifically the programmers and manufacturers, it's almost deep enough for (redacted). To think a program and tool made by a man will have no flaws, no accidents, and not randomly break down just because the bag of meat isn't in control is as foolish as thinking the world is resting on the back of a turtle.

If you want to remove the "human element", design an AI that can learn, and let that create and design everything with magical unbreakable materials, otherwise, you're at the whim of designers.

Plus, already seeing government agencies trying to turn autonomous cars into weapons should be enough to tell the normal person this is a terrible idea all around, but that's another variety of off topic.

In reply to Chris_V:

People hate the real numbers. Much like the firearms debate, the actual numbers just don't play out for the amount of energy put forth. But numbers and facts have never gotten in the way of a good agenda before, so I see no reason for it to now.

According to this... there were 262 million registered cars, 13 million accidents, 1.7 million accidents with injury and 32,000 deaths. So, the starting numbers we have been quoting were already overly inflated... and the deaths per mile is insanely low. Don't forget this also includes pdestrians and bicyclists that will still pull out in front of 2-ton moving vehicles (5300 in 2015, dropping the overall down to 27,000) and another 1,000 bicyclists. So no we are down to 26000 motorists, of 218 million LICENSED drivers.

Chris_V wrote: 0.0001% of the population (or less) are involved in fatalities.

I think you need to multiply by 100x to get to a percentage on that fatalities.

Driven5 wrote:

Chris_V wrote: 0.0001% of the population (or less) are involved in fatalities.

I think you need to multiply by 100x to get to a percentage on that fatalities.

No... I don't think so. It works out to .00119% by the actual numbers. I know, I know... it doesn't fit the agenda. But I'm not going to cook the books so you can feel better about your argument.

EDIT: that is also the actual percentage for registered drivers..... go with total population like Chris is talking and that number drops to .00076%

also, deaths per mile... that gets funnier. US DOE shows the total miles driven per year in the US at over 3 trillion. 26000 deaths.... 3 trillion miles. I'd say we're doing pretty well.

You have a knack at taking a what could be a reasonable argument and just turning a phrase or two in order to make it sound really unreasonable.

Bobzilla wrote: In reply to Chris_V: People hate the real numbers. Much like the firearms debate, the actual numbers just don't play out for the amount of energy put forth. But numbers and facts have never gotten in the way of a good agenda before, so I see no reason for it to now. According to this... there were 262 million registered cars, 13 million accidents, 1.7 million accidents with injury and 32,000 deaths. So, the starting numbers we have been quoting were already overly inflated... and the deaths per mile is insanely low. Don't forget this also includes pdestrians and bicyclists that will still pull out in front of 2-ton moving vehicles (5300 in 2015, dropping the overall down to 27,000) and another 1,000 bicyclists. So no we are down to 26000 motorists, of 218 million LICENSED drivers.

Bobzilla wrote: No... I don't think so. It works out to .00119% by the actual numbers. I know, I know... it doesn't fit the agenda. But I'm not going to cook the books so you can feel better about your argument.

Cook the books = basic math?

Chris was off by 100x, and you're still off by 10x.

32,000/262,000,000 = .000122
.000122 * 100 = .0122%

And while it's lying with statistics to only include motorist fatalities as being affected by this technology (many pedestrians/bikers are killed while by driver error while having the right of way, they don't all just illegally walkiin front of cars that have the right of way and can't stop in time), and even using a motorist statistic that include more than just licensed drivers as being potential fatalities but then dividing by the number of registered cars or licensed drivers, I also don't think it changes the percentage much. There would be reductions across the board. So it would most closely be approximated by total road fatalities divided by total number of Americans...Which is still (32k/320M*100) = .01%...Not .001% or .0001%

Which is still pretty good for framing it in a positive light, as a form of pushing a contrarian "agenda".

A few things:

-Someone listed crush zones and seat belts as nannies. I don't see it that way. A nanny actively intervenes in your behavior. Those two things are passive.

-I haven't seen it noted that accidents are in fact up - ending a very long run of safer and safer accident/mile ratios. I believe that's primarily a function of distracted driving, which suggests that a major part of driving automation is to make the world safer for idiots.

-It'll be a long time before I'll trust the things. It's one thing to navigate a pleasant, well-ordered freeway or suburb. But big cities are incredibly complex. Can you imagine a Tesla trying to wind its way through some Asian cities? If electronics could have a nervous breakdown, Ho Chi Minh City would do it.

-Some accidents will still happen. I'd much rather have a shot at saving my own bacon than leave it to a computer.

Put me in the not interested, no way, no how, never ever camp. I have no use for a self driving car.

Something for the pro side to consider: legal and moral implications when said self driving car must make the inevitable "kill or die" decision. Obviously these miracles of modern science will be designed to avoid accidents etc. what happens when it encounters another vehicle driven by a meat bag that puts it into a Kobyashi Maru? Where either the occupants it is carrying will die or the other vehicle/pedestrian/cyclist will die.

First off, do we as a society want to open that Pandora's box? Allowing a computer to decide which is better? If that is ok then is it ok to allow hospitals to decide that a patient isn't going to make it so stop treatment?

Even better, how will the courts decide fault? What if an autonomous car, in the act of avoiding an accident causes another?

The only way that these could be prevented would be if all non autonomous vehicles were immediately removed from all roadways and nothing but autonomous vehicles are allowed. Because I damn well guarantee that no matter how idiot proof we try and make these, there will arrive before us an idiot of unbelievable idiotness.

And on a last note...... meatbaga aka pilots have been out flying their replacements for a long time. Doing things that do not make sense but in the end is the right thing to do. Sully is one of those that did the unthinkable and did it without thinking. We don't have an AI yet that can free think like a human with tears of experience. That day may come. But until then, I'll take my chances.

You guys are looking at single year numbers...

The odds of a car occupant dying in a transportation accident were 1 in 47,718 in 2013; the lifetime odds were 1 in 606 for a person born in 2013.

http://www.iii.org/fact-statistic/mortality-risk

Chances of dying in any type of motor vehicle accident during your lifetime if born in 2013: 1 in 113 Those are pretty berkeleying high odds. It is the #1 cause of death by accidental injury.

Obviously this is apples to oranges but... commercial airline flights are near-autonomous. Automotive fatality rates are 750 times higher per mile traveled. Should we not be concerned about safety on planes anymore?

The number of deaths per passenger-mile on commercial airlines in the United States between 2000 and 2010 was about 0.2 deaths per 10 billion passenger-miles, [23] [24] while for driving, the rate was 1.5 per 100 million vehicle-miles for 2000, which is 150 deaths per 10 billion miles for comparison with the air travel rate.

https://en.wikipedia.org/wiki/Transportation_safety_in_the_United_States

The loss of individual autonomy is the elephant in the room. I don't think it's possible to fully account for the psychological and societal side effects that would come from removing the most tangible aspect of personal autonomy that most people in this country have. Spend enough time examining the neurotic thought processes of people who 'cant afford a car' and you will start to understand how effectively the perceived lack of choice can predict the choices you actually end up making. America as a society is more or less based on the concept of placating enough of the population with a perceived quality of life that they will allow and even contribute to the absolute debasement of anyone outside their monkeysphere. In my mind, there's a very good chance that removing the illusion of autonomy and freedom that comes from the daily subconscious gratification of a powerful machine responding to your whims will render our fractious, classist, racist, jingoist mass of other-hating citizenry down into one forcefully 'integrated' class who will be far more aware of their assigned roles as drones and less willing to go through the motions of the fruitless labors they used to happily pilot their own cars to and from. The whole house of cards could go down!

So if our society wasn't disintegrating fast enough, maybe autonomous cars will accelerate the process enough that i'll see the beginning of the next phase of humanity in my lifetime instead of just the slow painful end of the last one!

I love that there are so many people that can see the future with such certainty. Things like "there will be reductions across the board". Really? What if there's an increase across the board? You can guess all you want but in the end that is all it is. A guess. No better and no worse than mine.

My belief: we're already seeing an uptick in accidents. The pro-auto are attributing this to higher levels of distraction. What if it's the drivers waiting on the car to save them? "My other car tells me when there is another car beside me. I diidj t see it". "I thought the brakes would stop me even though I had the gas pushed".

The more we add, the more people stop driving and start allowing the car to do it for them. When those things break, they have accidents because they are no longer involved in actually driving. This is why I say take away the nanny devices. Active lane departure, brake assist, back up assist, self parking, etc. accidents are going to happen. Period. Making the vehicles structurally safer is smart. Making it drive for you is not.

In reply to ProDarwin:

How many people die of natural causes in the us every day? What's the overall, absolute cause of death in every single human being born in the last 10,000 years? I can tell you it isn't cars, guns, religion or famine. It's not plagues, cancers, or sickness. It's birth.

Every single person on this board has a 1 in 2 chance of waking up tomorrow. Those, my friend, are the true odds. Cars are our outlet. Take that away and that 50/50 E36 M3 for tomorrow starts to get iffy.

So sorry. These odds and percentages are laughable. I'm willing to bet more people die of natural causes every week than are killed by a car in a year. Whatever numbers you want to use.

Nick (Bo) Comstock wrote: Put me in the not interested, no way, no how, never ever camp. I have no use for a self driving car.

I can think of one or two...

Others not suitable for mixed audiences.

Bobzilla wrote: Every single person on this board has a 1 in 2 chance of waking up tomorrow. Those, my friend, are the true odd.

Schrodinger's statistics?...LOL!

In reply to Vigo:

Wow. Dude, I thought I had a dark view of the world. That's impressive!

In reply to Driven5:

Life's odds. We don't know what tomorrow brings or if tomorrow will even happen I'll take that .01% chance that I'll die in a car. Hell if odds were that good in vegas we'd all be rich

In reply to Bobzilla:

Well, I guess you win...I don't know anybody who can argue with somebody whose "life's odds" statistically have 50% of the current population dying each and every day.

Keith mentioned the car replacing the horse as an example. I think it may be more line air travel. Used to be kind of exclusive and a nice way to get around. Now that it's available to the masses it has its convenience, but also kind of sucks as an experience. Hmm. guess the same can be said of commuting now. The other day I'm making my way home cruising with traffic at about 70 and Joe Doushenozzle goes by outpacing traffic by 30-40 mph doing hard crosses of three lanes of heavy traffic, using the emergency lane multiple times and being an overall rolling hazard leaving havoc in his wake. As much as I hate the idea of autonomous cars I can see the appeal of not having to share the road with that guy at the helm of a 5,000 lb suv.

Stats are interesting for populations. On an individual basis things are binary

KyAllroad wrote: In reply to Driven5: As long as berkeleying feels good there is ZERO chance that the birth rate will fall. And since we are all so very very wired to obey the pleasure center of our primate brains, that's not going away either. The biological drive to reproduce is just ridiculous and (one could argue) completely against our self interests as a species. Look at parts of the planet which have been wracked with disease, famine, drought, and war for literally generations. Like Somalia for instance, you'd think that a sensible person would look around and say "no way I'm bringing my potential offspring into this environment". And yet the place is positively awash with kids. And to fully steer this thread into the weeds, global famine or disease is the only realistic way to get the numbers back into the territory of "sustainable". The trick is something bad enough to make a real difference but not so bad that systems collapse and the population reduction becomes self sustaining and we regress to a pre-industrial condition.

There some new technology called birth control, so you can get that good feeling without anyone getting pregnant.

So yes, there are regions where birth rates are still high, but the majority are declining.

If the data is accurate: Fertility rates are declining

I think this is an entirely interesting post and Im a gonna think about it- Economists who study this stuff have some sort of metric to do with 'fake luxury', or something like that. It has to do with buying 'Gucci' or 'Kayane' (the dumbE36 M3) branded crap from Ross (where I have bought most of the clothes in my adult life) and ypu get to pretend that you are the 1%. The fake freedom of driving where ever the hell you want unthered in your trusty steel steed is absolutely essential to the rapidly evaporating 'American Dream'.

Vigo wrote: The loss of individual autonomy is the elephant in the room. I don't think it's possible to fully account for the psychological and societal side effects that would come from removing the most tangible aspect of personal autonomy that most people in this country have. Spend enough time examining the neurotic thought processes of people who 'cant afford a car' and you will start to understand how effectively the perceived lack of choice can predict the choices you actually end up making. America as a society is more or less based on the concept of placating enough of the population with a perceived quality of life that they will allow and even contribute to the absolute debasement of anyone outside their monkeysphere. In my mind, there's a very good chance that removing the illusion of autonomy and freedom that comes from the daily subconscious gratification of a powerful machine responding to your whims will render our fractious, classist, racist, jingoist mass of other-hating citizenry down into one forcefully 'integrated' class who will be far more aware of their assigned roles as drones and less willing to go through the motions of the fruitless labors they used to happily pilot their own cars to and from. The whole house of cards could go down! So if our society wasn't disintegrating fast enough, maybe autonomous cars will accelerate the process enough that i'll see the beginning of the next phase of humanity in my lifetime instead of just the slow painful end of the last one!

Aww, geez...this E36 M3 again? (Walks away from what could have been a decent thread.)