Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Two remarks about taking the booth lady as an example:

1) Her job has already been automated, thus validating the automation thesis. Her job has/used to have two aspects: ticket sales, and information. Tickets are now being sold by automatons, and she only has to handle the edge cases of people unable to use the machines, or the machine failings. As for information, most people will now use convenient smartphone apps to navigate BART, and she will only handle the few people who don't --tourists or people unwilling to use smartphones.

2) She was a terrible example in that you cannot automate her, since she is precisely what is left after the automation of her job. The work that comes to her is what did not fall into the automated use cases.

The value of some jobs is the human contact, or human interface they provide, and she falls into this category. This is also why we won't automate, say, barmen, or waitresses.

But what CAN be automated about BART, are train drivers (absolutely unnecessary in our day and age), and train dispatchers. They don't do consumer-facing human interaction, and machines would do their jobs better. And in time, they will get automated.



At the time of this comment, there is a post on the HN front page titled "More Connected, Yet More Alone" (link: https://news.ycombinator.com/item?id=6312076 ). One thing that automation like this misses is the much needed human interaction. Yes, humans are not as efficient or dependable as machines (not yet). But we are now entering a rather depressing period in history where human interaction is set aside in favor of digital and mechanical efficiency.


Y'know, I may be missing something but I don't feel like I get much out of the fleeting interaction I have with checkout people. They are almost always "tuned out" and any kindness they dispense is completely rote. If we're worried about human interaction, fire the checkout people and hire more greeters.


Oddly enough, I find the "tuned out" nature of most checkout people to be an opportunity. At least for the past fifteen years, I play a game with checkout people. My goal is to get them to smile. I make some comment about something, anything, with the goal to get them to respond and hopefully smile at me, and take me as something other than one of the hundreds of people they contact each day. I'm about 90% successful, but I keep trying for 100%. Every once in a while, either I'm really distracted or I see they are so distracted that I don't try. I have an ulterior motive: as I polish my skill at making a random stranger smile, I get better at tossing out a verbal gambit on very short notice to an attractive stranger.


Yes, whether they are completely disinterested or overly cheery the majority of the time they appear that they couldn't care less about your transaction.

Which is fine, it just feels less weird of an interaction when it is a machine that doesn't care.


http://xkcd.com/1227/

Your complaint is nothing new. People will always be able to romanticize the past.


The present is better (for humans) than the past. But that doesn't mean that everything is better.

When I go to Safeway, and have to go through a self-checkout, I do feel less connected to the people around me. When I order delivery, it's online through GrubHub, not over the phone. It's fairly obvious that automation is decreasing the amount that we have to interact with one another, and some people dislike that.

This is a specific problem that's because caused by automation, even though automation solves many other problems. It's a problem that should be thought about and fixed. And your attitude of "things are generally better now, so don't complain" is utterly counter-productive.


I absolutely reject your assertion that there is any net harm to be mitigated. Automated customer service reduces those moments where human interaction is required, but in the process frees up time for us to devote to more meaningful forms of human interaction. It certainly doesn't prevent you from interacting with other people, just frees you from the obligation of some fleeting encounters. And I personally don't find those encounters to be all that rewarding on average - a good customer service experience doesn't brighten my day much, but an avoidably bad one really kills the mood.

You should probably learn to accept the fact that some people want the option of having a business transaction be strictly business, without the overhead of social conventions of politeness. This does not make a person anti-social, it just makes them not so pathologically extroverted that they need human interaction (error-prone, and still relatively impersonal over the phone) just to order a pizza.


I never said there was any net harm; quite the opposite. Nuclear power is incredibly useful and relatively good for the environment, but there is the risk of meltdown. While it's a net positive, there are still issues that should be discussed and dealt with.

The fact is that much of our lives is spent doing things like buying things, eating out, et cetera. And for many people, spending your time doing those things without any human contact is lonely. It may not be for you, but you don't speak for everyone.

I absolutely believe in automating as much labor as is possible. Nonetheless, I think there is room for improvement in how society functions so that we get both efficiency and, for those who want it, human contact.

I do, however, reject your premise that automated customer service is better than in-person help. I am not an expert at checking out groceries in the same way that a cashier is. I go much slower through the process, and have to focus more on that. Right now, automated customer services largely serves to cut down on costs, not to improve service.

(I also never said anyone was anti-social. I do think that "pathologically extroverted" is needlessly insulting, and "need human interaction" is a strawman. I don't need a cloth napkin either; I just prefer it.)


I think this was excellent observation - one that I did not make the first time around.

I am unsure if I find it compelling that there is some utility in having a person there for the human contact piece. If you were able to remove the person and retain the usability within a non-human interface (to an equally useful or better replacement) would people say "I wish there were a person here?"

I would like to say no for most cases, I think there are a few people who would prefer that almost in spite of the efficiency gain.


Often the staff you see "standing around" or doing menial jobs have more complex roles than one might appreciate at first.

People say "I wish a person was here" in emergencies and on their worst days. When a dress gets stuck in the escalator, when the disabled need assistance. When people are impaired or confused or just looking really distressed. When the station needs to be evacuated.

Some of this can be outsourced to the general public (although I find it a little callous to do so), some cannot.


The whole reason her job remains is that it is very difficult to have the same usability in a non-human interface, in a manner satisfactory to human users. So, from a practical point of view, such an "if" is kinda pointless. Ask yourself if you prefer IVRS interfaces on the phone, or talking to a human.


Ask yourself if you prefer IVRS interfaces on the phone, or talking to a human.

Neither, I prefer a decent mobile-ready website. Why should we be limited to a dumb audio pipe?


> This is also why we won't automate, say, barmen, or waitresses.

Hold on there -- it's already being automated in Japan (https://www.youtube.com/watch?v=s0Z1EaFFICI ), Germany (http://www.engadget.com/2007/08/27/fully-automated-restauran... ), etc. Here in America there'll be some resistance but I'm pretty sure we'll go through automation in restaurants/bars as well. The driving force is economic benefit, sooner or later when these robots keep get less expensive, it'll start to make more economic sense for restaurant owners to have a robot server instead of a human server (no wage to pay, no health insurance to pay, no pension to pay, no risk of accidents or inconsistencies, no fear of getting things stolen, no fear of tension between BOH and FOH (common problems in a restaurant)). Savings will likely be passed on to customers, they won't have to tip, and so ultimately a large subset of consumers will eventually accept it. Except of course, a very small minority that is interested specifically in the human interaction... so, much like the lady in the article, the server job will not completely disappear, but it will be heavily minimized. And like there are tools with which customers can buy tickets by automation, customers will be able to order on things like E la carte ( http://elacarte.com/ YC-funded by the way). A server's job is cut in half with E la carte. Restaurant owners will realize that servers are doing half the work, so you need half of them now anyway (or alternatively, they now need half the pay... which will make the server job half as attractive).


One novelty restaurant != all-around automation. Computers are unbelievably, mind-bogglingly stupid, compared to pretty much any human capable of putting his pants on without outside help. So I wouldn't count on robots replacing humans in actual customer service - on the contrary, I expect companies that do not provide proper human interface start losing customers to those who do.

>>> Except of course, a very small minority that is interested specifically in the human interaction.

It's not like pubs/bars are universally known as places where people go to socialize... No, of course nobody actually needs any human contact, everybody is a hikikomori nowdays.


I agree with much of what you have said -- I think you misunderstood my original post, I'm not claiming these jobs are going to be completely eradicated from society come another 20 years, I just simply think that they'll decrease in numbers by a considerable amount. So a restaurant that once used to have 10 servers working at any given time will need maybe 3 or 4 with automation going on.

> So I wouldn't count on robots replacing humans in actual customer service - on the contrary, I expect companies that do not provide proper human interface start losing customers to those who do.

Right, but that's for the /edge/ cases. Like when you're going through automated calls and the robot voice has you go through a process of giving information in various ways -- either by saying something out loud and it'll try to determine what you said, or you just simply type it through the keypad... and then when you fail to communicate properly you get transferred to a real human being.

I can tell you that a very high amount of my friends/family relatives (middle-class folks) cite the reason of their not going to restaurants to be the cost of eating out. Automation will help reduce that cost, so I don't think you have a very strong argument here that people will just stop going to restaurants. And, there is more to a restaurant service than just the 'server experience' -- there is the decor, the fine view, the washed dishes, the nice walk around the town it takes to get there and so on.

As for pubs and bars, yes they are the places that people go to socialize, but not all of that socialization is with the bartender or the server, it's with other people too who're there to also socialize. Bars and pubs will get help from automation, just lessly so than restaurants.


>>> So a restaurant that once used to have 10 servers working at any given time will need maybe 3 or 4 with automation going on.

Or, more likely, same 10 working part-time to avoid government-mandated expenses for full-time workers. This is what is happening right now. Keeping people on the job becomes more expensive all the time.

>>> either by saying something out loud and it'll try to determine what you said,

Oh how I hate those. I usually press 0 until it gives up and connects me to a live human.

>>> the reason of their not going to restaurants to be the cost of eating out.

See above - employing people gets more and more expensive. You have to abide by a hundred of regulations, have the proper paperwork to prove it, get dozens of licenses, pass inspections, and then some lawyer drive-by-sues you for not being ADA-friendly. And then comes the local union and demands you to double the wages because they say so. Of course it'd be expensive.


Restaurants maybe, but I think more than any other food service job, bars see people who just want to talk to somebody. The bars know it too- you can tell, when they learn your name after just two or three visits.


>This is also why we won't automate, say, barmen, or waitresses.

For rich people, obviously not.

The direction of the economy is to eliminate the middle class. Has been for multiple generations. A horde of middle class machinists and bean counters can easily financially support a waitress at a lower class level of income. What happens when all those middle class jobs go away and they're unemployed or people greeters? There is no way the same people, now much poorer, will be able to support the waitress unless her income also shrinks by more than half, which can't happen given the fixed costs of living. Therefore either most restaurants will go out of business other than those serving the few remaining rich, or the now poor ex-middle classes will be served by robots aka fancy vending machines.

One other remark you may have missed is there is a huge, and growing segment of the population who can not and never will be able to afford a smartphone. And this fraction of the population has been growing both as a quantity and as a fraction of the population for generations as the economy permanently declines.


It seems that smartphone ownership (which is steadily increasing) is more of a factor of age than income (though both are a factor): http://www.nielsen.com/content/dam/corporate/us/en/newswire/...


An interesting counter-example to your point is the DC Metro system. There are human train drivers, but the driver operates a single switch varying speed/brake. The system automatically starts and stops when operating normally, and the centralized management system can operate the trains more or less against the will of the conductor.

In summer 2009, the outdated trains had a massive crash killing eight people, and there was a great amount of debate as to why, even if the driver engaged the manual break, the automatic control software failed to stop the train when a head-on collision was imminent. [0]

The fact you do not reference this event makes me think this was not headline news outside of the Beltway (where I lived at the time). But I do presume this will help support Luddite, anti-automation views in transportation industry, especially in other places where similar accidents occur.

[0] https://en.wikipedia.org/wiki/June_22,_2009_Washington_Metro...


It seems driverless trains will be held to a much higher standard than ones with a driver. As with driverless cars we are willing to accept a certain percentage of accidents, the second a driverless car runs someone over though it will be a massive controversy.


BART was originally designed to run without train drivers. That was rethought after the automation system launched the "Fremont Flyer" into a parking lot.

You really need drivers for safety reasons.


My uncle worked for a large railroad in Chicago, in their switching office. "You need a train engineer because that means someone not in management or the switching office is the guy whose hands were physically on the machine right before poor Mrs. Milly died. There is no other purpose to have them on the train. People like to think that having eyes on the ground stops accidents, because they think eyes on the ground stops physics, but if you are on the train track and the engineer can see you then physics says you better get off the train track because neither the engineer nor anyone save God Himself will stop the train from reaching your current location."

The train engineer's job is to hit the start button when the switching office says to. It's drive-by-wire, and he's not only just a wire, he's statistically speaking the most bug prone wire on the train.

He also had some sharp words for management on the topic of encouraging engineers to think they were responsible for accidents, claiming that it caused preventable suicides after unpreventable ones.


> You really need drivers for safety reasons.

One incident does not a point prove. There are many completely unattended trains in operation, several of which are in the United States.

http://en.wikipedia.org/wiki/List_of_driverless_trains


None of the systems listed are very complex. One or two lines, or a loop of track, with very few level-grade crossings. Most of the ones in the US are just airport shuttles. Sure, that's easy to do.

I wouldn't move the subway systems in New York City to UTO operation, though. Not a chance.


It's already done. The docklands light railway in London is more complicated than the metro systems of most cities and it's been driverless since the 1980's http://en.wikipedia.org/wiki/Docklands_Light_Railway#Map. On the other hand, it still has a staff member, 'Passsenger Service Agent', on every train to deal with customer service. They are also trained to drive the train at very low speed (I'm guessing <10mph) when the automation fails, which it does occasionally.


"Passenger Service Agents" are on some, not all trains. And on about a third of the trains they are on, they are manually operating the train at full speed (generally for "diagnostic" (read as sanctioned joyriding) purposes, I am informed by a friend at TFL).



What's the difference between an automated train failing to stop at the end of a line in a system with 200 stations to an automated train failing to stop at the end of a line in a system with two stations?

Of course in the rest of the system if can be incredibly complex to ensure safety, but that's not where the accident you're relying on to make your case occurred.


So we're on the cusp of having driverless cars and driverless airplanes, but driverless trains are impossible?

I've ridden the driverless train systems in Singapore, Tokyo and Copenhagen and they're all great.


The DLR in London operates fine without drivers:

http://www.visitlondon.com/traveller-information/getting-aro....

I'm sure there are plenty of other examples.




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

Search: