I see a similar problem even in software jobs. Employers frequently advertise for highly specific skill sets that almost nobody has. Then when nobody or only fraudulent people apply, they reject them all and claim a skills shortage. The problem seems to be a basic misconception about how transferable software skills are. An excellent programmer with no experience in Python will be out performing a poor Python programmer in a matter of weeks, even though they have had to learn an entirely new language. This idea, however, seems entirely lost on most HR departments, and the result is an almost entirely "fake" skills shortage.
It's not just HR. I did an interview with a YC startup that obsessed over my lack of Rails experience. I have several years of industry relevant professional experience in Java, Python and PHP, but only ever dabbled in Ruby and Rails. Their offer (which did come through, grudgingly, through many repetitions of "so will you commit to teaching yourself rails before starting?" (No, I won't.) was quite a bit on the low end, but I'd probably have turned it down anyway for that strange lack of comprehension of programming skills.
I think a core problem is "HR" departments with absolutely no understanding of the positions they are supposed to be hiring for, even for technology companies where technology is supposed to be a core competency.
The only thing stopping me from calling out HR departments in general as a sign of organisational cancer, since I am yet to encounter one that pulls even a tenth of its weight, is that they tend to be comprised of the only cute girls in the whole company. Even on Friday (2 days ago) I was out drinking with a bunch of tech guys and a single cute HR girl. That is pretty much the sole reason I would oppose eliminating HR as we know it from all technology companies.
Agreed. Further, I personally think this is the issue at the core of the "data science shortage" right now.
"Data science" means many things to many people. Some shops want BI folks who can code. Others want people who can keep up in deep learning threads (like the one on HN's front page) and know about Hinton's work. Others still want HDFS badasses who can distinguish between Cascading and Cascalog (lol).
Shops classify all of these things as "data science" when they're really very different (respectively: Excel/scripting BI analyst; ML researcher with a focus on deep learning and neural networks; data infrastructure developer with broad exposure to MR frameworks)
This is stupid. I worried about checking off these boxes until I realized recruiters are doing the 1990s Java thing all over again. This time, though, the buzzwords are Hadoop, information retrieval, and Andrew Ng. I think it's best to ignore this, keep working hard, and grow your employment options.
A similar problem is HR departments abusing the "knobs" on the position's requirements. For a software-based CRM company, they required something like 7 years of Java EE development experience for the lower level development positions, and 14+ for the team lead positions. How many 14 year Java developers are there? Furthermore, do you really need 14 years of experience in a particular language for this position? I think the HR desk is lamenting the lack of good applicants and turning up the "years of experience" knob, which is in all likelyhood counter-productive.
Yes, it's similar. To explain the manufacturing situation in our terms, they're advertising for the equivalent of "Java programmer, 1 year of experience" and paying $10/hr. No one applies. Then there's a phony "skills gap". But also, no one wants to study to go into the field which actually creates a real skills gap.
Often times the absurdly specific skillset requirements are an H1B game. To get an H1B, you have to prove that no citizen is qualified for the position, so when companies find a foreign candidate they like, they'll put out a job ad with exactly the skillset that that candidate has, and then when nobody responds or all the candidates are unqualified, they can hire the person that they really wanted to hire in the first place.
Economist Dean Baker comments on this: That Shortage of Skilled Manufacturing Workers is Really a Shortage of Employers Willing to Pay the Market Wage
News stories have been filled with reports of managers of manufacturing companies insisting that they have jobs open that they can't fill because there are no qualified workers. Adam Davidson at the NYT looked at this more closely and found that the real problem is that the managers don't seem to be interested in paying for the high level of skills that they claim they need.
Many of the positions that are going unfilled pay in the range of $15-$20 an hour. This is not a pay level that would be associated with a job that requires a high degree of skill. As Davidson points out, low level managers at a fast-food restaurant can make comparable pay.
It should not be surprising that the workers who have these skills expect higher pay and workers without the skills will not invest the time and money to acquire them for such a small reward. If these factories want to get highly skilled workers, they will have to offer a wage that is in line with the skill level that they expect.
I think it is important to note that the fast-food jobs with comparable pay are low level managerial positions, not entry level ones. You can start working for minimum wage at a fast-food job without any previous experience and/or skills. These jobs require very little training, and allow employees to add value almost immediately (and well before promotion to the $15-$20 an hour positions).
Contrast this to the skilled manufacturing jobs which require up front experience. Though many blue collar fields offer entry-level positions with on the job training, apprenticeships, and opportunity for advancement, this doesn't appear to be common practice in manufacturing. Why not? I think the main reason is that it is very hard for a low skill worker to add value to a manufacturing company. There aren't any comparable entry-level positions that allow the employee to learn while still being productive.
Because of this, hiring an unskilled employee for the purpose of training them is a huge risk, since it requires a significant investment. And since this industry is already very unstable with razor-thin margins, it's not something many employers seem willing to do, which is unfortunate.
So maybe the solution is coming up with better training programs, so that manufacturers can hire new employees without taking on such large risks?
Classic: according to the article, the definition of a qualified new hire is one with the qualifications to get accepted into university engineering programs.
Ya'll have probably read some of Prof Peter Cappelli's editorials (he's been making the rounds to promote his book "Why Good People Can’t Get Jobs"). One of the things he's pointed out is that HR use of resume databases encourages people with poor search habits to believe there is no choice. Basically, it's a failure consider the Bayesian math of nested filters.
Consider a company, in a city of a million people (don't want to pay for relocation), which wants someone with a bachelor's degree and five years of experience. The US unemployment rate for people with bachelor's degrees or higher is ~3.8% and about ~7% of the population is between 25-29 years old. Rough and tumble numbers put the pool of candidates around 2500-2700 before including any subject-specific knowledge. Ask for a specific discipline, like "Computers, mathematics, and statistics" for which 4.2% of bachelor degrees were awarded, and you can cut your your pool of candidates down to about 100 (1/10,000). If you also filter to require mastery of PHP, javascript, and Marqui, 10 years experience with Server 2008 and fluency in a Romance language...
tl;dr: the absence of time travelers in the applicant pool is not evidence of a skills shortage.
"The so-called skills gap is really a gap in education, and that affects all of us."
That closing line doesn't really sound like the gist of the piece at all. It sounds like manufacturing employers just don't want to or don't feel they can afford to pay works a fair wage.
Exactly. This article makes it sound like manufacturers somehow aren't restricted to the supply and demand of their labor.
If laborers are leaving manufacturers for better pay at fast food companies, workers need higher pay. Its not that complicated. If the company can't afford higher wages, maybe the demand for this skilled labor isn't what we think it is.
If these skills are actually in high demand, they will pay the bills.
Here's the thing, though: You have to assume that they are correct. If one believes that they are incorrect, he is effectively stating that he (someone thinking abstractly in the absence of any hard data) is better at pricing the value of the work than the manufacturer (who is directly affected by the outcome and has a wealth of hard data).
Several commenters have basically claimed that the manufacturers are behaving irrationally; a claim like that requires some powerful arguments to credibly support. I haven't seen much of that yet here.
I suspect that the truth is that manufacturers would be willing to pay to train workers if they could be reasonably sure that the investment would have time to pay off. But thanks to Moore's Law and everything that comes with it, manufacturing technology is changing far more rapidly than it used to when everything was more stable.
It seems likely that it's more cost-effective for the manufacturer to lower that risk by training lower-paid labor elsewhere and letting those workers go (or retraining them, if it's cheap to do so) when they're no longer suitable, than it is to do the same thing in first-world countries for much higher cost.
I'm glad you caught that line because I thought it was confusing
Who are we supposed to understand is under-educated? The employers? Or was this a vain attempt to pull a fast one and shift blame to public schools for...? What? Pumping out kids with the mathematical literacy to understand that $14 is greater than $10 therefore McDonalds pays better than manufacturing?
I also thought that line was unfair. Sure, one can go and get a 4 year degree to run your machines, at their own personal cost of $100,000 to $150,000.
It's not just about "fair wage" at that point - enough to have a place to live, a car, food in your belly... but that "fair rate" also includes being paid enough to afford the $500-1,000/month student loan payment that is the outcome of getting such a degree.
Agreed ... if the applicants are so uneducated, why are they choosing "rationally" (as the article puts it) to avoid these low-paying, high-skilled jobs?
This is fairly speculative, but I have been thinking this paradox over for a while.
A high demand for scarce highly skilled workers must drive
salaries up in competition.
Unless the skills are not actually high skilled, but obscure skills. I predict that "high skilled manufacturing" is a morass of proprietary solutions to
highly specific process needs - that is, the metal fabrication plant alluded to at 10USD per hour has highly specific machinery doing a fixed task, and that the skill is mostly one of repairing the proprietary code.
From a software perspective, that mostly means the machinery is "legacy" - and impossible to refactor. So it can never be improved upon, only replaced.
I am not a 3d-printer fanboi - the advantages of additative manufacturing are hugely over hyped (at the moment), and likely to remain elusive for, lets say, a generation, before it becomes obvious to all we throw away factories and their jobs and build millions of thing-o-matics.
However in that generation there will be huge opportunity for semi-general manufacturing - robots with sufficient flexibility in parts and software that they can be re-purposed easily as part of a (virtual) conveyor belt.
This sounds truly skilled work - flexibly adapting as processes and customers change.
if providing semi-general manufacturing machines is uneconomic compared to proprietary simpler but "obscure skill" machines, then we cannot expect a productivity
premium for actual highly skilled workers and should expect
10USD phour jobs to limp along till Shanghai takes their lunch.
If however semi-general machines can be made to adapt to
different manufacturing requirements, then our whole manufacturing base may be in want of replacing.
Again something emerging economies will have an advantage in.
So, overall, the West should view itself as Great Britain was at the end of the 19th Century - a pioneer whose advantages had run out, and without wholesale massive investment will simply enter a managed decline.
Disruption opportunities - Development of semi-general robotic manufacturing solutions that can quickly be re-purposed. And proving that it is both ecomnomic for a greenfield site and an installed base.
WRT the skill of "repairing the proprietary code," there isn't any repair of code (in the literal sense). It is actually the skill of clicking the right buttons (often in a horribly designed GUI) to make a proprietary machine do what it needs to do.
Come to think about it, it isn't obscure skills, it is obscure knowledge. The problems are (a) there is really no skill involved, (b) the knowledge involved is specific to a proprietary machine and proprietary program, and (c) the training courses are training specific knowledge, not the principles and cross-training that would apply more universally.
Mapping to our (geek) world, training obscure knowledge vs. actual skills is all the "office productivity training" that goes on which really is "which buttons to click in Excel"... leaving the trainee helpless in LibreOffice because the button is in a different spot and possibly labeled differently. Note that the trainer is not entirely at fault - I've been amazed at the number of trainees that cannot generalize.
Both you and gvb seem to perhaps lack first-hand experience of machine shops.
It's not just "repairing code"... Oftentimes, small job shops will be taking a dimensioned drawing and write their own G code to actually generate the objects.
This will probably involve some back-and-forth with the machine stepping through the part manufacture, and must be done carefully to guarantee that both the machining is done efficiently by the machine and that the program doesn't result in a catastrophic failure damaging the workpiece, the tool, the machine, the operator, or some combination of the above.
This is, in fact, skilled work. Moreover, experienced operators know how to check parts for proper dimensioning (QC at the bench), perform maintenance on their machines and change tools, and watch a part in progress to make sure something really dumb isn't about to happen.
Moreover, the job doesn't stop at the machining center--you may still have to go and clean things up manually, do welding, do polishing, or any number of finishing touches.
Keeping this all working together is not something to laugh at.
I haven't even touched on the logistical challenges of retooling or upgrading equipment, or operating with machines that (unironically) still accept paper tape.
Part of Isbister’s pickiness, he says, comes from an avoidance of workers with experience in a “union-type job.” Isbister, after all, doesn’t abide by strict work rules and $30-an-hour salaries. At GenMet, the starting pay is $10 an hour. Those with an associate degree can make $15, which can rise to $18 an hour after several years of good performance.
I'm not a fan of unions or work rules, but those are terrible wages. How does the guy expect to get 21st century skills for 20th century wages?
Fortunately the market tends to weed guys like him out if he refuses (or is unable) to raise salaries to meet his required quotas, unless he's just clever enough to get government handouts that keep his company afloat indefinitely.
Hire 5 guys in China? He literally can't pay more or the cost of doing business overseas wins. The golden arches has to pay more because the night manager can't live on a different continent.
If an Entrepreneur can't pay for the resources needed to build/ensure a product/service at their market prices, he should and eventually will, close business.
There is nothing wrong with that. That is the economical system in the works.
If no one is willing to pay for his product/service at the price he needs to charge, it is because either people do not need it, or they can get it cheaper somewhere else. In business, if you can't compete, you shouldn't.
In this specific case, these jobs are getting oursourced to china or replaced by machines. This is wonderful. Not for infimal number of people that lose their jobs, but to the huge number of people who will be able to afford that product/service cheaper.
I can't set up an ice-cream kiosk in antartica and whine about how business is hard and what will be of the antartica ice-cream industry.
What does that mean for us though, the "skilled workers" when everything is being shipped overseas? A friend of mine who works as a systems engineer for Verizon said their dumping 30% of their US staff and shipping it to overseas. This seems to be the pattern (anecdotally) as of late for a lot of large companies. When there's someone with comparable skills on the other side of the world who will do your job for a fraction of the cost, what's to stop a race to the bottom?
Okay, so if there is a problem, what's the solution? Sure enough, highly skilled labor in Taiwan, for instance, that is willing to do the same work for a fraction of the cost, isn't going to stop working just because US employees don't find $10/h attractive. It's a free market and if we want to make things in US, we have to keep the costs comparable.
Now, assuming blue collar is done, as robots can take care of most of the tasks, could someone explain why we need workers to oversee the computers?
The way I see it, you need a really expensive robot, and you need a really expensive programmer to make it work; everything in the middle is cheap. Software should be able to distill anything that is happening on the robotic side and present it in such a way that a trained mechanic would understand and be able to fix it. If that's not the case, then we need better developers (and in this case, perhaps, better designers and human-factors engineers) to write better software.
Now, a few folks mentioned that what you need is really a mixture of experience in the mechanical side of things and understanding of the software - presumable that allows you to react and solve problems whenever something goes astray. Well, that could probably be solved by having a few engineers on staff who would help when needed.
That leaves us with yet again, fairly simple mechanical labor. Perhaps then, the article is right, we have a serious education problem. Those who attain enough education, leap forward and presumable learn more to subject themselves to mundane mechanical tasks, while those who would be greatly fitted to do the jobs, actually don't have enough education to understand even the most basics.
Someone's suggested comparing the math needed at Mc'Donalds with the math needed at one of these factories. I'd be curious to know too; although I suspect that in McDs all the calculations are done by a computer and all humans need to do, is simply not to f-up. Even then, when humans fail to add 2+2, all they lose is an occasional McFlurry, while at a factory they could impact tens of thousands of dollars at once.
So, here we have a conundrum. We need labor to work the $10 jobs, but the pool of employees is simply atrocious. At the same time, qualified labor has better things to do with their time. Now, back to my questions - what's the solution?
I cannot fathom how the author takes the side of the employers at the end of this article. So workers should reduce their expectations so that certain manufacturers can succeed while paying them below-expected rates? Does the author legitimately expect individual workers to look out for the country's employment best interests above their own?
The candidates smart enough to realize that they could earn 2x-5x as much after college won't take these jobs, and most of the people who can't earn those wages elsewhere aren't qualified to do these jobs.
From the article: "The secret behind this skills gap is that it’s not a skills gap at all. I spoke to several other factory managers who also confessed that they had a hard time recruiting in-demand workers for $10-an-hour jobs. 'It’s hard not to break out laughing,' says Mark Price, a labor economist at the Keystone Research Center, referring to manufacturers complaining about the shortage of skilled workers. 'If there’s a skill shortage, there has to be rises in wages,' he says. 'It’s basic economics.'"
Agreed. That is the basic problem. If a worker can produce hundreds of widgets a day after specialized training with new computer-controlled machinery, but the worker could make just as much money per hour right after high school flipping hamburgers at the local fast-food restaurant, there is no reason for the worker to go through two years or more of specialized training, especially at the worker's own expense.
Much of the rest of the article discusses the overall rationality of workers seeking jobs that they can obtain with the least investment of their own time in training for a given income. Of course. Part of the problem is that if companies hire on the basis of course completion certificates rather than on the basis of demonstrated competence, they will miss out on good workers, and yet hire some lousy workers, and thus be reluctant to offer competitive starting wages. (It's expensive to hire a worker who can't do the job, and it's also expensive to let go workers who don't learn on the job and to hire their replacements.)
In what I think has become my best-liked comment on HN, I've collected references other participants here helped me find about company hiring procedures. Companies need to hire on the basis of actually being able to do the job, not on the basis of what classes workers have attended. The review article by Frank L. Schmidt and John E. Hunter, "The Validity and Utility of Selection Models in Personnel Psychology: Practical and Theoretical Implications of 85 Years of Research Findings," Psychological Bulletin, Vol. 124, No. 2, 262-274
sums up, current to 1998, a meta-analysis of much of the HUGE peer-reviewed professional literature on the industrial and organizational psychology devoted to business hiring procedures. There are many kinds of hiring criteria, such as in-person interviews, telephone interviews, resume reviews for job experience, checks for academic credentials, personality tests, and so on. There is much published study research on how job applicants perform after they are hired in a wide variety of occupations.
EXECUTIVE SUMMARY: If you are hiring for any kind of job in the United States, prefer a work-sample test as your hiring procedure. If you are hiring in most other parts of the world, use a work-sample test in combination with a general mental ability test.
The overall summary of the industrial psychology research in reliable secondary sources is that two kinds of job screening procedures work reasonably well. One is a general mental ability (GMA) test (an IQ-like test, such as the Wonderlic personnel screening test). Another is a work-sample test, where the applicant does an actual task or group of tasks like what the applicant will do on the job if hired. (But the calculated validity of each of the two best kinds of procedures, standing alone, is only 0.54 for work sample tests and 0.51 for general mental ability tests.) Each of these kinds of tests has about the same validity in screening applicants for jobs, with the general mental ability test better predicting success for applicants who will be trained into a new job. Neither is perfect (both miss some good performers on the job, and select some bad performers on the job), but both are better than any other single-factor hiring procedure that has been tested in rigorous research, across a wide variety of occupations. So if you are hiring for your company, it's a good idea to think about how to build a work-sample test into all of your hiring processes. If the job you are hiring for involves use of a computer-controlled machine tool, have the candidate put the machine to use making sample parts (advertise the job in a way that makes clear a work-sample test is required, to screen out people who have no clue how to operate such machines). Hire the able, and pay them what they are worth.
Ask yourself about any hiring process you have ever been in, as boss or as applicant: did the applicant have to do a work-sample test based on actual work results expected in the company? Why not?
I agree with most of this but perhaps not the conclusion.
The key point being that there is a skill 'valuation' gap which is being forced on the manufacturers by competition in distant countries. So piecework manufacturer in Minnesota has to compete with piecework manufacturer in Shenzen, and the former pays employees more and uses more automation and the latter has traditionally been able to throw many more humans at the task. Between the slowly rising demands in these countries, and the challenge of working with your factory that far away, we're just about to the point where its going to flip. I would have called it for this year but the great recession seemed to push everything back 2 - 3 years. So perhaps in 2015. Small manufacturers will either go out of business or require a higher price per good delivered, which supports effectively doubling the wage paid to their line workers. And those line workers will be essentially programmers programming tools, not doing the actual tooling.
The switch will be enabled by "high resolution" workstation automation. Today the last unassailable benefit of the 'tool and die' shop is that you've got a machinist and a machine tool which can be dynamically scheduled for any kind of widget. They are 'soft tooling' in that all of the gear that makes stuff doesn't need a dedicated fabrication line. The benefit is realized because setting up 'hard tooling' has a huge upfront cost and while the dedicated line is much more efficient, it also takes space away from other lines.
The calculus goes something like this : 0 - 1,000 widgets a 'precision' shop can make them; 1000+ to 50000 widgets a 'sweat shop' (primarily labor driven shop) can make them, and for 5000+ widgets you start to get the benefits of hard tooling amortized over your widgets.
As we get to the point where we have machines that have 'high level' sorts of capabilities (some early machines have been demonstrated by the Japanese and others) where your machine can take a dumped out pile of parts and assemble the widget, or 'partial hard tooling' where a machine creates intermediate tools to maximize efficiency "automatically" then it flips back to the place where a well equipped precision shop can make 10,000 widgets a month or arbitrary design and amortize over all 10,000 which switches the economics question back to transportation and re-spins (interaction with the original designers).
When I see robots that can pick up a cell phone, toss it and then catch it again at an arbitrary angle, you realize you can do material transfer by having workstations literally toss partial assemblies from one station to the next. That kind of flexibility is going to make for a very different kind of factory, and of course factory worker.
>If you are hiring for any kind of job in the United States, prefer a work-sample test as your hiring procedure. If you are hiring in most other parts of the world, use a work-sample test in combination with a general mental ability test.
Why would you use a mental ability test outside of the US, but not in the US? Is it illegal in the US or something?
The name of the article is misleading. It is all about the greedy fat cats that want mice to go to the mousetrap, but moaning about high prices for cheese, so they are unable to charge the mousetrap properly. Good for mice, bad for cats.
I wonder what kind of skills are lacking.
Claims of deficiencies are qualitative.
For example, how much math does somebody need to know for this kind of job? Is it trigonometry to find hypotenuses? Are they writing programs or entering in data to notepad?
This is an important distinction because we need to compare these skills to those of a shift manager at McDonald.
For example, at a hotel the shift manager must supervise the employees, handle money, and keep journal entries. The later two involve rudimentary mathematics and computer skills. There is a good chance that this is a fair wage when compared to similar jobs.
From my experience of college level instrumentation design, if one want to design parts and run CNC to its full potential, he/she should be decently proficient in curve drawing, understanding gear design, etc, and the math required for it can be quite involved. This is after you know the broad design detail. Some difficult things involve how to best approximate a curve on given CNC constraints, figuring out best fit for parts, etc; and then write appropriate programs for it. IMO, there are only a very few people who would have skill set like these, but not enough to move on to higher level engineering jobs.
Although, a lot of this thing is getting automated, and there is a huge push towards figuring out how you can feed a CAD design directly to CNC and let it figure out best way to implement it.
PS: I cannot expect that you can ever hire enough good people in this category for $10 an hour. They would probably be writing VB scripts and earning more money. Working in a fast food chain would be totally a waste of potential of decently skilled workers.
This industry is more experience based. Sure, theory will teach you that under certain conditions a plasma cutter ( or welder)'s arc is going to be certain shape and the speed need to be certain inch/minute for minimal dross, however, it is never 100% accurate. Therefore, the actual skills lies in the know how to adjust.
Moving on, with CNC automation, you need experience in both industry to understand how to program a CNC machine. With full automation becoming the trend, I see a huge lack of skilled workers who can fit this wold. Afterall, why would you go into manufacturing of metals if you understand computers.
...highly skilled manufacturing jobs — the ones that require people who know how to run the computer that runs the machine. --
Thinking meta for a moment here: How far does this concept go of machine process. Could we program machines that can run the computers that run the machines? And how about machines that program those machines?
No we can't. I base that on being a software architect who sat through endless useless marketing meetings from IBM & their ilk promising the imminent advent of Doodleware (have a business person with domain knowledge & no technical skills draw a process on this monkey-easy UI tool & we'll generate a working application from it.)
I maintain that being a programmer will still be a highly paid skill in demand for decades to come for one simple reason : You can't bullshit a computer. You can't sweet-talk a machine into accepting your algorithm because of your great personality , looks or any other intangible except for knowing what you're doing.
if you want a parallel, look no further than the sorry state of machine translation & that despite the enormous amounts of datasets they've been playing with. Human intelligence & judgement, even on a very very narrowed use case is not something computers are close to replicate.
TL;DR: Probably not on legacy industrial equipment. Or legacy people. But maybe if you start from scratch and only involve computers.
I was given a good piece of advice when I started work: "A process engineer's job is to make themselves obsolete." I've striven to make sure that I did everything I could to make sure I wasn't needed anymore.
And I'm happy to say that the processes will keep me employed for as long as I'm willing. Not from any self-serving malice, but rather that in the end it's terribly messy. The sheer inelegance of industrial manufacturing seems to defy strong automation at the meta-level. And not just because the machines are all only halfway obsolete, new, deprecated, upgraded, documented, up-kept, cleaned, and run (halfway to spec).
It's because to a computer, this mess quickly degrades into special cases. Lots of them. There actually are some shockingly clever software packages out there capable of the early stages of what you're describing, and - this is the kicker - they all require specialized engineering support to create, install, train, and use. The end product can be handled by a floor worker or management, but at some point you've got to explain the situation to a computer somehow, and that ends up being as complicated as programming.
For a much better written punchline, I'll defer to the short story Profession by Asimov [1]. I think it's cynical to think we'll never be able to solve the 'original thought' problem, but I also think it's easy to underestimate it. Things really are different today than 30 years ago, so perhaps a few more iterations will lead to tape-machines that can program the tape-machine programmer tape-machines. [2]
[2] Perhaps the worst part is we sorta already have this. There was just a post on the LLVM [3], which discussed how a computer translated and optimized a division command into machine code. That used to take a lot of work, and it took Real Programmers to optimize that code before this sort of automation. So it's not inconceivable that your ideal comes true. But I am cynical enough to think sami36 is right [4]: Doodleware is past the horizon. But computers that can talk to computers is probably in sight.
> Could we program machines that can run the computers that run the machines? And how about machines that program those machines?
Given the history of progressive rounds of optimism and failure on this idea, so far: no. Before you can replace humans, you need to ... um ... replace humans.
However, the general principle of progressively more specialised machinery and skills is not new. It's called "the division of labour"; it really ought to be called "the specialisation of process" or perhaps "the specialisation of outputs", given that capital too becomes progressively more specialised to particular tasks.
People have run software that manage mechanical turk campaigns that manage people performing tasks such as filling out online forms to create email accounts and such....
When applying for programmers I'll refuse to hand in a résumé without being able to demonstrate my skills and provide sample work first.
The hiring process that most companies use is backwards, and I think mostly due to people that go to school to learn HR and taught this impractical approach.
Companies are simply offering wages that are comparable to some worker in China would get. It only makes sense in a flat world. Question is: do we want that?
In the IT industry, specially IT support and break-fix development the same thing happens. Local people don't want to accept low salaries? Let's look what India or the Philippines are charging. Problem solved.
I read about this so often, that I can't think anymore that the problem really is the low pay for high skilled jobs. There must be a reason why in so many countries for so many years nobody increases their hourly rates and hopes to get well educated people for the current rates.
[+] [-] zmmmmm|13 years ago|reply
[+] [-] mseebach|13 years ago|reply
[+] [-] enko|13 years ago|reply
The only thing stopping me from calling out HR departments in general as a sign of organisational cancer, since I am yet to encounter one that pulls even a tenth of its weight, is that they tend to be comprised of the only cute girls in the whole company. Even on Friday (2 days ago) I was out drinking with a bunch of tech guys and a single cute HR girl. That is pretty much the sole reason I would oppose eliminating HR as we know it from all technology companies.
[+] [-] achompas|13 years ago|reply
"Data science" means many things to many people. Some shops want BI folks who can code. Others want people who can keep up in deep learning threads (like the one on HN's front page) and know about Hinton's work. Others still want HDFS badasses who can distinguish between Cascading and Cascalog (lol).
Shops classify all of these things as "data science" when they're really very different (respectively: Excel/scripting BI analyst; ML researcher with a focus on deep learning and neural networks; data infrastructure developer with broad exposure to MR frameworks)
This is stupid. I worried about checking off these boxes until I realized recruiters are doing the 1990s Java thing all over again. This time, though, the buzzwords are Hadoop, information retrieval, and Andrew Ng. I think it's best to ignore this, keep working hard, and grow your employment options.
[+] [-] ivany|13 years ago|reply
[+] [-] logn|13 years ago|reply
[+] [-] nostrademons|13 years ago|reply
[+] [-] chrismealy|13 years ago|reply
News stories have been filled with reports of managers of manufacturing companies insisting that they have jobs open that they can't fill because there are no qualified workers. Adam Davidson at the NYT looked at this more closely and found that the real problem is that the managers don't seem to be interested in paying for the high level of skills that they claim they need.
Many of the positions that are going unfilled pay in the range of $15-$20 an hour. This is not a pay level that would be associated with a job that requires a high degree of skill. As Davidson points out, low level managers at a fast-food restaurant can make comparable pay.
It should not be surprising that the workers who have these skills expect higher pay and workers without the skills will not invest the time and money to acquire them for such a small reward. If these factories want to get highly skilled workers, they will have to offer a wage that is in line with the skill level that they expect.
http://www.cepr.net/index.php/blogs/beat-the-press/that-shor...
[+] [-] clarkm|13 years ago|reply
Contrast this to the skilled manufacturing jobs which require up front experience. Though many blue collar fields offer entry-level positions with on the job training, apprenticeships, and opportunity for advancement, this doesn't appear to be common practice in manufacturing. Why not? I think the main reason is that it is very hard for a low skill worker to add value to a manufacturing company. There aren't any comparable entry-level positions that allow the employee to learn while still being productive.
Because of this, hiring an unskilled employee for the purpose of training them is a huge risk, since it requires a significant investment. And since this industry is already very unstable with razor-thin margins, it's not something many employers seem willing to do, which is unfortunate.
So maybe the solution is coming up with better training programs, so that manufacturers can hire new employees without taking on such large risks?
[+] [-] icelancer|13 years ago|reply
[+] [-] opendna|13 years ago|reply
Ya'll have probably read some of Prof Peter Cappelli's editorials (he's been making the rounds to promote his book "Why Good People Can’t Get Jobs"). One of the things he's pointed out is that HR use of resume databases encourages people with poor search habits to believe there is no choice. Basically, it's a failure consider the Bayesian math of nested filters.
Consider a company, in a city of a million people (don't want to pay for relocation), which wants someone with a bachelor's degree and five years of experience. The US unemployment rate for people with bachelor's degrees or higher is ~3.8% and about ~7% of the population is between 25-29 years old. Rough and tumble numbers put the pool of candidates around 2500-2700 before including any subject-specific knowledge. Ask for a specific discipline, like "Computers, mathematics, and statistics" for which 4.2% of bachelor degrees were awarded, and you can cut your your pool of candidates down to about 100 (1/10,000). If you also filter to require mastery of PHP, javascript, and Marqui, 10 years experience with Server 2008 and fluency in a Romance language...
tl;dr: the absence of time travelers in the applicant pool is not evidence of a skills shortage.
[+] [-] molsongolden|13 years ago|reply
That closing line doesn't really sound like the gist of the piece at all. It sounds like manufacturing employers just don't want to or don't feel they can afford to pay works a fair wage.
[+] [-] evv|13 years ago|reply
If laborers are leaving manufacturers for better pay at fast food companies, workers need higher pay. Its not that complicated. If the company can't afford higher wages, maybe the demand for this skilled labor isn't what we think it is.
If these skills are actually in high demand, they will pay the bills.
[+] [-] achompas|13 years ago|reply
I'd say our education system is working out fine if we're skipping the high-barrier jobs that don't pay a premium wage.
[+] [-] davidp|13 years ago|reply
Several commenters have basically claimed that the manufacturers are behaving irrationally; a claim like that requires some powerful arguments to credibly support. I haven't seen much of that yet here.
I suspect that the truth is that manufacturers would be willing to pay to train workers if they could be reasonably sure that the investment would have time to pay off. But thanks to Moore's Law and everything that comes with it, manufacturing technology is changing far more rapidly than it used to when everything was more stable.
It seems likely that it's more cost-effective for the manufacturer to lower that risk by training lower-paid labor elsewhere and letting those workers go (or retraining them, if it's cheap to do so) when they're no longer suitable, than it is to do the same thing in first-world countries for much higher cost.
[+] [-] opendna|13 years ago|reply
Who are we supposed to understand is under-educated? The employers? Or was this a vain attempt to pull a fast one and shift blame to public schools for...? What? Pumping out kids with the mathematical literacy to understand that $14 is greater than $10 therefore McDonalds pays better than manufacturing?
[+] [-] xyzzy123|13 years ago|reply
I found a long-running thread where they discussed average wages for machinists / cnc operators, and it actually made me sad. It's here:
http://www.cnczone.com/forums/polls/42247-what_average_hourl...
When I compare this to what even mediocre programmers get, I can't help but feel that there's real blue-collar discrimination going on.
One thing though - there's a real variance in the responsibilities (and a abilities) of a "CNC operator"...
[+] [-] rpwilcox|13 years ago|reply
It's not just about "fair wage" at that point - enough to have a place to live, a car, food in your belly... but that "fair rate" also includes being paid enough to afford the $500-1,000/month student loan payment that is the outcome of getting such a degree.
[+] [-] smoyer|13 years ago|reply
[+] [-] lifeisstillgood|13 years ago|reply
A high demand for scarce highly skilled workers must drive salaries up in competition.
Unless the skills are not actually high skilled, but obscure skills. I predict that "high skilled manufacturing" is a morass of proprietary solutions to highly specific process needs - that is, the metal fabrication plant alluded to at 10USD per hour has highly specific machinery doing a fixed task, and that the skill is mostly one of repairing the proprietary code.
From a software perspective, that mostly means the machinery is "legacy" - and impossible to refactor. So it can never be improved upon, only replaced.
I am not a 3d-printer fanboi - the advantages of additative manufacturing are hugely over hyped (at the moment), and likely to remain elusive for, lets say, a generation, before it becomes obvious to all we throw away factories and their jobs and build millions of thing-o-matics.
However in that generation there will be huge opportunity for semi-general manufacturing - robots with sufficient flexibility in parts and software that they can be re-purposed easily as part of a (virtual) conveyor belt.
This sounds truly skilled work - flexibly adapting as processes and customers change.
if providing semi-general manufacturing machines is uneconomic compared to proprietary simpler but "obscure skill" machines, then we cannot expect a productivity premium for actual highly skilled workers and should expect 10USD phour jobs to limp along till Shanghai takes their lunch.
If however semi-general machines can be made to adapt to different manufacturing requirements, then our whole manufacturing base may be in want of replacing.
Again something emerging economies will have an advantage in.
So, overall, the West should view itself as Great Britain was at the end of the 19th Century - a pioneer whose advantages had run out, and without wholesale massive investment will simply enter a managed decline.
Disruption opportunities - Development of semi-general robotic manufacturing solutions that can quickly be re-purposed. And proving that it is both ecomnomic for a greenfield site and an installed base.
[+] [-] gvb|13 years ago|reply
Come to think about it, it isn't obscure skills, it is obscure knowledge. The problems are (a) there is really no skill involved, (b) the knowledge involved is specific to a proprietary machine and proprietary program, and (c) the training courses are training specific knowledge, not the principles and cross-training that would apply more universally.
Mapping to our (geek) world, training obscure knowledge vs. actual skills is all the "office productivity training" that goes on which really is "which buttons to click in Excel"... leaving the trainee helpless in LibreOffice because the button is in a different spot and possibly labeled differently. Note that the trainer is not entirely at fault - I've been amazed at the number of trainees that cannot generalize.
[+] [-] angersock|13 years ago|reply
It's not just "repairing code"... Oftentimes, small job shops will be taking a dimensioned drawing and write their own G code to actually generate the objects.
This will probably involve some back-and-forth with the machine stepping through the part manufacture, and must be done carefully to guarantee that both the machining is done efficiently by the machine and that the program doesn't result in a catastrophic failure damaging the workpiece, the tool, the machine, the operator, or some combination of the above.
This is, in fact, skilled work. Moreover, experienced operators know how to check parts for proper dimensioning (QC at the bench), perform maintenance on their machines and change tools, and watch a part in progress to make sure something really dumb isn't about to happen.
Moreover, the job doesn't stop at the machining center--you may still have to go and clean things up manually, do welding, do polishing, or any number of finishing touches.
Keeping this all working together is not something to laugh at.
I haven't even touched on the logistical challenges of retooling or upgrading equipment, or operating with machines that (unironically) still accept paper tape.
[+] [-] anigbrowl|13 years ago|reply
I'm not a fan of unions or work rules, but those are terrible wages. How does the guy expect to get 21st century skills for 20th century wages?
[+] [-] Jach|13 years ago|reply
[+] [-] stonemetal|13 years ago|reply
[+] [-] Soarez|13 years ago|reply
There is nothing wrong with that. That is the economical system in the works.
If no one is willing to pay for his product/service at the price he needs to charge, it is because either people do not need it, or they can get it cheaper somewhere else. In business, if you can't compete, you shouldn't.
In this specific case, these jobs are getting oursourced to china or replaced by machines. This is wonderful. Not for infimal number of people that lose their jobs, but to the huge number of people who will be able to afford that product/service cheaper.
I can't set up an ice-cream kiosk in antartica and whine about how business is hard and what will be of the antartica ice-cream industry.
[+] [-] goostavos|13 years ago|reply
[+] [-] kirillzubovsky|13 years ago|reply
Now, assuming blue collar is done, as robots can take care of most of the tasks, could someone explain why we need workers to oversee the computers?
The way I see it, you need a really expensive robot, and you need a really expensive programmer to make it work; everything in the middle is cheap. Software should be able to distill anything that is happening on the robotic side and present it in such a way that a trained mechanic would understand and be able to fix it. If that's not the case, then we need better developers (and in this case, perhaps, better designers and human-factors engineers) to write better software.
Now, a few folks mentioned that what you need is really a mixture of experience in the mechanical side of things and understanding of the software - presumable that allows you to react and solve problems whenever something goes astray. Well, that could probably be solved by having a few engineers on staff who would help when needed.
That leaves us with yet again, fairly simple mechanical labor. Perhaps then, the article is right, we have a serious education problem. Those who attain enough education, leap forward and presumable learn more to subject themselves to mundane mechanical tasks, while those who would be greatly fitted to do the jobs, actually don't have enough education to understand even the most basics.
Someone's suggested comparing the math needed at Mc'Donalds with the math needed at one of these factories. I'd be curious to know too; although I suspect that in McDs all the calculations are done by a computer and all humans need to do, is simply not to f-up. Even then, when humans fail to add 2+2, all they lose is an occasional McFlurry, while at a factory they could impact tens of thousands of dollars at once.
So, here we have a conundrum. We need labor to work the $10 jobs, but the pool of employees is simply atrocious. At the same time, qualified labor has better things to do with their time. Now, back to my questions - what's the solution?
[+] [-] tomjen3|13 years ago|reply
The solution is to adapt or close the factories down, and go work at McDonalds (or get a better skilled job elsewhere).
[+] [-] kosei|13 years ago|reply
The candidates smart enough to realize that they could earn 2x-5x as much after college won't take these jobs, and most of the people who can't earn those wages elsewhere aren't qualified to do these jobs.
[+] [-] tokenadult|13 years ago|reply
Agreed. That is the basic problem. If a worker can produce hundreds of widgets a day after specialized training with new computer-controlled machinery, but the worker could make just as much money per hour right after high school flipping hamburgers at the local fast-food restaurant, there is no reason for the worker to go through two years or more of specialized training, especially at the worker's own expense.
Much of the rest of the article discusses the overall rationality of workers seeking jobs that they can obtain with the least investment of their own time in training for a given income. Of course. Part of the problem is that if companies hire on the basis of course completion certificates rather than on the basis of demonstrated competence, they will miss out on good workers, and yet hire some lousy workers, and thus be reluctant to offer competitive starting wages. (It's expensive to hire a worker who can't do the job, and it's also expensive to let go workers who don't learn on the job and to hire their replacements.)
In what I think has become my best-liked comment on HN, I've collected references other participants here helped me find about company hiring procedures. Companies need to hire on the basis of actually being able to do the job, not on the basis of what classes workers have attended. The review article by Frank L. Schmidt and John E. Hunter, "The Validity and Utility of Selection Models in Personnel Psychology: Practical and Theoretical Implications of 85 Years of Research Findings," Psychological Bulletin, Vol. 124, No. 2, 262-274
http://mavweb.mnsu.edu/howard/Schmidt%20and%20Hunter%201998%...
sums up, current to 1998, a meta-analysis of much of the HUGE peer-reviewed professional literature on the industrial and organizational psychology devoted to business hiring procedures. There are many kinds of hiring criteria, such as in-person interviews, telephone interviews, resume reviews for job experience, checks for academic credentials, personality tests, and so on. There is much published study research on how job applicants perform after they are hired in a wide variety of occupations.
http://www.siop.org/workplace/employment%20testing/testtypes...
EXECUTIVE SUMMARY: If you are hiring for any kind of job in the United States, prefer a work-sample test as your hiring procedure. If you are hiring in most other parts of the world, use a work-sample test in combination with a general mental ability test.
The overall summary of the industrial psychology research in reliable secondary sources is that two kinds of job screening procedures work reasonably well. One is a general mental ability (GMA) test (an IQ-like test, such as the Wonderlic personnel screening test). Another is a work-sample test, where the applicant does an actual task or group of tasks like what the applicant will do on the job if hired. (But the calculated validity of each of the two best kinds of procedures, standing alone, is only 0.54 for work sample tests and 0.51 for general mental ability tests.) Each of these kinds of tests has about the same validity in screening applicants for jobs, with the general mental ability test better predicting success for applicants who will be trained into a new job. Neither is perfect (both miss some good performers on the job, and select some bad performers on the job), but both are better than any other single-factor hiring procedure that has been tested in rigorous research, across a wide variety of occupations. So if you are hiring for your company, it's a good idea to think about how to build a work-sample test into all of your hiring processes. If the job you are hiring for involves use of a computer-controlled machine tool, have the candidate put the machine to use making sample parts (advertise the job in a way that makes clear a work-sample test is required, to screen out people who have no clue how to operate such machines). Hire the able, and pay them what they are worth.
Ask yourself about any hiring process you have ever been in, as boss or as applicant: did the applicant have to do a work-sample test based on actual work results expected in the company? Why not?
[+] [-] ChuckMcM|13 years ago|reply
The key point being that there is a skill 'valuation' gap which is being forced on the manufacturers by competition in distant countries. So piecework manufacturer in Minnesota has to compete with piecework manufacturer in Shenzen, and the former pays employees more and uses more automation and the latter has traditionally been able to throw many more humans at the task. Between the slowly rising demands in these countries, and the challenge of working with your factory that far away, we're just about to the point where its going to flip. I would have called it for this year but the great recession seemed to push everything back 2 - 3 years. So perhaps in 2015. Small manufacturers will either go out of business or require a higher price per good delivered, which supports effectively doubling the wage paid to their line workers. And those line workers will be essentially programmers programming tools, not doing the actual tooling.
The switch will be enabled by "high resolution" workstation automation. Today the last unassailable benefit of the 'tool and die' shop is that you've got a machinist and a machine tool which can be dynamically scheduled for any kind of widget. They are 'soft tooling' in that all of the gear that makes stuff doesn't need a dedicated fabrication line. The benefit is realized because setting up 'hard tooling' has a huge upfront cost and while the dedicated line is much more efficient, it also takes space away from other lines.
The calculus goes something like this : 0 - 1,000 widgets a 'precision' shop can make them; 1000+ to 50000 widgets a 'sweat shop' (primarily labor driven shop) can make them, and for 5000+ widgets you start to get the benefits of hard tooling amortized over your widgets.
As we get to the point where we have machines that have 'high level' sorts of capabilities (some early machines have been demonstrated by the Japanese and others) where your machine can take a dumped out pile of parts and assemble the widget, or 'partial hard tooling' where a machine creates intermediate tools to maximize efficiency "automatically" then it flips back to the place where a well equipped precision shop can make 10,000 widgets a month or arbitrary design and amortize over all 10,000 which switches the economics question back to transportation and re-spins (interaction with the original designers).
When I see robots that can pick up a cell phone, toss it and then catch it again at an arbitrary angle, you realize you can do material transfer by having workstations literally toss partial assemblies from one station to the next. That kind of flexibility is going to make for a very different kind of factory, and of course factory worker.
[+] [-] rsheridan6|13 years ago|reply
Why would you use a mental ability test outside of the US, but not in the US? Is it illegal in the US or something?
[+] [-] keryor|13 years ago|reply
[+] [-] jaggederest|13 years ago|reply
[+] [-] ommunist|13 years ago|reply
[+] [-] frozenport|13 years ago|reply
For example, how much math does somebody need to know for this kind of job? Is it trigonometry to find hypotenuses? Are they writing programs or entering in data to notepad?
This is an important distinction because we need to compare these skills to those of a shift manager at McDonald.
For example, at a hotel the shift manager must supervise the employees, handle money, and keep journal entries. The later two involve rudimentary mathematics and computer skills. There is a good chance that this is a fair wage when compared to similar jobs.
[+] [-] galadriel|13 years ago|reply
Although, a lot of this thing is getting automated, and there is a huge push towards figuring out how you can feed a CAD design directly to CNC and let it figure out best way to implement it.
PS: I cannot expect that you can ever hire enough good people in this category for $10 an hour. They would probably be writing VB scripts and earning more money. Working in a fast food chain would be totally a waste of potential of decently skilled workers.
[+] [-] Causalien|13 years ago|reply
Moving on, with CNC automation, you need experience in both industry to understand how to program a CNC machine. With full automation becoming the trend, I see a huge lack of skilled workers who can fit this wold. Afterall, why would you go into manufacturing of metals if you understand computers.
[+] [-] sharkweek|13 years ago|reply
Thinking meta for a moment here: How far does this concept go of machine process. Could we program machines that can run the computers that run the machines? And how about machines that program those machines?
One of my favorite Twilight Zone episodes : http://en.wikipedia.org/wiki/The_Brain_Center_at_Whipple%27s
[+] [-] sami36|13 years ago|reply
I maintain that being a programmer will still be a highly paid skill in demand for decades to come for one simple reason : You can't bullshit a computer. You can't sweet-talk a machine into accepting your algorithm because of your great personality , looks or any other intangible except for knowing what you're doing. if you want a parallel, look no further than the sorry state of machine translation & that despite the enormous amounts of datasets they've been playing with. Human intelligence & judgement, even on a very very narrowed use case is not something computers are close to replicate.
[+] [-] HCIdivision17|13 years ago|reply
I was given a good piece of advice when I started work: "A process engineer's job is to make themselves obsolete." I've striven to make sure that I did everything I could to make sure I wasn't needed anymore.
And I'm happy to say that the processes will keep me employed for as long as I'm willing. Not from any self-serving malice, but rather that in the end it's terribly messy. The sheer inelegance of industrial manufacturing seems to defy strong automation at the meta-level. And not just because the machines are all only halfway obsolete, new, deprecated, upgraded, documented, up-kept, cleaned, and run (halfway to spec).
It's because to a computer, this mess quickly degrades into special cases. Lots of them. There actually are some shockingly clever software packages out there capable of the early stages of what you're describing, and - this is the kicker - they all require specialized engineering support to create, install, train, and use. The end product can be handled by a floor worker or management, but at some point you've got to explain the situation to a computer somehow, and that ends up being as complicated as programming.
For a much better written punchline, I'll defer to the short story Profession by Asimov [1]. I think it's cynical to think we'll never be able to solve the 'original thought' problem, but I also think it's easy to underestimate it. Things really are different today than 30 years ago, so perhaps a few more iterations will lead to tape-machines that can program the tape-machine programmer tape-machines. [2]
[1] http://www.abelard.org/asimov.php
[2] Perhaps the worst part is we sorta already have this. There was just a post on the LLVM [3], which discussed how a computer translated and optimized a division command into machine code. That used to take a lot of work, and it took Real Programmers to optimize that code before this sort of automation. So it's not inconceivable that your ideal comes true. But I am cynical enough to think sami36 is right [4]: Doodleware is past the horizon. But computers that can talk to computers is probably in sight.
[3] http://eli.thegreenplace.net/2012/11/24/life-of-an-instructi...
[4] http://news.ycombinator.com/item?id=4827459
[+] [-] jacques_chester|13 years ago|reply
Given the history of progressive rounds of optimism and failure on this idea, so far: no. Before you can replace humans, you need to ... um ... replace humans.
However, the general principle of progressively more specialised machinery and skills is not new. It's called "the division of labour"; it really ought to be called "the specialisation of process" or perhaps "the specialisation of outputs", given that capital too becomes progressively more specialised to particular tasks.
[+] [-] Evbn|13 years ago|reply
[+] [-] andrew_wc_brown|13 years ago|reply
The hiring process that most companies use is backwards, and I think mostly due to people that go to school to learn HR and taught this impractical approach.
[+] [-] gtirloni|13 years ago|reply
In the IT industry, specially IT support and break-fix development the same thing happens. Local people don't want to accept low salaries? Let's look what India or the Philippines are charging. Problem solved.
[+] [-] erikb|13 years ago|reply