Looking to the past record is invalid in this instance.
Jobs that require knowledge were the kind that was created in the past, and are being eliminated in the present.
I agree with you completely about the One Percenters, especially the ones in the US. The milk of human kindness is unknown to them. But when automation renders almost everyone unemployable, it will start to affect the bottom lines of the corporations that are the engines that create One Percenter wealth, and in a very negative way. That’s where I pin my hopes for Basic Income being enacted, and at levels high enough to be more than a sop to the social justice types. For Basic Income to work for the One Percenters, it will have to be enough so that its recipents can buy all the things a middle class income allows permits. Otherwise, the corporations will still be hurting. That’s why I think it will be a bipartisan cause, eventually.
This is the part I can’t get Deeg to see. All of the past waves of automation have involved specific kinds of machines displacing specific kinds of manual labor of some sort or another. But the current one involved infinitely adaptable software and increasingly adaptable hardware that displaces both manual and mental labor. Almost EVERYTHING that human beings can do is at risk here.
More than half of the jobs needed to keep people fed are already long gone, and yet the unemployment rate is not particular high. Why is that?
In fact, most of the jobs that have ever existed have been eliminated by technology. Why aren’t most people unemployed?
Because they moved on to find other jobs in the service industry, etc. Now, why can’t robots and automation do those new jobs, too, better than humans? If you want to posit new fields of human endeavor that can’t be replaced by automation, you have to explain why robots and automation can’t displace them.
We have “robots and automation” right now.
But the US economy added a quarter million jobs in December, more than three hundred thousand in November, and new data on previous months revised their numbers upward. If robots can do every job better than humans, why is the economy adding jobs?
The answer is the same that it has been for the last two hundred years. If you keep asking the same question in thread after thread, you will receive the same answer. People and machines work together. We are not substitutes but complements. People and machines together produce much more than machines separately. US industrial production is higher than it’s ever been, as is total US employment. We have more machines today, and more workers, than we’ve ever had before. One hasn’t replaced the other. Both are increasing because both are valuable, and they’re even more valuable when working together.
You’re the one one arguing that “This time is different!” You’re the one saying that centuries of the Luddites being wrong will suddenly reverse itself. The burden is on you to demonstrate that. You’ve given your reasons for this belief, and they are not ridiculous. There is an absolute kernel of truth to what you’re saying: it is possible that we will, eventually, have a machine that can do literally any job better than humans.
This is not an incremental change. This is that singularity thing that people talk about.
We are not anywhere close to this point.
You seem to think that robots and software are so easy to design and build that we can plop the things down anywhere, at essentially no cost, and then toss the hyu-mons to the trash heap. But the economy is still adding jobs. The hyu-mons are still valuable. We have more robots and more workers than ever before. Employment is still going up. More complex machines require big costly investments, they’re hard to update, they take literally years of planning. Software is shit that needs continual monitoring and updating. Human wants are essentially limitless, and our society always learns to take advantage of the workforce it has available, because that workforce is the most valuable resource we have. When humans are replaced in one context, their valuable skills are available in other contexts.
A person can be told, “Do this!” and they understand. The only kind of automation that can replace the human ability to adapt to any given work circumstance is full Artificial Intelligence. Nothing else will displace workers, on net, because anything short of that requires expensive human programmers and expensive human engineers and expensive human mechanics, working over years and years of planning to make sure the new machines really are better than the old machines. Meanwhile, the workers can just filter to new jobs like they’ve always done in the past two centuries. As long as machines can’t repair themselves, as long as factories need humans to keep the pieces working, then the machines require an expensive, extensive human infrastructure around them in order to function.
Anything short absolutely full flexibility in automation will mean machines must continue to be designed for discrete tasks and constantly supervised in those discrete tasks, and that programming, engineering, maintenance, and all the rest – although very efficient when properly utilized – will continue to be much slower to organize and arrange and much less flexible than just hiring a bunch of dudes and telling them “Do this!” People are clever and can learn things on the fly in a way that an engineered device cannot. This flexibility is priceless, and it will continue to be priceless right up until the fateful point when a fully autonomous machine can be built. Only then does the picture shift.
Think: Real Estate Value Collapse.
I think you’d get very different answers if you asked people, “Do you think YOUR job will be automated in the future?”
It’s very easy to assume a job will be automated if all you know about that job is the simplistic surface description of it. For instance, “Truckers will be gone, because we have self-driving trucks.” But then if you ask a trucker, he might have a different viewpoint because he understands the complexity of the job. For example, truckers check the truck for damage and safety risks. Truckers check to make sure the load hasn’t shifted and is properly strapped down. Truckers deal with roadside inspections. Truckers manage the problem when the truck breaks down. Truckers deal with the people at either end of the delivery - making sure the proper goods get loaded in the right away so they don’t shift or cause imbalances, and then they help offload at the other end. Truckers act as agents for the company, getting signatures, answering questions, etc.
And I’m not a trucker. I’ll bet if you asked a real one he could come up with hundreds of tasks he has to do in the course of his job or when there is an emergency or an exception.
I am, however, a software engineer, and I saw that at least one person thinks that software development will be automated. That is a complete laugh. For example, a big part of software development is simply requirements gathering - understanding what the customer wants. You have no idea how hard this task is. It’s so hard that poor requirements gathering is probably the primary reason for software projects that fail or go over budget or do not make it in the market place. And it simply can’t be automated because the process is one of engaging the customer, digging for information that the customer doesn’t even know he has, understanding the environment the software will be used in and the goals of the people using it, etc.
And because requirements are never fully complete or perfect, even low-level programmers have to make judgement calls constantly.
Realtors seem to be another category that everyone ‘knows’ is going to be automated because there are on-line real estate markets. But what does a good realtor do? The last time we bought a house, our realtor screened numerous properties for us to avoid wasting our time. He then dealt with the home-owners of the candidate houses and arranged times for a viewing. When we put in an offer on the house, he had to take it to the home-owners and negotiate with them on our behalf. When they wouldn’t agree to the price, he talked to the other realtor and both of them agreed to cut their fees to get the price in the range we would accept. In other words, he used his brain to find a way to make the deal happen.
But that was just the start. Our realtor then arranged to get us a lawyer to close the deal, he handled getting the proper legal documentation, he found us the best price for a mortgage through a broker, yada yada. And I’ll bet he had to do a lot more stuff that we’re not aware of as part of the job.
Good realtors add value because they have knowledge and special information. They know who the good lawyers are and who the bad ones are. They know where corners can be cut. They have better knowledge of the market and what houses are actually selling for rather than what they are being advertised for. To a large extent they are knowledge workers. Only a small part of their job is actually finding houses and showing them to their clients.
And so it goes with most jobs. Even store cashiers. I’ve been a cashier, and while some of it was rote work, a lot of it wasn’t. I had to stock all the magazines, chocolate bars, and other small goods you find at a cashier station. I had to clean off the produce scale and wipe down the belt when it got dirty. I was also there to act as a theft deterrent, since people had to walk past me to get out of the store. I also had to call other people when the customer had a problem or needed something and couldn’t find it. I had to clear jams in the cash register, and deal with customer ATM cards that wouldn’t read properly. I had to deal with the times when the customer didn’t have enough money after I rang everything through and bagged it. I had to handle the goods that customers would decide at the last minute they didn’t want.
I had to bag the groceries - has anyone ever seen a robot that could bag groceries? A surprising amount of judgement goes into it to do it really well. I mean, for a human it’s trivial, but for a robot, not so much. How much weight should go in a bag (varies depending on whether there are sharp corners on the products, whether the customer is a large man or a little old lady, etc). You have to know which products needed to be individually wrapped (meats, soap, products that are leaking or sweating, etc). You have to make sure that crushable objects are not under heavy objects. All trivial choices for people, but really hard for machines.
Have you noticed how many stores have automated teller machines that go unused while people line up to see a real cashier? That might be because they don’t work very well. I use them when I have one or two items that can be easily scanned. But the risk of something going wrong goes up as you add more things. You put your lettuce on the scale and punch in the number, and it doesn’t work. Now you have to wait for a human. You put a product under the scanner, but the barcode is unreadable. Now you have to wait for a human. You forgot to enter your discount card at the right time - now you have to wait for a human. Robots are good at handling routine, normal, well defined cases. They’re bad at exception handling because they do not have judgement.
Another impediment to automation is regulation. A significant amount of regulation we have today is job protection for special interests. I remember a couple of decades ago a university came up with a way to build houses that involved the walls being a composite with plumbing and wiring baked right into the wall structure. These walls could then be plugged together like lego to make houses, and they would not require electricians or plumbers on the job site. Much cheaper, great for the poor.
The designed worked, and they wanted to set up a company to construct them, but ran into a major stumbling block: Every state has different trade rules, and almost all of them mandate that plumbing or wiring connections be made by a tradesman. Some states had actually codified requirements like screw-in fittings instead of press-fit fittings, specifically to prevent that type of construction and to protect the jobs of people in the politically powerful trade unions. The maze of regulations was so thick that the whole project was scrapped, so we still put houses together the way we did 50 years ago.
My prediction: In 30 years there will be more people in the workforce than there are today. Automation will make inroads in some fields as it always has, but the increase in productivity and the shifts in what we value and what we do will create jobs in areas we can’t even conceive of today.
If the labor participation rate is lower, it will have more to do with demographics, wealth, social programs and social mores and will have very little to do with people being unemployable because of automation.
I was thinking of starting buggy-whip business, too soon?
With Carbon footprints being an issue, using more horses might or might not be a good plan.
Post gasoline?
A better plan than it might seem.
A few manufacturers are still in business, and the whips are called carriage whips nowadays.
I agree with you that automation is not going to eliminate a lot of these jobs. However, the number of people required to fill these jobs is going to be a lot less than if automation did not exist. If we can create new categories of jobs we’ll be fine - that is what we have always done. However we seem to be creating lots of relatively low skill low wage jobs. For instance, there is and is going to be an immense need for health care workers to wipe the dribble off the faces of people like me. But no one is going to put a kid through college on that salary, at least as it stands today.
But one trucker is going to be able to handle this, not two. And for double trucks also, which reduces the need for drivers. Heck, the interstate highway system probably has done a good job in increasing trucker productivity.
If we were back in the Assembly Language/Fortran days, how many programmers would be required to write the software we write today? More than we got. Requirements is a dealing with people job that doesn’t seem to get automated well. We write about the same number of lines of code a day as always, but those lines do a whole lot more. And we have lots of packages that eliminate the need for programmers. When I started large companies developed their own EDA tools because none were commercially available. Now no one does, we buy them, and EDA departments do a lot less programming and a lot more support.
Now this one I totally agree with.
In my grocery store there are many self-check lanes, where one person monitors six lines, and where people bag their groceries themselves. They are terrible, but they are the wave of the future I’m sorry to say. People do use them because they are more for small orders, and because there are more of the self-checkout stations open than regular stations. When they get good and can handle stuff. (The other day I needed a person because I was buying cough medicine which had alcohol and so couldn’t go through self-checkout. But there is always a line for them.)
He is talking about the future, not today. Do you think the robots we have today are as far as we can go? There is a big push in developing robots that can deal with people (one was on the cover of the latest IEEE Spectrum) but they are not commercially available yet.
A fast food place redesigned for robot counter staff is going to have serious advantages. Better customer interface, fewer mistakes, and no turnover. The robots will have to be cheap enough, but there will be a hell of a market, so I suspect it will happen.
As for adding jobs today, sure, but we’re playing catchup still. And we are just now finally seeing some wage improvement - a good sign that employment has been tight even as we’ve added jobs, and a sign that the jobs we’ve been adding have not been great ones.
We need lots more people doing nursing - but do you think those jobs are going to pay well? Even though they won’t be automated for a very long time.
Dang, I waited too long.
There’s the future, and there’s TheFuture, and they are not the same thing.
Put it another way. We have two facts here that we have to reconcile.
Fact Number One: The economy is adding jobs. We have better technology, more machines, and more workers than we’ve ever had before. The trend is upward, for both workers and machines, and this long-term trend is centuries old despite business cycle fluctuations.
Fact Number Two: There is a theoretical machine that we can easily imagine that would be superior to a healthy human worker in any job and across every conceivable dimension.
So here’s another way to describe the same point. Some people think machines will start to eat away at what people can do, where the “red” of automation slowly replaces the workers, starting with the least skilled workers first, looking something like the following. These workers have literally no work options left because they can do literally nothing better than a robot. The green dots represent skills that require levels of human ability that cannot yet be fully automated.
------50IQ-----75IQ---->-100IQ-----125IQ-----150IQ------>
I think this is a fair visualization of people who fear an 80% unemployment rate.
Unfortunately, such a belief shows a regrettable lack of appreciation for how valuable human beings actually are, and how similar our skills sets are. If robots get around to replacing workers, on net, it will look like this. (You’re going to have to scroll to the right. Very very far to the right.)
------------------------>--------------------------------------------------------------------------------------------------------------------------------------------------------------X------------->
That X is the entire spectrum of human skills. All of us collectively.
You can’t cite single industries. We all used to work agriculture, and now we don’t, but there are more jobs than ever. Why? Humans beings are clever. Human beings are flexible. Human beings can quickly adapt. These are valuable skills. That’s why other people build machines. If human work weren’t so damn valuable, there’d be no incentive to build machines to replace that work. Human beings can quickly adapt, but robots can’t be quickly programmed or quickly engineered or quickly implemented.
Basic human intelligence, of the kind all healthy people have, is valuable.
If any single fully functional human can be totally dominated by a robot along every conceivable dimension of economic production – where the worker can’t just hop to a new industry like workers have been doing for literally centuries – then the entire species is irrelevant for production the very next day. For healthy humans, we’re simply not that different. One robot better than the worst of us means every robot better than the best of us.
We can tell a human worker, “Do this!” and then be confident they can start doing it without the program crashing and needing to be debugged over the next couple weeks. That is an astoundingly valuable skill, and robots aren’t anywhere close to replacing it. The thread title is “professions that will cease to exist”. The more interesting question is “professions that will spring into being”. But the more interesting question is the harder question because we can’t predict the future course of technology.
This is job catchup from a cyclical recession caused by a financial crisis.
It’s not catchup from a technological displacement.
People have a hard time dealing with confounding variables. During the worst part of the contraction from the Great Recession, we had posters blaming all the robots (even though it was a money problem) and we had posters blaming all the Chinese (even though it was a money problem). Job growth isn’t being crippled by offshoring. Wage growth? Yeah probably. It’s fair to look overseas for the main causes of wage stagnation (as conventionally measured) in the developed world. China and India joining the world labor markets added two billion people to those markets. Flexible, clever, valuable human beings hit the market, and at roughly the same time US wages showed relative stagnation. Hard to believe that’s a coincidence. When they become developed countries, unskilled labor markets in the US will tighten again.
Another very difficult problem is understanding how to make a change to a complex existing system (which is 90% of activity for enterprises).
Most of the changes in any moderate to large enterprise requires deep knowledge of the enterprise data model (e.g. special pricing will need to get a new promo code field), the enterprise usage of the data model (e.g. some customers are configured to use special pricing, some use a different module for price resolution), all human processes and interactions that generate or consume that data, all external systems that generate or consume that data, how the humans and systems will be impacted by the changes, etc.
I think the only way this could be automated short of human level intelligence, would be to have a computer-usable store of information about the functionality of the enterprise system including the capabilities of the software, the specific implementation/configuration of the enterprise, the non-computerized processes and actions the humans take and the external system dependencies.
To create that would require a massive amount of effort and would require very disciplined processes by the groups that add/change technology and processes once the knowledge store is created.
Today, in my experience, new functionality and new dependencies due to the tools and platforms used for each project are added at such an incredibly rapid rate in any growing company that doing the above would be too expensive, nobody would (or does) do it.
I think the only way to get there would be if this type of information was baked into commercial products, but it’s kind of a chicken-egg thing, requires money from vendors with no immediate selling value.