Hacker News new | comments | ask | show | jobs | submit login

We expect people to be frightened of robotic cars for the same reason they are scared of dying in a plane crash: Some deep-seated fear of dying in a manner that isn’t our own fault. Thoughts:

1) Marketing/PR for autonomous vehicles needs to really drive home their safety, so when you hear “robot car,” you think “… saves lives.”

2) Is it hard to imagine the opposite fear in children born 10 years from now? Having seen mostly robotic cars in real life, and human-driven cars getting in accidents on TV and in movies, might the child of the future react with terror when the robotic chauffeur intones, “human driver detected, approaching from rear”?

3) Loss of jobs. When we're heading to a future of robotic anything that will replace an entire profession, drivers in this case. What will be the economic effect of it?

I don't know why you're being down-voted, it's a valid question.

Computers have eliminated a lot of jobs over the last 30 to 40 years. For example VisiCalc, did a lot of what many tax accountants charged for. It used to be that humans, mostly women, switched (or routed if you will) phone calls.

Automation certainly does eliminate many jobs. But interestingly it also seems to make previously impossible things possible. For example, if we still used humans to switch phone calls, we'd need far more humans then are available. And I don't mean looking for work, I mean everyone could be working a switch board operator and it would not be nearly enough.

On the down side, many people speculate that our tax laws have become as complicated as they have, partly because tax prep software makes it possible to still do your taxes for relatively cheap.

Driver-less trucks will certainly suck for truck drivers, I can just see that History channel show taking a sad turn as more and more of its real life stars are slowly replaced by self-driving trucks.

But what possibilities will it create? Could cars, or for that matter trucks, become like a subscription? You just subscribe and then if at any point and any where you need a ride, you just dial a number and a few minutes later a dirverless vehicle arrives to take you anywhere.

The same could work for goods. Request as much trucking capacity as you need, minutes before you need it. The trucks just show up.

How many hours total (everyone's added up) hours sitting in traffic could be eliminated if a clear majority of the cars are intelligent and networked? Thousand? Millions? All that previously unproductive time, now productive.

All the lives saved from accidents, will they create jobs?

I think you make some great points. If we look back in history, almost everyone was a farmer. Farming was the job you did to survive, whether you liked it or not. Along came tractors and other farm machinery, and now something like only 3% of the population are still employed as farmers.

Interestingly, farming has been moving to driverless equipment out in the fields for the last couple of decades. At first it was not well received. Who would want to give up control of their machines? But in the last couple of years, the technology has really exploded. People are starting to see the benefits to their business.

I expect driverless cars will be the same. It is going to be a long time before it is accepted by the masses, even when the long list of benefits are present. Hopefully we don't outlaw them before they really come into their own.

> now something like only 3% of the population are still employed as farmers.

That's about exactly accurate: it was 2.8%, as of 1998, and is projected to dip down to 2.4% by 2018. [1] I happened to be looking at this just a few minutes ago, linked from http://zamfi.net/blog/one-day-we-will-all-be-programmers

[1] http://www.bls.gov/emp/ep_table_201.htm

Where do farmers use driverless equipment? I live in a farming community and nobody uses driverless equipment.

When I say driverless, I'm not precluding the presence of an operator.

John Deere's iTEC Pro, for example, is capable of controlling the tractor, but still requires an operator to be in the cab, just in case. Although here's a fun video of someone who disabled the safety sensors: http://www.youtube.com/watch?v=zU4liQvrcm4

I don't know of any large operator that isn't using at least autosteer, which, for things like cultivating essentially is driverless. For what it is worth, I come from an area of grain farmers.

Thanks for the link. I should ask around to see who is using these. edit: sounds like i'm wrong, they ALL are using automated equipment. they just sit on the tractor, too.

Personally, I look forward to self-driving sleeper cars and a new type of motel that is little more than a secure parking lot plus some nice bathrooms. What is the feeling of "where you live" if you switch cities each time you sleep?

Also, in your hypothetical, if we are all routing calls as our day job, doesn't that mean that at any one time, 1/3 of humans are asleep, 1/3 routing calls, 1/3 possibly on calls. With at least two people per call. How can that not be workable?

  With at least two people per call. How can that not be workable?
He was partially referring to the internet. What would the web look like if two servers that wanted to move data between each other, (say, in a stock exchange) had to use circuit-switched telephony?

Switching is hierarchical, so for every non-local call placed there would be several human operators involved.

So out of all of humanity only 1/3 would be making calls? Everyone else would be supporting these phone calls? What on earth would they be talking about?

"But what possibilities will it create? Could cars, or for that matter trucks, become like a subscription? You just subscribe and then if at any point and any where you need a ride, you just dial a number and a few minutes later a dirverless vehicle arrives to take you anywhere."

I've heard of this idea before in the context of driverless cars. The thing is, this service already exists and is known as a "taxi". Since driverless taxis would likely be cheaper to operate, this means another source of job loss.

Exactly the same thing can be said of anything that has become automated; any machine manufacturing has offset the physical labor jobs. You can even argue that things like nail guns and power saws resulted in massive loss of jobs since one person can suddenly do the job of 6.

I think that in the next human generation or two we will end up seeing all menial and physical labor jobs being replaced by machines, leaving only entertainment and intellectual jobs.

I also am not sure whether "loss of jobs" really is some sort of absolute concept. Imagine that tomorrow all jobs could be done by "thinking" machines, even software development and architectural design and whatever. Suddenly everyone in the whole world lost their job; does that mean that everyone is out starving on the streets and there are just empty foreclosed on houses across the country? Obviously not; all of the same material wealth still exists to be divided across the same number of people (assuming the thinking machines don't care about material comforts).

Presumably cities will need to be rethought too. Cars and trucks will become more systematic in their movements.. Will most streets as we know them become grassy walkways for human transit? Do we really need everyone to have a car, and thus access to their building by a street? Just imagine a city like new york with next to no (large vehicle) traffic.. smaller vehicles to ferry packages from larger vein streets..

I love your vision of New York, but it's dangerous to use New York as a model. The population density of New York makes it unique (at least in the US).

Haha.. True. However, I live in Tokyo, and can see it working here too !

Marshall Brain wrote the very interesting (IMHO) "Robotic Nation" in the exploration of your question:


And, to save time, here are the standard HN responses:

1. Marshall Brain commits the Lump of Labor Fallacy (http://en.wikipedia.org/wiki/Lump_of_labour_fallacy): The lump of labor fallacy holds that the total amount of work is fixed. This is considered fallacious because increasing productivity also increases the size of the economy, thus creating new jobs in the process. IOW, the total amount of work is not fixed.


2. The industrial revolution destroyed all sorts of jobs (of say, craftsmen), but ended up employing as many or more as factory workers. Similarly with the information revolution. Etc... All of these out of work truck drivers will become... massage therapists.

My critique of the lump of labor fallacy argument is that it seems to hold as long as our ability to consume can keep up with productivity gains. But what if there's a natural ceiling to our ability to consume; what happens when we produce more than we can consume? Or, when luxury ephemeral goods become trivially cheap (say, watching YouTube videos).

I'm not an economist, so I strive to remain agnostic on this question, but it still seems far from settled to me. Even if you believe that things will eventually shake out and workers displaced by automation will find new jobs, we still may be in for an uncomfortable period of high unemployment and disruption as the economy adjusts. (AKA, productivity shocks.)

I've argued that the Lump of Labor fallacy fails because it considers workers to be more fungible than they are. It is possible to create jobs for which no one in the existing unemployement pool is qualified to fill. That is a structural unemployment of 0%.

When that happens competition for the qualified labor goes up, and salaries rise in response. While the structurally unemployed[1] remain so and remain poor. This leads to a building disparity between rich and poor and that the problems such a disparity creates are quite well known.

[1] 'structurally unemployed' refers to a person who is unemployed and unqualified for all of the employment opportunities in their area. It comes from a similar concept 'structurally fired' where a company removes the positions that are held by current employees without creating other opportunities.

Curiously enough, though, all of Europe and the US has seen more or less the same kind of development of this kind in recent history. Yet the disparity between rich and poor differs widely between those countries, and it has also gone up and down quite strongly over time.

This is very strong evidence that the kind of disparities you mentioned are not truly structural. The "structural argument" is just a decoy used to prevent the disadvantaged from asking inconvenient political questions.

Sweden actually has higher unemployment than the US, but less wealth disparity, because they have a better social support net for the unemployed.

Exactly. It is about political choice, not about structural inevitabilities.

Actual numbers: http://www.wolframalpha.com/input/?i=unemployment+rate+swede...

It looks like Sweden merely caught up with the US in terms of unemployment in the last 20 years.

Look at the quite drastic ups and downs in unemployment in Sweden in the 1990s. I know that's not exactly what you wrote, but does that look like something that is driven by "structural changes"?

You're picking some bad examples for future growth. Instead, look at the problems people really want to solve.

I'd pick health care as something that's just going to keep growing. As long as people have health problems, there will be a market for attempts at solutions. As the rest of life becomes more affordable, more money can be spent on this.

Advancing technology might eliminate some diseases, and make treating others trivially easy.

Note that nobody dies of smallpox anymore. That disease no longer exists in the population, it has been eliminated. IBM Watson will (eventually, kinda does now) automate diagnosis. Respirocytes[1] would eliminate decompression sickness, and make it pretty hard to bleed to death, with a medic handy. A Vasculoid[2] would eliminate bleeding entirely, as well as blocking all blood-borne diseases, by the simple expedient of replacing blood with something else.

1: http://en.wikipedia.org/wiki/Respirocyte

2: http://jetpress.org/volume11/vasculoid.html

Well okay, in the really long run, perhaps health care will someday be a solved problem. But that's not going to happen without lots of market growth and investment in the meantime.

Well, we've already seen entire professions and daily tasks disappearing thanks to technology improvements. (e.g. blacksmiths, or washing clothes by hand)

I think the sustainable future is one where we all work significantly less than today, and where more people do things where humans are much better than machines, such as arts/music and care (for the elderly, children, sick etc.)

While I agree with this, I think one of its less desirable side effects will be to increase the ranks of the poor.

Automation will replace a number of things: 1) Driving. 2) Wait staff (waiters / waitresses) 3) Piloting 4) Checkout 5) Reception services 6) Restocking / Inventory management 7) Fast food cooking staff 8) Hair styling / cutting 9) Surgery 10) Mining 11) Logging 12) Farming 13) Masonry 14) Security / Monitoring

These changes will be gradual, like in Machining where it used to be a machinist, his tools, and a mill. Now the mill is CNC and he's more of a programmer. The guy running the tool is more of an operator.

As automation capabilities get better we'll see robots like the DaVinci surgery robot for things like Hair Salons, where a single operator/stylist can style hair for 10 customers simultaneously. But that stylist will have more technical training than stylists of today do. So most stylists will be out of work, the smaller and smaller percentage of people who are both creative stylists and have the technical ability to run these salons will displace them.

I see the following mechansims at work in the US today which is pushing us toward that future.

1) Efficiency (or competing on cost vs quality) is 'winning' (or displacing) quality service products on the market.

2) The education and capacity of the both 25th percentile from our youth is incapable of basic problem solving.

3) The risk from employee theft (shrinkage) works against #1 making automation solutions justifiable to employers.

When possible transferring the risk/responsibility for service delivery to the consumer is more cost effective than training under-educated employees or hiring people with health benefits.

It might seem implausible but I remember when scanners were introduced into supermarkets. The canonical low margin business. They helped margins because less experienced cashiers could check out customers as quickly as more experienced cashiers did. It also resulted in fewer errors (by the cashier). I remember joking that at some point you would just walk behind the register and check yourself out.

Well once the automation around collecting cash and other payments became good enough, that is exactly what has happend. There are several stores in my area where you have the option of checking out yourself. On a recent trip there was one 'real' cashier and all the self checkout lines were open (8 stations) with one supervisor.

The technology enables the automation the economics drive the adoption. Things like pay kiosks now are quite able to take any of cash, checks, credit or debit cards. That gets hardened at ticket buying machines, then jumps to cash registers, I expect it to be in taxi cabs next.

So the 'social justice' question is what do we do with people who are structurally unemployable? Sure they have some responsibility for that (education and basic problem solving skills are lacking) but once they get past their rebellious youth and realize that nothing they will do can prevent them from growing old, and now they want to go back and 'fix' those lacks, how can we best facilitate that?

The advancment of automation and the elimination of jobs that can be done by people is neither a 'good' thing or a 'bad' thing. It just is. And I don't see us changing the economics or the technology which would inhibit it. Advances will be made, applicability to wider and wider problem sets will be possible, by and large its an economic 'win' to have a machine do something rather than a human.

You've probably seen the 'buy American, look for made in America' programs that people run. No doubt we'll see 'for made in America by People' campaigns as well.

It used to be that the difference in productivity in most (manual laborers, artisans) was at most a small multiple. It seems inevitable under the capitalist system that increasing automation in all fields will widen the wealth gap considerably, as the most productive people become many times (perhaps many orders of magnitude) as productive as the least. The difference between someone who performs a task and someone who can program many cheap machines to do the same thing will grow much, much wider as machines become cheaper to buy and operate. The owners of those machines will become extremely rich unless wealth redistribution becomes more common. Large social changes or revolution seem like they'll become required.

At some point, though, once every type of work can be done by machines, wealth could probably be redistributed in such a way that noone had to work very much. That would probably require human level intelligences, though, and then cue the great robot uprising...

Parkinson's Law says this is unlikely.


It would be nice, but I can't see it happening, unfortunately. Too much of our society would need restructuring, and there are those that would fight tooth and nail against that.

I sure hope that there will be a jump at some point allowing life quality to increase, and not bureaucracies, as Parkinson's Law seems to suggest (life quality hasn't improved during the last 30 years in the developed world, referring to general happiness and well-being). I am optimistic that brave politicians will make the necessary changes at some point - the world is too rapidly-transforming for huge changes in our lives not to happen.

There is no fundamental law prohibiting this rapid change from being accompanied by widespread death and destruction.

I think the sustainable future is one where we all work significantly less than today

Given the large number of people I know who spend half or more of their time surfing the web at the office, I suspect that future is already here to some extent.

Generally: Unless our demand for goods and services can grow at a pace that matches worker productivity as amplified by automation, there eventually simply won't be enough work to be done to keep everyone fed & housed.

This is paradoxical, since our productivity will be at an all-time high. Eventually we'll arrive at some sort of alternative economy but at least in the US there will be some serious unpleasantness along the way.

That's not the way to look at it.

See creative destruction: http://en.wikipedia.org/wiki/Creative_destruction

e.g. the car being invented did mean less work for horse-carriage "drivers", but eventually everything was better than before.

Unfortunately Google is incredibly bad at sales and marketing so don't expect them to understand that their main challenge will be the public perception of driverless cars.

I think they understand this. What I'm not sure of is if they will have the sales ability to do anything about it.

Google is very good at PR. Their image as being "naive and bad at marketing" might be part of their PR.


Yeah, that's why no one uses their products.

Google is good at entering a market just as it's catching on with a meaningfully-differentiated product, and the product markets itself. See Search, Gmail, and Maps.

When Google doesn't do this, they have to rely on marketing. See Wave or Buzz.

Every big company has its string of failed products. I'm not convinced that Google is really worse in this regard than Apple, Microsoft or Yahoo.

Has Apple had any failed products in the last 10 years?

The XServe line was shut down (although you can still buy the OS X server edition).

The Cube was a failure, although I'm not exactly positive it's exactly within 10 years.

.Mac (or MobileMe) hasn't been as successful as I think apple believed it would be.

iPod Hi-Fi, buttonless iPod Shuffle. But that's a pretty good track record.

None of them has any failed products if you set convenient enough window.

Google has only existed for 12 years, so it doesn't seem like it would be fair to talk about Apple failures from 1988 in this context.

There were enough failures with Apple (constant string of failures in fact) until late 1990s. Also, if we take Yahoo!, a successful business by any measure, it is hard to assign any two years where they didn't fail at something. Mind you, Yahoo! is a prefect example of marketing-driven company, much more so than Apple (who actually prides itself more at solid design and technology).

I would honestly have a lot of difficulty naming a Yahoo! product that launched in the last 10 years, period.

I understand that Apple had more than enough failures in the 80s and 90s, but honestly they are basically a different company since 1998 when Steve Jobs came back and completely changed the trajectory of the company.

They have managed to have basically no significant failed products at all in the entire term that Google has existed at all as a company. Google and Microsoft have had several extremely high profile failures, and dozens of failures as minor as the most significant failed product that Apple had.

The stock price of the company went from 3.5 in 1997 to 350 in 2011. I just don't think that the Newton is at all relevant when talking about how careful the respective major companies are to avoid large public failures.

You mean web search, gmail, android and other crap they produce?

I think they were going for sarcasm.

I think you've got the reason people are afraid of flying wrong. People aren't afraid of dying in a bus crash, or on a train, even an automated one, like they are of flying.

Indeed. It is perhaps due to the 700mph, 6,000 feet in the air part that makes them especially jittery. True or not, I think most people feel safer in a car & probably feel like they are more likely to survive, due to 60mph 12 inches off the ground.

(While flying is safer in many ways like deaths per mile or per passenger, I would bet you are a lot safer in a car than a plane in the event of a crash. Car crashes are dangerous, but plane crashes are nearly guaranteed fatalities)

No. It's hundreds of people dying at a time with nothing that any of them can do to mitigate their fate. People are really poor at assessing risk, but marvelous at concocting horror stories.

(I can see that you are not from California. At 60mph, you are in serious danger of being rear-ended.)

Given the nature of public transport I think we can discount the "out of my control" factor. We can even remove the human element; driverless trains are found in quite a few places and nobody is screaming in terror on them.


So what is it about flying? Probably the whole "trapped in a metal tube travelling at 500mph a mile above the ocean" thing.

Plus trains are more or less locked into a path.

Yep. I hate flying, but would pay a lot of money for a driverless car, and sleep like a baby while I commute to work.

It may be the fact that the airplane is in the air that makes the difference. Being 20,000 feet in the air is quite different than cruising along on the highway in a bus.

That being said, I wonder what the driverless cars would do for the blind/visually-impaired. It could be quite liberating for them. They wouldn't have to live near public transit/take jobs with public transit access. Neat.

I think it is more about getting the media and not the public at large behind the safety and potential of autonomous vehicles. The reason people are so afraid of flying is because plane crashes are so hyped by the media outlets. If the media is behind it, then the people will follow. Also, it is about control. If Google allows a fallback feature where the driver can take control in the event of a malfunction like cruise control currently works, people will perceive autonomous cars to be safer because they CAN be in control.

People are afraid of planes because if the power goes out, you die. On the ground, if the power goes out on your driverless car, you cruise to a stop.

And if you cruise to a stop in the middle of a train crossing?

Or if the steering shuts down while on a long curve next to a cliff?

Trains travel slower through populated areas. If you're stuck on a crossing jump out. Chances are the train will stop if the driver sees you in time and has the distance to stop (about 1mi/1.6km).

Steering doesn't shut down, it just goes from easy to hard. If the car is in motion it requires very little energy to turn.

Volvo's City Safety is marketed in the same way you describe: http://www.youtube.com/watch?v=O5MR2S2K5iM&feature=relmf...

Perhaps we could better aquatint people with what a meat grinder American roads are. About 40,000 people die every year in America in traffic accidents caused by humans. More than 10,000 of those deaths are caused by drunk drivers, and it seems like most those deaths would simply go away if they could auto-pilot their way home.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact