Meh, I expect there is a whole world of transition involved, and improvements to be made in all sorts of areas, but the incentive exists to make them and they are purely incremental technical improvements, so its virtually guaranteed they will be made. It isn't going to take 300 years. 100 years ago cars could barely make 25 MPH and even then only on a few specially maintained roads. I'm sure plenty of people thought they wouldn't catch on for '300 years' either.

NPR had an April Fools Day story where they had an automated car that did that, cursed the other drivers, used the horn, etc. It was kind of funny.

but there's no HUGE change that occurs when a train is automatic. Yeah, its cheaper, and probably somewhat safer, but trains are REALLY safe already, an order of magnitude more so than automobiles, so the gain is much smaller. On top of that nothing much changes. The train still goes to the same place with the same schedule no matter how it is operated.

With cars they wouldn't operate in anything like the same way. Just think about this, if cars were truly driverless and could be summoned as needed why would you care what car you use at any given time (beyond you have some specific need, say for a bulky cargo or etc). This means that the entire paradigm of cars can change, they could simply be a service where they circulate around as needed and pick people up, preposition themselves, etc. They could also achieve MUCH higher utilization, 10 cars could do the work now done by 100 cars, which all mostly sit idle all day anyway.

Its not just a minor detail for cars, its a TOTALLY DIFFERENT SYSTEM that isn't interchangeable with the existing system. That's why I compared it to the change between horses and automobiles, not the change between manual and automatic trains, which really amounts to a detail, not a paradigm shift. This change is therefor inexorable because it brings vast efficiency gains and increased functionality. Once it happens the change will quickly propagate across the world, with any holdout regions being instantly marginalized and their citizenry left looking rather foolish as they spend 10x more than everyone else for a worse service.

It seems inevitable to me. Its going to take a lot of work and time and increased CPU power and better sensors, etc to fully realize all that, but we will. The advantages are so large that any nation 'Luddite' enough to refuse will be stone age compared to the rest of the world 30 years later.

Well, we will have to disagree. I think the change is now inevitable and won't be seriously impeded just by a few accidents all that much. The cars being sold 10 years from now will largely be capable of self-driving in a wide variety of situations, and that will just snowball. Once the idea of the car as a utility that can simply drive itself to where it is needed and carry out simple tasks on its own takes hold driving will die a very swift death.

Actually the number has declined pretty sharply in the last 10 years, to 32,719 in 2014. Its still a huge number, and the number of injuries is staggering, numbering in the several millions. I didn't look for an economic loss estimate but at the rate of 90 deaths per hour in the US it has got to be quite expensive. Every person in this country is effectively paying something like $1000/yr just to pay for the consequences of crappy human driving.

Yeah, we could lower our medical costs drastically, by 300% anyway, but it would still be 1/3 of a huge number is a huge number.

I didn't mean to suggest otherwise, sure, it will start with 'highway mode', but that won't be a long phase. For obvious reasons people will desire the ability to drive on surface roads. Actually parking will come before surface roads, Ford already has a system that can autonomously move the car and park it, driving at low speeds in a dense environment. It will only be a matter of time, I still say about 5 years, before these capabilities have effectively merged and we have a car that can deal with most situations. It might not be allowed to just drive without a 'supervisor' for a certain time, but there will be a huge push to get rid of that limitation.

Think about it, an autonomous car can go to school and pick up the kids, it can go to the store and be loaded with groceries, it can go to the dealer and get maintained, it can come and fetch you, obviating the need for a second car in many cases, etc etc etc. People may think they want to drive right now, but really they don't. They want the power and convenience of instant effortless transportation.

And once that day comes, then they won't even care about owning a car, the concept will be silly. The entire edifice of the car culture will vanish like a puff of smoke into nothing. Cars will simply become a ubiquitous utility, a service you pay to access and which supplies you with however much personal transportation you need or can afford. And of course then it can be supplied at different grades, you can ride with other people going where you're going and its cheaper, you can ride a 'car' that stops at various stops and picks up and drops off lots of people, gosh its a bus! Everything is about to change. The entire edifice upon which most people's reasoning about autonomous vehicles is built will be quaint nostalgia in 20 years.

Those are all considerations, but 50,000 people are killed in automobile accidents in the US alone every year. When that carnage is reduced to 2500 the naysayers will have zero ammunition, especially since the COST is huge and thus the savings also huge. Just as people accepted the hazard of cars over that of horses to gain advantages so will they accept driving by machines. The more economically sensible alternatives pretty much always win out over time.

I think its one of those things that once it comes to fruition everything simply changes. Its like automobiles. Most people laughed and insisted that horses would be around for another 100 years and cars were 'a fad' or 'a toy for the rich', etc. Once Ford made the first cheap car horses were done in 10 years flat, off the road.

It will be the same way. Safe, automatic driving will free up people's time, it will reduce costs greatly, and it will start a whole series of changes in the transportation infrastructure that will snowball. That's how I see it. In 10 years people will start to balk at buying a car they have to drive for themselves, and eventually they won't even care to own one anymore, it will be trivial to summon up what you need from whomever you contract for that service. The whole fetish of car ownership will go up in a puff of smoke. There will of course always be a few nostalgics, hobbiests, collectors, people that drive in some specific situations perhaps, but not much.

Its not possible to say exactly how long the transition will take, but in 5 years automated driving will be much improved over its current state, which is already pretty good, if limited. Within 10 years it will probably be accepted, like cruise control is now, and somewhere down the line, probably within our lifetimes, certainly well within the lifetimes of younger people, it will be ubiquitous. I'm guessing about 20 years, after that a human driver will be an oddity, if not an outright hazardous situation to be dealt with.

Not hardly. I've driven in NYC, Boston, many other cities all over the US, as well as in more rural areas, in the NE, on bad roads, snow, ice, etc. I think I know all about driving in the US. Most people just want to get from A to Z. Once automatic driving is here it will rapidly kick the humans off the road, nobody will be stupid enough to ride with you.

Most drivers accidents are in places they are familiar with. Its not entirely clear if this is simply due to mostly driving in such places, but it is commonly asserted that over familiarity often leads to inattention. I know this is true for me, and so presumably for many others. The fact remains, human drivers have a high error rate, and so far all the automated driving systems being tested in the US fall far below that number, even given that they drive in controlled circumstances.

It won't even amount to a statistic. Yes, there's plenty of work to continue to do, but even today cars can drive well enough to avoid the vast majority of accidents that humans get into because they're attentive and alert at all times, and they have much better adapted sensors, deployed in more and better locations, etc. The driving record of the existing autonomous systems is really incredibly good. Yes, they may only drive in conditions they're already prepared for, but as I said, most driving is of that sort anyway. Given another 5 years of polish and good heuristics on when NOT to use autonomous driving I think the accident rate of these vehicles will be a fraction of a percent of what it is for humans.

that 99.9% of the routes that cars drive today are on the same few miles of road? Do you know that its really easy for different computers to exchange information. In other words if its pre-mapped, its pre-mapped for EVERYONE, and you really don't need to pre-map a ton of routes hither and yon to get everyone doing 98% automated driving. That's 98% fewer accidents.

Yeah, its hard to believe fixed routes won't be entirely mapped soon. Aren't there city buses that can already do that? I believe so. Because its a small fixed route it can be completely mapped and analyzed to the point where there aren't any surprises except what normally isn't there, you can pick it all out, you already know all your navigation decisions, etc. It will still take a couple more decades for the whole thing to get routine. I doubt truck drivers are losing TOO much sleep yet.

I use a number of products like that which really don't get much in the way of updates. Even so, at least your project has a maintainer that probably answers the very infrequent question and obviously addresses any bug reports in some fashion. Lots and lots of sourceforge projects never ever release code. I expect many die at the "I had an idea" stage, but others just never really sort out the organizational or "marketing" issues (IE getting people interested and trying the code so that something grows).