I forgot to mention the possibility of work you never have to do because you are just rich and other people will do the work for you. But there still has to be an expectation of someone working in the future on your behalf, or it all falls apart
We haven’t reached peak *unemployment, so no replacements are happening?
I would argue that the number of people leaving the workforce longer than the unemployment statistics are tracked is far less interesting than the workforce participation rate. It fluctuated steadily at 66-67% from 1989 - 2003, then flatlined from 2002-2009, when it began to gradually decrease down to 63%.
There are 205 million working-aged adults in the U.S. - this 3% drop since 2009 constitutes 6.15 million fewer workers in the U.S. Over an 8 year period. Will this trend continue? It’s impossible to say. But it’s clearly an abnormality - you’ll note that since 1950 there has never been a dip more than 1% in the workforce participation rate and the overall trend before the 2000’s has been a gradually increasing workforce participation rate. The only other time it has flatlined was in the late 70’s/early 80’s.
It’s worse when you consider that when the downward trend began there were only 195 million working-aged adults. Percentage of workforce participation has continued to decline even though there are now 10 million additional people who are working-aged. Looking at this chart below at maximum is fairly revealing:
You could argue that all of the people who’ve left the workforce and haven’t returned are pursuing school; deciding to stay home with their families; and have already earned enough money and are retiring early. But this seems overly optimistic to me, especially considering it lacks precedent.
If your argument hinges around red and green IQ lines made up to represent the areas automation will probably be capable of replacing… I’m sorry, but it’s a weak argument.
Sure, there are LOTS and LOTS of skills that humans have that automation won’t, for a very long time. But there’s a very limited market for flute-players, early-universe cosmologists, ballet dancers, etc. Automation doesn’t need to do a lot of diverse things in order to replace most human work. All it has to do is common tasks, well enough that is cheaper to use a machine to do the job than it is to use a human to do the same job.
Anything not covered by the automated process becomes a task on a pile for a human, whose job increasingly becomes just filling in the gaps for the hoardes of robots who aren’t yet able to handle that particular task yet. I’ve used my own job as an example; I am quite literally proofreading the work of an algorithm, performing the tasks it doesn’t yet know how to do. The fact that I still have a job, doesn’t change the fact that the same workload 5 years ago would have required thousands of workers, and now is handled by me and a few dozen other people who handle the problem cases by the hundreds, while the algorithm automatically resolves tens of thousands more and only routes it to a human when it faces a situation it hasn’t seen before. As we work on the errors it presents, we are training it on the correct method for addressing that error, so that next year even fewer people will be needed.
Your assertion that unemployment won’t increase before General A.I. because automation can’t do *this and humans can, and it can’t do *that but humans can, strikes me as a technological “God-of-the-gaps” argument.
Are there jobs that require a breadth of skill, experience, and creativity? Of course. Is it typical? No. And we don’t require an A.I. that can handle those jobs before we run into serious problems. We’re already seeing the beginning of a major change in society, now. The more advanced and cheap our automation becomes (the latter being something you don’t seem to be getting - technology becomes cheaper over time but you keep calling automation *expensive), the more pronounced the employment crisis is going to become.
I’m not sure who you are ascribing this particular strawman to since you don’t specify, but since you are responding to me with the rest of your post, I have to say I am exceptionally offended by this characterization.
I am not describing a problem with human beings. I am describing a problem with our system, and I am certainly not describing people as worthless fuck-ups. Also note that from the moment I entered this thread I have made it clear I AM one of the people who will be displaced. I am backed into a corner where there is nowhere left for me to go within our current system once my current job is gone. If I’m a misanthrope, I’m apparently a self-hating one as well:
http://boards.straightdope.com/sdmb/showpost.php?p=20526445&postcount=680
Why does one group of humans have a buffer protecting them that others don’t have? I could write a book, but if you actually give a shit about this question, for a start Google generational poverty, student debt, rising college costs, the explosion of the prison population, and recidivism rates which are largely due to inability to find work due to requirements to disclose criminal backgrounds, IMO.
From your mouth to God’s ears. The only question is how long are we going to ignore the necessity of such before we finally do it. A fundamental culture change is needed; the debate at this point is in how much we should be *cutting social programs, not how much we should be *increasing them.
But if there are 4 burger flippers each of whom do 3/4 flipping and 1/4 miscellaneous stuff, a machine can get these done with 1 person.
I got my CS degree in 1973. Since that time massive amounts of work we had to do has gotten automated in a sense. Lint for one. Libraries and open source modules which mean you don’t have to write intricate code.
Now, if you are one of those who defines the work, you are a lot safer than someone who is a coder. But you are going to be hard pressed to find people who take reasonably high level specs and write assembler these days.
While I did almost everything on my last project before I retired, from research to specs to user interviews, there is no way I could have done it without ten people without this type of automation.
Indeed. But an article in the Times yesterday I think was about unions asking businesses who backed the proposed tax cuts - which are being sold as job creators - to promise they will use some of their windfall to hire. I don’t think they are expecting an answer. Money from productivity can go to jobs, or it can go to execs and stockholders.
The low unemployment rate shows that jobs are there. But the problem is the well paid factory worker now working in nursing homes as a caregiver, at much reduced wages. A lot of the high growth jobs which don’t require special skills and training are not very well paid.
Now when we get robots changing sheets and emptying bed pans, things will get even worse. Though the quality of care will go up - I hope - when the people have more time for the patients.
Well, manufacturing output is up, but manufacturing jobs are down. Cite. So these new jobs are not inside the factory.
And I hope you are not saying that VP of Social Media is a bullshit job. Besides it being a way of getting a company name and message out there, if the company starts getting dissed on social media it helps to have someone who knows what to do.
QFT, and I agree that it was a strawman to accuse you, or me, or anyone here that I see, of acting like the reduction in the need for human work is something for only “those people down there”. Maybe some people see it that way, but I haven’t seen it in this thread.
Your points about workforce participation are also spot on. When you look at just male participation (since feminism and the introduction of married women to the workforce obscured the trend to a degree), it’s more starkly apparent. * The Atlantic* discussed this in 2015:
An important point to consider in this argument that I think you and I have tried to press while feeling like we are getting strawmanned:
**Hellestal **dismisses or ignores this possibility, insisting that the economy will continue to work as it always has, until the moment strong AI is cheap, at which point it will flip (and, as s/he says, we may have a lot more issues to deal with than jobs). I like your “god of the gaps” retort to this line of argument.
Continuing with the Atlantic article:
That “moat” sounds familiar. :dubious:
ETA: I’m past due to fulfill my weekly pledge to “unplug” from email, social media, and message boards over the weekend. So I’ll check back Monday afternoon or evening. Have a good weekend, everyone!
The God of the Gaps argument means holding to the same faith-based belief by putting off the lack of evidence for their position, pushing back the place where evidence might be found, in defense of the notion that they’re still right anyway despite their complete lack of evidence because evidence will potentially exist in the next available “gap” or location.
This might in fact be the single worst analogy you could have possibly chosen.
I’ve been citing facts here, which have largely gone ignored. I’ll point out again that job numbers are at their highest levels in history. That is the actual fact of the matter. What faith-based belief does is ignore current facts in favor of dogma. So here is the actual God of the Gaps argument with respect to history of automation. It really does exist. Here’s the first gap that I personally know of, Queen Elizabeth refusing a patent to inventor William Lee for his knitting frame.
But that quote is hard to confirm and source, so it’s easier to hit the modern era. Here’s a convenient modern list about the fervent belief in the God of Automated Job Loss, whose followers are always looking in the next available gap.
NY Times (1921): “Will Machines Devour Man?”
NY Times (1928): “World Asks Whether Man Will be Enslaved by the Instruments he Creates”
Turns out that was wrong. The Great Depression was a monetary (gold standard) problem, quite strikingly similar to the current eurozone monetary woes – a fixed exchange-rate regime. But people scapegoated the wrong issue, then as today. “Technological Unemployment” became a common phrase in the 1930s, even as job recovery begins, as people scapegoat the wrong issue.
NY Times (1940s): “Does Machine Replace Men in the Long Run” and “Is the Robot Beginning To Think?” with “Employment Drops”
The next gap where your predecessors in faith looked for their God of Automated Job Loss: the creation of the digital computer.
Hal Boyle (1949): “Machines are Laughing at Men”. “Machines have taken the second step to supersede their creators — mankind. They can think. They have mechanical brains that can solve some problems quicker than a thousand brilliant men.”
1950s: “Workers see ‘Robot Revolution’ Depriving Them of Jobs”.
A Congressional representative suggests government study to “avert mass unemployment”.
Still no sign of your God. Next gap.
1960s: “Automation Linked to Jobless Count” or “Automation Looming Large in Labor Picture”. And there are, of course, more calls for government inquiries into the tech fear.
But this is my favorite: “Automation Might End Most Unskilled Jobs in 10 Years”. Hey, that one sure looks familiar.
The Secretary of Labor under Kennedy voices concerns that labor might end up on the “slag heap” because of machines. Or hell, here is president Kennedy himself from a separate cite:
These fears come in waves. The late 60s and early 70s died things down a bit, but then the transistor starts going gangbusters and people start looking for your God again in the next available Gap: the dreaded microchip.
1978, “PM Acts to counter ‘chips’ that kill jobs.”
1980s: “A Robot is After Your Job”
Search for the next gap becomes hot and heavy for a while. More 80s stuff:
“Technology called threat to 5 million jobs”
“Imminent ‘Robot Age’ Awakens Concern”
“Unemployment will end Honeymoon with robots”
A couple centuries of history at this point is now a “honeymoon”. Fears cooled a bit in the 1990s, but they’re up and running again today. People still have perfect faith in their God of Automated Job Loss. He’ll show up. Just after this next gap, in just a few more years.
The “God of the Gaps” accusation shows breathtakingly little awareness of the history of this debate. This might be the most narrow, parochial thing I’ve seen in this entire thread. The entire absurdity of this whole conversation is that the people who are speculating about massive job loss are the very ones with all of history against them. The Chicken Littles have been Chicken Littleing for a century and more. People always project their own anxieties into the misty fog of the unknown. Fears about the present bleed into speculations about the future. That’s just human nature. That’s why we get the continual claims that “THIS TIME IS DIFFERENT!!” God is always around the next gap.
The facts, of course, have been otherwise up until now.
There are more jobs, at higher wages (globally), than there have ever been before. The net change has been positive, more machines and more human workers being more productive than ever before. Literally the entire historical record points in one direction. But people still look for their God past the next gap. Next year. Or the year after. Or the year after that. People point out their own particular industries, where they personally suffer massive anxiety from their (perhaps less than ideal) choice of career. And hey. That sucks. No getting around that. It’s a stressful life.
But logical problems begin when people try to extrapolate narrow experience onto broad societal trends. The result is massive selection bias. There’s not even the slightest attempt to get a representative sample, not just from today but from historical perspectives. People think that how the world seems to work in front of their faces must be how the rest of the world works, too. Well, automation has certainly destroyed a lot of jobs. Possibly even the majority of jobs that ever existed. But never on net. There have always been more jobs created, on net, than the jobs that were lost. It’s just that the jobs are different. In almost every single historical case, the new jobs that were created not only were not previously imagined, but could not have possibly been imagined by people at the time.
None of that matters to people of deep faith. Their God is always around the next gap.
This is what I meant by the “burden of proof”. There is a long, long history of people saying not just similar things to what people say today, but quite literally exactly the same things. But the God of Automated Job Loss keeps his parishioners. They keep looking for Him in the next gap, in the next few years. It’s the classic “Seen vs Unseen” problem as described by Bastiat. It’s very, very, very, very easy for a worker in a particularly sensitive (perhaps ill-chosen) industry to see first-hand how jobs are being replaced. But job creation? That’s invisible. Very few people have first-hand experience with the creation of a brand new type of job, or even a brand new industry. It’s unseen.
The human bias is to emphasize what is right in front of our faces, rather than what we can’t see. This kicks into high gear for anxious people working in (perhaps ill-chosen) industries. The “THIS TIME IS DIFFERENT!” argument is especially compelling in technology, too, because technologies really are “different” in the sense that the same thing can’t be invented twice. New inventions are genuinely not the same as previously invented tech. That’s why they’re called inventions. So in literally every single previous era, people felt justified in saying that the tech they were seeing – which was legitimately different from the tech that had appeared before – would result in an effect on the economy that was different from the tech that previously existed.
Except every previous time, they were wrong.
The fear was always, always, always the same. “Mass unemployment”, “millions of jobs lost”, etc. All wrong. Despite the jobs lost to automation, there was never net job loss to automation. The Job Creation Mechanism of the economy keeps working. This Job Creation Mechanism is simply the fact that the world is not perfect, and we try to remedy its imperfections with our available resources. That’s what jobs are. Human beings happen to still be a valuable resource.
And that’s the “argument” we face here.
“THIS TIME IS DIFFERENT!!”
Well, it could be. But is that really the reasonable way to bet at this point? Look at the burden of proof. Look. At. Every. Previous. Era. Literally the only change in the argument is that “machine learning” and “Moore’s Laws” are the current terms-du-jour being tossed around, rather than “chips” or “vacuum tubes”. People in sensitive industries will keep losing their jobs, just as they have for centuries. And the Job Creation Mechanism will also continue to work. The Zero-Marginal-Productivity (ZMP) worker is a unicorn that has never, ever showed up in the past. It’s just another God of the Gap, where the “gap” is the future.
Millions of jobs will be lost in only 10 years time… and that’s been true for at least a hundred years.
It is not literally, logically impossible that this time really is different. But if it were different, I’d like to see an argument from that from someone with more than a modicum of awareness of what the actual history here is. The issue for me personally is that I outlined, in some length in the posts above, the kinds of things I’d like to genuinely see in any serious argument for a large mass of “unemployable” ZMP workers.
No one has even begun to try to answer those questions. Ignored, all of them.
[spoiler]I’m trying to limit myself to one topic a day. Here are points I haven’t yet addressed.
[ul]
Job loss vs net job loss, in more detail.
[li]An extended list of job statistics.[/li]
[li]The idea that “technology” drops in price (conflating new inventions with previously invented doodads).[/li][/ul]
If there is something else (substantive) I’m missing, it can be pointed out and I can add it to the list.
[/spoiler]
Yes, sentient beings can easily switch tasks. I’m not disputing this. Robots can do tasks indefinitely and perfectly consistently without rest or wages. For the vast majority of jobs currently being done (by the percentage of the population they employ), the latter is far superior to the former.
I am familiar with basic economic principles, and I understand that a more productive robot will grow the economy. My point is that given the high capital investment required to automate, this economic growth will be concentrated among the wealthy, who can afford the large initial capital investment required to develop automation. After the initial investment, they can then take advantage of lower wages and thus lower operating costs to consolidate their position.
So what incentive does the robot-owning moneyed class have to pay for all these new jobs that the economy will create? This seems to me to be the trickle-down economics argument version 2.0, and trickle-down economics is bullshit.
When I feel like wearing my pretty tin-foil hat, I speculate that the next major violent uprising in western culture will result from this issue in the end. I don’t see any other way that this situation will resolve itself other than blood and fire in the end.
And I take exactly the opposite view.
I think in today’s modern, service-based economy, there are a minority of people whose job involves putting tab A into slot B as quickly as they can.
In my office of 120 people, I doubt there is anyone who could tell you exactly what they’ll be doing this time next week. It’s constantly shifting, and every day is different, based on the requirements of our customers and projects.
To make money.
Again, this is the strange concept of work that many have, particularly in the US, where the employer is doing the employee a favor.
But in fact a private-sector job only exists where a human can contribute more economic output to the company than their salary; and in fact it’s usually vastly more – several times their salary.
In this thread, some people have argued that, temporarily at least, there won’t be anything useful for humans to do, and that’s a valid position to take. That’s what we’ve been discussing.
But if you agree that the economy will create jobs for humans to do, it doesn’t make sense for companies not to provide those jobs. They’d be shooting themselves in the foot (and be at a competitive disadvantage).
We may disagree on the near or medium term prospects for automation, but we certainly agree on this. What you are referring to is what Marx called “surplus value”, which is one of the reasons I always chafe at rich business owners being called “job creators” (another being the fact that many new businesses destroy just as many jobs as they create, like a new restaurant opening in a town where that market is already saturated).
My first thought was to say that this will continue as long as human workers provide surplus value. But maybe if their surplus value is not as great as that of a robot rival, it’s not the case. I do think, again, that in many job fields there will be humans involved but increasingly leveraging their labor manyfold by the AIs and robots they use as their tools.
We should of course note that there are ways jobs may be preserved even if the cost of their wages/salary outstrip their value. One example was snarkily raised by a libertarian economist in a podcast I was listening to. He said that he expects when driverless trucks become safe, efficient, and widespread, Congress will pass a law requiring they carry drivers anyway, and we’ll see fleets of trucks on the highway with “sleeping Teamsters” aboard. (Less cynically, this could actually be necessary or at least useful in a transitional period when driving down the highway is easily automated, but the last couple miles to the supermarket or fast food joint, and the unloading of the trucks, are not.)
That may be true for your office. It is not true for the majority of the workers in any of the offices I have worked at over the last three decades. I’d guesstimate that between 30 and 60% of all office work is repetitive enough that they could be automated without requiring true AI, depending on the specifics of the industry. It’s not cost-effective to do yet yet, but it will be in my lifetime.
Also, I’m going to point out white collar bias, because this idea that humans are better than robots will absolutely not be true for any of the blue collar industries I have worked in. Every single one of them would be more profitable with increased automation. The easy example here is a self-checkout. That is an automated worker that takes jobs from human cashiers, and retail sales followed by cashier are the two most common jobs in America … for now.
I agree fully.
The bolding is mine, to highlight my issue. Why are you assuming that these new jobs are being created for humans to do? A minority of them might be, but I suspect the majority of those jobs will contribute more economic output to the company when done by robots instead. Isn’t the whole question on this thread what will happen when the average robot contributes more economic output than the average human?
There have got to be a lot of people employed as security guards. Maybe not for much longer:
Again, this falls into the category of the kind of thing where you’d still have a certain number of people involved: a few in control rooms, and some others who are waiting in patrol cars to go where they are needed. But with these robots, you could vastly reduce the numbers needed to actually walk or drive a beat where they spend most of their time in boredom, watching over an area or waiting for someone to do something. (The humans on call after these robots were fully implemented would presumably be kept quite busy, being sent from one trouble spot to another in fairly short order.)
Wow.
Someone tell me again how humans will win this labor battle against technology?
In this case, I’m not so impressed, actually.
The robot rolls around on the sidewalk, filming, and if too many homeless people appear in the camera, it calls a human guard.
How is this different than just mounting security cameras overlooking the sidewalk, with a human guard to montor a bunch of screens.?
And the robot costs an hourly fee…cameras are free after installation costs.
The article said that people have disabled the robot by throwing a blanket over it and spreading barbecue sauce on its lenses.
It’s a nice story, but not a game-changer.
At least not yet… 
Indeed, in my opinion it’s an over-engineered solution. I have the home version of this: https://www.alarm.com/ because my husband is a software architect for the company and he uses our home as a test bed. I couldn’t find the technical details outlined on the site, but the system can be set up to monitor a room or hallway. The user draws a line or a box somewhere in the room, like a tripwire, and if something crosses the line or enters the box, the video is saved and an alert notification goes out. The really fancy part they don’t talk about is that they are working on the video analytics so that the system can detect if the movement is a human, a dog, a horse, leaves blowing in the wind, etc. They can already identify humans vs animals. Random things like leaves is harder to tackle.
Which means you don’t need a team of guards staring at computer screens or walking around, and you don’t need a silly robot rolling around, either. (Unless you want the robot as a visual deterrent, of course.)
A big part of the story (that chappachula may have missed) was the whole “deterrence by it’s presence” factor.
And the systems described are still replacing human jobs even if they don’t look like robots.
Yeah. Not to mention video cameras don’t do so well if the perps wear masks and are in and out before the cops get there. Or don’t realize there is a camera. Lots of people in my neighborhood have cameras - stuff gets stolen off their porches anyway.
Put a towel over the robot and the guards will come in a second. Anyhow, your typical moron crook might thing the robots have cattle prods or something.
BTW, I was reading the piece in the current New Yorker about Estonia, and one of the people interviewed has a robot lawn mower. I bet that is on the way. Lawn services do other stuff, but they won’t have to come as frequently if your robot does it. I’d buy one in a second. Assuming they were not Tesla priced, that is.
The Federal Reserve “beige book” reports that a fast food chain has found that people order more food when they use a self-serve automated kiosk than when they order from a person. (Presumably because they feel less embarrassed?) So as if fast food chains didn’t already have an incentive to go this way (not having to pay wages), they can actually sell more? Expect these to catch on very quickly. (Funny that the automat really does turn out to have been the idea of the future.) And with them, perhaps a correlated rise in obesity, diabetes, heart disease, and cancer.
OTOH, with the increased profitability from lower labor costs and higher volume, they may be able to offer more nutritious foods at a reasonable cost.
One of the main reasons for obesity and all of that is because eating healthy is expensive, difficult, or both.
I wold love it if it went this way, but my sense is that this is unlikely to happen simply via the free market. OTOH when government has tried to give a nudge (as with the soda thing in NYC, or Michell Obama’s efforts to get children to eat healthier), people seem to lose their damn minds. 