Why is this job not handled by a machine yet? | Hacker News Reader

Why is this job not handled by a machine yet?

venturebeat.com

Counterpoint: a few years ago, having been stuck in abnormally high traffic on the way to the airport, I was running right on the cusp of the arbitrary check-in deadline for my flight. I began the automated checkin service, which was comparitively extremely slow on that day also. It let me start the process, but threw an unspecified error when I got to the final stages to print my boarding pass and bag tags. Upon retrying, I was past the check-in deadline. Buggery.

Walking over to the check-in counter, a polite explanation of the situation I found myself in (and, possibly, my moderate-level status with the carrier but I don't fully know) made the service attendant remove the problem in about 30 seconds, guaranteeing that both I and my luggage would make the flight with no problems despite being 'technically past the deadline now'.

Flesh-based interaction points, by their very nature, have a tendancy to span a far broader range of service quality than machines. The bad are far worse because machines don't gossip and hate their jobs. Conversely, machines have little hope when things go south compared with a good service employee, who'll do their level best to go above and beyond their basic responsibilites to make whatever issue you have either go away or minimally impact you.

My experience is that one of the most apparent examples of this is if you ever find yourself significantly up the greasy ladder of the airline status game for whatever reason. I postulate that because carriers have realised that 'bending the rules' is a relatively 0-cost commodity to exchange for brand loyalty and this is precisely factored into the training and job directives of their employees.

So why was the human able to fix the problem and the machine not? Perhaps the machine has an arbitrary deadline whereas the human measures the progress of the actual boarding process, in which case the fix is to build that into the machine. Perhaps they use differential "deadlines" as a way to subtly discriminate against undesirables customers, in which case the clear light of automation probably improves matters overall.

Perhaps it is because machines are based on pure reason and strict logic, and passenger terminals are apparently based on the exact opposite of that.~

I put forth a guess that airline passenger terminals have not been made more efficient with automation because that would expose just how inefficient the airline is away from the public areas of the terminal. The ground crew service has to be automated first, in order to actually see any benefits.

For instance, boarding procedures are intentionally inefficient, for two reasons. If the passengers boarded quickly, and in an orderly fashion, they would just have to wait even longer for the ground crew to prepare for departure. And the first-class passengers have to be guaranteed space in the first-come-first-served overhead bins. That's really the only reason to board early. If you don't have a carry-on bag, you might as well wait until just before they disconnect the boarding ramp, because they will still be loading bags and running checklists at that time.

The baggage claim areas are intentionally hard to reach from the embarkation areas, because people will have to wait for their bags when they get there anyway.

A lot of the inconveniences you experience as a passenger in the airport are there by design.

So I would further guess that the arbitrary check-in deadline exists as a method to resolve overbooking conflicts. If that particular flight had been overbooked, the human could not have done anything other than offer a conciliatory apology and some compensatory voucher, depending on their mood and your demeanor. The deadline conceals the fact that the airlines sometimes sell the same seat to multiple people.

If the anti-labor argument were taken to the extreme, the explanation should be that the machine was programmed by humans, who gossip and hate their jobs. At the end of the day, the machine is still a human touch point, albeit an abstract one.

That sort of thought has interesting philosophical implementations. You're positing that when our creations interact with others, we are unconsciously the ones interacting with others.

To a deeper extent, when you create and then raise a child, and your child grows up and interacts with others, that is also a case of you unconsciously interacting with others.

Follow the chain of creative identity backwards in time to the beginning, and you end up with the cause of the universe, whatever that may be, unconsciously interacting with itself, through everyone and everything.

It's not very relevant, and perhaps not even important, but it's an interesting perspective :)

Through our eyes, the universe is perceiving itself. Through our ears, the universe is listening to its harmonies. We are the witnesses through which the universe becomes conscious of its glory, of its magnificence.

- Alan Watts

I would imagine that fewer causal links would indicate higher relevancy. That is to say, chains of different length shouldn't always be considered the same.

To the extent that we can automate the programming of the machine, we should (and the history of programming is full of that - e.g. on the earliest computers assembling was done by hand, but this was repetitive, mechanistic work so people wrote assembler programs instead).

Something tells me that by the time that machine gets fixed the guy will have already lost his flight.

Building the actual boarding process into the software probably isn't as easy as you think, and airline ticketing is a long list of edge cases. I'm not a particularly frequent flyer, but the following scenarios from my own recent experiences spring to mind:

- "Your computer systems somehow completely failed to notify me in advance you cancelled my flight. Your airline has a legal obligation to book me on an alternative route, and pay compensation in the event of a significant delay."

- "I've followed your suggestion and been all round the airport looking for somewhere to print my ticket confirmation details, but can't find anywhere, and it's 11pm. Look, it's on my laptop screen."

- "Would you like a complimentary upgrade? You'll need to change your shoes first to comply with our business class policy, mind you"

- "We're overbooked, would you be interested in the option of a £500 and a hotel stay overnight? Any questions?"

- "You don't have an onward ticket? Not even a boat or bus ticket yet? OK, despite experiences you may have had with other airlines, we're strict on enforcing that country's visa policy, so you'll have to buy one. Here's the cheapest we sell, and this is the process for obtaining a refund after landing."

- "Oh, you've got a ticket with that airline, have you? We've got a codeshare agreement now - would you like us to put your bags through?"

- "The flight is delayed for five hours, here's your complimentary meal"

- "Look, I understand that your 777 has significantly fewer seats than the 747 it's replaced at short notice, but we're some of the first people to arrive, why would you bump us off? We're travelling on business so we need to be there and aren't at all interested in your compensation package. Yes, we have to be there on that day, for a trade show. No, we're not frequent flyers with your airline, but if it makes a difference a colleague of ours just got headhunted to a VP position in your airline. Yes, he sits next to us for the next month. Oh, there are seats available now?"

Needless to say, even where software can be written to deal with these edge cases, keeping it up to date with the latest issues isn't cheap, especially when you need (inexpensive) staff hanging around your kiosks just to deal with people that don't understand how to use them anyway.

Discrimination tends to happen in reverse, with special privileges created for customers that have a particularly strong need to travel, customers that have suffered from the airline's cockup, situations where customers need extra help etc. It's difficult to argue that "Computer says no" is an overall improvement on that.

> Discrimination tends to happen in reverse, with special privileges created for customers that have a particularly strong need to travel, customers that have suffered from the airline's cockup, situations where customers need extra help etc. It's difficult to argue that "Computer says no" is an overall improvement on that.

I think it is, overall. Fixing problems may take a bit longer but they'll get fixed for everyone, rather than ad-hoc fixing it for upper-class folks who are friends with VPs and hanging the regular folks out to dry.

But the point is that the problem of the last-minute substitution of a smaller aircraft for maintenance issues isn't something the kiosk software is going to fix. Hell, the kiosk software can't even attempt to remedy simple problems like there being a lot of people needing the flight which departs in thirty minutes stuck in the queue behind the people on the flight which departs in three hours, never mind figure that it's probably not fair to prevent someone who's missed the cutoff time by thirty seconds as a result of the queue from trying to run to the gate.

Disgruntled business people missing part of the trade show they'd bought the flights for, the couple of club scheme members the algorithm prioritises boarding the flight instead of being offered the option of cash compensation and masses of free points for flying tomorrow they might have actually preferred and a VP getting a flea in his ear for the worst airline service ever probably leaves everyone strictly worse off than before. Not to mention how much more incredibly rude it appears to the average person if an airline informs people they can't board the flight they're waiting for via the passport scan machine rather than a sympathetic human that's just going to ring a couple of people to check whether there's anything that can be done.

To be fair, if the machine also tells me it deposited the cost of the ticket back in my account, plus $500, and gives me the option of either 1) A hotel stay and a ticket for tomorrow, printed right there, 2) A list of other flights going to the same destination today, which I can book/check-in from the terminal.

The pain is a lot less, though possibly still slightly less unsympathetic, depending on the mood of the attendant you'd otherwise have gotten.

> You'll need to change your shoes first to comply with our business class policy

Which airline, if you don't mind my asking?

Singapore Airlines. I was in flip flops at the time, tbf, and they didn't mind the boardshorts :D

The other thing is the interaction needs to be fast and practical. Like many IT departments, the internal IT support at my employer likes to follow the latest fad more than being useful, and replaced the dial in menus in the phone helpline with a voice recognition system where you have to describe your problem to be directed to the right area. Of course you don't know what kind of sentence it expects, particularly for a complex problem, and half of the employees are non native english speakers and the damn thing doesn't do foreign accents...

If you work for ANZ, godddamn. Even if you don't please stop it. This is probably one of the most annoying processes I've had to go through and made a 2 minute call take about 15 minutes while I argued with a robot that can only handle very specific phrases.

Whenever I encounter those I just say, 'operator' over and over (though I suspect saying anything it doesn't recognize works work too) until I get a person.

I've had to work on making these voice recognition systems before. We originally allowed for people to just say operator pretty much anywhere in the app, and the client got upset that so many people asked for an operator and got sent to their agents that they had us restructure it so the person had to try pretty darn hard and be overly persistent before it decided to transfer to a person.

I felt a little bad doing it, especially after listening to the actual calls where people would just sound so annoyed and frustrated as they continuously asked for an operator, but the client insisted that less people actually reached their call centers.

I'm fairly sure it was Bell Canada that would reply with "Alright, it sounds like you're having problems, let me transfer you to an operator.." when you swore. (Not sure if it still exists, it's been years since I've been a Bell customer).

It was actually brilliant because it was responding to real frustration in a useful way, and it was disarming because it was a humorous thing for an automated system to do.

These days, I've noticed hip phone menu systems hang up on you when you try to destroy America chanting this obvious codeword to commence the attack.

Automated telephone systems should be required by law to respond to "shibboleet" and take the appropriate action. Or, at the very least, should be required to let me press 0 for an operator.

Well, most automated systems are that way because the company doesn't really want to talk to you. You can't legislate good service if the company has decided that they don't want to provide good service but want (as evidenced by their actions in e.g. phone systems) to provide either cheap service or no service. I mean, having a person that says "fuck you, don't call us ever again" and hangs up the phone without listening would be a technically valid solution to a law that requires "to let me press 0 for an operator".

Honestly, I'll take that over the phone robot telling me "we care about your call" for the 152nd time. At least they're being honest that way.

With most voice-recognition systems if you blow a couple raspberries down the receiver (or rub your finger on the mic or otherwise generate some noise) they'll give up after a couple of tries and take you to a human.

Some just hang up on you when they give up. Quite awful

I usually just tell them "sales" (because sales treats callers the best of all the departments, as they usually want your money) and then wait for the actual person to transfer me to the appropriate department.

...and then wait for the actual person to transfer me to the appropriate department's automated voice system.

* fixed that for you

i have had similar experiences with check out machines in Target and large retailers. they puke if you dont follow the directions to a T.

Now I deliberately avoid any automated checkout and go straight to the human cashier. if you insist on replacing humans with machines, it has to be flawless

I prefer automated checkouts far more than human cashiers. It's much faster. Maybe there's some problem 20% of the time which is quickly resolved by an employee. (People who would say that the system still needs humans ignore the fact that one person can now look over six checkouts, thus cutting jobs by one-sixth.)

I find the automated checkouts to be much slower than human cashiers.

Humans can optimise the path from the conveyor belt to the bag so that the item just so happens to pass via the barcode scanner.

I could do that too, but the machine throws a hissy fit and doesn't let me scan the next item until I've placed it in the bagging area, and waited a second for the scales to stabilize. In the time it takes me to scan and bag one item on the automated checkout, an experienced human cashier can get 5-6 items bagged.

And that's only if things go as expected. Sometimes the scales bug out. Sometimes I have to navigate menus to specify what fresh produce I'm buying (while human checkout operators have memorised all the numbers). Sometimes I make a mistake and I'm not allowed to self-correct it. I have to wait for a human to come along and fix my mistake.

Not to mention, the machine is yelling instructions at me the entire time. "Yes I know I should use the pin pad to complete my transaction, in fact I did so while you were still playing the previous audio clip."

The only time the automated checkout is faster is when there is a queue at the real checkouts. And even then, the queue might be worth it just so I can avoid the automated checkout yelling at me.

I've had the exact same experience. I could scan the items a lot faster than the machine allows me to (sure, not faster than an experienced human cashier), but it insists me to bag every item, has problems a lot of times when it stops and asks me to call a human agent, can't use some of the gift cards that I want to use with it, and doesn't override the price when the store has advertized a discount and the machine doesn't detect that. Also, for things that don't have barcodes (fruits/vegetables etc.), you have to 'search' from a list and either select the image of the perfect type of pepper that you're buying (or enter a code) - which makes it even slower. A human cashier remembers all those codes and very quickly types in. For a machine to do that right now, it will have to use some advanced image recognition techniques to just find out which of the 100 kinds of Apple it is - and set the price accordingly. Sure, all these problems can be theoretically fixed, but so far, I guess we're just not there.

One question I have about the automated checkouts at my local supermarket is they seem to intentionally slow the user down. For example, if you go to pay and put in a card, it says 'please remove your card, press the -debit card- button, and pay with your card'

It seems like no reasonable person wouldn't notice this flaw, let alone the presumably several people involved in creating this software, so it leads me to believe that there is some reason the shops want to encourage not using these checkouts.

I wonder why?

It's those dreaded words: Minimal Viable Product.

Even if it's slower for the transaction itself, the fact that they generally replace 1 human cashier with 2-3 automated checkouts mean that waiting times are much shorter in my experience.

Although I did discover at my local supermarket that if I'm buying alcohol at a self checkout (which requires someone to confirm my age), it can be paradoxically faster to go when it's busy, because there's always someone at the machines, whereas they rely on wandering staff to see the alert when it's quieter.

Which also paradoxically means it's easier to turn up at a quiet time and steal the alcohol (even if you're underage!) than to buy it legitimately.

I don't mind automated checkouts that much because I worked as a cashier as a teenager and based on that experience I'm able to check myself out pretty quickly. When there's a line for all of the checkouts including the automated ones, though, I avoid them, because the average customer is generally awful at using them and the lines at manned checkouts move much faster.

I guess it depends on the implementation. At my store you bring a scanner and some bags along, and scan and pack each item as you go. At the checkout you simply return the scanner and swipe your card.

What store is this?

Waitrose or Tesco in the UK have these scan as you shop schemes.

Automated checkouts aren't about each individual machine being faster than one cashier, they're about being able to serve ten people while only one cashier watches over.

The slow part of human cashiers is waiting for the people in front of you to get their payment done and move on with their stuff.

The payment part is unnecessarily slow at these things. We should be able to wave something at the device and walk out the door by now. But instead it's hit I'm done, wait a second to be able to hit pay with card, wait for it to tell you to enter a card, enter it, wait for it to tell you to enter pin, answer if you want cash back, wait a few seconds for it to accept and make a buzz to remove your card, wait 6 seconds for the stupid question "did you enjoy your experience today? yes/no" to go away (that I refuse to answer because it shouldn't fucking ask me in the first place), wait a couple seconds for it to decide to print the receipt and the cash to dispense, finally get your bags into your cart, and then leave.

None of the steps are really that long by themselves, but in aggregate it's easily another 1-2 minutes checking out.

> The payment part is unnecessarily slow at these things. We should be able to wave something at the device and walk out the door by now.

Been using this for years in Canada. Most well known as PayPass (MasterCard) but also PayWave (Visa) and Contactless Payments (American Express). By far it's the fastest and easiest way to pay (including vs cash), and I use it whenever I have a choice. I'd say it takes me well under 10 seconds which includes taking out and putting away my card.

I've noticed the newest terminals in sit-down restaurants are even working with it now, by moving the prompt for the tip to the start of the process (before you tap or put your card in) instead of the end.

I prefer automated checkouts out of respect to humans - how terrible job that is! And also as a reminder, how terrible the software is. (I design and write software, not of this type though, for a living.)

Although I take my time to go through it. Damn, if it's done by machine, I can do it on my own pace.

So out of respect you put people out of work whilst increasing the profit of the already wealthy - in most cases - shareholders? Is it more dignified to be out work and penniless?

My position, having worked as a checkout assistant FWIW, is the opposite. Using the self-checkout is redistributing money from the poor to the rich, AFAICT: you're saving the company wages, paying the same amount, and doing the work for free.

If the out of work checkout assistants got a living wage then you might be on to something.

I am social democrat, and as such I am for all people to have living wages.. (and I am actually proponent of basic income and progressive taxes)

However, checkout assistants are not very well paid and the job is horrible (it's not the worst, arguably, but could be improved).

I do not however feel responsible for the bad moral decisions of other people. My action in itself doesn't increase profits of the wealthy, it's actions of other people. These people need to be made socially responsible (for example via progressive taxation), not me.

There was a time when you got good money to work in a grocery store, even in a checkout. I have (older) family members who decided not to finish college because they were making so much money at the grocery store, with pensions, in positions that would only pay minimum wage nowadays.

I worked as a grocery checker for a year or so. My dad did it for half his career. It used to be a solid, union job in most places in the us.

It's really not even that bad, if you like talking to people. The checking groceries part can be sort of meditative and relaxing.

I am not from U.S., so I can't compare. Here (Czech Republic) it's usually one of the worst types of employment, rife with abuse (there were cases where employees needed a permission to go to the toilet, had to stand all day for no reason, etc.).

> I prefer automated checkouts out of respect to humans - how terrible job that is!

I prefer automated checkout too - but I'm curious as to why you think the job is "terrible"?

Before my first software development gig, I worked for 6 months as a cashier. I won't say it was all rainbows and roses, but it wasn't "terrible". Hard on the feet (standing), but I was young at the time so it wasn't too big of a deal. Dealing with irate customers could be interesting at time (pass 'em to a manager if you couldn't calm 'em down). Occasionally you'd get someone trying to return an item bought 20 years ago (actually had this happen to me).

But there were also the "regulars": The 70 year old lady who came in every Wednesday to purchase a bunch of liquor - who'd then walk out, and hop into her brand new Vette. The weird cat dude, who always smelled, but was friendly - purchasing a cart full of cat food, and another of rubbing alcohol and swabs (poor cats). The guy you always carded for beer, because he looked underage - and even though you know he wasn't, you did it anyhow, and he complained - but corporate policy, ya know?

Then a break - go face the aisles, or stock new stuff, or pull the old stuff. Have some lunch or some dinner. Eventually close the store, sweep the aisles, clean up, shut down, and go home.

For me, it was an "after school" kind of deal - not a career. But it wasn't a terrible job. Just difficult on the mind and body (it was a bit boring at times). My fellow employees were all friendly, and management was nice and helpful. I always got paid on time, meager as it was.

It gave me a better perspective and understanding for those working behind the counter, because I know what they're going thru. Plus, it also taught me how to properly scan items quickly (though when I started, the place I worked at didn't have scanners - they didn't get those until a couple of months later). Even so, when an item doesn't scan, I find it a bit frustrating, but most systems have a way still to enter the UPC number...

...still, I wouldn't want to go back to those days.

I strongly dislike the automated checkouts; I prefer to wait in line to get a human.

I don't like the bagging system that has to ensure you put your item into a bag, and that has a quite noticeable delay in responding to your actions.

I also don't like the tiny area they have for the bags. What am I supposed to do? Put full bags back into my half-emptied shopping cart alongside items that haven't been scanned yet? It makes me uncomfortable, like they'll think I'm up to something.

I also tend to buy items in bulk, like 2-liter bottles, and it's nice to say "I have 8 of these" and pass one to the cashier.

Yeah, when I checked groceries and had the produce codes memorized, I could blast through a grocery cart full of food in the time it takes me to check out a dozen items in the self checkout. I always go to a person for anything other than the express line.

> I also tend to buy items in bulk, like 2-liter bottles, and it's nice to say "I have 8 of these" and pass one to the cashier.

This is really not the case the self-checkout lines are designed for.

I use them almost exclusively for those cases where I duck in and grab a couple things quickly, and want to get the hell out. It's always faster and easier than even the 12-items-or-less express checkout line, not to mention the regular lines, where you get stuck behind somebody with a cart rounded to the sky and $400 worth of groceries in it.

OTOH, you can make sure your 2-liter bottles and tortilla chips are bagged separately.

>It's much faster.

It depends on the store. At the grocery store a few miles away, I can scan an item and immediately put it in a bag. At the grocery store closest to me, you scan the item and place it on a belt, which takes it to the end of the counter. You bag everything at the end. By contrast, at that store, a cashier can scan an item an place it directly in a bag.

Those self checkouts aren't automated. Those are having human customer do the checkout work.

Automated would I walk out the door and rfid or whatever senses what I'm carrying and charges my account.

Or, I don't go to the store at all and shit just shows up on my doorstep, periodically or whenever I hit a button.

I feel like a lot of people miss this point. The technology has existed for years although clearly the cost to implement a touch screen console has also plummeted.Even accounting for that most self-service checkouts are just a tilled turned around.

I wonder how much of this is super markets jumping on to the fact people are conflating automation with using a machine and assuming self-service checkouts are proof of a technological revolution. At the end of the day the same amount of work is being done at the till as if a cashier was working but it is now being done for free and a company won't miss a chance to capitalise on an opportunity like that.

Amazon Prime Pantry seems to do the latter. We've been getting some of our stuff delivered that way the past few months.

I shop at Aldi's where the prices are low and the employees are well compensated. The owners, 2 brothers, have some unique ideas about efficiency. They haven't installed any machines because they are slow, customers are slow, and they don't handle edge cases well. Employees are few and far between. Generally 1 cashier and 2 on extra busy days.

The Target on the other hand has rows and rows of empty checkout tills. They try to get you to use the automated tills but that line is so much longer that it just boggles the mind. Any time savings is wasted on the machine having an error,

While I haven't used Target's, I do like Home Depot's self checkouts. They are pretty intelligent on their own, like just swiping a card to initiate the payment process, even if you don't press the "pay now" button. Then, when you do encounter a problem, there's an attendant for every four checkout stations that can help you from their workstation. Usually they know you're having a problem before you can even turn to tell them so, and by the time you do, they're waving you on and the problem has been resolved. My only issue with them is that they shut them down about an hour before closing, and waiting for a human cashier is even more annoying after the self-checkout experience there.

How come there is no remove scanned item option like the real cashiers have. It happens to me -every single time-. Prefer the human option, truly.

What if we had machine intelligence monitoring employees to decide which ones are going above the call of duty? Too Black Mirror?

No thank you. The good companies already strive to recognize it, the bad ones will use it as grounds to let someone go for bending the rules.

Arguably there are people who have it in them to do better, but feel that their extra contributions would not be recognised. Tipping is supposed to be a solution to this in some jobs, but there's not much of a culture of tipping airline check-in desk clerks in most cultures

Sure. But if their company is not already trying to recognize it, why would they pay for software to do it? And if they are, and are just not good at it, why would they recognize software as the solution? I just haven't seen any companies going "We sure would like to recognize our best employees, but we just have no way of knowing why they are". -Especially- in customer service land.

i have flown into JFK over 10 times and to this day i still dont know how those passport scanning machines work. i try various permutations of placing it on the machines but everytime i just give up.

I've heard that in the UK they have people controlling those automated passport gates - the matching of the photograph is not done by machine. Needless to say, mine never works and I have to go to the person who asks "is that really you?" and I reply "yes" and that's enough human touch verification to pass through

I've found that there are quite a few things you need to get right with the UK gates, and the UI doesn't do a great job of encouraging correct usage.

In particular, I was instinctively looking at the screen instead of the camera, despite the big sign telling you not to.

It might not be everywhere, I find I have different level of success at different airports in the UK. The machines at Gatwick never have a problem with my passport so maybe they have 'human help'. At Stansted however my passport has failed to work at least the last four times I've been through.

Mine used to not work, now, the same passport works everywhere without flaw. I notice I have to take my glasses off otherwise I does not, but that is not too strange as my passport pic has glasses. Not that it matters, most airports have manual checks after automated check anyway.

Where are the second checks? I've used the system often in Europe, mostly London, and never had a second check.

Heathrow, Malaga to name two.

I think some passports are just bad at this, but I have no plausible theory as to why. Mine never scans without at least 3 attempts at any scanner, in any situation, anywhere I've been in the world.

Try the new mobile passport app instead: http://www.mobilepassport.us/

The simple answer? Because humans are actually intelligent.

Find me someone in a job who has the exact same day at work every day with no exceptions and I'll show you a job you can automate. Exceptions are what define things that machines can't do well, and every scenario brought up in this story has routine exceptions.

More importantly, even in cases perfectly suited to automation, we can't pretend all the software we build is perfect or even remotely approaches the quality that can be produced by the brightest minds in the industry, either. Everyone wants to automate everything, and you can be sure most of it will be done poorly. Even the simplest human at least does a better job than a poorly built piece of software. I can rage 24/7 about how bad the software people write is. At least 90% of it. I rarely find myself annoyed at how incapable of doing their job the typical service worker is, and when I do it's virtually always fixed by their boss or a manager.

Automation done well is a tool. It turns many jobs into a few, where people handle exceptions and have actual administrative control. I've had tons of bad experiences with automation that has little/no human backup and it's by far worse than dealing with people. And this is coming from an introvert.

> Find me someone in a job who has the exact same day at work every day with no exceptions and I'll show you a job you can automate. Exceptions are what define things that machines can't do well, and every scenario brought up in this story has routine exceptions.

We're no where near machines replacing human operators 100%, but a 90% solution with humans handling the exceptions is still a 10x boost in productivity.

Where I'm at, we're big on getting our people to manage the exceptions rather than doing the work. If you're looking at the entire job to see if it's routine then of course you won't find much room for automation. But if you can automate even 10% of their day then that's a big win.

And yes, the software is total shit. There are no unit tests and it's written in RPG (half of it dates back to '95). When it breaks we go back to managing the exceptions, much like "talking to the manager" when the service worker screws up.

The company wouldn't be afloat without automation like this. We would be completely crushed by our competitors. I don't think the automation needs to be done well at all. It just needs human backup.

It's not so much that humans are intelligent as it as that machines are just really dumb, which is really just because humans are still too dumb.

You are entitled to be skeptical based on your personal experiences. That one smart guy has had terrible experiences with 100% automated pre-2020 systems is not a good enough reason to shy away from the idea of pure automation.

Bingo. Automation is cheaper, not higher quality. The only people that consider not talking to service agents or cashiers an improvement in service have social anxiety issues or something similar. A checker at a grocery store actually does all of the work for you. Objectively it is better in every way unless you consider it a negative to be involved in human interaction.

How about when they scoff at your having brought re-usable bags, slightly demanding of them to do something different? Or when they pack your goods in a way that damages them? Or puts colds next to warm, or highly heat conductive items? Or when they're clearly sick, and likely spreading that illness to everyone that comes through that lane? Or when they're at the end of their shift, hate their job, and give you that lovely dead-eyed stare that just makes you feel warm and fuzzy when you try to make small talk?

How about when they reek of cigarettes? Or when they make a mistake (and they do) and all the people behind you start eyeballing you and the cashier, clearly upset at the delay? How about when they ring up produce incorrectly, or double count items, overcharging you? How about when something comes up wrong, and then they argue with you, because they're stressed with a queue to process?

Human cashiers being objectively better in every way than self-service automated lanes is not correct, especially as you sample more and more cashiers. SOME cashiers are better. Many aren't.

> The only people that consider not talking to service agents or cashiers an improvement in service have social anxiety issues or something similar.

Or they are in a hurry. I will take the faster option every time and that usually means no human.

> A checker at a grocery store actually does all of the work for you.

What if there was no work to do? Amazon is exploring that concept and if it works as described (I haven't used it yet, but have walked by and it seems to) seems objectively better in every way.

https://www.amazon.com/b?node=16008589011

> Or they are in a hurry. I will take the faster option every time and that usually means no human.

Correct, but this is the result of the store intentionally routing you through the self checkout by under staffing cashiers to save money.

>What if there was no work to do? Amazon is exploring that concept and if it works as described (I haven't used it yet, but have walked by and it seems to) seems objectively better in every way.

Excellent point. I should have used "self service" instead of "automation". It was a poor choice of words on my part. Full automation to remove anyone from needing to do the work would certainly be objectively better. The article linked is more about self service than automation.

Sure, you can have bad experiences when dealing with humans. My best experiences are with humans and my worst? Always a machine at the root of the problem. Even the article mentions this point, that the humans were there to soften the blow of a machine that won't print tickets unless conditions are met. There are some jobs where people are so miserable by their meaningless, poorly managed job that I wish they had been automated out of existence but even where I'm agreeing with the author it's to allow the people to move more center stage. The places we're most agitated are places that only exist because the first time our species passed over the problem there wasn't a better way of doing it and replicating what's been done is easier than recommitting to the problem for a modern solution. Other agitations come where we don't want a rigid machine-like process, such as the ticket printer, where we want humans in control to make reasonable choices.

It's not good enough to just to get them out of the deep machinery of the planet, it has to elevate them somewhere or we're just replacing people from the instrumentation of the universe without good reason.

> Always a machine at the root of the problem.

I have a perfect example of this.

Years ago, my mom's coworker discovered that her bank card no longer worked. ATM's would error. Online systems would reject all transactions.

So she went to a bank branch to sort it out. After explaining to the teller her situation, the teller brought up her account details, looked up from her monitor and said, on a straight face: "You are dead."

What had happened was that she had had a full namesake who had also had an account in the same bank. When the other person had died, and her estate had closed the accounts, the person doing the account termination had only looked up account owners by their full name. Closed all of them, with the explanation "account holder deceased".

Human errors happen. So do computer errors. It only becomes a problem when humans rely only on what computers tell them. This particular bank teller didn't even realise that the facts stated by the computer were contradicting what was literally standing in front their eyes.

Critical thinking is a dying trait.

> This particular bank teller didn't even realise that the facts stated by the computer were contradicting what was literally standing in front their eyes.

It may be more complex though. The machine can be wrong, but so can your eyes. The previous customer coming to that teller with the same situation might have been someone who stole the identity of a deceased person. You can't just assume the machine is wrong because it disagrees with what you see.

This is also not the problem of machines per se, but of large systems - where different parts are handled by different people who don't know each other (or may not even be aware of each other's existence), and a set of fixed procedures (i.e. meatspace algorithms) are used to coordinate everything.

Isn't it exactly a human error that you are describing? The root cause was a distracted employee that input wrong information in the system, not the system being flawed...

Sure, the root cause was incompetence when closing down accounts but that is not the problem.

The real problem manifested in human blindly accepting what the computer told them. If there is no reason, or even NEED to apply critical thinking, why is there a human involved at all? This is not an indictment of automation over human labour. I would rather that humans are in the loop precisely because they have the ability to spot these kinds of errors - and help to correct them. Computers are good at repeating mundane tasks. Humans are not. We should be making most of their distinct abilities, not molding the two into same form.

Now, Temporal rightly pointed out that this might have been a case of attempted fraud. But if you are trying to spot fraud, spotting anomalies and being critical on what you accept as objective truth should be on the top of your mental map.

The root cause was a system that allowed a potentially incompetent employee to indiscriminately close bank accounts without at least a date of birth or social security number confirmation. It's a problem that could have been prevented with a four to eight digit input field and an if statement.

The system isn't flawed only in the sense that the bank doesn't give a shit. They get to reap the returns on your capital either way.

Well the problem was that of identity. The bank assumed that a namesake was the same person, rather than relying on a more unique identifier.

Are you sure the teller didn't just have a dry wit?

Or did they then proceed to refuse to deal with the living dead person?

So what happened to the money in the account? I mean, when you close an account(s) of a deceased relative at a bank (presuming you have the legal right to do it - it takes more than just "showing up" and claiming such), the money is generally withdrawn and processed as a cashiers/bank check (depending on the amount - for certain large amounts, there may be a wait period) - and given the person closing the account(s).

...so - what happened to your mom's co-worker's money?

The bank refused to admit error, at first. It took more than two weeks of constantly being on the phone and pestering before they reopened the account.

In the meanwhile, it was pure luck that she didn't have any critical expenses, such as taxes during the time window.

The bank never issued an apology.

Traditionally, the original paperwork machines were human-based and called "bureaucracy". They were well-known for being extremely infuriating, even before they were ported to computers.

The problem here is trusting the machine and not trusting the actual human and not giving human any leverage to do anything extra.

In all cases where machine and human are assigned to a job as a unit, human should have authority to override automated decisions.

There are two reasons why it isn't so: business doesn't trust its drones, and workflow tends to break down on unexpected input. But it should be.

With full audit trail of the overrides that makes sure the human gets sacked if they fuck up, sure. I dislike limiting human agency, but unfortunately, most employees aren't True Believers of the Corporate Cause, so you can't assume they'll be loyal and diligent all the time.

Full audit trail is the first thing you should put in when building a back office system.

Some good points you bring up.

This is not a response to your entire post, just a small one to "where we don't want a rigid machine-like process" - I think this is something that machines are perfectly capable of solving. We are not arguing about whether or not, in the ticket printer case, the machine should or should not print a ticket outside it's rigid specified rules (late by 1 minute etc). In fact I think that most of us would agree that as a matter of principle such grace-periods are a fundamental lubricant of a comfortable society. We just have (historically) had to plead our cases and rely on the grace of other humans to grant such meager exceptions to rigid rules. We are just arguing at this point about what the numbers for such grace periods should be, machines can easily implement them once we find a happy medium. My $0.02 anyway.

If we implement a 60 minutes deadline and a 10 minutes grace period, we might as well just set the deadline to 50 minutes and be done with it.

I'm an advocate for a hybrid approach. Get machines to do the bulk of all the work, and have a human handle problems, edgecases, and "bending the rules for loyalty".

Arbitrary deadlines like this are heuristics that we give to systems or people who don't have realtime access to the full picture so they can make decisions that are probably, on average, more likely right than wrong. The human checkin clerk can, at their discretion, call the gate and see if boarding is delayed, or find out how the security lines look, gathering additional information to make a decision, but the one hour deadline is a rule of thumb that they can use to make an initial call. But even with the additional discretion they have, they are still guessing whether the rules are worth bending for this passenger, what the real value to the airline is of getting that customer on this flight or bumping them to a later one.

A computer replacement that doesn't have any additional data, and isn't programmed with the discretion to make those calls to find out additional information, is going to just make the call that checkin is closed. It is probably an inferior solution to the human.

A computer replacement that has information about current security wait times, gate status, as well as information about all other flights, the location of the inbound plane, the estimated lifetime value of the customer standing in front of the kiosk, what competing airlines are doing, and the exact current company financials, can put all that data together to make far better decisions than the human checkin clerk could, in the best interests of the airline, and it doesn't need to be given arbitrary rules like 'turn people away if they show up 59 minutes before scheduled departure'.

I totally agree: generalizing something always has a cost. Not every experience with a machine is smooth, not every experience with a human is smooth. Therefore we need both! And so far only humans are able to improvise, understand, use emphaty... But machines are fast, scale easily, never stop working (haha)... Both is the wise choice!

Great post. I hope we get over the stigma of "destroying jobs" when it comes to ones like these that machines can easily do. Whenever I walk around the streets of New York I can't help but think about this. Be it the miserable cab drivers sitting in traffic staring at red lights their entire life; or guys hauling dusty insulation out of a demolition site and throwing it into the back of a truck while coughing up a fit; or delivery guys biking up and down the hostile avenues of manhattan in whatever weather to deliver Chinese food.

To me it will be a great tragedy if we never end up filling these necessary jobs with machines, be it for political reasons or something else. There's nothing stimulating or fulfilling about these jobs. It's first world human misery.

> There's nothing stimulating or fulfilling about these jobs. It's first world human misery.

Tell that to the workers who actually depend on those jobs to live. It's easy to get rid of "unnecessary" jobs when they're not your job.

This is the stigma I'm talking about. My argument is that humanity in general needs to ascend and let machines take care of the dirty work. It's not a law of the universe that there must be people who depend on shitty jobs to live.

But I'm one of the crazies who believes that in an ideal world, there would be no 9-5 norm and it would be possible for people to spend most of their time with their family and taking care of their own physical and mental health, rather than sacrificing their waking hours to keep the unemployment rate down. These people are already acting as fleshy machines.

I'd like to live in the same world you describe. I work 9-5 and of course, I'd rather be doing whatever I please instead of going to work. But longing for the destruction of these jobs without a realistic alternative (e.g. basic income is still a pipe dream) is too near-sighted.

Problem is, there isn't a lot of pressure on the political forces to implement any sort of basic income, since there's still a high enough demand for meaningless human labor that it's easier to keep focussing on giving a job. Automation is potentially going to finally push us over the edge where trying to find menial jobs for every worker is no longer a tenable solution. With the current political climate, we won't ever get past it being a "pipe dream" without that being a the case, or at least it being a tangible threat in the near future.

>Problem is, there isn't a lot of pressure on the political forces to implement any sort of basic income

You mean a completely unproven system that most economists believe cannot possibly be funded without increasing spending to unheard of levels and no you don't save enough by cutting the "inefficiencies of the current system" and has a good chance to quickly die due to inflation?

Can we please stop bringing up UBI as a silver bullet to all problems in every automation article?

> You mean a completely unproven system that most economists believe cannot possibly be funded without increasing spending to unheard of levels and no you don't save enough by cutting the "inefficiencies of the current system" and has a good chance to quickly die due to inflation?

I could have sworn I've heard that somewhere... oh wait, it's the same tired argument used against every progressive policy since the New Deal. You'll have to forgive UBI supporters for not taking economists seriously given their track record over the last century, during which the developed world increased spending to "unheard of" levels about a dozen times.

The US income tax, which required a Constitutional amendment, is only 104 years old after all.

> Can we please stop bringing up UBI as a silver bullet to all problems in every automation article?

I don't see anyone seriously claiming that UBI is a silver bullet, just that it's the only bullet we have left against our future robotic overlords.

Especially since you can't solve supply-side problems by manipulating demand.

Universal Basic Income cannot meet the minimum survival needs of the population as long as it does nothing more than transfer money.

The vision most people have when thinking about UBI cannot be realized without some form of command economy separated from the profit-motivated consumer-goods pipeline. If all you do is give poor people cash, their landlords raise rents, their grocers raise prices, and their bus lines raise fares.

Any UBI that I'd actually expect to work is nowhere near as simple as just cutting checks.

Sorry, I never meant to imply that I hope we do it hastily and without an alternative. It's an economic challenge for the ages, all I'm saying is it's worth taking on.

Strictly looking at everyone's selfish interests we will head to destruction of jobs which can be automated because it's the companies that decide which people to employ. If a law is passed that people must be employed, automation will happen elsewhere and companies from abroad will be more successful.

In other words, it's inevitable.

> I'd rather be doing whatever I please instead of going to work

Not the same thing, the argument here is for more meaningful jobs, not just fewer bad jobs.

If there was less expectation of getting a bad job, there would be more support for getting good jobs.

The problem is, we can't ascend until we fix the underlying "if you're out of a job then you starve" problem - put UBI infrastructure in place, and then we can let the machines do the dirty work.

That's a pro not a con. If everyone was guaranteed food/shelter etc there'd be an epidemic of people who did nothing with their lives - there already are a lot of people who do nothing while working a shitty job.

Basic needs are motivation for doing things, but problematic too - they can also be demotivating.

The problem is that mobility & opportunity is costlier than food & shelter & small luxuries. As such, if you give people enough resources to do something with their life, they can just use it for basic needs & sls and do nothing else; why risk a comfy, if depressing, lifestyle on opportunity?

It's a difficult problem, how to ensure people always have opportunities available, while 'punishing' lack of ambition. Maybe the nuance here is the government becoming more aware of the impact of psychological factors, but this hasn't turned out well in the past...

Why do you care if other people are happy being unambitious? Most people aren't horribly ambitious. How are 7 billion people all supposed to change the world?

Ambitious doesn't mean trying to be president. It means doing more than eat, sleep, play xbox. Most people do more than that.

To me the problem here is the transition from Western capitalism to the [socialist] nirvana you describe.

It requires the rich and powerful to give up their lofty positions .

Not clearly socialist to me, But why do the rich and powerful need to give up wealth?

Because otherwise the dirty peasants will invent rationales for dragging them out into the streets and murdering them.

It isn't always easy to predict when it will happen, but horrifying, deplorable violence seems to be the only way for the lower classes to effectively reduce wealth inequality.

Some people are getting nervous enough to buy land in New Zealand or condo space in hardened apocalypse-survival bunkers. They know whose heads will roll if the guillotines ever get reassembled. They don't actually want to voluntarily give up their wealth, because they know how their peers treat those that don't have it, and because being constantly surrounded by luxuries is addictively awesome.

So we're frequently stuck in the situation where the poor think they're poor because the rich are too rich, and the rich don't care about what the poor think, because they're too busy kite-surfing in warm turquoise waters, buying recreational estates in Kauai, pitching giant museums to major cities, and generally living the good life.

The rich don't actually need to give up their wealth, but they might want to consider paying closer attention to the zeitgeist, and backing off on meddling too obviously in the lives of others at the appropriate times. Just like you might closely monitor the weather on your giant yacht, and make for port whenever a storm is brewing, you monitor the political climate and shelter your wealth when necessary.

Sounds like you should keeps your wealth and buy uzi-armed guards. Giving it away to peasants will allow them to buy uzis.

> There's nothing stimulating or fulfilling about these jobs. It's first world human misery.

That's really nice of you to make that determination for those miserables. They're probably so uneducated they don't even know how bad they have it. They are lucky that's people like you can come along to take away those horrible bill-paying, medicine buying, sources of personal satisfaction.

You could have made the same inane argument for all the uneducated factory laborers who have been replaced by machines. Nowhere did I suggest these people should be broke and without health care instead. And you can't argue with me that these people I mentioned genuinely, deep down, enjoy what they do for a living.

>You could have made the same inane argument for all the uneducated factory laborers who have been replaced by machines.

There's nothing inane about the argument. Have you ever been to Detroit? I've lived there for years. It's a horrible thing for these people to have been replaced by automation in their lifetimes. They didn't see the benefits in the increased productivity like their employers did, they just lost their livelihood and had to scramble for jobs that paid much less while simultaneously trying to retrain into other fields, if they were enterprising. The entire region is obviously devastated from the effects of it. Their prior automated work being "unfulfilling" and therefore terrible is a) debatable and b) nothing compared to the struggle of having no income at all. One guy I knew ended up homeless after he lost his job, then with a leg amputation because of frostbite from sleeping outside. When I met him he was making a living ticket scalping in his wheelchair all day. The short term consequences are real and terrible for these people.

Doesn't this just outline the terrible dependence these workers had on their jobs? No compensation of duty of care exists from employers; any number of things could have caused the workers to lose their jobs, they were always at risk even when not struggling.

as an analogy: Crack addicts going through withdrawal don't conclude from their pain that the drugs were a good thing. OP wasn't suggesting workers should suddenly have their jobs taken away, but that the whole situation is bad. Changing the situation, and how to do it, is an entirely different topic.

Is this sarcasm of some kind or a legitimate point you're trying to make?

>Doesn't this just outline the terrible dependence these workers had on their jobs?

You're effectively saying people shouldn't rely on their jobs to support their livelihood. What other option do they (rather, we) have? Maybe factory workers should have had the foresight to see the industry changing toward automation, but even if they did, then what? Keep working to support your family 40 or more hours per week but learn an entirely new craft on top of that? Does that actually seem reasonable?

>any number of things could have caused the workers to lose their jobs, they were always at risk even when not struggling.

Yes, but the industry still existed. Say you worked for 10 years at Ford and got fired. Maybe GM or Chrysler has an opportunity for you, and they'd probably love your experience.

Does that sound familiar at all? Say you work for 10 years at Google and get fired. Maybe Facebook or Amazon has an opportunity for you, and they'd probably love your experience.

My point is that putting the blame on the employee shows a complete lack of empathy for someone because they "chose the wrong path in life". Does the employer have a duty to support the individual who became obsolete? No, but we should all collectively have a duty to make sure the lives of these people aren't completely destroyed, because as unlikely as it seems right now for those of us in the software world (and I'm sure it seemed unlikely in 1995 for those in the auto industry), we very well may be next.

> What other option do they (rather, we) have?

Better jobs. What options do a crack addict have? Very few. That doesn't mean they are better off now being deprived off it, they would be better off never being dependent.

> Keep working to support your family

Why did they have a family when they were at risk? Either didn't perceive the risk, they operated under a false sense of security; or no other options were offered because the jobs that existed were seem as suitable.

If there were no jobs they wouldn't be inclined to settle down. I note that turning down a job can be grounds for losing some welfare benefits.

> Maybe GM or Chrysler has an opportunity for you

speculative. In the context we are talking about there are dangerous correlations: If one worker loses their jobs, the chance that another will increase (due to the possibility of a common cause, in this case automation), not to mention the chance of greater competition from other out-of-work workers (or lower compensation).

> Maybe Facebook or Amazon

Are you still talking about bad jobs? I'd love programming/tech to be more automated, somehow.

> putting the blame on the employee shows a complete lack of empathy for someone because they "chose the wrong path in life"

that's your strawman,not my argument.

> Does the employer have a duty to support the individual who became obsolete?

Do you mean at current, legally? No.

legally or morally, in the future? yes, maybe, who knows. Corporate tax pays for some social welfare. Things won't change if you're not allowed to criticise the current system.

> we should all collectively have a duty to make sure the lives of these people aren't completely destroyed

I disagree, this isn't enough. We have a duty to make sure the lives are full of this kind of risk too, "completely destroyed" is too low a bar.

> the lives are full of this kind of risk

should be:

"their lives aren't full of this kind of risk"

That is horrible, and I understand that it's a big socioeconomic challenge. I don't claim to know the answer, but I hope we find it as a species.

Why not? I worked on an electronics assembly line for a while doing electro-mechanical assembly. I enjoyed it and got satisfaction out of creating a physical product. It was kind of fun, not too stressful, sometimes monotonous, but also not too difficult. And after 8 hours I went home and didn't have to think about it until the next day.

I'm not saying that everyone enjoys that, or that we should keep those jobs around because I found it fun. But don't assume that just because a job seems menial to you that it must be horrible for everyone who's doing it.

Our life force diffuses over the options we perceive as viable and available. Many years from now, when we are all gone, kids will look back and wonder how humans could have ever been doing such jobs. Much in the same way we look back and find it hard to imagine how humans could spend the entire day chasing down prey and running from lions. It's that much of a leap that is required.

Much in the same way we look back and find it hard to imagine how humans could spend the entire day chasing down prey

Lots of humans today pay good money to get a chance to go out into the woods and spend an entire day chasing down prey.

For entertainment. Not because they'd die if they didn't.

Sure. But I also know people who still romanticize the whole hunter-gatherer lifestyle and wish they could live that life (whether or not they'd actually want to keep living that life after first month is of course an open question).

Who knows, in a couple of hundred years we might have people romanticizing spending all day in a cubicle working at a mid-sized accounting firm :)

Having worked retail and customer service jobs firsthand, and having friends (and obviously coworkers) that have as well, I'd assure you that the statement you quoted is accurate for the majority of those workers. Most people leave for any other job the moment they can, and those who can't get old and bitter.

<There's nothing stimulating or fulfilling about these jobs

I object. I can't tell what someone else feels is satisfying and I believe you can't either.

Have experiences and relationships with all types of people and jobs. I have and agree with Mike Rowe. The happiest group of people he ever worked with were those he filmed and performed dirty jobs with.

After freeing myself from the stigma of a professional job I gained more happiness with much less responsibility and much more free time. Like those jobs you mentioned. If I had known that when I was on call awaiting for the next system to crash only to get interrupted time and time again I would have intentionally changed careers faster.

> or guys hauling dusty insulation out of a demolition site and throwing it into the back of a truck while coughing up a fit

Those guys/gals may actually enjoy manual labor - some people just can stand the thought of being cooped up behind a desk in a stuffy office (or at least that's how they may envision it).

Though I will say that anyone dealing with insulation - especially in a place like NYC - who doesn't wear a minimum of safety equipment (goggles, gloves, respirator or dust mask) is just begging for asbestos-related health problems down the road (heck, even regular insulation can cause respiratory health problems over time).

A few years ago, I worked at a California university where 5 people manually keyed in all transactions for the day. Something an export from the bank could do in minutes.

I cannot fathom at all why anyone, other than top 10% or so whose wealth depends on everyone else's labor, is fighting automation. Or wants to keep a job. I would gladly let machines do the work so I can play fetch with my dog all day.

Instead of converting productivity into less work we turned it into more profit. That's a bad trade.

When the machine does your work, and the bosses keep your wage for themselves. What are you going to eat, or indeed feed your dog, how well you past for it without a wage?

Main issue are the laws wrt land ownership etc, e.g. the old "means of production".

Except I don't consider the products that allow production (factory, machines etc) to be the means, so much as access to those products on the market, lines of credit, land ownership on which to put a factory etc.

production needs to be achievable for anyone, otherwise you have an effective monopoly. Means of production should be considered akin to a basic utility.

Why wouldn't playing fetch with his (and other people's dogs) not become a perfectly viable job in a world where consumer goods are dirt cheap and human labor relatively scarcer?

Why would labor be cheap?

I don't know. The premise was precisely that labor would become more expensive.

The idea is that if we can automate "everything" (i.e. everything that can be automated), then we'll be left with human-only services. Since humans are scarce, and there will be rising demand from the income effect, people can still perfectly live off exchanging services.

I'm not sure that makes macroeconomic sense, so if someone wants to clarify misunderstandings for me, I'd be happy to hear it.

Well, what makes you think that humans are scarce? Unemployment is a thing already. For example, you can take a look at what e.g. waiters or Walmart greeters earn right now - that's a price set by supply&demand. Automation moving more people to such jobs means the supply of labor would increase, driving down the price for services.

Assuming the jobless people will have less income than if they worked, and more time to do stuff themselves (some of the services industry is essentially something that you'd prefer to make yourself but you're paying to save yourself some time) demand for service jobs would decrease, again driving down the price for services lower than it is now.

> Automation moving more people to such jobs

I don't think the set of possible jobs is limited to either existing jobs, or in particular waiters and Walmart greeters. That's kind of the point. We invent new jobs all the time. Won't we come up with jobs that cannot be automated, or that we'd rather be performed by humans?

> Assuming the jobless people will have less income than if they worked

Well, that's more of a tautology than an assumption, and fails to engage with the scenario highlighted above. The point is that people will get other "jobs." (I put it in quotation marks because I'm pretty sure that people will get paid to do things we'd never thought of as "jobs" at some point).

Of course, S&D still works the same way. But human desire is pretty much infinite, and we can always come up with new jobs and new markets. The idea that all of them will be desirably automateable seems unproven to me.

What's preventing those new jobs from being done right now? We already have all these unlimited needs and infinite desires right now, but we aren't willing to pay a minimum wage for them at the moment or we would be doing so. Also, there clearly are all kinds of personal services possible which are bought by a few of the rich but not en masse - there is not a demand for them, and for the average person probably cannot be, i.e. if a service is nice but people would value 1 man-hour spent on it as 0.99 man-hours of their time, then that service can ever be in demand only for a wealthier-than-average minority.

Yes, it's possible that some new habits will create some new professions in time, however, for the short term (e.g. the next 5 years) such new occupations won't create millions of jobs (e.g. because unless right now millions of people would be preparing for them, the people won't be ready so soon) while automation will put millions of people out of these jobs. Who knows to what equilibrium it will come out eventually, but in the short term the labor supply will be dominating over demand and the price (and perceived value) of labor will continue to fall.

> Since humans are scarce, and there will be rising demand

Seems to me that humans will be in surplus? It depends on what price must be payed for the automated things, which I assume won't be free?

Not free, but presumably much cheaper than they are currently.

quote my self since you didn't see it.

> Instead of converting productivity into less work we turned it into more profit. That's a bad trade.

This guy frankly sounds like a monster, willing and eager to put company profits above all else.

That said, he raises valid points. If a machine can do a better job for less money, why shouldn't a machine do it? But he fails to mention hidden costs. What do we do with the workers who lose their jobs to machines? Eventually there will be no work to retrain them to do.

I can envision a future where machines build other machines and there's just a few executives remaining at the top of some companies, collecting ever-increasing profits as workers are eliminated.

I want to say that a universal basic income will alleviate the problem, but are there other options? How do we live in a world where work has been mostly eliminated?

I don't know how far it will get anyone, but I like to think that craftsmanship will be greatly valued.

I truly enjoy things that someone made by hand. The quality isn't even always important. I buy chests and trunks and shop-built tools even when they're not particularly good quality. I enjoy them, they're unique, and they have a story that you can sometimes decipher from the tooling marks and clues left in the wood.

I think such things will always fetch a premium, and perhaps when robots can take care of all the immediate issues--manufacturing, delivery, food production, whatever--then we will see a sort of reversal and people will start to pursue some more intimate, small-scale creation.

Off the top of my head, I can think of quite a number of things I enjoy that robots can't provide for me (yet?): home-made baked goods, pasture-raised eggs, hand-created furniture/kitchen items/artwork, learning, creativity (with a personal story, especially!), well-trained animals, etc.

That said, I'm not a typical consumer, and those things I'm interested in are also things I am interested in producing myself.

Still, I would not be surprised to see a barter type of economy spring up alongside a basic income.

I agree and like those things too. Sadly the economics of producing those things in the current environment typically lead to subsistence levels for the crafstmen.

In the American economy of 2017 if you don't have scale, pop, leverage, access to the discount window etc... you are going to be poor or close to it.

Its not company profits, he clearly makes the point that customers in many faces prefer to interact with machines than humans and its cheaper to have machines than humans.

The work of human in the author's world is designing machines and automate the next step of tasks that haven't been done or spend time on doing more creative tasks and be more inventive.

It's both. Note, I did say "do a better job for less money". As someone who is a bit of an introvert with bad hearing, I often dread simple interactions where there's a potential for confusion or a language barrier, so I love a self checkout or a touch screen ordering experience. Cheaper or not, this is usually a better experience for me.

Still, that doesn't solve the problem of how do people put food on the table? We can't all build and design machines.

https://en.wikipedia.org/wiki/Social_ownership

This doesn't work so well when goods are scarce, but...

Basic income is a start. Then they figure out what to spend it on. Ideally PhDs. Of which you use that to work in that specialization; solving problems, figuring things out. Patenting discoveries, selling patents, patents will go up as companies rich in cash by not having employees will bid on said patents .. or copyrights, thus those will increase and enrich those who filed such to pay for more research.

Partly I feel one of the best things we can do is redesign grade school, say from 7th-12th grade. The majority of it would be the same I guess..but there would be a strong orientation towards teaching individuals how to invest and trade; let them get hands on experience in actually doing it and so forth. We'd develop a universal basic income that people would start getting after graduation. They can invest/trade with it. This is not going to be the sole income point, but instead they can use that to pay for school.

Some wont want to get an education, which I understand. So they do other things that young adults attempt to do, start bands and what not. Many will fail at that particular attempt; but hey they get basic income and know how to trade with it so they at least wont starve.

I think skills like crafts will be important too despite robots. One if you know how to make a good hand made suit you are set, the rich are your clients. Blacksmithing odd items could be something in demand. A hobby of mine is working with metal, I make a lot of cool stuff, sometimes I sell or trade, but usually I keep and forget about it.. usually until someone offers to buy it. Either or, I think the future is not so bleak, we just need basic income, maybe fix the schools a little bit, encourage people to get PhDs, and also spread y combinators about.

Short version: how about we just literally give everyone a job that needs one?

Long version:

As Hyman Minsky* points out, financial capitalism is inherently unstable, in the sense of having perturbations that are big enough and frequent enough to transform the lives of individual human beings. Our impending wave of automation is one of those perturbations.

We're probably not going to get rid of capitalism anytime soon, and even if we did that through violent revolution, the resulting perturbations would almost certainly be worse. So we've got to adjust what we've got. Plus, capitalism does have some good features.

We could literally ban robots from taking some kinds of jobs, but that's a temporary fix, and do we honestly want people driving trucks all night, hopped up on drugs, in order to feed their family that they never see? We should get rid of shitty jobs where we can.

I'm not sold on universal basic income. It's really expensive, and it eliminates one of the best features of capitalism, which is working for your money. Working for your money gives people meaning, purpose, and a network of human relationships, which help promote a healthy society where people feel valuable and valued.

So how about this: universal basic employment. The government literally guarantees everybody a job. At a living wage, for a reasonable time commitment (less than 40 hours a week, I hope). As the capitalistic side of the economy shrinks, government automatically creates jobs to pick up the slack.

Capitalism coupled to an adaptive socialist job-creation program.

BUT WAIT, you scream, THAT WILL JUST EXPAND FEDERAL BUREAUCRACY AND EVERYONE WILL END UP EMPLOYED IN A SOCIALIST HELL!

Devil, details. I would argue that a related deep issue in American society is the breakdown of our local communities. So let's kill two birds with one stone: build an adaptive hierarchy of taxpayer-funded job creation driven from the bottom up.

Local councils meet with their neighbors, figure out the kinds of things that need to be done, built, whatever in their communities, match their needs with the local unemployed talent pool. Counties monitor the local needs. States monitor the county needs. And the feds feed money as needed and operate with a light touch.

All these local community jobs pay the same, and it's a living wage. This serves as a soft wage floor in place of a hard minimum wage. It's a bit like an Americorps that pays enough to support you and your family, all the time, for anyone that needs a job. It has a lot of features of the New Deal, actually.

* Underappreciated economist, newly resurgent since the 2008 financial crisis. If you took some economics classes in college, you should be able to understand his writing, but he's not a great writer. I'm just barely getting familiar with standard economic terminology, so his stuff has been a really hard slog.

Agreed. A minor nitpick about "I'm not sold on universal basic income".

I'd argue that what your are proposing is both "universal" and "basic income", it's a variation of the usually proposed UBI, that in my opinion tries to deal with one of the main risks of such a system. A percentage of people (even if perhaps small) just doing nothing and the moral effect of that on the rest of the population.

I once made a comment in HN along the same lines, local community work..., it made me smile to see in your comment something so similar to what I tried to express, however I'm not a native speaker and the comment was probably not clear enough.

Yeah, that's true, it is a mechanism for ensuring universal basic income. It tries to avoid the societal hazards of people not working as well as costly subsidies to people that don't need them.

And yours made me smile. I hope some of these ideas catch on.

> So how about this: universal basic employment.

The problem this has is that it fails to provide the one thing that UBI (and, a number of the existing benefit programs that UBI is proposed to replace) provide or would provide, support for effective retraining and other support for labor market mobility.

Well, I mean, the problem beside the fact that it's top-down central planning in the Leninist mold.

I find this post extremely one sided. It seems that the author is too busy to praise the machines without trying to determine the bad side of using all automatic systems in such a sensitive place like an airport.

Imagine a hack attack or bug (it happens) that causes all terminals to stop working for 12 hours on the airport like JFK. What would be the losses? Would it be still worth it?

In my opinion AI should complement humans, not substitute them. The word is not working in A and B scenarios only, there is always more options.

The problem with this argument is that there is already dependency on the systems. If a system is down, travel may be halted regardless of if there are human staff members present.

> Imagine a hack attack or bug (it happens) that causes all terminals to stop working for 12 hours on the airport like JFK. What would be the losses? Would it be still worth it?

We already get hours-long mainframe outages, not to mention weather shutting down airports entirely. The system survives just fine.

> We already get hours-long mainframe outages

I often wonder if those outages aren't sometimes caused by hackers to begin with. Publicizing them as hacks would only serve to cause panic, and let other hackers know how vulnerable their systems are.

We also already get strikes.

Maybe it's a European thing, but I've spent more time delayed in airports due to humans striking than I have due to computer systems being down.

Most POS systems aren't very good from a UI perspective. Notice the display on a checkout stand device that takes credit cards. With "cloud-based POS", you may see messages at the wrong time - "insert card" comes up before the system is actually ready for the card, and may flash again during the transaction at moments when card insertion is inappropriate. On some systems, you get past some prompts with the touchscreen, and others with the keypad. False "chip read error" messages are common if the card is inserted at the wrong point in the cycle, and "wrong point" may be out of sync with the displayed messages.

Somebody probably tried to solve a real-time problem with a web client, and the result sucks. There's certainly enough CPU power locally to do this locally, but no, it has to be "in the cloud" now.

What was the business using? Clover, Rebel, NCR? On the pinpad itself, every vendor has limited control, and as you get into the smaller vendors, it comes down to where they just have to take what Verifone & Ingenico provide.

> According to the International Air Transport Association, moving to a fully automated check-in process could save the airline industry $1.6 billion a year. Which raises the question: Why haven’t we?

Because you need the people with the domain knowledge to handle the exceptional cases anyway.

Many years ago I was flying Continental from MSP to CLE. Cleveland proceeded to get 2 feet of snow over about 5 hours. The humans at Continental rebooked me onto the next 3 Continental flights, and then the following 2 Northwest flights before CLE threw in the towel for the day[0].

You don't want to expose that level of functionality to the lay person, and if you automate it, you run the risk of creating a system so hide-bound by its rules, it can't respond appropriately either. Just like the one that wouldn't let the author check in 57 minutes before his flight[1].

At this point in time, humans still have the edge on having local knowledge and the good judgment to apply it to the exceptional situations.

The real purpose of the flight attendants is to get your ass of the plane in an emergency; serving drinks and snacks just justifies their presence the rest of the time. Similarly check-in and gate agents are there to fix problems when they arise, in the last 10 years or so, the ability to check you in is pretty much secondary (modulo oversize bags, etc, which are just another example of exceptions).

[0] We got home the next day on the first flight. I'm sure the folks in Cleveland had a much rougher time getting to their destinations.

[1] Been there done that on an LGW to CLE flight back in the day. Also watched my original flight board from across the concourse where I was booked on a flight through EWR that departed a mere 10 minutes after the CLE flight.

One time I accidentally went to the wrong airport (Newark instead of JFK). A stern but kind lady at the checkin counter gave me a disapproving look, and proceeded to put me on the airline's next flight (leaving about the same time as from the other airport) without batting an eyelash. Being so used to battling the rigid bureaucracy of international flight and US immigration, I was floored by how flexible at least this one piece of the system turned out.

I get they are there to handle exceptional cases. But how about we only staff positions where they actually can handle exceptional cases? So many times I'm interacting with someone who is interacting with a machine and are actually incapable of making the human and compassionate choices that exceptional cases require.

If there isn't a system for overrides available then no person is required.

The last few times I've had a cancellation on United, the app automatically rebooked me onto the next available flight. No human required. Your example has actually already been automated out of existence.

I've had the opposite experience at an airport - I arrived 20 minutes before the flight departed, so a member of staff at the desk had to call ahead and ask if it was still physically possible for me to board. They then accompanied me through the airport to fast-track me through security so that I could catch my flight. They even kept the gate open for me so that I could do the awful walk of shame into the plane, but at least I was able to take my flight and get home for Christmas.

I believe the official policy is that you must "check in" more than an hour before your flight. So if you check in online before going to the airport, you're fine to arrive 20 min early.

Since the author uses the example of airline travel: I almost always use the automated kiosks -- but I want to have human agents available... because every few years I'll do something crazy like wanting to check in for a four-segment itinerary combining parts of three separate tickets on two different airlines. The automated kiosks don't know how to do this, nor should they -- it's not worth the cost of writing code (and UI) for such weird situations -- but the humans simply ask "who the hell booked this trip" and go ahead and handle it.

The trick is to have humans available for the weird situations while not wasting their time on routine matters; the approach some banks use, of "ATMs are free, but there's a fee for any teller-assisted transactions which you could have done at an ATM" seems to work pretty well.

Generating code + UI for some calculated topology of possible object/ticket interactions?

My job will be automated as soon as managers will be able to express themselves precisely enough and will know exactly what they want.

The moral of the author's "arbitrary check-in deadline" problem is the opposite of what he thinks it is. That's a problem caused by poorly-designed and non-overridable automation that could have been solved by giving more power to the human attendant.

Yep, has nothing to do with machines. Delegations of differing levels of authority applies to human only dynamics too; I once worked in a store were regular shop floor staff had to repeat company policy, but management only could override. I'd often tell customers they could not return an item due to company policy, only for them to go to a manager and get a refund anyway - I wasn't allowed to suggest myself they ask a manager, or allowed to suggest a manager might override policy.

Agreed. I have more commonly experienced problems in recent years caused by excessive automation (whether actual automated systems or rules that effectively automate humans) than the reverse.

> I then had the pleasure of experiencing some of these much-lauded “human touch points” first-hand on my travels back to NYC from abroad

> the clerk, who had apparently been craving some human touch points herself, had left her post to converse with the other clerk on the far side of the store. I ventured across the store to give them my money and was made to feel like I had barged into their private living room, demanding to be served

Was "abroad" France? Sounds like it (I'm French).

Anyway, the problem with AI isn't that we'll miss the so-called, and probably non-existent, "human touch". It's that it doesn't work.

(Examples given in the article, of vending machines and automated checkout, don't have much to do with AI).

See translation for example. It's not that automatic translations are bad, it's that they're incomprehensible: you couldn't guess what the text means if your life depended on it.

In Europe, Amazon offers to translate your listing in all marketplaces' languages (English, German, French, Spanish, Italian) when you create it for just one; that produces gibberish and results in zero sales, until you use an professional (human) translator.

Google translate, despite their recent bragging about a revolutionary update, is no better. Small example: in French, "coup de coeur" means something you love at first sight ("j'ai eu un coup de coeur pour cet appartement" => I fell in love with this place at first sight). But Google translate translates "coup de coeur" with "heart stroke"

https://translate.google.fr/#fr/en/coup%20de%20coeur

which is beyond wrong (in no context ever can this meaning ever occur). I noticed this because on some forum a user was describing his blogging activity as writing about his heart strokes and I wondered how many he had had and how it was possible to still blog about them.

Back to human interactions: machines are great when the outcome is discrete and completely predictable; using a vending machine, buying things from Amazon, checking in at the airport, etc.

But when some creativity is needed, machines can become super annoying; calling customer support and finding oneself interacting with a robot makes you want to murder someone (but there isn't anyone left to murder).

>> I made my appeals up the chain of command, but it appeared the real boss — their system — was digging its heels in on this one.

Eh, seems like the problem here was not humans, but a machine that had been programmed in a terminally inflexible manner.

Which is what most people are worried about, when handing over control to machines, isn't it? With a human, there is always the chance that you might be able to negotiate some exemption from the rules, to make your life easier. With a machine you're stuck with whatever rules are programmed to it (or it has learned etc). Even the best adaptive systems we have available are not very flexible, unfortunately.

This is really a huge problem that's not going to be solved by sneering at the older generation who "need to ask their grandchildren what a tweet is" 'cause they just don't get it etc. I mean, maybe let's pause here and remember we'll all grow old too, one day, right? If we keep making machines that need a young person's familiarity with technology to operate correctly, we'll never get systems that make everyone's life easier.

As a counterpoint, I rather dislike self-checkout at grocery stores. It invariably ends up being a fight between "place item in bagging area" and "remove unknown item from bagging area".

No, you damn robot, I just want to use my goddamn cart. Please shut the everloving-hell up about it instead of making me fumble for some "skip bagging" button on a shitty resistive touchscreen that doesn't register half the time.

The only reason I bother with them is because the lines are shorter. Otherwise, I'd rather just go to a regular checkout line, with a checker who's way better at scanning items than I am and (depending on the store) a bagger who's way better at bagging than I am.

The human in the story couldn't fix the authors boarding pass issue because management limited the discretion available.

Customer service is all about flexibility and resolving problems. That means giving humans discretion to act.

That manifests in different ways. One night I was stuck working until the wee hours and walked into my hotel (a Hampton Inn, not a posh place), checked in and headed up to my room. I hadn't eaten, was soaked by the driving rain and was generally miserable.

The night auditor sent a maintenance guy up with extra towels, a frozen TV dinner and some sleepy time tea. That isn't a full service hotel, and she didn't need to do anything. But she did.

People hate arbritary rules. When Target treated people returning things like criminals, they angered lots of people and lost customers. Meanwhile Walmart is pretty hassle free, employees can only say "no" in specific circumstances.

Self checkouts for grocery have been around for a decade. But they rarely take over -- there's a reason.

> An experiment at McDonald’s found that customers using self-service kiosks were more likely to supersize their orders, resulting in an average increase of $0.30 per order. It appears customers ordering from a human were more self-conscious of the employee’s judgment, which I found shocking.

I like and use the self-service points at McDonalds mostly to avoid:

1) groups of school kids who can't decide who pays for what/who wants what/must order as a group and dealing with the money distribution... when there's a 12-people-line behind them

2) Language barriers: Let's just say that (in Germany, at least) fluent German is not always a job requirement. Or fluent English - especially bad on fastfood shops in train stations.

3) when ordering for groups, it's easier to be able to directly type in what you want instead of hoping that the cashier correctly gets what you want (see #2)

While those refusing to order on a self-service station still are in the line and wait, I'm happily munching burgers.

There's one situation where I want a human checkout experience rather than a machine: when I have a stroller with a sleeping infant and bags strewn about its basket and the shopping precariously balanced on the hood because I didn't see where the shopping baskets were when we entered and a pre-schooler whining that we didn't pick up the magazine they really wanted.

In that situation, I really want someone else's help with scanning, paying for, and bagging everything. I imagine people with permanent impediments would also prefer human help.

Doesn't the authors argument go the other way: make labour cheap enough (with a guaranteed minimum income) that every patron can have a dedicated shopping assistant & the business can save money on expensive software and inflexible machines?

I love the way you write, very entertaining.

As a consumer, as long as there is one good self-service option I'm happy. When I'm in the car with two kids and need to get petrol, I'll drive the extra distance to pay at the pump rather than having to go into the store.

I don't think all companies need to go down this route. There's value in segmentation for those customers who can't or won't embrace technology. Take the recent OECD survey on computer skills [1]: only a third of people can complete medium-complexity tasks.

[1] https://www.nngroup.com/articles/computer-skill-levels/

> I'll drive the extra distance to pay at the pump

Another data point is New Jersey, where gas attendants are required: http://mentalfloss.com/article/74549/why-cant-you-pump-your-...

New Jersey is the worst. I always have to remember to fill up before I cross state lines into that place, so I don't have to deal with people yelling at me if I absent-mindedly jump out and start gassing up.

Also, one very minor gas station-related automation point: if you run a gas station, and take the clip out of the pump nozzle, so that your customers are forced to stand there and squeeze the trigger the whole time they are fueling, rather than setting it and letting it click off when the tank is full, you should burn in hell.

I wonder if you couldn't make your own 3D printed replacement for that. Just a little piece of ABS that would get dislodged when your car is full.

There's already a shutoff in the nozzle that prevents overfilling when the level in the tank reaches the end of the nozzle, at least on any pump I've used in the past decade.

A Leatherman is just about the right size to jam in there and hold the trigger open if the metal clip has been removed.

Use your gas cap.

This article is mostly from a service job viewpoint, but I regularly have the same thought about factory type jobs. More and more I see things being sold as "hand-made" and I nearly always think to myself that that's a negative attribute. Why would I want something more expensive and less consitently made?

I recently watched a TV show in which someone toured a factory that made crisps (potato chips). At one point they were frying in a fryer roughly the size of a childs paddling pool. Someone with a rake (I believe it was a standard garden rake) stood beside it and occasionally stirred. They stated this was what let them put "hand made" on their packaging.

To add a data point regarding self checkouts (I'm wondering if anyone else experiences this). Often when I'm waiting for a human checkout, I'll see the snacks and drinks, and throw one on the belt. But when I go to the self checkout I'm so involved with "a job to do" (that is, doing the job of a checkout clerk), that I've never considered adding a candy bar to my purchase. It is only after I've completed and paid, that I think "Gee, I really could have used a Snickers". But then it is too late, I've already paid and there is someone behind me waiting to use that checkout machine.

That's just bad design. The self-checkout could easily prompt a few questions before you check out, based on what you scanned.

"Hey, you got milk. Why not add a 4 pack of Oreos for 5% off? They are right behind you!"

That would be incredibly irritating. Just like when a fleshy cashier does it...

Computers should be seen as turbocharged bureaucracies. They're great for simple, repeatable, and well-defined tasks. Once you venture past that, however, everything breaks down and refuses to work.

It's sad that the author resorts to blaming old people for the lack of adoption. As if the only possible explanation could be that there is a defect in the people making the decisions.

Left out of these discussions of the joys of automated checkouts is that the illiterate and/or obstinate population makes up a large segment of consumers in certain areas. McDonalds has enough difficulty training their staff to use the checkout machines in particular areas that you do not want to see what happens if you throw the lunch rush to their own devices. Interfacing with such machines is a skill like any other and people like the author dramatically overestimate how common it is.

McDonalds have actually started rolling this out in some locations. A giant wall mounted iPad type deal to order meals through.

Here's a burger blogger with a video review of him trying one out: http://burgerlad.com/2015/02/mcdonalds-touch-screen-ordering...

This is the primary reason I want places like McDonalds and Dunkin Donuts to turn their POS systems around and let me key in my own order. My success rate on the staff behind the counter putting an order in correctly has been plummeting for years, and is hovering around 50% these days.

I like the solution employed by many Grocery stores and some big box stores. lines for both solutions, whether at check out or elsewhere within the store.

there is room for both because the capabilities of people varies and some do need assistance. if the unmanned position is not being used frequently or has complaints then business needs to jump on it quickly to make people understand it will work and work well but at teh same time, keep a few people to humanize the area

Bill Burr sums up perfectly why some people hate self checkout. Additionally when companies implement these systems customers tend to get less service and with none of the savings passed on to them.

Check the video, it's very hilarious.

https://www.youtube.com/watch?v=FxINJzqzn4w

It's interesting that so many people are so opposed to automating jobs. When I checked out my groceries myself, a neighbour also did the same. I said: «Hello are you also trying out new things?» and she seemed embarrassed. A week later she returned to the traditional way of buying groceries.

I always wonder how come crab fishing cant be mostly automated whenever I watch one of those Discovery shows, surely the most dangerous bits can be done by machines by now, or is it a job protection thing?

Never heard of this deadline before. I've never arrived an hour early to a flight in my life. Is this a real thing?

systems are not perfect and prepared for all scenarios, when i tried to choose row for infants in airplane their system didn't let me to choose it, only human at check in could do it for me

truth they could fix their system, but there are other rare scenarios not worth implementing

Interesting idea. But how does that one gets implemented among all the noise ?

After reading this article, because of the tone it was written with I think it would be easy to dismiss the authors entire premis of replacing customer service agents with AI to improve efficiency and cut costs as a cold, heartless rant from an elitist that doesnt understand or worse even care about how important those jobs are to maintaining our social fabric... but I won't, the fact is those jobs are being handled by machines right now and the author is somewhat correct, the bad employees need to be replaced but there are valid reasons why people are used instead of machines and I will try my best to give a few of them to address the authors points. Here goes.. Airports are very stressful for a certain amount of people, these people tend to get upset when say, they arrive late for a flight check in, it is much better for that person to be able to vent their frustrations towards a trained rep than another passenger without the social skills and training to de escalate a situation, who most likely is also upset as this would most likely result in further problems requiring police, courts, prisons etc. Airline reps also however little it may seem are an athority figure and as such require passengers to treat them with certain level of decorum either by law or just social convention which tends to make othe rise unruly travelers toe the line. AIrvine reps are much more effective at moving people along than you may realize. There is a reason why during big events the police are called out to direct traffic and pedestrians instead of relying on the light controllers with all of their algorithms and that is because things run smoother, granted the example of an empty airport may not support that argument but all airports get busy and you cannot expect people to be available only when things get busy. (well you can but then you get bad workers) When it comes to customs agents, I agree that being questioned is annoying and can make you feel like they are searching for a reason to arrest you, but that is their purpose and the fact that you felt that way is actually a good thing, yes scanning a passport is much better for the travelers but not necessarily better for everyone, just as swiping your drivers license into a kiosk might be better for the majority of people going through a drunk driver checkstop it would defeat the whole purpose. The cashier that was too busy talking to their co worker to serve you should be fired. But there are vending machines in airports for everything now, they do tend to breakdown, get vandalized, run out of product. Yes you can solve these problems and as I said there are machines already doing this. Finaly the McDonald's example, yes people do tend to up size their orders at kiosks just like how slot machines tend to encourage people to keep gambling, as a matter of fact though there was a certain movie called Super Size Me that attacked McDonald's for exactly what you see as a virtue but that is another argument all together. I live in an area that has McDonald's with these kiosks and they provide a value added service called create your own burger which allows you to mix and match McDonald's products together to create hybrid burgers like a quaternary pounder fillet o mac!! but again for some reason they do not accept cash and considering many of mcdonald's customers either prefer or are forced to use cash due to their financial situations there are still problems that need to be solved. In conclusion I tend to agree that there are many benefits to automation and computer services, I myself use the self checkouts at my local grocery store but that does not mean that we can discount the benifits of real people, both systems have value and smart companies realize this and as such use both, choice is always better.

Is there a machine that can help the author get to the point?

Yes, there are some good reddit bots that can automatically summarize an article quite well.

These machines will only help readers after the article is finished. chrismealy however was asking for a machine helping the writer while the article is in writing. But I also think the question was meant a little sarcastic.

No, the author could run the summarise-bot, then use the results to focus the writing towards the salient points.

Sarcasm was certainly intended.

Reply on news.ycombinator.com