Slashdot Mirror
Workers: Fear Not the Robot Apocalypse (wsj.com)
An anonymous reader shares a report: For retailers, the robot apocalypse isn't a science-fiction movie. As digital giants swallow a growing share of shoppers' spending, thousands of stores have closed and tens of thousands of workers have lost their jobs. The brick-and-mortar retail swoon has been accompanied by a less headline-grabbing e-commerce boom that has created more jobs in the U.S. than traditional stores have cut. Those jobs, in turn, pay better, because its workers are so much more productive. This demonstrates something routinely overlooked in the anxiety about the job-destroying potential of robots, artificial intelligence and other forms of automation. Throughout history, automation commonly creates more, and better-paying, jobs than it destroys. The reason: Companies don't use automation simply to produce the same thing more cheaply. Instead, they find ways to offer entirely new, improved products. As customers flock to these new offerings, companies have to hire more people.
productivity gains result in job losses. They have in every industrial revolution. Then in 50-80 years tech caught up elsewhere and there were new jobs. In the meantime there were two or three lost generations living in abject poverty because in America if you don't work you don't eat. And it wasn't the New Deal that fixed that (it helped, but wasn't nearly enough) it was a global war, 80 million dying and basically the whole world getting blowed up and needed to be rebuilt.
History is basically the working class trying and failing to pry money out of the hands of the ruling class. Why the hell people don't see this is beyond me.
Hi! I make Firefox Plug-ins. Check 'em out @ https://addons.mozilla.org/en-US/firefox/addon/youtube-mp3-podcaster/
Here's an unencumbered link to the article:
http://www.cetusnews.com/busin...
When you hear hoofbeats, think horses, not zebras
>> Throughout history, automation commonly creates more, and better-paying, jobs than it destroys
The aggregate job count is rarely the complaint of existing workers and their families. It's that the new jobs get created somewhere else and often require skills that the original workers don't have, and that the workers don't feel like moving, don't want to retrain and/or are considered too old to retrain or hire. See "West Virginia" or most of America's near-inner cities for examples...
When we have burger flipping and order taking, shelf stocking, part picking robots, where do the people who do not, can not, or will not get an education work?
The fact that it all seems to be going in the right way - for now - does not mean it will continue. Many e-commerce jobs for humans will be destroyed in the next few years as e-commerce gets more and more automated. Yes there will be jobs, but for far fewer and better qualified/skilled people. If you are a relatively unskilled worker - in my view - your prospects are not going to be good. And, what is worse, it will be people like us that are facilitating this.
Don't even start me on what robotics are going to do to the trucking industry...
Has led to less overall jobs in the auto industry. I'm pretty sure that the jobs created with supplying and maintaining these automation tools did not completely offset the jobs lost...
Jobs don't matter. Jobs have never mattered. There are, and will always be, a means to an end. That end being not starving to death. In the past, jobs were a means of divvying up the limited resources for that goal, but as the resources become less limited, something like a UBI will become necessary.
This should be a good thing, but we've got such pig-headed ideas about economics that we're taking the blessing of not needing labor and turning it into a curse.
This is my signature. There are many like it, but this one is mine.
Only if the new offerings are not produced by robots. That's where the breakdown happens, why the advent and adoption of general-use robotics and algorithms isn't another example of historical automation.
Buggy whips went out of demand, so the people went to build cars. Cars started to get automated, so people went to build the increasingly-intricate car parts. But now car parts can crafted wholly by robots (or, for a continuously-expanding class of parts in general, "printed"). Automation in the past was about very specific processes for very specific outputs; you couldn't take a line used to make cars and easily change it to one that makes bicycles (or soup.) But soon we'll have a robot chef that works mostly by mimicking human actions, so if it can cook it can assemble.
The "creative" jobs will hold out longer, but algorithms will replace many of these, too: IBM's Watson has made a movie trailer. A lot of marketing these days are applying set rules to things (certain colors evoke certain responses in certain demographics, etc.) A lot of music is based around similar setups. Hell, Japan has a popular singer who's not even a real person.
The only question I see is: how fast will this happen? If it's extremely slow then make-busy work might fill in the gap as robots and "AI" take over most regular production. If it's very fast then we'll have a lot of robots producing things that most people can't afford to purchase, and "things" will eventually include food.
Let's just take a look at Amazon for a moment. I watched some news articles about them. Their warehouses already feature a lot of automation.
The only thing they need humans for is to take stuff off shelves (that robots bring to them) and put in boxes to fulfill orders. And you can bet your wallet as soon as Amazon figures out how to automate that, those jobs are gone. Poof. And yes, they definitely do intend to automate that, they're working on prototypes and ideas as I write this.
Automation is going to be a very huge disruptive force, and as it starts to happen, it will accelerate ever faster, just like computers did from 1980's to now.
Right now, we haven't reached that critical point where automation is going to displace workers. There's enough 'other' jobs to offset what automation replaces. But you're kidding yourself if you think that's going to hold. It's going to flip to the other side very soon.
Jobs and technology are mostly independent issues in the big picture. An economy is basically systematic bartering: you want stuff I have/produce, and I want stuff you have/produce. We trade to get what we want. Money, banks, etc. are simply tools to make such trades easier and scale-able.
The problem is that politics, geopolitical complexities, and herd mentality create boom and bust cycles, and inequality where the winner takes all and the losers get tossed. These boom and bust cycles are probably as natural as the spiral arms in the Milky Way galaxy. Capitalism bubbles have been occurring for at least 400 years, long before the USA and the Federal Reserve existed (since many blame the FR).
And capitalism does NOT guarantee reasonable equality. For the past the 40 years or so, the rich have got much richer while the rest mostly stagnated. E-commerce hasn't reversed this trend. The economy produces much more, but it's not trickling down. There seems to be a feedback cycle where the owner of machines and real-estate get yet more machines and real-estate, creating a winner-take-all economy. (It's similar to Marx's prediction.) Automation may be part of that, but it's also because the rich can buy up land and companies during slumps. The middle class is usually trying to make ends meet during slumps and so don't have enough spare cash to play that game nearly as much.
The article is paywalled so I cannot see it, but I am skeptical of claims made by the WSJ. They are often biased.
Table-ized A.I.
Not exactly surprising that the WSJ is not thinking terribly far ahead on this.
The issue is not what has happened up until now. It is what happens in the decades to come. Because we do not have general-purpose AI yet, nor do we have sufficiently advanced robotics yet.
The agricultural revolution displaced massive numbers of workers, but they could be employed by the industrial revolution.
Now, let's try to apply that to a post-general purpose AI and advanced robotics world. You're a farmworker, and your family has been farmworkers for generations. Someone builds a new device, and you are now redundant. So you move to the cities....and there's no industrial revolution work to be done because its being done by general purpose AI and robotics.
"You could get a job building the robots!!" No, why would you use a human to build the robots? You'd use another robot.
Getting through this is going to require a lot of rethinking how society works. Since the dawn of civilization, we have defined and supported ourselves via work. Work will no longer be possible for a very, very, very large portion of the population. We need to start talking about this massive transition now if we want to make that transition without massive bloodshed.
People to not peacefully starve to death.
I think we hit the peak usefulness of the "retrain for a better job" advice back in the early 90s when large corporations got around to destroying their "legacy" white-collar workforces. This was driven by computerization of clerical office work finally reaching a point where permanently fewer humans were needed. Back when I first graduated from college (around 1998), one of my first IT jobs was with a huge life insurance company. According to some of the old-timers I was working under, the then-sparsely populated headquarters was jammed wall-to-wall with various clerical workers up until the 80s or so. The HQ took up two Manhattan city blocks, plus a huge tower uptown, plus they had tons of large regional offices around the country. It was apparently so full that the company staggered start times so that crowds didn't overwhelm the elevators and escalators in the building. Maybe some of those clerical workers got better jobs, but I don't think that's going to happen this time around.
The fundamental problem that needs to be solved is this: If you want to continue with a consumer-based society, you must find some way to allow everyone to sell their labor for a price that allows them to continue consuming and keeping businesses alive. This includes everyone -- not just STEM graduates, CS people, programmers, data scientists, etc. The economy only works when the majority of people can afford to participate at a level appropriate to their skills. Large fully automated cloud data centers employ 20 security guards, HVAC techs, disk-pullers and rack-and-stackers...not 5000 system administrators. Automated warehouses employ a couple of robot-minders. Automated trucks employ zero truck drivers.
If we still have to cling to the idea that everyone has to have a job and earn money to be worth anything in society, how do we avoid nasty problems that crop up when the majority of people are unemployed and desperate? Technology people tend not to understand this, but look outside the tech bubble and see what kind of work the vast majority of people do all day. It's repetitive, automatable and may go away very soon.
Workers are hired until the marginal value of their labor equals the marginal cost of that labor. You hire Alice at $50,000 salary because she'll bring in $100,000. Bob, because of diminishing returns, will only bring in $75,000, but at $50K hiring him is still a no-brainer. But if Carol will only bring in $50,000, you won't hire her unless you can get her for less than that.
What this means is there is no general economic law that connects changes in worker productivity to a particular kind of change in employment levels. It depends on what you do with that productivity.
Imagine a world in which computers were laboriously assembled by workers on breadboards using prototyping techniques. Let's say it costs you $10,000 to assemble a computer this way. In that case you'd only sell a small number of computers because they'd be highly specialized machines. Now suppose you introduce modern assembly techniques, with printed circuit boards and wave soldering. Now the computer which cost you ten thousand dollars to assemble can be made for well under $100.
If you continue to sell a very small number of computers at high prices, you'll lay off most of your workforce. On the other hand if you start selling your computers for $180, you'll end up adding to your workforce. Both scenarios turn increased worker productivity into increased profit, but in different ways.
Now let's imagine an entirely different scenario: a fast food restaurant. It's hard to imagine selling a lot more Big Macs because you drop the price. Nonetheless the same principle applies. If you can find a way to make money off the newly surplus labor, employment wont' go down. If you can't, you'll let people go.
Post may contain irony: discontinue use if experiencing mood swings, nausea or elevated blood pressure.
The ATM graph doesn't appear to show what the author claims it shows. It looks like ATM's greatly delayed the growth of bank teller employment. Without ATM's, it looks like the total tellers would be roughly more than double the current quantity (although the chart doesn't cover enough years to get a good feel for the pre-ATM rate). The rate of teller growth may be finally going up again, but that's probably despite ATM's if we look at the pre-ATM rate. Using that chart, it appears ATM's indeed did take a big bite out of overall teller jobs.
Table-ized A.I.
It's fair to say that efficiency CAN make the economy better, but the benefits of that betterment often don't get distributed well in practice. That uneven distribution is largely why Mr. Trump won. The slightly-Democrat-leaning rust-belt leaned toward Trump this time, bucking the trend, because they've been hit hardest by automation and outsourcing. The distribution problems can be both geographical and by class (since the rich are still getting richer).
This map shows the delta of the voting pattern per last election. The red-shifted areas fall predominantly in the rust-belt. The distribution problem has been an economic stumper of late.
I don't fully believe in Trump's claimed solutions to the rust belt, but he did focus on and popularized the issue better. If you have a back-ache, you'll probably pick the doctor who talks more about backs, even if their solutions seem nebulous.
Table-ized A.I.
Throughout history technology and innovations have solved specific labor problems and opened the way for more relatively unskilled labor to be done. The wheel meant it was possible to carry heavier things and people were free to work on other things. The cotton gin required less work to process cotton, opening more jobs for relatively unskilled work. Same with the printing press, etc.
What we have here is different. Cheap automation, over the next 100 years, will be able to do any work a human can do, but far more cheaply and 24-7-365 (366 on leap years). AI will be able to replace every white collar job, from help desks to engineers to lawyers - and do it far cheaper and 24-7-365.25. They will be able to be trained quickly and efficiently should a new task be desired and the way it's going only a few people will reap thier rewards. Please explain to me how even skilled workers are going to be able to compete with this because soon we will have AI and generic robotic automation, deployed rapidly to novel situations and its like nothing this world has seen before.
Forget everything we taught you in 101. In the real world, business owners barely know what they are doing and will instead do whatever they think will meet their personal goals. There is no economic law that connects business behavior and rationality or forethought. You can generally bet though that productivity gains will be more likely funneled into owner profit while employment grows as slowly as possible.
It is unwise to apply historical precedent when the difference may be qualitatively profoundly different.
Robots with AI are not mere machinery, they are much more like slaves. And there were serious problems economically for non-slaves because of competition from slave labor. This was a big deal in the Roman empire.
As usual, the elites owned almost all the slaves---and one reason Gaius Julius Caesar was assassinated by the oligarchy was because he favored restrictions on slavery in order to benefit the wages of free Romans.
Now robot AI slaves are unlikely to spontaneously revolt, and a major category of these will of course be armed guards, Unsullied Machines, who will prevent democracy from imposing restrictions on the elite's slavery.
No ! Companies do not have to hire more employees due to more products being sold. They simply need more automation and also need to get rid of employees as human workers are cost negative. Further, when the company makes more money that have no reason at all to spend or invest it in the US.. In other words the notion of trickle down economics is a sick joke.