California Allows Truly Driverless Cars + Passengers in them

[Vinyadan]

https://www.nytimes.com/reuters/2018...alifornia.html

Since April 2nd, a law from February allows test driverless cars to be really driverless: no back-up drivers. And now a new law allows passengers in driverless cars. Passengers must be over 18 and the lift must be free of charge. No airport trips allowed.

While 50 driverless car companies operate in California, only one of them has applied for road testing without a back-up driver. https://sf.curbed.com/2018/4/3/17193...ornia-uber-dmv

So, what could have been a huge shift will probably come to pass with a very soft pedal.

[Blithesome]

I can’t properly express my enthusiasm for this! We’re finally entering into that glorious stage of the future where nobody has to waste a sixth of their waking life staring at traffic, and I’m unutterably delighted.

[monomer]

I just want to note that Waymo has been operating completely driverless cars in Arizona since November 2017. California has much more stringent regulations when it comes to driverless cars, but I wouldn't be surprised if the single company is Waymo.

[Max_Killjoy]

Great news for people who think the Axiom was a utopia.

Soon your chairs will be self-driving too.

[Blithesome]

Great news for people who think the Axiom was a utopia.

Soon your chairs will be self-driving too.

I think that people should do precisely as much tedious manual labor as they prefer to, and that this is a step in that direction.

People went into hysterics when washers and dryers were invented, too, and at the advent of the sewing machine, and whenever a new labor saving device was produced; this is nothing more than another instance of the same pattern.

If chairs become self-driving, it’ll be a huge step forward for the severely disabled- just like this is- and it’ll be a huge step forward for technological progress as a whole. I welcome the possibility.

[Xyril]

Driving is a funny skill that many people develop not because they want to, but because they have to, at which point they develop strong opinions about the adequacy of their own skills and the inadequacy of everyone else's.

I don't know if perfect driverless cars are possible--I don't care, really. If we can plausibly get the safety rating up to that of the average human driver, then we might possibly change our culture to the point where people who are terrible no longer feel the need to pretend otherwise.

I enjoy driving, but that's my choice, and it shouldn't foreclose others from making a different choice.

[anjxed]

People went into hysterics when washers and dryers were invented, too, and at the advent of the sewing machine, and whenever a new labor saving device was produced; this is nothing more than another instance of the same pattern.

If chairs become self-driving, it’ll be a huge step forward for the severely disabled- just like this is- and it’ll be a huge step forward for technological progress as a whole. I welcome the possibility.

Sewing machines, Dryers, Washers and other labor saving machine doesn't move. Whereas cars do. The closest thing we have to autonomous moving machine would be planes. Even then there must be a pilot keeping watch at all times. This is something very new and people have a right to be concerned

[Khedrac]

One of my major concerns with driverless cars is "will they break the law when they need to?"
Last year I was twice stuck at temporary traffic lights where the other light had failed. Being properly designed the working light turned red (safer to stop the traffic when there is a problem in the system). So far so good.
The problem was that both occasions were on non-working days so there were no workers on site to check and correct the lights (since both were fixed by the time I returned over 8 hours later I assume that someone was summoned); so how long does one wait before (correctly) concluding that the lights are broken and need to be ignored? If in a self driving car which won't break the law I could have been stuck for hours...
Not a directly life threatening situation - but one that could be if the self driving vehicle was a medical transport.

Now, if the engineers have asked this question and come up with an answer (I remember reports that Google's cars have an "aggressiveness" setting so that may cover this) then well and good; the problem is that I am a software tester so I have a fair idea just how poor engineers can be at asking enough questions.

Edit: Count me as considering this a step forward, but equally very glad it's not being trialled anywhere near me.

[Strigon]

If chairs become self-driving, it’ll be a huge step forward for the severely disabled- just like this is- and it’ll be a huge step forward for technological progress as a whole. I welcome the possibility.

This is exactly why I want self-driving cars; I couldn't possibly drive a car, and all forms of public transport are, for one reason or another, problematic at best. Sure, there's a significant gap between self-driving cars and self-driving wheelchair accessible cars, and then another gap between those and versions of them that are affordable, but if they are someday available my autonomy will simply go through the roof.

Heck, throw in one of those personal care robots you occasionally here about on the news, and I'm practically as independent as you foot-people!

[Douglas]

I'm sure there will eventually be lines of self-driving cars specifically targeted for people with disabilities that make driving impossible, and wheelchair accessibility will be an obvious important feature for a large portion of those. It may take a while for them to become easily affordable, though.

[Max_Killjoy]

I'm sure there will eventually be lines of self-driving cars specifically targeted for people with disabilities that make driving impossible, and wheelchair accessibility will be an obvious important feature for a large portion of those. It may take a while for them to become easily affordable, though.

And when they come, they'll be a great thing for those who actually need them.

[Chen]

Sewing machines, Dryers, Washers and other labor saving machine doesn't move. Whereas cars do. The closest thing we have to autonomous moving machine would be planes. Even then there must be a pilot keeping watch at all times. This is something very new and people have a right to be concerned

Yeah pilot error causes more air crashes than any other factor. Overall human error is even higher (it includes maintenance and controllers). Automating planes would be a good thing from a safety point of view but people are irrationally not accepting of not having the pilots fly the plane. Despite the fact they're more likely to kill you than mechanical, electrical or computer failure is.

[Max_Killjoy]

Yeah pilot error causes more air crashes than any other factor. Overall human error is even higher (it includes maintenance and controllers). Automating planes would be a good thing from a safety point of view but people are irrationally not accepting of not having the pilots fly the plane. Despite the fact they're more likely to kill you than mechanical, electrical or computer failure is.

Aircraft automation is actually starting to become a factor in incidents due to negative impact on pilot ability.

https://www.nbaa.org/ops/safety/2013...r-aircraft.php
https://flightsafety.org/asw-article/use-it-or-lose-it/
https://www.newyorker.com/science/ma...rds-automation
https://airfactsjournal.com/2011/09/...lot-addiction/
http://www.thedrive.com/travel/1692/...e-pilots-worse
https://labs.psych.ucsb.edu/schooler...ed_cockpit.pdf

More generally:

https://news.mit.edu/2015/no-driverless-cars-1013
https://www.cnet.com/news/googles-vi...says-mit-prof/
http://www.businessinsider.com/toyot...ng-tech-2017-1
https://www.computerworld.com/articl...erved-too.html
https://www.theverge.com/2014/9/23/6...nst-automation

[theNater]

Aircraft automation is actually starting to become a factor in incidents due to negative impact on pilot ability.

https://www.nbaa.org/ops/safety/2013...r-aircraft.php
...
https://labs.psych.ucsb.edu/schooler...ed_cockpit.pdf

Did you mean to say that automation's impact on pilot ability may become a factor? Because that's all these two links, at least, support. (And the first one is explicit that automation has helped improve safety).

[JeenLeen]

One of my major concerns with driverless cars is "will they break the law when they need to?"

I'd hope that driverless cars would have a manual override mode, where you can take over and drive, or at least issue a manual command like "go forward if no incoming traffic" to bypass those situations.
I think what I would prefer most is a car where I usually drive it, but can set it on autopilot if I wish (such as if exhausted and worried about falling asleep at the wheel.)

If I trusted the software, I'd probably be okay with a driverless car as a means of efficiency/laziness. I have some misgivings (e.g., losing skill at driving for if needed, not learning directions as well since not needing to know how to get from A to B, etc.)... but some of those aren't really important if I'm not driving. It'd be great to take a nap, or read a book, or be able to get home if I had too much to drink.

[Max_Killjoy]

Did you mean to say that automation's impact on pilot ability may become a factor? Because that's all these two links, at least, support. (And the first one is explicit that automation has helped improve safety).

No, I mean, as clearly stated and supported, that automation is already starting to erode pilot skill and awareness.

In the first article you maintain the link to in the quoted portion of my post, the FAA is in no uncertain terms recommending that despite general safety gains from improved systems, that pilots take manual control more often in order to maintain their "hand flying" skills. The automated systems are not and never will be perfect, and the pilot is there both to oversee and direct the automated systems, and to take over when necessary.

In the second, they are making a clear claim that over-reliance on the automated systems is degrading the cognition and decision-making skills of the pilots.


The problem with statistics is that they're aggregate. What is "statistically safer" can easily be deadly in a specific instance.

We also see the "more is always better" trap caused (in part) by the general human adoration for absolutes -- "If some automation makes things safer, then more automation will always be better and will always make things safer!"

(And never mind that "safety" is not the only consideration...)

[Max_Killjoy]

I'd hope that driverless cars would have a manual override mode, where you can take over and drive, or at least issue a manual command like "go forward if no incoming traffic" to bypass those situations.
I think what I would prefer most is a car where I usually drive it, but can set it on autopilot if I wish (such as if exhausted and worried about falling asleep at the wheel.)

If I trusted the software, I'd probably be okay with a driverless car as a means of efficiency/laziness. I have some misgivings (e.g., losing skill at driving for if needed, not learning directions as well since not needing to know how to get from A to B, etc.)... but some of those aren't really important if I'm not driving. It'd be great to take a nap, or read a book, or be able to get home if I had too much to drink.

Self-driving cars could be literally "perfect" in terms of their safety, ability to handle the unknown / unexpected, etc... and I would still never own, rent, or use a self-driving car. Such a machine is of absolutely zero utility or interest for me personally.

[Chen]

Aircraft automation is actually starting to become a factor in incidents due to negative impact on pilot ability.

https://www.nbaa.org/ops/safety/2013...r-aircraft.php
https://flightsafety.org/asw-article/use-it-or-lose-it/
https://www.newyorker.com/science/ma...rds-automation
https://airfactsjournal.com/2011/09/...lot-addiction/
http://www.thedrive.com/travel/1692/...e-pilots-worse
https://labs.psych.ucsb.edu/schooler...ed_cockpit.pdf

More generally:

https://news.mit.edu/2015/no-driverless-cars-1013
https://www.cnet.com/news/googles-vi...says-mit-prof/
http://www.businessinsider.com/toyot...ng-tech-2017-1
https://www.computerworld.com/articl...erved-too.html
https://www.theverge.com/2014/9/23/6...nst-automation

This is an argument for more automation and less pilots. It does not counter the idea that pilots cause most of airplane crashes at all. It does show the risk of semi-autonomous systems. I do agree these are MUCH worse than either full or no automation. We saw that in the Tesla autopilot crash. People think it's a real autonomous system and get complacent. Except it's NOT a fully autonomous system and thus they make things more dangerous.

You're just drawing the conclusion the other way towards the less automation side instead of the more automation side. For planes especially whether its a pilot in the cockpit or an automated system the impact on the people should come down to safety numbers. There's not even the recreational or personal freedom aspect in planes. Hell in many now you're not even allowed to see the pilot except prior to take-off and after landing. Put the proper regulation in (and aircraft software regulation is already incredibly strict) and I'd be way more comfortable knowing there's a 10^-9 chance per flight hour of a catastrophic software malfunction (current regulation) rather than rolling the dice on the pilot being overly tired, drunk or just plain the bottom of his class in the event of an emergency.

[Max_Killjoy]

This is an argument for more automation and less pilots. It does not counter the idea that pilots cause most of airplane crashes at all. It does show the risk of semi-autonomous systems. I do agree these are MUCH worse than either full or no automation. We saw that in the Tesla autopilot crash. People think it's a real autonomous system and get complacent. Except it's NOT a fully autonomous system and thus they make things more dangerous.

You're just drawing the conclusion the other way towards the less automation side instead of the more automation side. For planes especially whether its a pilot in the cockpit or an automated system the impact on the people should come down to safety numbers. There's not even the recreational or personal freedom aspect in planes. Hell in many now you're not even allowed to see the pilot except prior to take-off and after landing. Put the proper regulation in (and aircraft software regulation is already incredibly strict) and I'd be way more comfortable knowing there's a 10^-9 chance per flight hour of a catastrophic software malfunction (current regulation) rather than rolling the dice on the pilot being overly tired, drunk or just plain the bottom of his class in the event of an emergency.

You appear to be making the mistake of thinking it's an argument for no human crew and no human oversight.

Until the automated systems can handle these sorts of situations better than the human aircrews did, I want a human crew in the cockpit:

https://en.wikipedia.org/wiki/Northw...ines_Flight_85
https://en.wikipedia.org/wiki/United...nes_Flight_232
https://en.wikipedia.org/wiki/US_Airways_Flight_1549
https://en.wikipedia.org/wiki/Americ...ines_Flight_96
https://en.wikipedia.org/wiki/Reeve_...rways_Flight_8
https://en.wikipedia.org/wiki/DHL_sh...ent_in_Baghdad
https://en.wikipedia.org/wiki/Philip...nes_Flight_434
https://en.wikipedia.org/wiki/Air_Transat_Flight_961

[Chen]

You appear to be making the mistake of thinking it's an argument for no human crew and no human oversight.

Until the automated systems can handle these sorts of situations better than the human aircrews did, I want a human crew in the cockpit:

https://en.wikipedia.org/wiki/Northw...ines_Flight_85
https://en.wikipedia.org/wiki/United...nes_Flight_232
https://en.wikipedia.org/wiki/US_Airways_Flight_1549
https://en.wikipedia.org/wiki/Americ...ines_Flight_96
https://en.wikipedia.org/wiki/Reeve_...rways_Flight_8
https://en.wikipedia.org/wiki/DHL_sh...ent_in_Baghdad
https://en.wikipedia.org/wiki/Philip...nes_Flight_434
https://en.wikipedia.org/wiki/Air_Transat_Flight_961

To be fair there are items there that do have automated systems that could be done without human input (items 2 and 4 have better systems created to avoid needing manual throttle control).

Also https://en.wikipedia.org/wiki/Catego...by_pilot_error there are plenty of the opposite cases to look at. The point, like autonomous cars, is that once you can bring the accident rate lower than that caused by humans, its rational to switch to the new method.

[Max_Killjoy]

"Safety" is not the only consideration, here or otherwise.

There are real questions of what "we" are willing to give up to be (or even just feel) "safer".

[Strigon]

"Safety" is not the only consideration, here or otherwise.

There are real questions of what "we" are willing to give up to be (or even just feel) "safer".

Keep in mind, of course, that safety is only one reason to buy a self-driving car. I suspect that, for most people, it's not even a particularly important reason because not many people think of themselves as unsafe drivers.
I think the biggest factor will be convenience and novelty. Maybe as they become more popular, there will be more and more uses for them that we can't even imagine right now; that happens quite often.

[halfeye]

"Safety" is not the only consideration, here or otherwise.

It isn't?

I am aware some drivers like to speed, and that there's a thrill to that, however I don't think there's a right to that thrill that overrides other people's safety.

[137beth]

Most people I know who want self-driving cars don't even cite safety as the main reason: usually they want self-driving cars so that they don't have to drive. I get that some people find driving fun, but there are other people who don't, and would be much happier not driving if it was feasible for them to avoid it.

[Algeh]

The main reason I'd want a truly self-driving car would be that it would increase viable "weekend road trip" length because I could sleep while the car was driving me to and from places. (I currently use trains for this, but the cities I visit by train around here only have one train a day, starkly limiting my choices for arrival/departure times.)

I wonder how many other people would suddenly find 300-500 mile each way road trips viable on weekends if they could start them Friday night after work and wake up when they got there, then do the same again on Sunday night. (Or even longer - pretty much anything under 8-10 hours away seems somewhat reasonable if you're doing most of the travel during your sleep time.) I'd probably be much more tempted to start going to weekend events in other cities in that case. I'm pretty good at sleeping in the car now (while not driving - these kinds of trips currently would involve two people driving in shifts), and if I bought a car optimized around that it would presumably be more comfortable to sleep in than my current one.

This would not be good from an environmental or congestion perspective, of course, but I suspect I'm not the only one who would start doing it.

I also eagerly await but also fear the day cars become able to find their own parking spots and everyone just gets out of their cars at building entrances and lets the cars circle for blocks trying to find a parking place, then summons them again when they're done. If not regulated under things like the "anti-cruising" laws places used to have back when teens had cars instead of phones to get in trouble with, I could see some downtowns having the congestion mess that results from it being cheaper to spend the gas to have your car circle around downtown on its own for 3 hours than it is to pay for 3 hours of parking...

[Strigon]

The main reason I'd want a truly self-driving car would be that it would increase viable "weekend road trip" length because I could sleep while the car was driving me to and from places. (I currently use trains for this, but the cities I visit by train around here only have one train a day, starkly limiting my choices for arrival/departure times.)

I wonder how many other people would suddenly find 300-500 mile each way road trips viable on weekends if they could start them Friday night after work and wake up when they got there, then do the same again on Sunday night. (Or even longer - pretty much anything under 8-10 hours away seems somewhat reasonable if you're doing most of the travel during your sleep time.) I'd probably be much more tempted to start going to weekend events in other cities in that case. I'm pretty good at sleeping in the car now (while not driving - these kinds of trips currently would involve two people driving in shifts), and if I bought a car optimized around that it would presumably be more comfortable to sleep in than my current one.

This would not be good from an environmental or congestion perspective, of course, but I suspect I'm not the only one who would start doing it.

I also eagerly await but also fear the day cars become able to find their own parking spots and everyone just gets out of their cars at building entrances and lets the cars circle for blocks trying to find a parking place, then summons them again when they're done. If not regulated under things like the "anti-cruising" laws places used to have back when teens had cars instead of phones to get in trouble with, I could see some downtowns having the congestion mess that results from it being cheaper to spend the gas to have your car circle around downtown on its own for 3 hours than it is to pay for 3 hours of parking...

Actually, if this started happening, I suspect it would actually help traffic. Suddenly, more people are driving at night rather than day, spreading out the load. It wouldn't be an enormous difference either way, but that's my take on it.
And I suspect most self-driving cars will be extremely efficient, simply because I think there's an enormous overlap between people interested in self-driving cars and people who consider fuel economy to be very important.

Finally, why would you have your car drive around for three hours? Unless you have a ridiculous commute, why not just send it back home to your own personal parking space, be it in a parking garage or your own driveway?


But I agree with your general idea; long rides overnight are a fantastic use of self-driving cars. Heck, those could probably be done (perhaps not legally) fairly soon, if not now - driving on a highway for a few hundred miles without getting off is something a car's autopilot can do today, isn't it? Maybe not as safely as we'd like yet, but it's certainly more viable than a fully autonomous car.

[halfeye]

Actually, if this started happening, I suspect it would actually help traffic. Suddenly, more people are driving at night rather than day, spreading out the load. It wouldn't be an enormous difference either way, but that's my take on it.
And I suspect most self-driving cars will be extremely efficient, simply because I think there's an enormous overlap between people interested in self-driving cars and people who consider fuel economy to be very important.

Finally, why would you have your car drive around for three hours? Unless you have a ridiculous commute, why not just send it back home to your own personal parking space, be it in a parking garage or your own driveway?


But I agree with your general idea; long rides overnight are a fantastic use of self-driving cars. Heck, those could probably be done (perhaps not legally) fairly soon, if not now - driving on a highway for a few hundred miles without getting off is something a car's autopilot can do today, isn't it? Maybe not as safely as we'd like yet, but it's certainly more viable than a fully autonomous car.

Motion sickness might be an issue with that. I like the idea in general, but I'm not sure how much of a problem motion sickness would be, it's triggered by not relating felt acceleration to movement of the world. Smoothness certainly helps, trains and planes are usually very smooth movers these days, how smooth would autonomous cars be?

[Algeh]

Actually, if this started happening, I suspect it would actually help traffic. Suddenly, more people are driving at night rather than day, spreading out the load. It wouldn't be an enormous difference either way, but that's my take on it.

Yes, but I think people would also take more trips. For example, last weekend there was a gathering of some filkers about 350 miles away from me (the city in question is about a 5-6 hour drive from here if you don't stop, about a 7 hour drive if you're me and need breaks). I was invited to go, but decided that I did not want to spend all weekend driving to and from that city and rent hotel rooms for both Friday and Saturday night just to spend Saturday afternoon/evening filking. Instead, I drove many fewer miles to go to a local gaming convention, even though I like filking better than gaming and probably would have had more fun at the other thing once I got there.

With a self-driving car, I could pack it up the night before, set an alarm for stupid early o'clock Saturday morning, get in the car, go back to sleep, and get to the filk thing later that day. I could then get back in the car later that evening and get back home overnight, turning what would have been a 3 day trip into one long day. I suspect I'd take a lot more of those than I would 3 day weekend trips. In my personal case I could see it happening on the order of once or twice a month rather than once or twice a year between friends in other cities and regional geek things. Those trips would generally be using the roads at off-peak times when they have capacity, but it would be quite a bit more gas than I generally go through in a weekend. Long trips like that, particularly at night, are the hardest use case to do with batteries, so we might see gas consumption increase if those kinds of trips become a thing that people can reasonably do.

Finally, why would you have your car drive around for three hours? Unless you have a ridiculous commute, why not just send it back home to your own personal parking space, be it in a parking garage or your own driveway?

I suppose that's a good point for 3-4 hour or longer things. I could see suburbanites going to dinner downtown where it actually wouldn't be enough time to send the car all the way home and back deciding it's cheaper to tell the car to wander around the city rather than pay parking, though. More likely, if the car is smart enough to understand all of the assorted parking rules, they'd tell the car to drive until it hit free 2 hour street parking somewhere, and those neighborhoods would see an uptick in random cars parked there. (I could also see people who live in apartments downtown without free parking similarly telling their cars to drive to wherever the nearest non-hour-limited neighborhood is and park there rather than rent a parking space. All kinds of interesting yet annoying things become possible when you don't have to walk to and from where you put your car.)

[gomipile]

Sewing machines, Dryers, Washers and other labor saving machine doesn't move. Whereas cars do. The closest thing we have to autonomous moving machine would be planes. Even then there must be a pilot keeping watch at all times. This is something very new and people have a right to be concerned

And subways, and slidewalks in large airports and shopping malls, and trams, and elevators, and escalators...

[Vinyadan]

Driverless cars used instead of public transportation would increase traffic, no doubt about that. However, they also would allow building multistorey car parks that don't take into account the need for humans to get off the car, so they can squeeze more cars in less room.