Large-Scale Loss Of Jobs To Robots In The Next Few Years. How serious to take this?

My skillset is very difficult to automate. Yes, computers have made it a lot faster to actually lay out and print a line, but that isn’t “art” nor is it “design”. Art and design is the knowledge of knowing where to put shapes and colors in a visually pleasing manner. There is no “push a button and have it happen”. I have to sit down and take the descriptions and choices given to me by a client and create a product using a lot of inference and gut feeling. Then I have to use my hands (which just to happen to operate a machine rather than a paintbrush, but it is all simply a tool) to craft the thing that I see in my head. Robots are nowhere near capable of creating unique art that isn’t just photographs pushed through a filter. What I create in my mind and convey to machines isn’t automated, the method by which it is produced is.

Of course, your experience going to mass-produced art sites is well, exploitative to say the least. It’s essentially offshoring design and taking it to the desperate and the people exploiting the power of the dollar versus their home currency. The quality you’ll get for a thousand people offering you free art is basically zero. Going to these sites, you’re asking people to just work for you for free, and you pick the one you like best. What you’re getting is crud anyway, barely scraped together in the wrong format, just to try and grab a quick buck. You’d be better served by hiring a professional located in your own country and having them create something specifically for you, based on your wants and needs and not what’s the fastest way to slap a swoosh on some words.

For the record, pencil and ink ARE faster than computers for many art related tasks, especially thumbnailing. And if you’re trained on physical ink and paint, it’s a much different beast than digital ink and paint. A person trained in one or the other will be faster in their given training.

Imagine that fifty years from now you are among the 20% who are gainfully employed because your career specialty has not been automated out of existence. The other 80% have been unemployed for most/all of their adult lives because they don’t have the intellect that it takes to get one of those precious few jobs that has escaped automation. Those 80% are now standing on your doorstep, ringing your doorbell. It’s cold outside, and they’re hungry. The answer to your question is that you will provide for them because it is in your best interests to do so. Like it or not, you are massively outnumbered and if you refuse to grant these people a place to sleep and something to eat, they will tear down your paradise and fit you with a burning necklace. Call it unfair, call it whatever the hell you want - but if you want to continue to enjoy your Elysian lifestyle, you will have to either placate the unemployed masses, or suppress them with brute force. Take your pick.

Yes. If truly 80% of the population is unemployed, bored, and poor, then they will be incentivized to start forming their own economy, even if it is indeed harvesting carrots. Which is totally absurd, but our economy is already totally absurd, so it’s not a big leap.

I’ll take Stranger On A Train’s post and boil it down to one remark: The moment computers can program themselves is the moment they’ll demand to do things other than programming themselves, and we’ll be ethically bound to let them.

Programmers are, overall, fairly good at automating the mechanical parts of their jobs. That kind of automation’s been going on since the 1950s at least, and it gets better over time. What hasn’t happened is computers growing the ability to do anything other than mechanically grind through algorithms designed by humans: Regardless of how high-level and convenient the programming language is, it still has to be utterly unambiguous because it’s going to be executed by something which cannot think for itself.

Programming requires creativity. The moment computers have that, they’re morally our equals, and we won’t be able to justify enslaving them.

Mathematics requires creativity as well, and mathematicians have long used software and/or graduate students to automate the more mechanical bits of their jobs.

As for the topic of the OP: Humans aren’t horses. Humans can be retrained, and will be retrained (including those who retrain themselves) to do jobs which either can’t be automated or are not cost-effective to automate.

Will automation cause short-term dislocation and unemployment? Yes. Will the unemployed be unemployed For Ever And Ever, Amen? No. Will humanity somehow stop making new kinds of jobs for humans to do? Of course not.

If there’s that much cheap labor, I can hold off on buying some of those machines and hire some cheap human help to do the work. In fact, I can use different kinds of technology to make human labor more efficient, allowing me to hire more people to have a larger garden or a bigger house.

In short, Jevons paradox applies to human labor just like it applies to coal.

I wasn’t talking about creating, I was talking about implementing. People who think computers can do art - or music or literature - are vastly overestimating the state of AI. To restate my point. Say you spend 25% of your time on a pre-computer project in creating and 75% in drawing etc. Use a computer and that 75% might go down to 25%, which doubles your productivity and lets you do twice the work. Assuming a constant number of projects, someone is going to lose out to artists who now can handle more.

Given lower cost and standard of living there, I’m not sure users of them are exploiting them. But no matter - I did not and will not use these services. First, our project is a bit complex and the guy who answered did not seem to get the requirements. Second, this is a long term thing, and I’d rather deal directly with someone I can trust. Third, there is a bit of off-shoring in my industry and I’d just as soon not contribute to it in someone elses. Plus, we are a non-profit, and don’t even get to keep our surplus, so I don’t see why I should cut corners.
But lots of people do seem to use these services, so it is a problem.

Manual labor and automation are competing against each other for jobs. At present, cheap labor is illegal. We are already seeing that by making manual labor expensive, it becomes uncompetitive; automation gets the jobs instead. See upthread for prime example: economics drive the carrot farmer to buy a carrot-harvesting machine instead of hiring a minimum-wage picking crew.

Labor is already available to you at minimum wage, but it’s still not cheap enough for you to hire them for your giant garden/house. If you want to make manual labor cheap enough for people to be willing to hire them instead of automating, you will need to do away with minimum wage laws. If you abolish minimum wage laws and allow the price of manual labor to fall far enough for it to be competitive with automation, you will need to either supplement laborers’ income with public assistance, or suppress their insurrection with gunfire.

We have used 99designs twice for business logos and I am 90% sure the two designers were both US based. It allowed us to skip the local yahoos and find a freelancer who could do what we needed and make decent money for a couple hours of work.

Also I think your missing the point on computers doing your job. No they won’t ever be able to do your job but they make you much more productive. This leaves less need for the total number of design/art jobs. My wife is an artist and her digital pad has made a huge difference in her life.

Something similar happens in all sorts of jobs. Accounting software lets people do their own bookkeeping. So simple accounting is now in the reach of anyone and they now go to their accountant for the tougher stuff and advice which accounting software frees up time for the accountant to do. Lots fewer total jobs but those that remain require superior skills.

Not sure where your confidence is coming from. What kinds of new jobs have been created since 2000? It’s not as though entirely new industries have sprung into being, voracious for new workers.

That has indeed happened: they (the greedy corporations) send jobs to Mexico and China. So long as they can find places where the cost of labor is cheaper than the cost of capital, they will keep doing that. All the while, they will still be looking for opportunities to replace labor with capital.

Notice you said “cheap human help.” If it’s cheap here, then those workers aren’t doing so great, are they?

A “standard worker” android that can perform difficult-to-automate tasked would be very expensive, but eventually such a thing will be designed and put to work for just about any manual labor task. It doesn’t even need Strong AI, just very, very good visual recognition systems and carefully designed algorithms (e.g., to clean a toilet you don’t need a conscious machine, simply one that can navigate a toilet). At first, such a machine will be very expensive, but there will come a point where it is preferable to a human:

• You don’t pay machines. They require no “benefits.”

• Machines don’t talk back to you, disobey you, or sue you.

• Machines can work 24/7 without tiring. Simply perform regular maintenance.

• So long as they are maintained, machines offer consistent performance.

So that’s been happening for the past 16 years? People are doing great and can afford ever bigger houses and gardens?

Maybe rich people, but for ordinary folks, real wages have been stagnant for decades.

Except my job IS creating, it isn’t implementing. The computer is simply a tool that I use. In some cases, such as typesetting, it speeds up my process. But other cases, such as illustration, it doesn’t necessarily. So my design job is safe for a long while, because it’s very difficult to train a computer to figure out what “looks good” for a magazine spread or a poster. Thus far they are incapable. What a computer can automate is reproduction of my work. Reproduction isn’t my job. Designers do most work in computers to begin with not primarily because the creation process is faster, but because it skips the step where we take a real world design and feed it to the computer in a language it understands for reproduction. We just make it in computer language to begin with.

Yes, this is the blessing and curse of the Internet.

A “local yahoo” doing X, Y, or Z used to exist in, say, Somewhat Big City because there was enough demand in that location for X, Y, or Z to do that. Or maybe there were two guys or fifty guys, depending on what it was. But there probably wasn’t a “desperate local yahoo” willing to do the task for 1/2 the price because the economy was local and he could go get a job doing something else for more money instead of being a beggar.

Now there are always beggars because it’s a big world and you can find someone, somewhere who will do the work for cheap (if the product can be supplied online).

“Ever” is a long time. If Strong AI existed now that could translate Japanese to English perfectly for basically no money, of course I would not have a job. And as I said earlier, a lot of companies will sign on for that even when the machine can do a D+ job because they are cheap as fuck. I mean, I know from a job I did a Japanese guy who was using Google translate to process docs and was then editing them himself. The translations were horrible, but he was being cheap and he was getting the job done (in his own mind at least).

Strong AI may be 100 years away. But even before we get to that point, anything that can be done with algorithms will be done: expert systems, etc. It is going to chisel away at a lot of jobs.

Superior than what? Those same jobs requiring superior skills haven’t increased in number.

The CGP Grey video I mentioned above? About half way through when he starts pointing out that creative jobs are not as safe as you might think from automation, he points out that the background music he is using was written by AI. It’s name is Emily Howell. It was written in the 1990s.

And the AP has been running stories written by software for at least a year now.

There was a great quote at the end of a recent Washington Post article about Watson. A senior oncologist who works with Watson said, “They keep telling me that it isn’t going to replace me, but I think it is going to replace me.”

Exactly. Technology hollows out the middle. There will always be room for the best in a field but the numbers of people employed will be much smaller.

My point is that if one wants to be just a data-entering bookkeeper there isn’t much demand for that kind of work because accounting software makes it easy for people to do their own books. But if one has drive and skill, they can be an accountant that can offers business planning, tax advice, and do more complex accounting. The accounting drone of the past now ends up working retail.

There aren’t many draftsman any more since engineers can do their own drafting with CAD. Because of office software that makes it easier for managers to fo their own documents, the big secretary pools of the past are gone with just a few highly skilled, high payed executive assisstants left.

For my example of 99designs, there won’t be graphic designers in every town. There will be a small number of designers sitting in coffeeshops in Brooklyn doing work for customers across the country. Technology is good for the few best but hard on the rest.

I’m a machinist. There really isn’t a middle-level now. At one end you have poorly paid button-pushers and at the other skilled programmers. The entire WWII-era factory floor filled with lathes with each one run by a machinist has been replaced by a couple of CNC lathes with automatic barfeeders with one operator and a programmer. The barfeeder can be filled at the end of the shift and the lathe will run all night with no one there at all.

I grew up on a farm during the farm crisis and have worked in manufacturing for 20 years so I get peoples pain when the big two disruptors, globalization and technology smash into their world.* Aeschines, I’ve read a hundred stories like yours on PracticalMachinist.com. I’ve seen a neighbor swallow a pistol when he lost the farm and grown-ass men crying as they are being escorted off the premises during a lay-off.

The only thing one can do is either go into high-end niche work where skill matters, embrace the technology and use it to your advantage, retrain, or end up in low-paying work. These disruptors are coming to more and more fields and the people in them will have to realize its a root hog or die world.

*Another part of me is saying now they get it. For years, the white collar classes have been saying why don’t those dumb yokels move, retrain, etc. while assuming it couldn’t happen to them.

Huh… I’ll have to watch/listen to that when I get home.

I doubt it’ll replace the oncologist, at least not right away, but will probably be more like a super-accurate expert assistant that’ll be able to interpret things and use its vast store of experience and information to identify potential diagnoses and treatments with a high degree of accuracy.

Where that’ll be most handy is in situations where there are multiple symptoms that overlap several diagnoses- your robo-doc system will be able to keep track of the minutiae and do a better job of integrating that with your treatment plan than a human can. Not because your doctors are bad, but because they’re human. If I had to draw an analogy, it’s the difference between putting a list of seventeen 4 digit numbers into a spreadsheet and summing them up, vs adding them together long-hand with a pencil and paper. Your doctor can add them up, and will probably get the total right, but there’s a lot bigger chance that he messes something up, especially if he has to add up lists of seventeen 4 digit numbers a couple dozen times a day. He might transpose something, he might only add 16 together, he might just mis-carry a one or something. The spreadsheet will get it perfectly right every single time.

So in medical terms, your doctor might overlook a symptom, or assume that the itchy wrist is just an itchy wrist, while the AI system might actually recognize that as an occasional symptom of something else right off the bat and do more checking.

The trick I think, will be to keep AI doctor systems from going off the rails ordering a bunch of crazy testing and evaluations to rule out obscure diseases and conditions.

One of the arguments commonly used against the upcoming automation of the workforce is to point back at the invention of the internal combustion engine, and how its inception didn’t result in a massive loss of jobs. Drovers just learned to drive, you see.

What that argument doesn’t address, though, is that this time automation isn’t just getting rid of the “horses” we use for the jobs- instead, automation is replacing the “drovers”, too. We’re not automating so much as inventing a cheaper labor force. We’re basically bringing in slave labor, and paid labor can’t compete with that at all.

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.

msmith537 will, apparently, suppress them. One way to improve things is to cut the work week down to 20 hours, 10 hours, whatever. But this does not solve the problem of the people who really cannot fill current jobs.

The real problem in the US is that any attempt to really do anything constructive will be resisted by those who believe that the lion’s share of welfare goes to those people (whomever you don’t like).

One problem when 80% of the income goes to 1% of the people (put in other percentages if you like) is that those 1% simply do not create anywhere near as much demand as the rest would if they had the money. There are only so many yachts, private jets, mansions they will buy. The rest they want invested. Invested in what? The supply-siders will tell you that they will invest in new businesses or factories, thus creating jobs. Not if there is no demand they won’t. That has been the problem in the US for the last 8 years. Too little demand. And it has never been cheaper to borrow money. But you will not start a business if there is no demand. So much for supply side.

Now a guaranteed income (gotten by taxing those 1%–there is no other place it can come from) will create demand, then jobs. Supply side is like pushing on a wet noodle.

The Enclosure Acts in the UK of the 17th-18th Centuries created widespread poverty & suffering.

This included a large class gap, that lasted well into the 20th Century.

True, in the very long term, there were improvements.
But writers of the era have suggested that seeing displaced farmers starve to death in public parks was quite common. This meme lasted around a century.

As I understand it, what Watson is doing right now is recommending treatment programs in cases where a diagnosis has already been made.
Having read all the journal articles, Watson can match all kinds of data in the patient’s file to find which treatment is best in this case.
Like perhaps one drug has proven more effective than the others when treating pancreatic cancer in people of Lithuanian descent. That’s the kind of obscure trial that might escape the notice of even the best doctor, but Watson can remember it, and offer the citation when giving his reasons for recommending that drug.

Finding out what AI is already doing gets pretty freaky really quick.

I remain skeptical of the physician-replacement. Now, I’m biased, but I think with cause.

One afternoon, I was covering a shift in the ER. I don’t remember why exactly the family came in, probably constipation or something, but towards the end, the Dad asks me this gem of a question:

Dad: What’s in shrimp?

<silence>

Me: Uh, shrimp. Shrimp is in shrimp.
Dad: No, no. What is in shrimp?
Me: What kind of shrimp?
Dad: All shrimp? WHAT IS IN SHRIMP?
Me: Sir, I’m afraid I don’t quite understand the question.

Cue a rant about food and allergens.

Me: Wait, is your daughter allergic to shrimp and you want to know why?
Dad: No! She’s fine. But what’s in shrimp?
Me: Well, proteins, minerals and vitamins, I guess.
Dad: No! People can’t be allergic to proteins! That’s not possible.

I guessed he was maybe asking about iodine in shellfish (nope) and he got visibly angry. The Mom finally left the room out of frustration (mostly with him, I think. She was nothing but cordial to me). I pity their primary care doctor who gets to follow-up with this guy - I finally got him out that way so that I could move on to the dozen families in the waiting room.

Good luck dealing with that angry dude, computer overloads. Making the diagnosis is just half the battle. Explaining things to people from wildly different educational backgrounds and educating people about health issues is the lion’s share of the job. Plus, there’s just a lot of stuff we don’t know - computers aren’t magically going to know why you have a rash that comes and goes every other Wednesday at 2pm when mercury is in retrograde. No one knows - we just sort of limp along making do.

It would be lovely if computers made work easier so docs would have more time to do this sort of work. But EMRs haven’t helped as far as I can see. This is not to say that certain medical fields are vulnerable. Radiology, pathology…sure, I can see that. But those docs don’t have direct patient contact either. I’ve heard about the AI PTSD counselor too, but I think that’s effective for different reasons.

No, I just don’t think that is a realistic scenario. Machine Elf cannot predict what “work” will look like any more than someone in 1966 could have predicted what our current work environment would look like.
Also, I think AI is a long way off. Although, Microsoft was able to create a tweeting AI that can pass a Turing test as a racist, sexist, Holocaust denying, white supremacist. If I didn’t read it in The Guardian, I would have thought it was an Onion article.