Video Author Peter Wayner Talks About Autonomous Cars (Video) 50
Tim: Hi. We are talking today with Peter Wayner, who is an accomplished author, and he has been a longtime Slashdot book reviewer as well. By the way, Peter has been thinking a lot about autonomous driverless cars, and he has a book upcoming on this. We will talk today a little bit about some of the issues about driverless cars – what they mean for everybody.
Peter Thanks Tim. There are a lot of issues that will be interesting to everybody on Slashdot because not only are there the typical issues that confront everyone like, is it nice to be able to go back from a bar late at night without having to worry about DWI but there are a lot of technical issues, issues about privacy, issues about designing computer systems; questions about how we are going to build the infrastructure, and in that sense what we are doing is we are taking a lot of the ideas from the internet and we are stretching them out in the meatspace, and all of a sudden we are going to have cars that will take us places in much the same way that websites take us places.
Tim: You mentioned privacy. Privacy I think with cell phones, I don’t want people eavesdropping on my conversation, now with cars, when I get my car right now, I don’t think of privacy as a central issue. So explain how privacy comes into it.
Peter Well, right now if someone wants to follow you, they have to follow you, and it is one of the classic parts of every cop drama and every kind of movie and there are always chase scenes of people trying to evade a tail, and lately the police have gotten these, they have built these nice GPS locating devices that track people without actually having humans on the tail, but those have come under a lot of legal surveillance itself where people have asked whether that this is really a legitimate thing, and what kind of warrant should be used. Now all of a sudden, we are going to add yet another layer on top on this, because these autonomous cars are going to be filled with computers and computers always collect data, and this data is going to be sitting around somewhere.
Tim: And it is comprehensive sort of data too.
Peter It is going to be comprehensive and it is going to be comprehensive for everybody, it is not going to just be the person who is inside the car, because the cars themselves have sensors all over the place, they’ve got jammers looking forwards, jammers looking backwards, they could easily keep track of everything that goes by. And when you saw what happened with the Boston marathon bombing and they were able to collect all these random snapshots that people were just taking of the marathon and piece together what’s going on, well can you imagine if every car in the city kept track of everything it saw, everywhere it goes. And if you work through the math, you realize that it is not that big a number, it won’t require that much disk space to record everything coming out of every car in New York City.
What’s going to be interesting and what’s going to be different is that a lot of people aren’t going to own their cars, and just as we have these kind of questions about privacy in the cloud right now, and privacy about who owns your email if it is stored with an email and a website, there are going to be questions of who owns the data when you take a cab ride that’s driven by an autonomous car. Is it going to be you? Or is it going to be the cab company? And the cab company could quite easily just disclose it in response to a subpoena. Or the cab company may do all kinds of crazy things with it. They may use it to sell you ads, they could do all the different things that the websites are doing today. So you go to a store, you go to a particular store, you go out late at night, the car will notice, and the car will know where you are going, and all of a sudden it won’t just be your web actions that drive the advertising coming to you, it is going to be all of your actions.
Tim: It is a lot more than a store loyalty card, they know what store you go to, but they know what store you go to, where you came from, where you bought the gas, or in future, where you plug in your car. So it is all that stuff. Now, you mentioned a cab you described it as a cab as if it is a car that you don’t own, it comes and picks up, is that what we see the model of driving in the future, it is really more like a cab company that urban drivers or urban residents are now much more used to?
Peter Well, I don’t really know what is going to happen. I don’t think anyone does. But the cab model is really economically interesting. Because if you think about the car that I own, or the car that you own, they are both sitting in the driveway rusting right now. And really the average car rusts for 23 maybe 23½ hours a day, and when you start to have an autonomous car, well, when you are not using the car, the autonomous car can be helping someone else. And then, if you kind of take that to the extreme, you think wow! It is just going to be much easier to have a cab and you dial up on your smart phone, the way you do at Uber right now, you dial up a ride and boom! The cab comes to you. When you work through these systems, you realize that there are all kinds of neat things you can do – for instance, you can have car pools, so the central reservation system can be sitting there, and says, well Bob wants to go downtown, and Chris wants to go downtown, and Pat wants to go downtown, and they are kind of right in the way, right along the same route, how about if I them send them a coupon, and it says if you guys all ride together, you get it for half price. And the autonomous computer, the central reservation computer can do all kinds of crazy things like that.
Tim: You know the city I live in, I live in Austin, and here we’ve got both Car 2 Go and Zipcar which have some things in common in that they are a shared ownership model, where you are buying a service but you don’t have to call someone because the cars themselves are distributed. I can see that happening too with driverless cars, is that they will be all around the city in a cluster, and when you want to go, the nearest one may activate itself and pop over.
Peter You are right. It could just kind of zip right over and I am sure, the Zipcar folks are thinking about it, and the Car 2 Go, those are already kind of amazing models and they are really useful. I use Zipcar all the time now, and I think that we are going to see even more just because the economics of owning a car are kind of prohibitive when you look at the fact that it can be so much cheaper if you just share it with your friends.
Tim: One thing, when it comes to certain decisions, when you buy a car now, there is a lot of individual decisions that go into it, because you buy a car for its style, for what consumer reports say or perceived safety advantages, if you are getting the car comes at the top of the list, when you order your robot car, I wonder if you feel funny about getting a safer or less safer when it comes to your driveway.
Peter Well, those decisions are going to be made by the central, if we move to some kind of shared system, those decisions are going to be made by whoever runs the buying process. So like right now, there is some guy in a back office of Zipcar who says ‘I like this car better’ for whatever reason, and so the car manufacturers are going to have to make those guys happy, and they are going to have to make the end user happy too but they won’t be the real deciders, and so if the central company can somehow keep enough of their regular customers happy enough then there is a lot of opportunity for making central decisions.
Tim: You know, I hope that it doesn’t boil down to two or three providers in a city, but I hope it is not like an ISP where you sometimes your decisions between worse and worser
Peter I think you are absolutely right. I think I‘ve been talking to a couple of my friends, a couple of other journalists I have been working with on this book, and I have been showing it to them, some of them are really optimistic and they think that this is going to be wonderful. But you know, one of them at least is incredibly pessimistic and he thinks he is going to have exactly the same problem that you are talking about there, where you just don’t have these choices with the ISP because there is going to be one best provider that is going to come in and roll everybody up, and then you won’t have choices anymore.
Less talk, more action (Score:2)
We keep hearing about these things, but so far there is not even a Tesla of Automated cars available on the market or nearing it.
How far off are we from drinking in the car again?
Re:Less talk, more action (Score:4, Funny)
How far off are we from drinking in the car again?
Today. It's called a limousine.
Re: (Score:2)
I can't seem to afford to take one of those everywhere I go. I meant in my car, with all humans able to drink or read or play on their phones.
Never gonna happen (Score:3)
They'll never be trustworthy enough. Autopilot in relatively uncrowded skies at different altitudes, with a pilot close at hand, is one thing. On complicated, crowded roadways it's quite another. You'll never be able to get through the social suspicion, the legal liabilities, and government red-tape, etc. It is and will remain an experiment which periodically pops up, but remains "20 years away" forever.
Re: (Score:2)
Re: (Score:3)
And we can create even more tools that offer a gradual evolution. We already have a database of all of the roads. With a bit more precision, we could build a device that could tell whether you're following a common path that others have taken before or if you're drifting into the way of oncoming traffic.
There are some, though, that suggest that gradual evolution may be more dangerous than jumping directly to fully autonomous vehicles. As the humans have less and less to do behind the wheel, their mind drift
Black Swan (Score:5, Informative)
http://en.wikipedia.org/wiki/Google_driverless_car [wikipedia.org]
In August 2012, the team announced that they have completed over 300,000 autonomous-driving miles (500 000 km) accident-free, typically have about a dozen cars on the road at any given time, and are starting to test them with single drivers instead of in pairs.[14] Three U.S. states have passed laws permitting driverless cars as of September 2012: Nevada, Florida, and California.
More miles than most drivers rack up in 20 years, and without having caused an accident. Laws passed in 3 states.
This is a lot closer than positive net output fusion, for example.
When you Scale Up is where the issues pop up (Score:2)
When you Scale Up is where the issues pop up and they have drivers on hand to take over right away.
Re: (Score:2)
Yes, scale up is a big question mark. But on the other hand, computerized cars should be able to communicate with each other. They can ask permission to change lanes or at least warn each other with better regularity than the humans I see on the road. They will be able to swap plans with each other and that should help them do a better job than humans. They'll have more information.
I won't be hard for robots to do better than us (Score:3)
Oops! (Score:2)
It won't be hard for robots...
And, no (Score:2)
Re: (Score:3)
MAHD - Mothers Against Human Drivers
Once the MADD group realizes there almost double the number of people getting killed by human drivers than by drunk human drivers their focus will change quickly.
They have had 30 (40?) years of experience and with modern social media the campaign to migrate people into autonomous vehicles will be swift and vicious. Most likely within five years it will be (in large urban centres) about as socially acceptable to drive your own car as driving drunk is today.
Re: (Score:2)
When you Scale Up is where the issues pop up
I don't get it. You're saying that having more computer controlled cars on the road is going to make it harder for computer controlled cars to function? Why?
and they have drivers on hand to take over right away.
It's true, and it would sure be interesting to know how often that happens. But 300K miles is a pretty awesome track record for any driver under any circumstances.
Re: (Score:3)
By the way, a computerized car isn't magic, I'll bet it also couldn't have stopped in time though I will admit it might have minimized the damage by braking sooner
Re: (Score:2)
Yes, But, scaling up at the accident rate they are currently experiencing (i.e. even though we don't know exactly what it is, we know that it is likely less than one accident in 300,000 miles for well maintained vehicles) would mean a lower overall accident rate than the human operated fleet manages.
So the tech doesn't have to get better. It simply has to get cheaper, more reliable (robust physically), integrated into the manufacturing process and rolled into the dealer and repair networks. No of which is a
Re: (Score:2)
http://en.wikipedia.org/wiki/Google_driverless_car [wikipedia.org]
In August 2012, the team announced that they have completed over 300,000 autonomous-driving miles (500 000 km) accident-free, typically have about a dozen cars on the road at any given time, and are starting to test them with single drivers instead of in pairs.[14] Three U.S. states have passed laws permitting driverless cars as of September 2012: Nevada, Florida, and California.
More miles than most drivers rack up in 20 years, and without having caused an accident. Laws passed in 3 states.
This is a lot closer than positive net output fusion, for example.
I'll believe it when they can drive in Boston traffic during rush hour with crowded streets, crowded highways at 65mph, one-way streets, jaywalkers, construction, snow, sleet, ice, short on/off ramps, non-standard intersections, traffic circles, etc. I've driven in San Francisco and it's a Sunday drive compared to Boston and New York.
Re: (Score:2)
Yes, you're right. Boston and NYC are nightmares. But then again computers can do certain things better than humans. They handle scale up more gracefully. A human might be able to process a number of pedestrians and dangerous items, but the human brain maxes out pretty quickly. If a computer can track one pedestrian, it can probably track 10,000 too. The scale up is just linear. You just add a bit more computing power. If the Google car can handle SF with a certain number of processors, I'm pretty sure it
there are other scale up isses that computer power (Score:2)
there are other scale up issues that more cpu power does not really help.
Like networking issues
compatibility issues and more
Re: (Score:2)
Can you elaborate? Which kind of networking issues? The advantage cars have is that they're only concerned with the cars that are nearby or about to be nearby. They don't need to worry about all O(n^2).
Re: (Score:2)
So what are the 'network' and 'compatibility' issues?
Re: (Score:2)
I believe that snow/sleet/ice are known issues, and google is avoiding 'em for now.
As for the difference between sanfran, ny, and boston - they don't feel all that different to me.
Of the things you mentioned, I have to think that construction is one of the oddest cases - and vegas is no stranger to that.
But the point is: they're on the roads, now. And they're doing really well. And it is a technology in its infancy.
To believe "it's never gonna happen" seems pretty foolish to me.
Re: (Score:2)
Exactly. Google is putting plenty of miles on their cars and they're finding quite a bit of success. The DARPA Grand Challenge cars are almost a decade old. We're switching over from science to engineering. Marketing won't be long.
Re: (Score:2)
I'll believe it when they can drive in Boston traffic during rush hour with crowded streets, crowded highways at 65mph, one-way streets, jaywalkers, construction, snow, sleet, ice, short on/off ramps, non-standard intersections, traffic circles, etc. I've driven in San Francisco and it's a Sunday drive compared to Boston and New York.
Why are you imposing a higher burden on autonomous vehicles than human drivers? :-)
Re: (Score:2)
Re: (Score:2)
States are passing laws. I have to think that the NHTSA has noticed. And even if they haven't said anything good about 'em, they haven't said anything forbidding 'em, either.
Will there be a min forced software updates time (Score:2)
Will there be a min forced software updates time periods let say all auto cars must have free software updates for a min of say 5 years No saying 1 year later we don't do updates for old cars any more want the fix to Bug X buy a new car or buy the 5K-10K computer upgrade.
moving to a rent a car system may not work that go (Score:2)
moving to a rent a car system may not work that good and some of the rent a car companies don't do maintenance and or really push it out.
also some let cars get to point where they wear down parts and then they make the last person to rent pay the cost (mainly done over seas with manual drive cars)
Re: (Score:2)
The problem I see is that if you're in a rental car or cab, there is some incentive to not abuse it. For the rental car, that's the guy inspecting it when you turn it in, for the cab it's the cabbie who at some point will kick you out for screwing up his cab. Remove those people and those checks, and the inside of a shared car will get pretty nasty pretty quick, and people will not want to sit in that shared car. For a car shared within a carpool group, there's the social pressure exerted because everyon
Re: (Score:2)
You're right about problems with rentals and shared things but the problems are slowly being solved. I've had great luck with Zipcars. People who abuse the cars are kicked out of the program. The cars of the future may have a video camera watching them at all times and the car company may just dig it up if there are questions about smoking or abuse. The privacy will suck but maybe people who want a clean car will choose to have the camera running.
The other sharing systems are doing a good job policing the i
Re: (Score:2)
In the day of 10 cent video cameras you don't think the average autonomous cab won't have about a dozen cameras recording everything and if necessary causing damage charges against your credit card (of felony charges being laid before you even have arrived at your destination?)
Secondary and Tertiary effects (Score:3)
But other things such as taxis and car rentals will become blurred as what exactly is the difference when there is no driver. Parking lots can significantly shrink if your car can park 10 cars deep bumper to bumper and negotiate to get out when you call it. (assuming you don't just rent/taxi when you need one)
Commuting patterns change when cars can utilize the roads at near perfect efficiency and you can either doze or do work (behind the wheel). Plus my assumption is that an all robot filled road will allow cars to go insanely fast and convoy bumper to bumper slip-streaming each other. This then changes the distance that people are willing to regularly travel by car.
Then you get other changes such as an all robot road system would have an insignificant number of accidents rendering much of the in-car safety systems worthless. So you can then chop the weight of a car if it has no airbags, crumple zones, bumpers, etc. This also implies that old fashioned cars will have to be banned from the roadways unless augmented with a robotic system.
Other non-obvious changes would be that the entire car-insurance industry will be decimated. If cars basically stop crashing you don't need much in the way of adjusters, liability premiums, comprehensive premiums, etc. You will only need insurance for theft (presumably harder in a smart car) and things like trees falling on it. A car that crashes would mostly be due to a manufacturing defect.
Also the whole accident related industry would be smushed. This doesn't just include car repair and paint/parts manufacturing but towing, fire/rescue, and even a significant reduction in many hospital's emergency trauma wards.
Then you have the impact on the auto industry itself. They will have a boom from getting to replace nearly every non-robot car on the road and then with the onslaught of improvements that will follow they will get to replace the first few generations of cars quickly. But some manufacturers will miss the boat and get sidelined. Others will not realize that many sales are driven through destroyed cars and mess that up.
Then there is the quality of life issues. Old people will maintain their freedom much longer than before.
But the one that I am most looking forward to are the struggles that lawmakers will have. They will keep thinking in old ways such as stupid speed limits that have no safety purpose. Things like stop signs will be an oddity and then there is the fact that if all the cars are perfectly driven then nearly all fine revenue will drop to zero. Lawmakers love punishing the sinful and this will be a huge sin tax that will be lost. I suspect that even drinking and driving laws will be slow to change as it is my belief that they are driven as much by religious temperance types as by safety concerns. In this regard I would predict that an unlicensed driver or even empty car will be allowed on the roads before someone can be passed out drunk in the back seat while automatically being driven home. Lastly this may very
Indeed. Why own a car at all... (Score:2)
I don't see your point about drinking and driving. It's perfectly legal now to be a drunk passenger, if the car drives itself I don't see how that changes. Our legal system isn't always logical, but
Re: (Score:2)
Many people, perhaps most, keep a lot of personal conveniences in their car; that would translate to having to lug a lot of things to the shared car for each trip. Nuisance.
Having to wait for a shared car is not convenient, and is a problem in emergencies.
The distance the shared car travels empty from its storage location to the user is wasteful of fuel.
Some people take pride in maintaining a clean, attractive car.
There is a class of people for whom the shared car technique works, some combination of low ca
Re: (Score:2)
Quite. There's more than the utilitarian side of things.
For many people there's a component of status. Usually it's a mixture of status and style and practicality.
I also like my own car. A minority like me like a car to require a lot of skill and attention in order to just keep it on the road. The safety comes from the driver. Skinny tyres, little grip, too little power , and stick. Too bad it's all automatical choke nowadays because having a car only you know how to start is also charming.
Of course practic
Re: (Score:2, Redundant)
Exactly. Those are great examples.
And that's why I started writing the book. The secondary and tertiary effects are going to be fascinating. Why put up signs if computers will use GPS to know where they are? There will be so much more freedom for everyone young and old. It's going to be a big change. Almost bigger than the Internet.
Re: (Score:3)
I look forward to your book as this is a subject that truly excites me. (And scares me with the across the board job losses).
Re: (Score:2)
Its called disruption. Autonomous vehicles will be fought tooth and nail by more than a couple of entrenched interests fighting to keep the jobs or industry alive.
There will be (for example) more than one jurisdiction where the unions manage to get laws passed to require drivers in (for example) mass transit applications to "ensure public safety" but in reality to ensure that the unions won't shut things down with strikes.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
If you check out some of the airports like Orlando or DFW, the trains are automated. It's largely a union and a political thing. They could be automated but the cities choose to create jobs instead.
Implications (Score:2)
Humans will not be allowed to drive (Score:2)
My prediction (which I have been making since 2012).
By 2060 it will be illegal for humans to drive in the USA.