Ricochet is the best place on the internet to discuss the issues of the day, either through commenting on posts or writing your own for our active and dynamic community in a fully moderated environment. In addition, the Ricochet Audio Network offers over 50 original podcasts with new episodes released every day.
The End of the Auto Era as We Know It May Be Approaching Faster Than You Think
Bob Lutz is a former vice chairman and head of product development at General Motors. And in this essay for Automotive News, he declares the end of the auto industry as we know it:
It saddens me to say it, but we are approaching the end of the automotive era. The auto industry is on an accelerating change curve. For hundreds of years, the horse was the prime mover of humans and for the past 120 years it has been the automobile. Now we are approaching the end of the line for the automobile because travel will be in standardized modules. The end state will be the fully autonomous module with no capability for the driver to exercise command. You will call for it, it will arrive at your location, you’ll get in, input your destination and go to the freeway. . . .
Most of these standardized modules will be purchased and owned by the Ubers and Lyfts and God knows what other companies that will enter the transportation business in the future. A minority of individuals may elect to have personalized modules sitting at home so they can leave their vacation stuff and the kids’ soccer gear in them. They’ll still want that convenience. The vehicles, however, will no longer be driven by humans because in 15 to 20 years — at the latest — human-driven vehicles will be legislated off the highways. The tipping point will come when 20 to 30 percent of vehicles are fully autonomous. Countries will look at the accident statistics and figure out that human drivers are causing 99.9 percent of the accidents. . . .
CNBC recently asked me to comment on a study showing that people don’t want to buy an autonomous car because they would be scared of it. They don’t trust traditional automakers, so the only autonomous car they’d buy would have to come from Apple or Google. Only then would they trust it. My reply was that we don’t need public acceptance of autonomous vehicles at first. All we need is acceptance by the big fleets: Uber, Lyft, FedEx, UPS, the U.S. Postal Service, utility companies, delivery services. Amazon will probably buy a slew of them. These fleet owners will account for several million vehicles a year. Every few months they will order 100,000 low-end modules, 100,000 medium and 100,000 high-end. The low-cost provider that delivers the specification will get the business.
Of course don’t forget the second half of the phrase “creative destruction”:
Published in Culture, Economics, TechnologySo auto retailing will be OK for the next 10, maybe 15 years as the auto companies make autonomous vehicles that still carry the manufacturer’s brand and are still on the highway. But dealerships are ultimately doomed. And I think Automotive News is doomed. Car and Driver is done; Road & Track is done. They are all facing a finite future. They’ll be replaced by a magazine called Battery and Module read by the big fleets. The era of the human-driven automobile, its repair facilities, its dealerships, the media surrounding it — all will be gone in 20 years.
Oh, man, that movie. While I agree with the Man Must Explore and Conquer the Frontiers and Combat Ignorance! message, the movie is such a load of tendentious, didactic, condescending tripe.
I can’t see people giving up owning cars, even if they end up self-driving. Think of all the stuff you have in your car. I couldn’t possible manage all that if I didn’t own the thing. I also don’t see people waiting at home even 5 minutes to summon a ride. So, if this conversion is to occur, the time it will take is the time that it takes for the whole US fleet to turnover, which I think I’ve read is somewhere on the order of 20 years. I also see it taking place much more rapidly in the big cities and on the coasts. Flyover country I predict will be much more resistant. I know you can punch in the GPS coordinates, but I can’t see using a self driving car to take me to the trailhead for Mt. Massive, or Elephant Hill trailhead in the Needles district of Canyonlands.
Possible? Probable. Predictable.
Yes, you’re correct about Moore’s Law. But anyone who has made a similar bet on AI has lost. AI research has broken more hearts than anything else I can think of. A functional AI capable of navigating the real world – dependent only on its own sensory and computational capabilities seem a long long way off. Perhaps modern breakthroughs on machine learning might have helped – but I dont know that any regulator is going to approve a black box driving system that nobody understands how it operates.
Deceptively simpler. The development between Deep Blue to the AlphaGo Zero represents a wholly different way of approaching AI, partially because Go is much harder than Chess.
https://www.xkcd.com/1002/ :
I can’t remember, but I think between one-third and two-thirds of the roads in the United States aren’t even paved.
…
Many of the tech geeks live in he parts of California where it rarely rains, snows, or sleets.
However, California already has among “the worst-maintained roads in the country, with two-thirds of them in poor or mediocre condition.”
http://www.ocregister.com/2016/05/29/a-diet-to-give-california-drivers-indigestion/
…
Furthermore, except for very large and cold Alaska, most of the states with the worst roads are controlled by Democrats.
STATES WITH THE WORST HIGHWAYS
50. Alaska
49. Rhode Island
48. Hawaii
47. California
46. New Jersey
45. New York
44. Connecticut
43. Massachusetts
42. Minnesota
41. Colorado
STATES WITH THE BEST HIGHWAYS
1. North Dakota
2. Kansas
3. Wyoming
4. New Mexico
5. Montana
6. Nebraska
7. South Carolina
8. Missouri
9. South Dakota
10. Mississippi
Source: Reason Foundation
http://www.mercurynews.com/2013/07/01/california-highways-among-worst-in-the-nation/
Well, I agree with anonymous here. This will incrementally come to us and we will take to it gladly because of the improvements to our lives. We needn’t pretend that because the perfect isn’t feasible right now (remember SDI’s crtitics) then we won’t see it ever.
Just look at these cars that have position monitoring for cruise control. My daughter’s car is fantastic for that feature. I loved it. Also, I drove a rental recently that had a little light on the side mirrors that came on whenever someone was in my blind spot. My other daughter has a new Suburban that vibrates the side of the driver’s seat when someone approaches from either side in the rear.
Then, think about how much we all already use our navigation systems in various forms. I was driving a few years ago into a new section of Redmond, WA (lot of new neighborhoods there) and using my GPS unit and I realized that I didn’t even know how to get out of the place because I hadn’t been paying attention to the route. It kind of freaked me out. Now I do this all the time and it doesn’t bother me at all.
Here’s the main thing: the day self-driving cars (JW’s point above) become safer than human driven cars OR a feature is added that makes driving much safer then we and our insurance companies will drive the country in that direction without so much as a law to do it.
Laws will certainly start soon for electric only vehicles in the main downtown section or perhaps it will be cities that will restrict driving manually in their downtown grid. So government will play a part in this but the things that stay and become permanently and gladly adopted will have a persuasion all of their own.
I have a Nissan Leaf and there’s a $3,0000 wireless charging system that’s available and it charges your car whenever you drive into your garage. If they ever get something close to a week long battery — let’s say between 500 and 1,000 mile battery — then game over for most driving and range anxiety will go away and these electric cars will take over and eventually be much cheaper to buy and maintain.
(Most drivers aren’t that mean… Many American drivers are nice.)
Then stick the autonomous vehicles in all the Third World countries where…
And if we don’t take to it gladly, let me guess: If you like your car, you can keep your car.
If that’s the case in my city, it will eliminate 99% of the traffic; a statistically insignificant people have electric cars, and the spaces for charging them are few. Municipal parking ramps could be refitted, of course, at great expense; since the cost will be great, it would be spread out in a general levy, rather than charging the electric-car drivers for the new paradigm.
Let’s say that people don’t have electric cars now, but will buy them of their own free will very soon once their manifest advantages are apparent, including the government’s promise of Free Gas through subsidized recharging stations. Imagine some sad bitter clinger who likes his gas-powered stick-shift vehicle. Dude’s like some guy saying he has a right to a bump-stock.
Just like Europe, which has everything figured out. If our city restricts manual driving, it means businesses leave for the suburbs, where the government does not dictate your preferred mode of movement.
Persuasion in the progressive sense of “a law, noncompliance with which results in confiscation of property.”
[Eye’m afraid my iPad has contracted that new Apple bug where strange things sometimes happen when you type the ninth letter by itself, thus eye have to use “eye”]
As a go player, eye have to jump in here.
Yes, James, go has fewer types of pieces (only one, actually, a little stone, though each player has a few hundred of them) and rules which are much fewer and simpler than those for chess. But the possibilities for where a game may go are far far greater with go than they are for chess. Both go and chess push human ability and intelligence to the limit, making them equally hard for humans. But the seemingly infinite possibilites inherent to go make intuition critical for grand strategies since no mind – computer or human – is able to comprehend all the possibilities. Computers have (eye assume) long been better than humans at fighting local battles within a go game, but they always loose the war to humans since humans (until recently) were much better at the intuition-driven grand strategies. Google kinda sorta found a way to teach its go program a computer-based style of intuition that allowedit to surpass humans much earlier than anyone expected. ‘Twas quite a shock.
Funny how the otherwise-very-reasonable Ricochet membership always loses its grip with reality when the subject of autonomous cars pops up. We love to use the buggy-whip maker as the perfect example of Luddite thinking. Yet I’ll bet most buggy-whip makers uttered nearly-identical predictions about the future of the ICE automobile 120 years ago.
anonymous and Larry Koler are spot on: this isn’t about whether the autonomous car can drive better than you, it’s about whether 1 million autonomous vehicles cause less damage than 1 million meatbags.
Yes, you might be smarter than a Google car when a kid in a white snowsuit runs across your snow-covered street in December in Alaska while you’re driving uphill both ways to get to work. But the Google car will definitely be safer than your 21-year-old daughter at USC who rear-ends some family in their station wagon because she was trying to put on lipstick while driving. Which scenario do you think happens more frequently on a daily basis?
Or let’s be more generous: you did an excellent job teaching your daughter how to drive. But what about that drunk driver on the road at 2 AM as she’s driving back from volunteering at the Hillsdale homeless shelter? In other words, even if you’re God’s gift to driving, that other guy on the road who isn’t could still kill you. I’d feel much safer driving on many of our roads knowing that everyone else was being driven by UberGoogle.
The incremental part is key.
As you pointed out, the takeover of driving by machines has already been underway for decades and on the whole we’ve been happy with it. Somebody above mentioned humans being able to drive better in winter conditions, which is patently absurd – even features like anti-lock brakes improve our safety by breaking the direct human-tire chain and augmenting our intentions. So we already drive more safely in winter thanks to computers.
And if development of fully-autonomous vehicles does indeed to prove a little thornier than optimists like Lutz predict (and I imagine it will), there are plenty of possibilities for slowly adopting individual features before going whole hog. It may be that in 10 years we’ll still have a wheel and gas pedals, but highway driving will be mostly autonomous and even city driving will be augmented by more warning systems and even automated braking if an unexpected threat is encountered.
By the time fully-autonomous vehicles are ready for prime time, we may barely even notice the change.
This is already very possible today.
When I drove to the Elephant Hill trailhead three years ago, I got there by punching “Elephant hill parking” into Google maps and followed its directions.
I’ve also used Google Maps to navigate the mazes of anonymous gravel Forest Service roads often found out west. It’s still often hit-or-miss, but I think that is due much more to the quality of the underlying map data (i.e. many of those roads are not properly inventoried) than any software issues. But it’s certainly not an insurmountable problem.
I dated a girl in Redmond back before smartphones and personal GPS was a thing, and I would have killed for some type of computer assistance in that navigational nightmare. Arrived 30 minutes late the first time I picked her up from her home when I accidentally went to “45794 36482th St.” instead of “45794 36482th Ave.” or somesuch.
I question whether life is really better with more technology. The law has to change also. If cars are self driving, it seems liability for accidents would transfer to the manufacturer if humans are forbidden to operate them. The horse was driven by a man, who was also responsible for it’s care and liability; so is today’s car for the most part.
My GPS every day tells me I should be driving around the block to get into the parking lot at work instead of taking the closer entrance. It’s smart enough to figure out I’m going to work in the morning and going home in the evening without having to push any buttons. But it hasn’t figured out there’s an entrance there.
It has gotten better about telling me to take an exit and then immediately get back on the freeway though.
I didn’t mean to imply there aren’t any kinks in the system. There are still many, and I’ll bet every one of us has a story to tell. In my case, Google hasn’t figured out that the street I live on was intersected in the middle by a construction site three years ago and that through traffic hasn’t been possible since then. As a result, it’s nearly impossible to get an Uber to show up on time.
But there’s a big difference between ironing out the kinks in routine daily issues and the type of “here’s a rare edge case that computers will never be able to solve”, which was the tenor of the original comment I was responding to (and many others here).
I suspect this will come to pass eventually but the stated time frame is too aggressive by far. A good deal of driving is by rote but there is still many situations where judgement is necessary and road conditions do not lend themselves to automation. Sure I can see metro users buying into the Uber automation model but the rural parts of the country can barely get cell service so I doubt this automated vechile service is going to be able to service rural road structures.
I hear what you are saying, but I see a distinction. The analogy is a little different here. We’re not (yet) talking so much about a change in the means of propulsion (horse-and-buggy to combustion engine to electric vehicles). We were originally talking about autonomous control. Whether I drive a horse-and-buggy, or a 1971 Chevy Impala, or a 2016 Nissan Leaf – the fact is I am the one driving. By all means, give me safety-enhancing technological gadgets to improve my experience. Give me ever-improving GPS to help me never get lost. But take away my ability to directly control the vehicle, and now you are changing something fundamental. It is one thing to surrender that control when I hop on any sort of public transportation, but I’m still – in most cases – surrendering that control to another human operator somewhere down the line. But with the exception of buses and the like, those are all operating on closed systems (trains on tracks, etc.). It’s not the same with the automobile and the open road.
As for the trading liberty for safety argument? Well, to each his own on that, I suppose. We’ve traded a lot of liberty in the name of safety and security, so naturally we need to trade a little more so that the roads will be “safer.” Self-determination, although a idealistic abstraction, still has a strong hold, even in a simple matter of driving a car.
The convenience of technology has its advantages, but it has also created a collective loss in fundamental skills and know-how. With autonomous vehicles, why bother ever learning to drive? It’s already happening, and that’s a shame. Learning to drive is a rite of passage, of independence, or it used to be. Now we want to make it yet another anachronism?
I don’t think I’ve lost my grip on reality. Driving is a cherished privilege, and one that I prefer not to have taken away from me just because someone invented a self-driving car. Am I a Luddite in this regard? You bet I am.
I can acknowledge the trend, and the strong potential of both the market and my preferences being legislated away by those who say they know what’s good for me and our driving society. Does not mean I have to openly embrace the inevitability, much less like it. But if someone wants a Googlemobile, by all means, let the market prove it out and make it available. I just don’t want my choice to the contrary limited.
I would assume that if/when fully autonomous cars become available, they will be able to function without needing constant (or even any) access to cellular networks. Even in urban areas, a network can become overloaded or otherwise out of commission frequently enough.
But this time, the AI guys have it all figured out.
Jim, my Luddites comment was directed toward the comments (which always show up in threads on this topic) saying that autonomous cars will never be able to surpass humans’ average level of safety/ability when driving. I think that claim is hogwash.
Your point is a different one – not that autonomous cars are incapable of surpassing our skills, but that even if they do maintaining our own autonomy is a worthy goal, as is the skillset of driving. You definitely make some good points, although there are also plausible counterarguments for each. For example, any time you go flying nowadays, it’s a computer doing the flying. Sure, there are two humans watching the computers and occasionally telling them what to do, but as far as I know no modern airliner has a manual override – even manual inputs still get interpreted and executed by electronics.
Back to cars, I would hope that human driving is never outlawed, but I can envision a scenario in which people slowly but willingly migrate over to autonomous cars. For people with long commutes (or long drives in general) it could be a godsend. So I can imagine a large number of early adopters.
Assuming the technology works, at some point enough people will have autonomous cars that insurance companies will charge you more to drive your own car. Then you can actually place a dollar value on how important driving yourself is to you. Would you pay an extra $1,000 in premiums per year? Perhaps you might, but I’ll bet a majority of Americans would eventually choose the option that allows them to save both money and the stress of driving.
Self-driving cars, smart homes, robots flipping burgers, self-flushing toilets…..pretty soon we won’t have to do anything but exist…
Bryan & Occ,
Just to add to this, the real meaning of Gödel’s incompleteness theorems is that the AI guys will never have it all figured out. A little less hubris would come in really handy here.
Regards,
Jim
That will soon be legislated out of existence, too.
As several of us have pointed out, they don’t need to have it all figured out. Just figured out at a level equal to or better the average person on the road. AI doesn’t even need to compute at our level to be statistically safer than we are, since it can compensate with advantages such as not getting fatigued, distracted, or drunk.
And that hubris can go both ways: a lot of people are nowhere near as good drivers as they think (or as they used to be 30 years ago).
Bottom line: AI doesn’t need to be anywhere near perfect to drive better than us, since we set a pretty low bar to begin with.
Drew,
Now you’ve hit on it. The problem is that damned Human! Let’s get rid of that damned Human and just leave their social security number and bank account. We can have AI keeping track of everything and making the decisions too. They will be born, purchase stuff, have medical and legal problems, and die, all according to government norms. The biochemical Human need never exist. They will strictly be a being in cyberspace.
Fabulously more efficient, yes. Drew stay with this idea. I think it’s a winner.
Regards,
Jim
It just seems to me that the driverless car default position in situations of your day to day metro driving odd ball stuff (ie: big pothole, large puddle in the street, unplowed heavy snow piled in the street, black ice, accident/construction lane closure, a car double parked or sticking out into the lane too far, etc.) will be to stop or slow down to a crawl and thus all the traffic behind the driverless car stops or slows down to a crawl …. so now you are far safer and less prone to accident/injury/fatality ….. but your usual brutal 20 mile one hour rush hour commute has just transformed into a absurd three and a half hour crawl that most humans would never tolerate on a daily basis.(7 hours of commuting to and from work?)
It would be a situation of driving behind the most ridiculously overly “safe” driver you immediately realize you need get around and away from now when you go out to do your daily driving tasks.
Mendel,
I never once used the word perfect. Anyone familiar with software development knows how many unexpected problems come up when the product is put out for beta test (actual users in real situations). Every one of these, in this case, will result in death. You will be placing your fantasy end result against the reality that you don’t have a system anywhere near the capability of a mean human driver and you should be made responsible for the results.
There’s no free lunch. That works for the moral world as well as the economic one.
Regards,
Jim
Actually, I was being sarcastic. My whole life, AI has been Just Around the Corner(tm) just like Fusion. Now, by 2050, AI will cause the Singularity and the End Times will be upon us.
I detect a lot of millinialism in Singularity talk. Asension for the athiests.