« Again | Main | Super Bowl Weekend Thread of Openness »

January 31, 2018

Comments

"Looking at the average hourly earnings, however, ignores at least three very important factors: expansion of non-wage benefits, fall in the price of consumer goods and rise in price of services, such as education and healthcare."
Cost of Living vs. Wage Stagnation in the United States, 1979-2015: Inflation-adjusted wages haven't gone up, the cost of living has for the most part gone down.

Notably, bonuses deductible in 2017 will reduce taxes at a 35% clip, whereas any bonuses deductible in 2018 only result in tax savings at 21%....

So it's the more or less the same rationale that led me to pay my 2018 property taxes in 2017. I should be a CEO (or at least a CFO)!

hey, charles, it's nice to know that a minimum wage worker can purchase a crock pot for 1/2 the real cost as opposed to 1979 prices.

I'm sure somebody who has been bankrupted by health care costs will really appreciate that.

Mmmmmm ... stew.

"Every Labor Day, you can count on seeing a spate of news stories saying that "real wages" in the United States haven't grown since the 1970s. That's true, more or less, but the reason for the stagnation might surprise you. It's a complex story, but it boils down to this: Blame health care costs."
Health Care Costs Are the Reason You're Not Getting a Raise

Inflation-adjusted wages haven't gone up, the cost of living has for the most part gone down.

It was a tenet of Classical Economics (Smith, Ricardo, Malthus, Marx) that the optimum efficiency wage rate was at the level of sustenance and reproduction and no more. You kept workers alive and multiplyin,' but you absolutely did not want them able to save or accumulate. As global productivity increases, if it has, and the prices of commodities and worker consumption goods decreases, wages must decrease in order to kept investment funds available.

Workers standard of living can perhaps be better examined by income share, national labour income share vs capital income share. In all these arguments I always look at both: if worker standard of living according to some claims has been reasonably steady, then where has the massive increase in capital income come from. Michal Kalecki is one of the better at this level of analysis.

But it is complicate. Looked up the Smith vs Bentham argument:Smith claimed that high interest rates caused speculation and malinvestment. Keynes agreed. Many now are saying the very low interest rates cause bubbles and financialization. Shrug.

Michael Roberts is my technical Marxian economist. Link is to an article on productivity. Roberts is a declining profits guy, still focusing on manufacturing. I understand little of his stuff, but he and his commenters are erudite and consistent.

Wages, including benefits, have not kept up.
See cite.

Call me crazy, but employee benefits are a less expensive way to raise the overall labor pay package vs a raise in cash pay....the cost can be expensed and no employee share of Social Security to pay. Win. Win.

The two more rigorous papers include employer provided benefits in their calculations.

"There is a widespread but mistaken belief that wage stagnation has been partially caused by a shift of compensation toward benefits. Benefits have grown far less than most people realize, rising from 18.3 percent of compensation in 1979 to just 19.7 percent of compensation in 2014"

Per the EPI paper

bobm,
I have bookmarked Roberts' page. Thx.

So, what to do? One idea is a universal basic income. Others include increasing the minimum wage, getting the fed to stop clamping down on inflation worries any time the labor market gets tight, among others (see the third paper). Whatever the solution, it is clearly not on offer from the current U.S. government

As a solution, raising the minimum wage seems to miss the mark. If you lose your job due to automation, and there is no alternative for the same reason, what do you care how much you would make if you did have a job?

Maybe your spouse still has a minimum-wage job. Maybe demand increases because people who will actually spend their additional income get additional income, so you get another job that’s now needed to meet that additional demand.

There is a widespread but mistaken belief that wage stagnation has been partially caused by a shift of compensation toward benefits.

I pay about twice as much for health insurance as I did last year. Same plan, exactly. BCBS.

And that's a win, because for a while my employer tried a self-insured deal that capital-S sucked.

In that same period I've gone from (my) industry standard unlimited time off to 18 days PTO total, which includes both sick time, personal days, and vacation. Get the flu for a week, need some time for medical or family stuff, and you're looking at one or two weeks vacation per year.

When I get sick I work from home on a VPN. I haven't taken a "sick day" in probably ten years. I do take PTO for in-patient medical stuff that involves sedation. My employer thanks me for this.

I'm not at the top of the heap, but I'm in the foothills of the top of the heap. Most people would kill for my deal. The only folks who have it better are folks in the real nose-bleed upper reaches of tech, ditto for finance, and non-profits like universities that have big endowments and insured pools.

The reason more $$$$ goes to benefits is because the cost of health care is insanely out of control in this country. "Increased benefits" just means you get the same, or typically less, but pay more for it.

You do not see a net upside in your pocket at the end of the month.

Just an anecdotal data point. But aren't they all.

Another reason, I think, that increased productivity has not resulted in increased labor compensation is that most increased productivity has come from technology, in one form or other.

I.e., capital investment.

Capital creates the added value, so capital wants the gravy.

See also: Uber.

My own feeling about this is that we should address this at the level of ownership. Being an employee of an enterprise should confer some degree of ownership, however expressed.

That used to take the form of things like pensions, which is if nothing else a financial stake in the enterprise even after you no longer work there. Those, of course, are gone, and in the modern corporate environment where companies are started and folded and passed around like trading cards, they'd be worthless anyway. They only make sense in a context where the enterprise itself is something that is intended to endure.

But profit sharing, stock options for companies that have stock, employee-ownership, all serve the same purpose. We just don't have the mind-set to do stuff like that.

Basically I think a lot of people are just going to be screwed, because the trend has been pretty much consistent for over a generation now, and I see no change on the horizon.

As a solution, raising the minimum wage seems to miss the mark. If you lose your job due to automation, and there is no alternative for the same reason, what do you care how much you would make if you did have a job?

Yeah, that is more of a solution to the decoupling of labor productivity from compensation, rather than being replaced by robots.

Could work for the lucky few of us who might make convincing buttlers...

As an aside, self driving vehicles will hit the mainstream much sooner than people generally think.
That's an awful lot of jobs - though it's likely to bring down the cost of personal transport quite significantly....

(*Could work for...)

*Maybe demand increases because people who will actually spend their additional income get additional income, so you get another job that’s now needed to meet that additional demand.

Once the robots learn to service humans sexually in addition to their day jobs bolting car bumpers on and cleaning out the gutters, the rich won't need the rest of us any longer.

"That's an awful lot of jobs .."

Doing what?

We'll have self-driving cars right after we all pay to rebuild all of the infrastructure that currently doesn't support self-driving cars all that well.

And I'm not getting how self-driving cars reduces the cost of personal transport. Are they gonna give them away?

it's likely to bring down the cost of personal transport quite significantly....

John Quiggin at Crooked Timber recently did a thread on self-driving cars. Some commenters, far better-informed than I am, expressed a lot of skepticism, especially about the time frame. Industry hype is dazzling, the reality is far more complex. As I have said before, my own car can't even keep time correctly, and it's not like that's some big new technology that no one quite knows how to fine-tune yet.

And I'm not getting how self-driving cars reduces the cost of personal transport. Are they gonna give them away?

And -- as usual -- wrs.

We're going to have self-driving cars eventually, in fact we'll probably be forced to switch to them. Just not as soon as the industry-driven headlines would have us think. I'm far far less sure that we need them, or that if we had any sense and ability to plan collectively for the long term, that's what we'd choose. But someone is going to make mountains of money off them, and more tech is going to save us, has always been going to save us, so...more likely than not, we'll have them whether we like it or not.

"That's an awful lot of jobs .."

Ah, place a "lost" on that. Get it now.

Am I the only one contemplating a hybrid driverless car, one that incorporates orgasmatron functions into the vehicle.

Advertising slogan: "It gets you there two ways."

Self-driving cars on the cheap using smartphones: Comma.ai.

Interview with the developer: Super Hacker George Hotz: I Can Make Your Car Drive Itself for Under $1,000 (YouTube)

Introduction to and text of the interview: Super Hacker George Hotz: I Can Make Your Car Drive Itself for Under $1,000: Comma.ai aims to bring plug-and-play autonomy to the masses.

Currently, Hotz's development is about on par with the smart autopilots that many cars already have, but doing it with a smartphone instead of the car's onboard computers.

Hotz's idea is to gather data on how human drivers handle each bit of road and then have his driving system mimic that, with minimal external sensors.

what happens when the data isn't available?

The guy in the lane to your left is texting. His car drifts over into your lane.

To your right are bunch of people standing on the sidewalk.

Go, computer! Do the right thing.

Somebody isn't paying attention and they step into the street in front of you.

To your right is a row of parked cars.

To your left is a line of moving cars.

Go, computer! Do the right thing.

When cars drive themselves I'm gonna start walking.

Just started Wolfgang Streeck's How Will Capitalism End?, this from the introduction discussing a Randall Collins essay. Collins' Sociology of Philosophies is an application of his Interactive Ritual Chains and recommended, for those who have the time and inclination to read 1200 pages on the history Indian, Chinese, and Islamic philosophy, besides the usual. Anyhoo...

"What exactly does this crisis consist of? While labour has gradually been replaced by technology for the past two hundred years, with the rise of information technology and, in the very near future, artificial intelligence, that process is currently reaching its apogee, in at least two respects: first, it has vastly accelerated, and second, having in the second half of the twentieth century destroyed the manual working class, it is now attacking and about to destroy the middle class as well – in other words, the new petty bourgeoisie that is the very carrier of the neocapitalist and neoliberal lifestyle of ‘hard work and hard play’, of careerism-cum-consumerism, which, as will be discussed infra, may indeed be considered the indispensable cultural foundation of contemporary capitalism’s society. What Collins sees coming is a rapid appropriation of programming, managerial, clerical, administrative, and educational work by machinery intelligent enough even to design and create new, more advanced machinery. Electronicization will do to the middle class what mechanization has done to the working class, and it will do it much faster. The result will be unemployment in the order of 50 to 70 per cent by the middle of the century, hitting those who had hoped, by way of expensive education and disciplined job performance (in return for stagnant or declining wages), to escape the threat of redundancy attendant on the working classes. The benefits, meanwhile, will go to ‘a tiny capitalist class of robot owners’ who will become immeasurably rich. The drawback for them is, however, that they will increasingly find that their product ‘cannot be sold because too few persons have enough income to buy it. Extrapolating this underlying tendency’, Collins writes, ‘Marx and Engels predicted the downfall of capitalism and its replacement with socialism’ (p. 39), and this is what Collins also predicts."

I remain slightly agnostic on the Labor Theory of Value but the Marxian prediction of our robot future based on the LTV says that although the robots can slap out commodities and stuff by the container load, since there will be ever decreasing human labor adding value to production, there will not be enough profits or surplus to fund or finance your beoootiful Basic Income or Job Guarantee or Universal Health Care. In order for Koch or Gates or Trump or Obama to stay wealthy their intangible assets need to be continually renewed by capitalist growth. Etc Etc.

Making stuff and making money are two different processes.

1st productivity diverges from wages, the productivity cannot sustain profits, then it gets burned down.

and I haven't even gotten into stuff like black ice, or wet leaves on the road, or sandy shoulders.

Computers, and all forms of rationalized systems, excel when conditions have been optimized for their operation.

It’s 2018, so where are the self-driving cars?: It’s time for a reality check

Is driving a car less, or more, difficult than judging a case in appellate court?

I should think strict constructionists would heartily approve of robot judges, if it could be done. And since reaction time is much less of a constraint on the bench than on the road, I say it could be done.

Passing lightly over fantasies of R2D2 arguing against C3PO in a case before judge HAL, I will repeat something I've been saying all millenium:

Robots make lousy customers. The Economy needs consumers just as much as it needs producers in order to be an Economy. The ultimate success of AI will be a "post-economy" world. Whether that world will be heaven or hell depends on whether humans (as sentient beings, not "economic actors") are smarter than we have reason to hope at present.

--TP

65mph, there's an object 50m ahead on the road. shape indeterminate. 5m x .5m, roughly.

it's moving! erratically!

what should the car do?

evasive maneuvers! STOP! the road is wet! skidding! object is moving into our path! panic!

it's actually a paper bag, blowing in the wind. any human would've driven right over it.

(.5m x .5m . f.u. overloaded dot operator.)

"I should think strict constructionists would heartily approve of robot judges."

Even though they believe the Constitution to be an unerring algorithm with no variables, where does it say "robot judges" in the Constitution?

Robots make lousy customers. The Economy needs consumers just as much as it needs producers in order to be an Economy.

Funny. A few months back some family members out west (all good capitalists - some owning their own businesses) were complaining on facebook about the new, higher minimum wage in Arizona (or Phoenix?). Not sure if it was the whole state, but whatever. Anyway, they all said it was going cause business owners to fire their employees (a la fully automated fast-food restaurants and such) and replace them with robots.

I mentioned that, once there weren't enough people with jobs to buy anything, they would have to create robot customers. But the good part was that they could be programmed to optimize the velocity of money to create maximum economic growth.

I was attempting slyly to introduce the idea that people at the bottom will spend their additional income far more readily than the people already making a good living, but I'm sure I was the only one who knew what I was hinting at.

Go, computer! Do the right thing.

It might be very difficult to take a fully autonomous car that can see 360° and whose attention never waivers by surprise. And, if taken by surprise, its reflexes would be magnitudes faster than any human's. Therefore, the car haveing to make decisions about who to save and who to harm might be very rare.

In those rare cases, the best approach might be that the safety of the car's occupant(s) comes first and the courts sort out the consequences.

And, if taken by surprise, its reflexes would be magnitudes faster than any human's.

at 65mph, reflexes take a back seat to planning because twitching is the enemy of momentum.

"Since self-driving cars will be safer than human-driven vehicles, the right thing to do is to encourage people to trust them so that their adoption and deployment will proceed as quickly as reasonably possible. Consequently, programming cars to preserve the lives of their passengers is the ethical thing to do."
Your Self-Driving Car: Minimize Casualties or Save You?: What algorithms will actually save the most lives in the long run?

I think self-driving cars will work best if all, or at least the vast majority, of the cars on the road are self-driving. You don't have to worry about the guy next to you texting and drifting into your lane if he's in a self-driving car. Drifting into another lane, or at least drifting toward another car, is the sort of thing self-driving cars won't do.

And I imagine they'll go with the flow of traffic, unlike some humans, who will do 50 when everyone else is going between 65 and 85, or who will try to do 100, weaving in between the other cars, under those same conditions.

They should merge in a more-coordinated fashion, or so I would hope. They shouldn't experience road rage, either. Nor should they try to speed through intersections before the light turns red.

I'm not blindly accepting of self-driving cars in the near term, but I can see a future where they'll be far better than humans, particularly if they don't have to mix with human drivers.

They should merge in a more-coordinated fashion, or so I would hope. They shouldn't experience road rage, either. Nor should they try to speed through intersections before the light turns red.

oh but i think they will.

some people aren't going to be satisfied with a car that patiently and politely navigates the smoothly flowing stream of traffic - especially if it's going to make them late. they're going to want a car that will take advantage of gaps. they'll want a car that will take advantage of other cars' well-behaved passivity, and drive aggressively. they'll want a car that will take chances because it will keep them from feeling like an average schmuck. they'll want to ride in exciting, fast, dominating cars.

and so car makers will make cars that will exploit the fact that the majority of other cars being well-behaved and defensive.

"When we first deployed a robot in a store, the associates were the people that understood it first. This boring, repetitive task of scanning the shelves—we have yet to meet someone who has liked to do that. Employees instantly become the advocates for the robot."
Walmart’s new robots are loved by staff—and ignored by customers: Bossa Nova is creating robotic coworkers for the retail world.

and so car makers will make cars that will exploit the fact that the majority of other cars being well-behaved and defensive.

So rich guys in beemers will still "drive" like dicks?

Pay-productivity gap nuance:

What's Behind the Growing Pay-Productivity Gap?: Pt 1, Hypotheses

What's Behind the Growing Pay-Productivity Gap?: Pt 2, Measurement Problems

If autonomous cars communicated with each other, a car whose occupant is in a hurry might negotiate micro-payments to other cars for them to defer to it.

If cars were drive cooperatively, whether by humans or computers, everyone would get where they were going faster. Being in a hurry and taking advantage of the lack of aggression in other drivers leads to a fallacy of composition - not that I can claim not to do it regularly.

Hypocrisy is the tribute vice pays to virtue, right?

A guy could make a living holding up traffic.

Maybe pedestrians could spend their day negotiating the payout of hundreds of micro-payments just to cross the street.

I, for one, will miss the regional differences in driving habits and behavior.

The New York minute will go the way of the dodo.

No more totally unpredictable demolition derby in
Boston, with the cutting in at 38 mph sans signals.

I doubt Manila drivers are going to cotton to uniform traffic patterns.

And the Italians. Well, the passengers in the autonomous cars will have BOTH hands free for insulting hand signals.

I wonder if it will be programmed for nickel rides?

"The patent, first spotted by Motor1, describes an autonomous police vehicle that would be able to detect infractions performed by another vehicle, either on its own or in conjunction with surveillance cameras and/or road-side sensors.

The AI-powered police car could then remotely issue citations or pursue the vehicle. Or (and this is where it gets really creepy), “the method may further involve the processor remotely executing one or more actions with respect to the first vehicle,” according to the patent."
Ford files a patent for an autonomous police car

Therefore, the car haveing to make decisions about who to save and who to harm might be very rare.

You go first, and tell us all how it went.

I think self-driving cars will work best if all, or at least the vast majority, of the cars on the road are self-driving.

In the meantime, Katy bar the door.

Leaving out the "guy in the left lane texting" thing, there are pedestrians, animals, bad road conditions, bad weather, and other kinds of vehicles like bicycles and motorcycles. There are flooded roads, tree limbs down in the road, and other occasional unplanned obstructions. In my neighborhood, there are frequent flocks of belligerent turkeys who will not yield to anyone for any reason.

We'll have to isolate self-driving vehicular traffic from the rest of the world to make it actually safe. We may have to condition the roadways and abutting surfaces and structures so that the ingenious robot driver can easily recognize their presence at speed.

Basically, this is going to turn into a society-wide re-engineering of the public transportation infrastructure.

Gizmos are fun, especially robotic ones, but I don't see this as a good idea. All in all, I'd rather have a bus.

If the cars communicate, they open themselves to malignant influences (=hacking). The 9/11 guys will look like pikers once some group (or even single person) does the first one million car crash. Or, less violent, what will happen when on election day all cars registered by members of one party will not start (after all polling places have been put out of walking reach)?

Ford files a patent for an autonomous police car

My wife and I were in Italy last fall. You don't see highway patrol cars in Italy. There are cameras on the roads, if you speed they take your picture and mail you a ticket.

Works great. Ask me how I know.

Leaving out the "guy in the left lane texting" thing, there are pedestrians, animals, bad road conditions, bad weather, and other kinds of vehicles like bicycles and motorcycles. There are flooded roads, tree limbs down in the road, and other occasional unplanned obstructions...

No doubt, and stuff will go wrong, but the systems will learn from that - and the experience of one autonomous vehicle is pooled with all the rest.
It's an interative process -
https://www.theatlantic.com/technology/archive/2017/08/inside-waymos-secret-testing-and-simulation-facilities/537648/

- cars that were doing 2mph on Alphabet's test facility roads drove 30k miles in California last November without a single 'disengagement', and will take part in a pilot commercial taxi service next year in Phoenix:
https://www.theatlantic.com/technology/archive/2018/01/waymo-maintains-huge-lead-in-self-driving-car-race/552059/

The autonomous driving system might be a much, much slower learner than a person - but it's driving orders of magnitude more miles, and that system will see many many more of those odd driving situations than any single person will experience in a lifetime.

I suspect we won't see much dramatic for a couple of years, and the pilot in Phoenix will probably have limitations and problems that will encourage the
skeptics. A couple more years and it will be conventional wisdom that these things are going to take over.

And I'm not getting how self-driving cars reduces the cost of personal transport. Are they gonna give them away?...

Assuming their ubiquity, it would mean the average household would no longer need to own a car - think a far more efficient and cheaper Uber.

The other part of the cost reduction will be the move to electric vehicles, which need far less servicing than ICE vehicles and will have a vehicle life of around 250k miles.

The economics for autonomous electrical goods vehicles are probably even more compelling for their operators.

(Not going to happen overnight, of course.)

and the experience of one autonomous vehicle is pooled with all the rest.

Much more efficient than training one hormone infused, alcohol saturated biological unit at a time.

russell: All in all, I'd rather have a bus.

There are people who argue that private cars, autonomous or not, will soon become unusable. Like proliferating bacteria choking on their own excrement, they will congest any plausible urban road network into immobility.

Yes, I know the old line about a restaurant getting so crowded that nobody goes there any more.

And I know that on-board AI, probably combined with an infrastructure of "intelligent" traffic management systems able to communicate with and even command the "autonomous" vehicles, could delay the terminal congestion for a decade or so.

But the worldwide trend toward urbanization seems relentless, so it's a race between AI and demography.

--TP

Regarding JanieM' car clock comment, crappy car electrical/electronic equipment standards are an issue that is being urgently addressed, as even without autonomous vehicles it's a huge safety issue:
https://semiengineering.com/rethinking-car-design/

Basically, this is going to turn into a society-wide re-engineering of the public transportation infrastructure.

More like a society wide re-engineering.

The US, despite leading technologically, is probably one of the toughest countries for autonomous vehicles to take over. It will probably happen a lot sooner somewhere like Singapore.
Some will welcome it. Most of my contemporaries learned to drive in their teens; a much lower proportion of my kids contemporaries drive at all.

and the experience of one autonomous vehicle is pooled with all the rest

So all these companies, madly chasing glory and profit, are going to share their data?

*****

Assuming their ubiquity, it would mean the average household would no longer need to own a car - think a far more efficient and cheaper Uber.

Two things being tacitly mushed together here: driverless cars, and the question of who will own them.

If people don't have private cars anymore, what happens when everyone still wants to get to work at the same time, but not quite, so having four passengers in the car isn’t any more palatable to the average commuter than it is now? There will still need to be a fleet of cars sufficient for peak times, with most of them sitting idle for most of the day.

*****

From Nigel’s first link:

At any time, there are now 25,000 virtual self-driving cars making their way through fully modeled versions of Austin, Mountain View, and Phoenix, as well as test-track scenarios.
[a picture caption]A piece of the world that Waymo has modeled for its cars in Chandler, Arizona, near Phoenix

So, at the moment, not Phoenix but a “fully modeled” version of Phoenix.

All aside from what “fully modeled” might mean, it’s no accident that this stuff isn’t happening in New England. I believe it was russell who mentioned black ice…….

*****

From Nigel’s comment: cars that were doing 2mph on Alphabet's test facility roads drove 30k miles in California last November without a single 'disengagement', and will take part in a pilot commercial taxi service next year in Phoenix:

From a commenter at Crooked Timber:

You can find dozens of links to a story on / around 7 November, all of which offer the headline announcements that 1) Waymo has removed drivers from its driverless cars, and 2) Waymo has launched a driverless car service in Phoenix.
Read the actual stories and you will learn that, in fact, the cars still have a Waymo employee in them; that, in fact, they have not launched a car service, but say they will “soon”; and that the test they’re running involves a geo-fenced area in the suburb of Chandler, not the city of Phoenix.

From such an article, dated 11/7/17:

Waymo is working on an autonomous taxi fleet, modeled after the pilot program it currently offers outside of Phoenix. The big difference here is that once fully launched, Waymo's service won't have a driver behind the wheel. Additionally, the company will up its fleet of self-driving Chrysler Pacifica minivans from 100 to 500. As you'd imagine, the entire ride-hailing process will be handled via a mobile app. While the initial service area will be limited to the Phoenix metro area, in a post on Medium, the company says eventually it will "cover a region that's larger than the size of Greater London." Waymo says that the public will be able to start taking autonomous rides "over the next few months" in the agricultural suburb, Chandler, Arizona.
Ars Technica writes that at first the taxis will have a Waymo employee in them so they can hit the "pull over" button in the event of an emergency.

(my emphasis)

So where is my George Jetson flying car? You mean to tell me the time has not yet come?

I love this part; I didn't even catch it at first:

At any time, there are now 25,000 virtual self-driving cars making their way through fully modeled versions of Austin, Mountain View, and Phoenix, as well as test-track scenarios.

*Virtual* self-driving cars.....in fully modeled cities......... Mixed in with human drivers? Dogs running into the road? Trucks pulling out to pass without any warning? Cyclists...doing what cyclists do....

Futurist - (a.)one whose concept of the future is unbridled by any social, economic, or physical constraint; (b.) a profligate pollyana; (c.) delusional techno-optimist; (d.) religious fanatic.

I have a lot of faith in the possibilities of self-driving cars, but not our current ability to shepherd our society into the future, even though it seems imminent. I suggest we start studying this technology. We are way too likely to be headed here.

I've read that virtual driving allows the researchers to create situations that have a low probability of occurring on real streets. And perhaps dangerous if staged.

I think you're missing the point of the first article, JanieM.

Simulation is one thing; real world tests another. The results from each feed back to the other.
Simulations aren't prefect, but they are a far quicker and cheaper way of testing changes to the system. So,e,of,which find their way back to the real world tests.

None of the objections you raise seem fundamental to me.
Black ice is just another engineering problem, once solved, solved (and most drivers in the U.K. these days have a problem with half an inch of slush); the Phoenix thing is just what I said - a pilot program - and the presence of drivers possibly regulatory.

In any event, we can see who is right in three or four years.

If people don't have private cars anymore, what happens when everyone still wants to get to work at the same time, but not quite, so having four passengers in the car isn’t any more palatable to the average commuter than it is now?

Peak pricing, I guess, which would at least mitigate the issue.
Clearly the biggest problem will be where there are the longest commutes, but it ought to at least halve the number of vehicles needed.

Thinking of my home town, the morning commute is between 15 and 30 minutes depending on congestion. There is similar commuter traffic in the opposite direction, and the morning and evening commute periods last a couple of hours. A wholly autonomous vehicle system could probably function with a quarter the number of vehicles.

So airplanes have pilots, but there's a lot of self-flying already, and has been for decades.

The reason I'm for self-driving cars is to give agency to the car companies for things like happened in that New Yorker article (did I discover it here? Was it you, Nigel? Sorry to forget).

I've had some non-human-injury accidents that were my fault, and I dread the idea that someone would ever be hurt by the possibility that I zoned out or couldn't see for a second or two.

Peak pricing, I guess, which would at least mitigate the issue.

Could work similar to the Uber model. Instead of one company owning enough vehicles to cover peak demand, it could contract with small companies and individuals who own one or more vehicles that they could clean and maintain during off-peak times.

Peak pricing would still be needed to reduce congestion and the amount of idle stock during off-peak times.

There have been tests where airliners have taken off, flown across the country and landed without anyone touching the controls. Autonomous aircraft would be closer to reality if US air traffic control is ever brought into the 21st century.

Peak pricing would still be needed to reduce congestion and the amount of idle stock during off-peak times.

Well, that's a bit contradictory. Peak pricing would encourage more cars to enter the fray. What would be needed is for the car owners to have to pay peak priced road fees.

So rich guys in beemers will still "drive" like dicks?

Not at all. They will pay big bucks for an app which will both operate their car like a dick is driving AND, more important, let them mess with the computers of the cars around them, so as to show the little people who is really important.

From Autonomous Vehicle Implementation Predictions, dated 1/26/18, by the Victoria Transport Policy Institute, a Canadian think tank (disclaimer, I haven't read the whole report yet, it's long):

This report explores autonomous (also called self-driving, driverless or robotic) vehicle benefits and costs, and impacts on transportation planning issues. It investigates how quickly self-driving vehicles are likely to be developed and deployed based on experience with previous vehicle technologies, their benefits and costs, and how they are likely to affect travel demands and planning decisions such as optimal road, parking and public transit supply. This analysis indicates that some benefits, such as more independent mobility for affluent non-drivers, may begin in the 2020s or 2030s, but most impacts, including reduced traffic and parking congestion (and therefore infrastructure savings), independent mobility for low-income people (and therefore reduced need for public transit), increased safety, energy conservation and pollution reductions, will only be significant when autonomous vehicles become common and affordable, probably in the 2040s to 2050s, and some benefits may require prohibiting human-driven vehicles on certain roadways, which could take even longer.

some benefits may require prohibiting human-driven vehicles on certain roadways

As is my wont, uninformed speculation on bigger issues. I can see a time when Americans, with their toxic mix of 'Because freedom!' and selective blindness at the impact of decisions, are going to reject the things that are required to make something like autonomous driving a reality.

https://www.japantimes.co.jp/news/2017/01/04/national/traffic-deaths-fall-4000-japan-first-time-67-years/#.WnPFjdKLSUk

In 2016, the number of people who died in traffic accidents dropped by 213, or 5.2 percent, from a year earlier to 3,904.
[...]
An agency official attributed the decline to [...] improved vehicle performance, like automatic brakes; [...]

I realize that there are lies, damn lies and statistics, but Japan has roughly half the population of the US, but the number of traffic deaths is 1/8th of the US. There are obviously a lot of factors (decent public transport, strong incentive to buy a new car at least every 5-6 years because of the shaken regime, a higher barrier to getting a license and probably other stuff I'm not thinking of)

I'm thinking about this now cause I've been in a discussion with a couple of people about medical testing results here in Japan. In Japan, large scale employesr aremandated to have health checks done on their employees once a year. There is a whole infrastructure (mobile vans, nurses, doctors) who come to large employers to give these health checks.

The people I'm talking to are adamant that their employers had absolutely no right to know their medical condition and have engaged in a long battle with them to not do the health checks. The compromise (which is built into the system) is that the person would go to a private office that specializes in these health checks, get the medical info and keep it on file at their home for 5 years, so if a medical issue came up, they would have some baselines.

Perhaps it is my recent medical issues that have me thinking like this, but I am to the point where I don't give a f**k who knows about my medical condition. I guess this is an old person's thing. And I have some colleagues who had serious conditions caught early, so are still working. So I'm wondering about people who refuse. I can see their point, but I think it has to bow down to the upcoming realities.

Now, I can see that some US companies, if they are paying for insurance, would be happy to strike off someone who is going to cost them a lot. I could see them negotiating with the insurance company to cut loose high risk employers in order to get a better package. Maybe I'm too trusting of Japanese employers. But to get the kind of savings that makes a government paid medical insurance system possible, you have to get a lot more data and process it automatically. I see the same thing with autonomous cars, in that the countries that will reap the benefits are going to be the ones that don't act like jerks about change. Which is going to leave out the US...

in 1998—approximately 194 billion labor hours.1 What this means is that there was ultimately no growth at all in the number of hours worked over this 15-year period, despite the fact that the U.S population gained over 40 million people during that time, and despite the fact that there were thousands of new businesses established during that time." That seems remarkable to me.

Could be. Factors to consider are labor force participation rate (declining) and labor productivity (plodding along).

think a far more efficient and cheaper Uber

Uber drivers work for short money, and they supply their own car. top that.

seriously, by the time this is actually a thing, i'll be dead or retired and riding my bike most of the time. my motor outings will mostly be taking my lawn waste to the dump.

so you kids have fun.

i'll take the bus.

Assuming their ubiquity

this reminds me of the old joke about the mathematician stranded on the desert island with a case of canned food.

and i loved the virtual car thing. when i need a virtual ride somewhere, i'll call one.

i'm sure we'll have driverless cars at some point. and i'm sure we'll all have to jump through hoops of all sorts to make them feasible, and i'm sure we'll end up doing it anyway, because (a) gadgets and (b) there's lots of money to be made.

i don't really see the upside, but wtf do i know. i'm old, i still use a flip phone. it seems like a lot of gee-whiz bullshit, to me.

have fun kids.

i'll take the bus.

Self-Driving Shuttle Buses Might Be the Future of Transportation

Autonomous Buses Will Revolutionize Public Transportation, but at What Cost?

China launches Alphaba public self-driving bus project

Self-driving bus starts first route in Germany

There have been tests where airliners have taken off, flown across the country and landed without anyone touching the controls

this comes up every time the driverless car things get discussed.

if we institute a federal regime where every vehicle has to declare its route of travel before embarking, including every street it will travel on and every on and off ramp it will take, and all entry to and from this network of roads is exclusively from some small number of embarkation and arrival points, and a medium sized army of specialist federal employees choreographs the departure and arrival of each vehicle along with the exact speed they will travel en route and the distance they will maintain from every other vehicle while en route, while traveling through a medium with no physical barriers except weather, then yes, drverless cars might be a more tractable problem.

the vehicles are not really the problem. given optimal conditions, i'm sure a driverless vehicle can het from point a to point b without incident.

if a roomba can do it, i'm sure a car can.

it's the optimal condition issue that's the problem. where i live, optimal conditions do not prevail. quality and type of road surface, weather conditions, whether the road that your GPS tells you to take is even there today, are all jump balls.

and until we get to the point where we can "assume ubiquity", masshole drivers are also a reality. world leaders in making right turns from the passing lane, preferably without signaling.

i had a mr robot when i was a kid, too. it was a great toy.

all of you kids please go do your in vivo driverless car experimenting in your own neighborhoods.

Self-Driving Shuttle Buses Might Be the Future of Transportation

i'd take a self-driving bus before venturing into a world of self-driving cars, anytime.

predictable, consistent routes of travel. dedicated lanes in some places.

an order of magnitude simpler case.

plus, much more conservative use of resources.

"predictable, consistent routes of travel. dedicated lanes in some places."

Might as well put in trolley tracks to everywhere.

At least trolley have enough mass that they can plow pass masshole drivers.

Could we install Robotic Death Lasers to deal with the masshole (or joisey) drivers? that might make it politically feasible.

Hey, trollies work a century ago. Around here, the Key System had great coverage, was cheap, etc. So what happened? General Moters bought the company and shut it down. So they could sell more diesel busses. Ah, capitalism!

Peak pricing would encourage more cars to enter the fray.,,

And passengers to leave it.

i'd take a self-driving bus before venturing into a world of self-driving cars, anytime..

Autonomous vehicles will likely blur the line between taxis and buses.

all of you kids please go do your in vivo driverless car experimenting in your own neighborhoods.

That's pretty well what will happen. Autonomous vehicles need regulatory clearance; they'll happen where they are allowed.
Like I said above, the US will likely be a laggard in adopting them; in somewhere like Singapore (a much more constrained, and far smaller environment; already heavily regulated; driving is a huge pain in the ass rather than an emblem of personal freedom...) they could become ubiquitous quite quickly.

So what happened? General Moters bought the company and shut it down. So they could sell more diesel busses. Ah, capitalism!...

This is more likely to shut a lot of traditional auto makers down in a couple of decades' time. They are late to the party, and the auto manufacturing business will be significantly smaller overall.

Might as well put in trolley tracks to everywhere.

buses need less up-front infrastrucure investment, but trollies are ok with me.

cars in general make sense in less densely populated areas. buses trolleys subways etc make sense in more densely populated areas.

cars are really convenient. you can go where you like, when you like, and they take you door to door. unless you are in a place where he parking lot is further from your destination than the bus stop. they are also a really large, expensive, resource intensive way to move a single digit number of people from one place to another.

there's a great case to be made for driverless cars for getting folks who can't / shouldn't be driving from one place to another. most of those needs can also be addressed by cabs, ride shares, or other services.

to the degree that driverless cars increase the number of cars on the road, i would actually see that as a significant downside.

but mostly i see driverless cars as a huge investment of intellectual, engineering, and capital investment, for at most a very modest upside.

people talk about them like they're gonna be some life-changing innovation. i'm not seeing it. the number of cases where there's a big advantage in having your car drive you versus just driving yourself is just not that large. and a lot of those cases - for example the sheer waste of human time spent sitting in traffic - would be better addressed by things other than self-driving cars.

the biggest win i see is that now folks will be able to go out, have a few cocktails, let their car drive them home, and not have to get a cab back to the bar the next day to pick their car up.

a social good, of sorts, but not the basis of an industry.

also not considered here is the fact that a lot of people would simply prefer to drive, because they enjoy it. driving is, or at least can be, fun. so it's likely that the massholes will always be with us.

mostly my feelings about self driving cars are:

that's kind of cool!
so what?
this is what we're spending billions and billions of dollars on?

i'm sure they are on their way, and in this country at least (the US) i'm sure they will mostly be a cool toy for upscale suburbanites, and a way to lay off people who drive trucks and buses and (maybe) cabs.

o brave new world that has such gizmos in it.

Low wage growth even in a tight labor market?

https://www.nytimes.com/interactive/2018/02/01/business/economy/wages-salaries-job-market.html

the biggest win i see is that now folks will be able to go out, have a few cocktails, let their car drive them home, and not have to get a cab back to the bar the next day to pick their car up...

How about you never need a parking space at work ever again ?

It's potentially a 24hr personal chauffeur, at a price anyone who now drives can afford.

Low wage growth even in a tight labor market?

Not surprising under conditions of market oligarchy?

Road traffic deaths per 100,000 population (WHO numbers)

USA: 10.6
Japan: 3.7
UK: 2.9

If only we had our freedom.

Typo there, should be 4.7 for Japan. LJ's numbers are lower: I suspect the WHO's are more comparable, but somewhat out of date.

How about you never need a parking space at work ever again ?

that'd be great.

i live in what is reputed to be the 6th most gridlocked urban area in the US. the volume of traffic from about 6AM to about 9:30 AM is extreme. ditto 4:30 to 7:00 PM.

the rest of the day and weekends, not so much.

even assuming you could cut the total number of cars needed to move that many people in half, or even to one third, there will be a lot of cars parked somewhere for a lot of the day.

your comment leads me to ask:

are we all going to share the same set of cars? can you smoke in the car? smoke a cigar? eat? make out with your spouse or date? who cleans the burger wrappers out of the car? do they do this after every ride?

does at least one passenger have to be an adult? do you want to get picked up by a car that just gave six eight year olds a ride home from the beach? or a young parent who just changed baby's nappie in the car? or a pair of very horny teen-agers? or some old geezerlike me who might not have complete control over their bodily fluids? somebody who is stunningly flatulent?

the great failure of technology is mistaking the whiteboard for real life.

Why don't all these questions apply to busses?

even assuming you could cut the total number of cars needed to move that many people in half, or even to one third, there will be a lot of cars parked somewhere for a lot of the day...

Some of them - the rest driving around providing those Uber like services throughout the day. Which is rather the point.

are we all going to share the same set of cars? can you smoke in the car? smoke a cigar? eat? make out with your spouse or date? who cleans the burger wrappers out of the car? do they do this after every ride?...

Who knows ? Several years to find out the answers to those questions.
Nothing to stop a fleet setting aside a percentage of their cars as smoking vehicles should they wish.
Who cleans the restroom in McDonalds ?
If you crap up the car, maybe it costs a bit more for the ride. Maybe you have utility cars for the rougher jobs.

Like the black ice thing, these are all solvable problems (if a little more complicated).

if we're sharing the cars, then we have taxis. we already have taxis. i don't know anyone who takes a taxi to work.

Cost too much, takes too long to arrive.
What if it didn't ?

From today's report:

Average hourly earnings rose 0.3% over the prior month and 2.9% over the prior year. This year-on-year wage increase is the best since 2009.

takes too long to arrive.
What if it didn't ?

it would. taxi owners couldn't keep enough taxis out there to serve everyone quickly during peak commute times.

i live 25 miles away from my job, in the very rural suburbs. there aren't going to be any taxis anywhere near where i live. we don't even have any Uber drivers in my town.

Average hourly earnings rose 0.3% over the prior month and 2.9% over the prior year. This year-on-year wage increase is the best since 2009.

the Obama economy keeps chugging along.

If we're talking about self-driving ride-service, the cars could be equipped with interior cameras, which would be known up-front. That would discipline most people against being messy and/or sexy.

Did someone propose that everyone, everywhere would only ride in shared cars?

Why don't all these questions apply to busses?

Some do, some don't. In general, there's a driver on a bus who keeps things between the lines.

Average hourly earnings rose 0.3% over the prior month and 2.9% over the prior year. This year-on-year wage increase is the best since 2009.

Thank you Obama!

Nigel and hairshirt, but especially Nigel, you guys seem very enthusiastic about self-driving cars. I for one will not stand in the way of this great leap forward in transportation.

Bon chance!

My point overall is that they will solve some problems and create some others. Like everything else.

It's still a bunch of cars driving around, just like now. And all of the things that are going to make it practical are going to require lots of corollary changes, which will be nice for some folks and not so nice for others.

If we're going to invest billions of dollars and thousands of man-years in R&D to transform public transportation, self-driving cars would not be at the top of my list. But it ain't my money, so somebody else will get to make that decision.

You said you'd take a driverless bus. Either way, humanity has survived the scourge of public toilets. We used to have telephone booths, too. They were neat.

Did someone propose that everyone, everywhere would only ride in shared cars?

russell asked "are we all going to share the same set of cars?"

If we're going to invest billions of dollars and thousands of man-years in R&D to transform public transportation, self-driving cars would not be at the top of my list.

I'm not sure it would be at the top of mine, either, if I had such a list. I mostly find the concept interesting, particularly if it ultimately involves cars working on a networked system creating optimal traffic flows (on the smaller scale - cooperative merging, intelligent traffic-signal timing, etc.) and routing (on the larger scale - arrival time, energy economy, wear-and-tear, or whatever else). That's all pretty future-y stuff.

What I can see in the nearer term is cars with more and more self-management, if not being completely self-driving. That's already underway. I don't think it would be difficult to have cars in more-urbanized areas made to obey traffic signals, prevented from deviating too much from the speed of other traffic, avoiding rear-end collisions and such. Basically letting people drive, but preventing them from doing the particularly stupid things people do. It's sort of like cab-code signaling on transit lines. The operator controls the train, but can't go any faster than the signal system allows, given the presence of another train ahead (or whatever other condition the system accounts for - switch alignment, broken rail, dragging equipment).

A lot of the articles speculating on the impacts of autonomous vehicles seem to assume that they're going to be a lot more expensive than current vehicles. At least at first. But there's the possibility of aftermarket kits that can convert existing cars to autonomous driving. If so, the switchover will be much faster and have a much greater impact than expected.

The comments to this entry are closed.

Blog powered by Typepad