A cursory examination of the mix of jobs that exist in Tennessee would strongly suggest that 50% job loss in two to three years is positively ludicrous.
I work industry and assure you there is zero chance of this happening. In most aspect of manufacturing robots are inferior to, and more expensive than, humans. We are more than two years away from replacing 50% of industrial jobs with robots - and directly industrial jobs are a minority of jobs.
[QUOTE=Lightnin’]
We’re going to end up with massive unemployment. Huge sections of our population won’t be able to find work, because there simply won’t be any jobs to be had.
[/QUOTE]
We have had this argument before, and the SDMB has some rather vocal proponents of your case, but it simply doesn’t make any sense.
Forget horses; what you are describing has already happened. Almost all jobs that have ever existed have been replaced by machines. Almost all people used to be farmers, but machines replaced them. Even after that machines have supplanted most jobs; hell, office jobs have to a large extent been replaced. Companies used to have platoons of women who just typed; all gone now, replaced by word processing. Telephone operators have been almost wholly replaced by switching equipment. There are no more lamplighters. Domestic servants used to be a very large part of the workforce; they have been almost entirely eliminated by machines.
The idea that machines will cause massive unemployment dates back to the Roman Empire, at the very latest, and has been a constant worry quite literally ever since then every time someone invented a labor-saving device. Yet it hasn’t happened. Why do you think that is?
You say I can’t predict the future, but if you are dismissing my predictions as being unrealistic, then it seems you yourself have definite ideas about what the future will or will not be. Are you of the opinion that automation will not continue to eliminate entire lines of work in the future (as it has been for some time now)? Do you think that living-wage jobs will continue to be available for people with limited intellect who have little to offer beyond manual labor?
Again, the phenomenon you describe has been going on since long before your great-grandparents’ great-grandparents were born and yet the labor force participation rate isn’t five percent. To date nobody has really explained why automating jobs now is really any different from the way they’ve been automated before.
As is so often the case in these threads there seems to be this assumption that the unwashed masses who aren’t smart enough to be on the SDMB are people with “limited intellect” to have nothing to offer besides grunt work.
That is simply not how the world actually is in real life. Most people are not idiots, manual labor is not the province of the stupid, and knowledge work is not entirely the province of the intelligent. People end up in jobs that don’t require extensive education for reasons other than being stupid, and people who end up in white collar jobs do not have to be geniuses.
We don’t know what the workplace will look like in 50 years, but one pretty safe prediction is that there will be a workplace for people. Technological change and automation has so far never created a giant unemployment problem, despite automating most jobs out of existence. This argument has been made again, and again, and again, since before you were born. It has always turned out to be wrong. Why will it be different this time?
I think there will continue to be service jobs where people would rather interact with a human than a machine. When I go to Del Frisco’s, I don’t want my steak dispatched from a vending machine.
And yes, I do think that those jobs will be available because many of them will be so automated, all they require is someone of limited intellect to push a big button.
But this is the root of the panic. It’s not that the manual laborers are going to be out of a job. Those guys have been losing jobs since they replaced flint sickles with bronze. What creates all these job holocaust think pieces is that white collar workers are seeing the prospect of losing their jobs. And journalists especially.
But people aren’t thinking of this correctly. Journalists didn’t lose their jobs because robots or expert systems can now write articles on meaning of the latest Republican primary. They lost their jobs because in the internet age there’s no need for 500 reporters to write the same story about the Republican primary. Now you just need two or three, and those two or three are enough for the whole country. Classified ads used to be a big revenue stream for newspapers. Now nobody pays for classified ads. It’s not because robots took over, instead Craigslist took over. The infrastructure needed to create and view a classified ad is so cheap that the marginal cost of an ad is zero.
Journalists believe in the coming robot job holocaust because the last two decades have seen a collapse in journalism jobs. Not a collapse of journalism, but journalism jobs. And not the guys in the basement who keep the printing presses working, but the guys upstairs writing the articles.
And so if the trends pan out for medical care it won’t be nurses and orderlies and janitors that are in lower demand, it will be highly trained specialists. You’ll still need a nurse’s assistant to get the patient to sit still while the scanner takes a picture, but you don’t need the radiologist with decades of experience to interpret the scan. You still need someone to change bedpans and get patients on and off the bed, but you might not need the expensively trained doctor who oversees the patient’s treatment.
People keep misinterpreting this as robots taking our jobs. But it’s really the same sorts of productivity gains that we’ve always seen that allow trained people to produce a lot more work for the same effort, and allows lower-skilled workers to produce the same quality of work as highly skilled workers.
That’s exactly the kind of thing I’m talking about- the AI might notice that the person has some specific genetic market, and then put together a half-dozen studies from four continents and recommend a very different diagnosis based on the symptoms, or recommend a very different treatment. And more importantly, it would be able to do this kind of thing for EVERY patient and potentially be up to the minute as to published studies, treatment results, correct/incorrect diagnoses, etc… for nearly every hospital, doctor’s office and medical journal/magazine in the world.
That’s just on the actual direct clinical end; I imagine for the near future, the real advantage will be in using the AI for more of a “big data” data mining purpose- identifying that brown-haired caucasian guys who weigh between 270-300 lbs and who do A, B and C in some proportion are X% more likely to have Y occur to them. This is the kind of thing that a normal doctor, or even medical researchers are unlikely to identify with any degree of accuracy, and it’s something an AI might be able to do very, very easily.
That’s funny, because I really don’t care if it’s some underpaid line cook broiling my steak, or if it’s the Steak-O-Tron-7000 with its integrated scale, 45 temperature probes, and steak vapor analyzer cooking my steak. If the AI does it right, it should be indistinguishable from, or even better than a human-cooked steak.
I’m paying for the service, ambiance AND food at a place like Del Frisco’s, and if their having a Steak-O-Tron-7000 in the back means that the food is more consistent and higher quality, then I’m all for it. But I’m with you otherwise, I don’t want a Protectron set in “Waiter Mode” taking my order and serving my steak. I’d rather have a real-live human.
Obviously, there is no “Protectron set in ‘Waiter Mode’ taking your order” but some restaurants are already experimenting with kiosks at which guests order their own food, or tablet PCs at each table.
Why do you think nurse/orderly/janitor jobs will be spared? Autonomous robots are still pretty primitive, but I don’t see a reason to presume that progress will come to a standstill any time soon.
A robot that can change a bedpan is likely to be a reality at some point, but it’s way in the future.
It turns out that mimicking the way human beings do things is really hard, which is why when we get machines to take the jobs of humans they do things completely differently. So the typing pool at the office wasn’t taken over by robot typists, it was eliminated by making people type their documents themselves. Repairmen of all sorts were eliminated not by making repairing things more efficient, but by making consumables so cheap it was more cost effective to throw out the broken item and get a new one than to fix the old one. You don’t have a robot in your kitchen holding a scrub brush in one robotic hand and squirting soap with the other, instead you have a dishwasher. And so on and so on.
So when robots come for the orderly who changes bedpans it probably won’t be a robot that walks into a patients room with hands that can pick up a bedpan. It will be by some process that eliminates the process completely.
I think that the lesson here is that any job that can be easily decomposed into parts that can be doable by an AI or machine will be.
So the good jobs left remaining are going to center around tasks that involve types of interpretation or creative tasks that machines can’t do. I don’t really see how even a really bright expert system AI would be able to take someone’s vision for a new building or something like that and translate it into a design like an architect might.
The way it would work is to give someone a set of desktop computer tinkertoys like in “The Sims”, the customer fiddles around with the design elements putting the bathroom here, the kitchen here, add this wall treatment, this level of insulation, window here, this many solar panels on the roof, and then click “save”.
The design app would know what sorts of designs are sensible, and won’t let the customer create something that won’t work or would violate code or would cost too much. Then the app forwards the design document to the builder and a bunch of driverless trucks haul the stucco and rebar to your job site the next day, the robots get to work, and you get the bill in the mail.
And the design app itself is free provided by the construction company, it’s treated as an advertisement.
Of course all the steps won’t happen overnight, but a 3D design suite that can be used by someone with very little training could be done today. Of course the designs in “The Sims” don’t have to follow code, or the laws of physics. But a more serious app could incorporate all of that. Don’t expect the construction robots to show up the next day either, just because you clicked the “purchase” button. But systems that know how to take a naive tinkertoy design and turn it into a real construction plan could happen sooner than you think.
That is where I see the trend going. Automation freeing up humans to do more human stuff with each other, rather than have our noses buried in screens or machines doing tedious and repetitive tasks we hate.
I’m surprised no mention of the Dilbert Principle or Putt’s Law. You’ll have a relatively small number of really talented geniuses actually inventing new stuff. But the rest of us idiots will be relegated to middle management and other BS jobs where we just sit in nonsensical meetings all day pretending to do stuff.
I feel like the world portrayed in Idocracy is more the result of widespread automation than dysgenics. The people who populated that world didn’t use their brains because they never had to.
“Plenty of 'tards out there living really kick-ass lives. My first wife was retarded. She’s a pilot now.”
I find this fascinating, mostly because of what it implies about your view of healthcare, and the difficulty of diagnosing things. What are the most common medical conditions? Diabetes, heart disease, hypertension, high cholesterol, allergies, arthritis, depression, acid reflux…
These are not conditions that you need a high-tech computer program to diagnosis. What exactly do you think the computers are picking up that doctors are missing? Weird autoimmune diseases? Rare genetic conditions? One-in-a-million immune deficiencies?
Doctors are essentially trained to recognize patterns. And in that you are correct - no doc will ever be as good as a computer that has the entirely of pubmed in it’s databases. But for the vast majority of healthcare, it’s not necessary. Common things are common. You don’t need a fancy computer algorithm to diagnosis diabetes, but getting someone to change their diet and take their medicine every day? Uh…really hard. Really, really hard for so many reasons. I remain hugely skeptical that a machine, even a brilliant AI, will be able to take that sort of task over.
Also, the rarer the condition, the less likely we are to have an effective therapy for it. So, even if you could say your “270-300 lbs and who do A, B and C in some proportion are X% more likely to have Y occur to them” what exactly are you going to do about Y? Let’s say Y requires an invasive procedure to fix - are you willing to have surgery earlier because a computer says you are 10% more likely to have Y in the future, maybe? Start a drug with all the associated side effects?
Weird undiagnosed medical conditions get a lot of attention because they’re ‘sexy’ and mysterious. It’s the House MD syndrome - the idea that rare things are much more common than they actually are. Unfortunately, medicine is just not that exciting. Outside of TV, it’s a lot more of begging people to take care of themselves, eat right, exercise occasionally, stop smoking, cut down on the meth, etc.
None of this touches on the legal, regulatory or reimbursement hurdles from any of this either, which are formidable. The practice of medicine is considered exempt from a lot of regulation - state medical boards take care of ensuring that doctors meet certain minimum requirements, but that’s all. You start bringing machines into it and it’s going to be a regulatory mess. I can’t imagine insurance will be thrilled to start paying for a bunch of esoteric tests because a single article on pubmed made a similar diagnosis in 1971 on a single dude.
Maybe a few people will get referred to AI MD to have a review of their bizarre nonspecific symptoms and tests. But the vast majority of people probably will not need this sort of service for the bulk of their routine medical care. And that is a good thing. You never, ever want to be the bizarre diagnosis in the hospital.
Other than designer babies, there’s the following options:
Etsy becomes the future. Almost everyone is occupied making one-off things for the wealthier, either in real life or (for example) in Second Life.
Personal servants become the big thing. The hiring of nannies, personal chefs, maids, etc. become a middle-class venture.
But in the long-run, and as the world moves into virtual reality, it’s possible that money itself becomes largely meaningless because so much of what we need is so cheap to produce in bulk that it’s not worth having a market around it any more. The only things that continue to have value are land and people. Everything else can be infinitely produced and copied (either by being very cheap to create or virtual), but those two will remain constrained.
If I may digress for just a moment, this reminded me of something I learned about a year ago.
Folks were conducting s study on repetitive stress injury in an office work environment. So they created a faux office where folks would do “work”. They wanted to make the “office” as realistic as possible (to control extraneous variables), so they included a coffee maker. And being good researchers, they tracked EVERYTHING.
Which is how they noticed that the people who drank coffee reported less pain than the people who didn’t.
The efficacy of caffeine in relieving pain from repetitive stress was not what they were studying, but was something their study turned out to be about.
“Experimenting with”?
At Panera Bread it is an option, but at Wawa it is the only way to place an order. Generally, the places that have gone to 100% touch-screen ordering are the ones running a very small labor force to begin with. I expect it will become the norm for fast food within the next 10 years.
I don’t see it coming to fine dining, or even Red Lobster, but those are places where you are paying for the experience more than the food.
That may not be the best example. I mean, those robots are not really being designed to tackle the kinds of jobs being discussed.
As has been said, it isn’t about the robots taking ALL of Job X, it is about robots reducing the workload for Job X to the point where fewer people are needed.
Like, how and industrial version of a Roomba (or Scooba, actually) could handle the sweeping (and mopping). Which means fewer janitors.
And something like Robear doing some of what an orderly does.
An interesting trend: increasing “automation” is not getting a machine to do the job, it is making a machine that can guide an average person through doing the job themselves. Remember when getting your blood pressure taken was something that required a trained technician?
There are vending machines than make keys now.
In the pilot episode for Emergency (made about 1970), there is an incident where a powerline worker dies, and a nurse says to the doctor “IF, when they got him to the ground, there had been a defibrillator there and a man trained to use it, that man would be alive right now. Tell me I’m wrong.” The rare thing then was not the defibrillator, it was the “man trained to use it”.
Today, there are a dozen defibrillators in the local mall, placed much like the fire extinguishers. Each was designed so that it can be used by anyone, although it helps if they can read English.