The Three Laws of Future Employment

Writing at New Geography Daniel Jelski offers a critique of (some of) Launching the Innovation Renaissance. We are in basic agreement about the laws of future employment:

Law #1: People will get jobs doing things that computers can’t do. Law #2: A global market place will result in lower pay and fewer opportunities for many careers. (But also in cheaper and better products and a higher standard of living for American consumers.) Law #3: Professional people will more likely be freelancers and less likely to have a steady job.

[But]…Laws #1 & 2 predict that there will likely be fewer STEM jobs in the future – they are both easily computerized and tradable. People will always be employed in STEM disciplines, many of them highly paid, but they’ll be paid for smarts rather than education. The disciplines will be much more competitive, with older and less talented workers left on the sidelines. Tom Friedman and Alex Tabarrok, reflecting conventional wisdom,  are mistaken in maintaining that increasing STEM education is a key to future economic competitiveness.

Jelski instead recommends English lit and psychology, at least if you are young and hot!  The logic–computers don’t write well and people don’t want to have sex with or be counseled by computers (yet!),–seems strong but wage rates and unemployment levels don’t support the argument. Jelski is correct about demand but forgets to take into account supply. Thus, the way to go is to be a hot engineer who can write well and get along with other people. (Jelski also forgets that my argument for STEM was in large part about the spillover effects).

I am in strong agreement with Jelski, however, that education is only the first step to success. Education is a tool; to truly succeed one must have skills developed with grit and applied with passion.


The big economic question I have is that we'll have this huge capacity to create stuff for consumption, but only a limited mechanism to agree to the dissemination of the goods to consumers.

In a world where few workers support many, many consumers, what is the mechanism by which these consumers, who far outnumber the workers (thus most are NOT workers), agree on who gets what to consume?

And, as a secondary effect, will being a worker be a good thing?

I am not so certain that wages in the future will, in fact, be lower. It may very well be the case that wages for those employed are higher, but that fewer and fewer people will be employed. Futurists 60 years go predicted that technological advancements in production would reduce the typical work week. They were wrong, because they failed to recognize that technology would enable highly productive workers to work more hours and would make less productive workers obsolete. The average number of hours may or may not be lower, but it is the distribution of hours that is more relevant to the typical person. Likewise, I do not believe that technology will necessarily lower wages for those who are working. If the current trend holds, it will only increase them. But what it will also do is displace numerous workers, perhaps turning them in to zero-marginal product non-workers. The distribution of wages will, like the distribution of hours worked, be more relevant to the typical person.

While the integral across the entire wage distribution may be larger, that fewer and fewer people will be employed means to me and to those that are not employed that wages have gone down, unless fewer and fewer are employed because of their independent wealth.

i see this as the big question -- will it all be service industry make-work so that consumers actually have some income with which to consume all the goodies being produced?

I'd prefer make-work over direct payments, but the total portion of the population gradually becoming ZMP workers could be massive over time.

If I'm not mistaken, Zero Marginal Products means their marginal production or output is equal to their wage. If you don't mind me nitpicking, they'd be negative marginal product workers.

I could be mistaken, but I thought referred only to the worker's output. For example, if I add a ZMP worker to an assembly line that produces 100 widgets a day, the line will continue to produce 100 widgets a day. Adding a negative marginal product worker would cause the line to produce fewer than 100 widgets a day. More important for the present context, if you were to layoff a ZMP worker, you would expect to see no decline in the number of widgets produced. In effect, this is what economists claim to have seen, with massive layoffs being accompanied by no decline in total output. I don't think that workers' wages are relevant to this observation.

We need to find aliens quickly to send all the ZMPs to fight...

There will continue to be new STEM jobs for a very long time. Computers are only just now starting to reach critical mass in terms of always being with you. For every self-driving car, there will be a mechanic to service it, a programmer to update it's GPS system, another to build interfaces so local restaurants can beam their menu directly to the car. Etc etc

Computers automate, but they also create new opportunities. Email marketing (not spam, but opt-in stuff that you get from Groupon or your broker) empowers new communication that wasn't economically feasible before. So even though, say, Charles Schwab spends less on postage per customer than ever, it employs many more designers and copywriters than it did before the web.

In this sense I tend to agree with the article. Self-driving vehicles will likely have robot mechanics (in the mid-term anyway). Repair shops will have salesmen and robot technicians, hopefully the same people. GPS systems will coalesce around a single standard and only a small handful of people will be responsible for every GPS system worldwide. This is nothing revolutionary, this just happens to be a rather visible example. In fact, I'll be surprised if my first car doesn't run Android or some market successor.

We are so, so far away from having robot mechanics. Either robotics would have to advance tremendously or cars would have to be totally redesigned. I don't see it happening in the next 50 years at least.

In 50 years, any repair-shop that has a salesman (which I take as suggesting it is aimed at the individual consumer) is likely to have human mechanics. Such shops will constitute a small nice market, catering to the wealthy and/or eccentric - and their non-mass-market demands may not be scalable or appropriate for robotics.

I agree that robot mechanics are not on the horizon. There's considerable manual labor in building today's cars because there are so many things a human can do quicker than a robot. This multiplies when it comes to fixing the same vehicles. It's much more likely cars will follow the same trajectory as TV sets, eventually becoming disposable rather than repairable.

My problem with this analysis is it's backward looking. Yes, mechanics may be automated, but there are a ton of new services and products that we can't even conceive of. Just off the top of my head I can think of a half dozen new business opportunities around self-driving cars:

1. "Auto tours" where we beam you some GPS coordinates along with an audio narration. Very useful in places like Yellowstone Park.
2. Drive thru without the drive thru - order food from a menu beamed to your car and automatically drive through a pick up station
3. "Like a boss tours" - rent a limo for a night that drives you and your entourage all over town

These won't necessarily succeed, but there very likely will be new services like this that emerge that we cannot even conceive of this. And all of these will require highly skilled people to create and maintain them.

Even though Google relies on a ton of "free" open source software, it nonetheless spends billions on data centers and the people who design, build and maintain them. Their infrastructure was not economically feasible just 15 years ago.

Think back to the year 2003. Imagine someone tells you about this new thing called social media marketing that will create thousands of new jobs, hundreds of agencies, and billions of dollars in spending. You would likely have thought this person to be nuts.

Since we have no idea what category of future employment will have the greatest demand, why not focus on getting as many high IQ people into the country as possible, since high IQ people do every job better except vegetable picking?




J'accuse: HATE-THINK!!!!!

Wait, wait, does he seriously think that computers can actually do STEM jobs... like say... programming? Let me know when this occurs, because it'll be a lot bigger deal than whose theory of job desirability is better. You should probably let Ray Kurzweil know.

"Further, computer science jobs are themselves being computerized. The job description for today’s computer scientist is only tenuously related to what they did 25 years ago."

No, no, no, no. This person has no idea what they are talking about.

I know that I am a little behind in my reading, but yeah, when did computers start programming themselves? And I mean not when they are coddled by a thousand IBM scientists?

Yes -- computers doe a fine job of much programming.

An excellent example is the "compiler." Once upon a time, one had to translate one's logic into "op-codes" in a process known as "hand assembly." I've done this. It's PITA that required HIGHLY skilled people. But starting in the late 50s, programmers first creatated "Assemblers" and then "compilers" the completed automate this task. Another great example is how software completely took over semi-conductor layout.

There will always be room at the top for a best designers; but the history of computers is that next two strata down being continuously automated away. That's how Google and Facebook can have such a massive economic presence with such small technical staffs.

Compilers and assemblers aren't really "programmer replacements" outside the sense that if your company has 100 assembly programmers, you can probably shift to a higher level language and fire a lot of them. From the perspective of someone writing code, they let them become more powerful, which increases the amount of things one can do, and therefore the things that others can demand for them to do.

As we've created higher level languages, the demand for programmers hasn't gone down, it's gone up -- because programmers are now more powerful (obviously combined with advances in hardware) they are now asked to do more things. It's like the (possibly apocryphal) comment about the IBM guy who said "I only see a market for a few computers in the world" -- sure, with the programming you can reliably do with assembly languages, there are a limited amount of things you want to do with it, and so if you find a way to automate some of those tasks with a compiler, there will be less need for programmers. But that doesn't remain the same: once we became more powerful, we learned we actually wanted to do more things. So many more things, in fact, that the # of job openings created by that demand was higher than those "eliminated."

That's the fundamental difference in the relationship between (programming and computerization), and (computerization and anything else). Programming is a low-cost substitute for certain types of other labor, and an input resource to computerization -- it feeds on itself so long as we have a demand for the outputs of computerizable labor. Yes, at some point in the future, when we've gotten to the point that every possibly computerizable job has been computerized so well that we can imagine no further improvements, then that last programmer will be fired -- but only after the code they wrote will have removed the need for every other computerizable job, and long after the point that millions of people who would've been factory workers have become programmers. Then, finally, Jelski will be right. While there are still massive gains from computerization (something we all believe will remain true for longer than the easily predictable future) still possible, he will remain wrong.

"Compilers and assemblers aren’t really “programmer replacements” outside the sense that if your company has 100 assembly programmers, you can probably shift to a higher level language and fire a lot of them."

In other words comilers and assemblers aren't really "programmer replacements" except in the sense that they replaced a lot of programmers.

... and then made many more programming jobs available, as the newly-more-powerful programmers could do more things with those higher level languages than ever before. Which is exactly what I wrote in the next sentence after the line you quoted.

I mean, look at it the history. When the high level languages were created, did every company just go "oh great, now we can just fire all but a few of our programmers!" or did they then realize that there were so many new things they could do with the new languages (and probably far more importantly, the new hardware) and end up having more programmers today? Do you know many organizations who had more programmers total in 1950/60/70 than today?

Those new programming jobs are much less difficult and do not require an advanced STEM degree. We're moving further away with natural language utilities becoming usable (even if very basic). A psych major could be a computer programmer in a decade without much training.

Also, the way you phrase " next two strata down being continuously automated away" it seems to imply that large numbers of strata have, and continue to be "automated away." The thing is, there are actual defined "strata" of the levels of computer code, if you're using the terms assembly and compilers

They are:
/ ---> compiled language (C++)
Machine code (1's and 0's) --> assembly (MOV)-
\ ---> interpreted language (perl)

Sure, we automated machine code and assembly and any programmer who only did that got fired. And certainly we have created packages/API's and stuff that automate sub-tasks of higher level languages. But do you really see us automating the next stages? What that would look like would be creating a layer that translated human language into a highlevel programming language, where you could "program" computers the same way you program people -- by speaking to them. The cool features of Siri notwithstanding, I think we're a bit away from that still -- and there are still many, many programming jobs to be had until then.

Argh, I botched my ASCII graph. Machine code leads to assembly leads to high level (compiled/interpreted) languages.

Compilers optimize algorithms. Programmers decide which algorithm to use and create it. The role of the UI designer has grown considerable over the last 15 years (and will continue to), but Facebook an Google still hire way more engineers than designers.

Moreso than entering code, software engineers design and implement systems that meet requirements from the business side of the company. My impression is that for a good (yeah, I'm projecting...) software engineer, the "programming" part of the job is a pretty minor and tedious part of the work. For me, writing code for a non-trivial project takes maybe 10-20% of my time. Understanding requirements (many of which are vague and conflicting) and translating them into a scalable, fault-resilient and maintainable design is where the bulk of the time is spent. Punching the keyboard and putting together the code is a rote part of this process; at that point I'm not really thinking very much, just implementing what I've already envisioned.

Saying that STEM jobs will go away as computing technology gets better is like saying writers will go away as typewriters, word processors, and voice-recognition software gets better. Better programming languages and libraries increase the ease with which programmers can translate their thoughts into something a computer recognizes. But they don't help with the engineering part any more than a typewriters helps you with plot and character development.

I wish the author had been clearer, but I believe his point is that the business analyst will eventually be the person interacting with the computer. So it's not that STEM jobs go away, they become melded with the existing staff. Of course, the author drops the ball on assuming that the liberal arts and business majors will magically be prepared to handle the technical tasks.

>Understanding requirements (many of which are vague and conflicting)

I believe you mean "Many of which are depressingly vague and inherently conflicting" -- add caps and exclamation points to taste.

I recall a Fortune 500 client sending us an RFP for something which was probably over $500K work; we ended up no-bidding it because no one, including the client's own project managers, could figure out what they wanted done.

Business analysts will increase in importance and they will have more and more powerful tools. But the continued growth of technology means that programmers will be plenty busy on even harder stuff.

Twitter processed something like 30,000 Tweets per second during the Super Bowl. Such a transaction volume was inconceivable not long ago. Ten years from now you'll probably be able to do something similar on a laptop.

But all that power will also create new opportunities that require more than spreadsheet chops.

A compiler does not "program". A high-level language makes programming (arguably) easier and could (arguably) reduce the number of programmers needed to accomplish a specific functional requirement, but that doesn't mean the compiler is programming.

Regarding VLSI, sure, computers have revolutionized the field. Have the number of people employed in VLSI design gone down as result of that?

> Have the number of people employed in VLSI design gone down as result of that?

I thought that they had, although I'd like to see a cite. Many applications for which one used to design an ASIC have gone over to FPGAs or plain old SoCs.

Hm...but this reduction is not due to the automation of VLSI design, correct?

To clarify, that isn't exactly a clear "as a result of."

One technology is being replaced by other technologies which are easier to use and don't require as many skilled employees. But the reason FPGAs and SoCs are easier to use isn't because of automated design tools. Among other things, it's because the debug cycle is faster. So as soon as the performance becomes sufficient to accommodate the application, ASIC users switch over.

Have the number of dedicated STEM jobs increased in VLSI? Or have the number of business jobs around VLSI increased?

One technology is being replaced by other technologies which are easier to use and don’t require as many skilled employees.

Does it really require less skill to write VHDL than to design an ASIC? Basically, I am somewhat skeptical that this is supporting the original contention that STEM jobs will decline due to computer automation. Clearly, if one looks at a specific task, we can say that automation improves productivity, so it requires fewer people to do the same thing. But what about the network effects, etc.?

> Does it really require less skill to write VHDL than to design an ASIC?

Most ASIC design is done by writing in VHDL or Verilog, and has been for decades. Decades ago, when people designed ASICs by drawing out circuits the skills leaned more towards traditional electrical engineering, and less towards management of complexity. It was certainly more specialized work.

> Does it really require less skill to write VHDL than to design an ASIC?

Did you mean to ask "Does it really require less skill to program an FPGA or an SoC than to design an ASIC?"?

Well, whole specialties go away. And there's much less pressure to get things right the first time since respins are not nearly so costly. As you move from full-custom -> ASIC -> FPGA -> SoC, you reduce the scope of problems you have to deal with.

As a firm you may choose to take that budget and schedule risk and apply it elsewhere, say to making the software snazzier. It's hard to get an apples-to-apples comparison.

I also think STEM will increase rather than decrease. Even though many existing services become codified, for legal and strategic reasons, rules are ever changing. STEM perspectives are needed to ensure new rules fit in gracefully with prior logic in all its complexity. Computers allow for more complexity -- this is why you see STEM types gaining over MBAs in strategic roles.

You see a STEM elite in these strategic positions, but don't need to be a genius to implement one more business rule X. You do need to have a firm logical grasp, knowledge of coding and an attention to detail, but these are common traits for STEM workers.

Security and convenience/design are also basically endless STEM sinks.

I don't find that English Lit promotes a style of writing that is frequently useful in business.

English Lit promotes a style of writing that is negatively productive in business.

Big vocabulary? No - use the identical term to refer to the identical concept all the time, to communicate clearly.

Term papers? (Smile)

STEM compensation will increase. STEM employment will stay stagnant.

My erstwhile field, Chem Eng, shows the way. Chem Es stand out amongst engineers for having the average highest capital per engineer amongst the engineering disciplines. Basically, chemical plants need to big to be economic, but it takes about the same number engineers to run a huge process or a ginormous process. Thus, those Che E's who are employed are *very* well compensated. But second rate chem e's aren't much use --their mistakes are too expensive. Thus there are plenty of Chem Es who never find employment or only stay in the field a brief time.

Software has clearly achieved this kind of scale as well; a few great programmers are more useful than a roomful of mediocre programmers. My firm (financial software) sorts through hundreds of apps per hire because of this effect.

As the calculational drudge get automated away and the ever-improving SCADA systems reduce the need for highly skilled people to the the near-process monitoring, other STEM fields will become more like Chem E and Comp Sci: massive returns to top tier, and group of highly skilled but moderately compensated "support" people, and not much else.

Only go into STEM if you have the passion AND the skill to reach for the brass ring. Otherwise, expect to be spit out by the time you're 30. (Mind you, that can still be a good life. Field Application Engineers have little trouble providing their families a middle class lifestyle years after their skills have fallen behind).

I can't speak for chemical plants, but as to software, we have not seen this trend exactly. There is some shift in skills emphasis, though. For example, much software is written in languages that support garbage collection, which (arguably) requires less skill than writing in a language without it.

Frankly, you're both correct. The GP is in the Financial Software business, where mistakes are costly. If the parent is an Android game designer, then certainly the hacker mentality will win. However, if a cultural shift occurs against hacking, expect the GP to be completely correct.

"to truly succeed one must have skills developed with grit and applied with passion."
and/or have natural talent, a high IQ, good looks, a trust fund, and a bunch of other inborn qualities that one has does nothing to receive or deserve. Just sayin.

But it's been that way since we evolved from apes. Just replyin'.

Not sure what it is you're saying. That only hard work should be rewarded? Or that every child should have the exact same starting conditions? the smarter ones made stupider and prettier ones made uglier?

Perhaps you think smarts and beauty are only used to steal from the stupid and ugly?

Talent is not inborn. No newborns are born with the ability to play tennis, write code, or compose a symphony. It's all about focused practice and desire - the "grit and passion" referred to above. Looks are not inborn. Everyone has the chance to improve their looks through self-discipline, diet, exercise, fashion experimentation, and scientific advancements. Someone with perfect features can still be perceived as a careless, idiotic slob. Maybe they'll have a better chance of getting laid every once in a while, but that's about it. Intelligence is not inborn. Hard work, exposure to challenging intellectual environments, proper diet and health, and other factors can all help one work to one's fullest potential. All of the above have some inheritable factors, but that's not the be-all end-all. (Don't know anything about trust funds other than the people who have them are just as much at risk at pissing away their life and accomplishing nothing as they are to succeed.)

Whether or not you choose to see life as hopeless because everyone else has these mystical qualities that you assume they're born with - well, that might be inborn in you, I don't know. But it's a useless way to view the world.

BTW I do think psychology *could* become an important major. Right now it's too abstract in the way it's taught at the college level, but applied psychology is useful for everything from interface design to hiring to project management.

Agreed. I chose to enter the most STEM-y of psychology areas, quantitative psychology. After 15 years of being involved in the field, I'm seeing less of an emphasis on quant psych people being good programmers/statisticians who are interested in measurement of human qualities, and more of an overlap across areas of specialty, where quant psych methods are being used more broadly across the field of psychology, and quantitative psychologists themselves are less focused on writing software programs and more focused on applied techniques and clear communication with those in other fields (medical education, clinical psychology, and so on).

"but they’ll be paid for smarts rather than education."

And what about the 49.9% of Americans who are below the median in smarts?

They'll be paid less than those above the median.

they'll have a choice to work more hours to consume the same amount of goods and services. As such, it's likely many of the smarter folks will prefer leisure to building up their employable skills.

And, if you believe people in general work smarter, harder, and longer when they can be compensated in proportion to their product, the less intelligent in society will benefit from trade with the smarter folk.

Smarts are what differentiates us from the machines. So equilibrium will be found when humans are only doing the things that require human thinking. Of course, customer service is one of these. So expect Wal-mart to use robotic stockers once energy becomes cheap enough, but Wal-mart greeters will always be there.

This is untrue. Labor versus capital costs are what differentiates us from machines. If it's cheaper to replace you with a machine, then businesses will do that. We do it all the time. The ATM is a good example. But go to a poor country, and companies don't replace people with machines to the same extent the do in advanced economies, because the capital/labor payback is insufficient. And don't assume this applies only to production equipment. For example, I owned part of a packaging company in Hungary. Our production equipment was largely state-of-the-art. However, we didn't use equipment for packaging our products. Those were packed by hand, and the largest source of employment in the company. We employed students, in many cases. Labor was cheap in eastern Hungary at the time. But if you went to, say, Switzerland, many of these functions were automated. So if you're being replaced by a machine, it means, almost by definition, that you are prosperous. Machines tend to replace people in rich countries; less so in poor ones.

"Lower pay and a higher standard of living". Sounds like 1920's, 1930's International Trade economists: You gotta buy more of the cheap stuff and less of the expensive stuff! What if I don't want to?

Law #1: Computers will allow people to undertake new jobs. For example, what is art today? Take a look at, say, the set in the video game "Assassins Creed". It looks just fantastic. There are a lot of very high quality, commercial art hours in that software. Similarly, have the number of analysts declined with the advent of Excel or Lotus 1-2-3? Just the opposite, I think. IT is the enabler, not destroyer, of many professions.

Law #2: If a country has been excluded from the global marketplace for many decades and is very large, then its entry into the marketplace will affect labor, product and service markets. China's entry into the global market means higher--much higher--wages for Chinese. It may mean stagnating wages in other parts of the world, at least for some sectors, until China catches up. But this should not be confused with permanently lower wages.

Law #3: The internet, telecommunications and other innovations affect economies of scale in production. Ideally, "an army of one" is the right scale, as it means all fixed costs have either been eliminated or are easily absorbed by the individual. People work in organizations when the benefits of fixed structures outweigh their loss in flexibility. Thus, freelancing is the ideal productive form (the theoretical optimum), assuming transactions costs are low. It also suggests a high degree of specialization, I think. (That is, a person will work outside an organization when the need for that person's services is less than one FTE. If you need a full time employee, you tend to hire one, precisely due to fixed cost issues. Specialized persons demanded less than 1 FTE by a given organization will be outsourced, and they will be called "consultants". Consultants reach economies of scale by working for multiple clients requiring the same functional skill.)

Law #3a: Many factors influence economies of scale over time. Don't assume that current trends will continue into the future. But it is true that higher wages imply a higher value of time, and therefore imply a trend towards flexibility and greater automony.

Law #3b: Don't confuse the social aspect of an organization with its business function. Freelancing can be associated with isolation and loneliness--but that's not necesssarily the right way to think about it. The work setting and the work function are not the same thing. Expect more flexibility in the work setting, particularly as the baby boomers approach retirement.

Law #3c: Freelancing does not necessarily mean "insecure". It can also mean "private practice", but if you think about it that way, you probably have staff as well. You just end up with a large number of specialized companies with relatively low headcounts.

All economies are not equal, or fair. Nor do they share the same dynamics and principles. Communism does not work like capitalism, and both are different compared to a plutonomy like the

This kind of prognosticating about "STEM" jobs by people who have never done a technical job in their life is really just beyond stupid.

Did total employment of engineers decline when the slide rule was replaced by the electronic calculator, because suddenly we didn't need so many engineers to man all the slide rules? It would be nice if people had any idea what an engineer does before making these sorts of predictions. Engineers don't carry out easily mechanizable processes. They think of ways to solve problems.

That's a good example that gets to what I was trying to articulate.

Maybe not so much engineers, but remember, the word "computer" used to refer to a person, not a machine.

And now look at how many people are employed in modern computing now that those human computers have been replaced.


I agree with this point, and I fail to see how STEM jobs are easier to automate than art, accounting, marketing, or being an Economics professor. Software development hasn't become less technical with the advent of compilers and debuggers; the focus has shifted though. CAD software (and other engineering software) has only been accompanied by increased demand for engineers. We've offshored a lot of engineering and software development jobs, but we've also offshored a lot of accounting and radiology jobs.
The last fifteen years of the internet have done more to reduce writing jobs than create them. I'll read Jelski's article, but so far the analysis of STEM jobs in this blog seems clunky.
I'm not saying the conclusion won't end up being right, but the analysis that leads there looks suspect and smacks of "the arts and sciences I understand will prosper but the activities that I don't understand must be easy for a computer to automate".

More simply--Law #1 might be right, but the Alex Tabarrok doesn't understand what computers do well sufficiently to apply it predictively. Neither does Daniel Jelski.

After reading Jelski's article, I largely like what he wrote, but I think he underestimates how tradeable and automatable fields outside of STEM are.

This post scarcity futurism stuff is reaching a peak these days. Most of the world lives in slums, the bottom half of the first world has been stagnant for 30 years, and the earth certainly can't support it.

I hereby rebut:

Isn't that rational optimist argument (potentially) akin to saying the dinosaurs were doing great up to the meteor, or my nan is 93 and has done great so far?

I don't see it that way. I think his gist is that our species has proven very resilient and especially in the last 300 years or so has improved its health and welfare exponentially, and we do so via trade and specialization and innovation, and he sees no reason that won't continue.

He cites numerous facts about how pretty much everything you can measure has improved in recent decades (and centuries), and how at every point in that timeline the smartest folks alive were pessimistic about any future progress. Every generation thinks the earth can't support us, and they are always wrong.

Now obviously if the population kept growing to infinity eventually that would be true, but population growth is already plateauing, with most demographers thinking we top out at 9 billion give or take, in about 50 years give or take.

I have no doubt we'll be able to feed and clothe and entertain them all. My pessimism is actually of the opposite sort. What happens to growth-based capitalism when the world's population stops growing and starts declining? Does the whole planet become a demographic nightmare like Japan?

It's true. Life has been pretty good for 300 years. And its not a terrible idea to say, "the trend is your friend." Then again, I'm sure Romans said the same thing about the first 300 years of their growing empire, but at some point things did start to stagnate, decline, and eventually collapse.

In the medium run I'm reasonably optimistic about improving living standards for most of the worlds population (assuming no big wars, which is a big if). I'm not sold that life is going to get better for the average first world person, and as I'm American I'm not super optimistic about things for the average American. It would not shock me if GDP went up, but all of that benefit and possibly then some went to a small portion of the population. Some new technologies will be cool, but having an iphone hasn't really done much to, say, keep marriages together or build communities.

You two seem to have restated my point. Just because things have been going well for a long time doesn't mean they will continue to.

Absolutely. The reality is that bad things happen all the time - most of life is just a disaster - unless you're lucky and upper middle class and then you can spout shit like the Rational Optimist

I live in graduate housing at a UC. If you see the hardworking immigrant students here who never let an opportunity escape their grasp, you'd realize that there are ways to create your own luck.

Will, with all due respect that's ivory tower bullshit. Just because some hard-offs make it it doesn't mean it's possible for everyone to.

Most of those hard working immigrant students are going to get dead end jobs and megacorps. I know because most of the Asians at the companies I've worked at get the shittiest jobs and advancement opportunities because they are seen as boring and uncreative. I'm sure they studied hard in school.

The best areas are those constrained by rules and regulations, medicine, law, finance.

"And what about the 49.9% of Americans who are below the median in smarts?"

We want more of them, too. Many more. The more the better. We want them in every school and in every neighborhood. Up with vibrancy!

(But also in cheaper and better products and a higher standard of living for American consumers.)

I wonder if the first statement might be true and the second one false.

Over the last 30 or 40 years many products have become cheaper and better, but as has been discussed ad infinitum on this blog the standard of living of a large percentage of American consumers has not risen.

and many on this blog think the notion that the standard of living hasn't risen is silly, given the abundance available to far more than previously, at lower prices (so wage stagnation doesn't mean standard of living hasn't risen).

What is wrong with crime as a career? O(ften profitable , full of colourful characters , guaranteed hospitality in times of bad business.

STEM jobs are not necessarily the object of STEM educations. How many good Mechanical Engineers are doing engineering even 5 years after hire? Not many...they rise in the rank to management due to having the math and logic background to create and understand strategy and finance. Likewise, how many EE or CSci majors are still doing programming as their main task? STEM is the signifier if higher IQ, which companies turn to as the qualification lacking anything else meaningful. Is the ME MS any "smarter" than the History MA? Nope...but one of the two will likely know statistics, advanced maths, and practical logic problem solving with contemporary software tools. That's what I see over and over. And I AM one of the non-STEM majors who works in tech due to getting in early as a kid. It must also be noted that this phenomenon is due to the incredible ineptness of Corporate HR...they spend 90% of their time avoiding potential lawsuits, only 10% trying to find smart people.

I'd say most CSers stay as programmers - Engineers tend to move up but CS majors are by and large hopeless nerds

He has it backwards. More automation allows/requires more engineers, not fewer. It means any specific project needs fewer engineers but you open up many projects which would not have been viable otherwise.

How are STEM jobs easily computerized? Surely before "we get to" STEM jobs we would have gotten rid of 90% of driving, retail, and service industry jobs?

I'd guess that part of the reason the service sector continues to take over the economy is that many of those jobs are complicated to "computerize" in a cost-effective way. It might be "easier" to computerize the job of a McDonald's cook than a computer programmer, but the payoff to computerizing the programmer's job is much higher.

Another way of looking at it is that the marginal revenue product of capital in a lot of service industries is smaller than in tech.

Professor Tabarrok is right in that education alone is no guarantee to success. I don't know exactly what he means by "skills developed with grit" but I agree that it must be applied with passion. Self-confident individuals who have a strong sense of self/a well-formed identity, a capacity to connect with others, drive and passion are more likely to succeed. Hot women with a humanities or arts background are ,however, more attractive than hot women with a math or science background. Perhaps this is because being sensual and emotional are to a degree feminine traits and they are better displayed in those with humanities/arts backgrounds.

". Education is a tool; to truly succeed one must have skills developed with grit and applied with passion."
Yea or a tenured position in an economics department. Because nothing says "I love competition" like joining a guild!


The people with the biggest returns from education are Doctors and Lawyers - and the size of the return is based HEAVILY on the fact that these professions have large regulatory barriers to competition

First thing we do, let's kill the ABA, AMA, and tenure.

Aren't law schools brimming with unemployable law school grads. Seems like a shitty guild.

Yes, but it's due to oversupply. Way too many law programs.

Whether it's the opacity/difficulty of measuring human capital or rent seeking that supports salaries, a competitive graduate from a top 20 law school can still field offers ~$160k.

Law #2: A global market place will result in lower (real?) pay and fewer opportunities for many careers - but also in a higher standard of living for (some?) American consumers..

It's the old Luddite argument all over again. Automatons will take a lot of what engineers do today. Yes, so what? The history will just repeat itself.

This guy is right STEM is a waste of time unless you're a superstar. I've been saying this since my very first comment on this blog though I guess you only listen to people with PhDs who write on trendy sounding blogs. Don't pin your hopes of returning to prosperous employment levels on STEM.

Counterexample: I graduated from a middling university with a 3.4 gpa, was offered $50,000 from multiple firms in San Francisco / Los Angeles. Granted $50,000 isn't much, but it certainly isn't a waste of time.

That's not really a counterexample - doesn't a 3.4 gpa put you pretty high up in the class in most STEM programs? And your right 50K isn't much considering most people who go into finance will be blowing past that 2-3 years in.

That's CBBB in a nutshell...if he can't earn like a finance stud right out of school then screw everyone you all suck.

I wonder if Andrew' would still hire him.

I wonder if Andrew’ would still hire him.

Huh? I don't get that reference.

No but I'm saying - and as I HAVE been saying - STEM is a crap major if you're not going to be a superstar - it's WAY harder to pull down straight As in a STEM program then if you say major in Economics or Finance or something. Tyler Cowen's big recommendation for employment is "we need more STEM grads" but not every Science or Engineering graduate is going to be elite and if there's only jobs for the top tier then it's a dead end and some SERIOUS rethinking needs to be done.

In a thread a while ago I mused your attitude was probably one big reason you are having trouble finding work. Andrew' (I think it was him) chimed in he'd hire you (if he ran a business). I'm curious if more of your postings have changed his opinion on that.

I think you make some good points about STEM majors and so forth, but your aggrieved tone is really off-putting. You need to stop whining about the injustice of it all.

TC and others are saying there needs to be more STEM grads because those kinds of skills are high value add, not art history and women's studies and so on. Those arguments aren't weakened by constantly carping how overpaid some in finance are.

And yes GPAs are lower for STEM majors...guess what, employers know that. And the guy above posted his 3.4 got him multiple job offers at a not too shabby STARTING salary. Not straight As, middling school (his term), not 'elite', and he's doing fine.

You got a huge chip on your shoulder and it's weighing you down, my man.

"And yes GPAs are lower for STEM majors…guess what, employers know that."

Ohh no they don't. They certainly don't - and what if you want to say do an MBA later or something? You're fucked you're not getting in over the Psychology grads with 90 averages. Most people, including older adults (the they who would make hiring decisions) who didn't go through a STEM program have no idea that average marks are lower - they just assume everyone can pull off straight A+s like in the humanities.

3.4 sounds high - my university just used straight percentages so I have no idea how the GPA system works - I think it's more an American thing, although some Canadian schools use it I believe.
But any way I think Tyler and Alex and totally wrong, there's huge disadvantages to getting a STEM degree if you're not top-tier and pushing this line about Science degrees has been dangerous for many people, not just me.

Either you, or Canada, or both are seriously messed up.

I'd like to add that I know electrical and computer engineering students from rank 40 and 50 undergraduate schools with 2.6 gpa's who are getting job offers ~$60k. They develop ipad apps. Stanford and Berkeley ECE grads start ~$80k. The employment data will likely back this up. There's a shortage for this talent and online communities like Reddit will confirm it.

I would argue, CBBB, that it's much better to be average in engineering and technology than average in softer fields. Small and boutique investment shops pay shit to analysts and grunts, and good luck if you're not in finance.

You and I are likely in a similar boat. I was not engineering - I studied statistics and econ (double major) with a programming minor. I applied to many places for internships to figure out how my skills are applicable. I researched the business, talked to graduate students/professors/people in industry to figure out how my skills and interests fit in to graduate programs and various industries. I tailored my resume and cover letters, applying only to positions I was interested in. I showed up prepared for interviews.

If you have a low GPA, it sends a signal to employers. My GPA was not the highest, but i was still able to convince people to give me a shot.

I suggest you take a look at placement rates and salaries for Masters in Financial Engineering programs. There's a good number of people at my firm (index research) that don't make too much more than $50k.

"Superstar" is relative. An inexperienced, not-so-good programmer can still make a higher than median wage simply helping people set up WordPress blogs.

To those who can't do that, such a skillset qualifies as superstardom.

Who's making above median wage setting up WordPress blogs?

Marketing agencies, PR agencies, social media consultancies, market research firms. A lot of what they do is translate a client's idea into a technical implementation. And a lot of that rests on setting up web sites. And a lot of those are built on WordPress. And most of them are very average in terms of technical sophistication.

I've worked with people in all these fields for over 10 years and they're doing just fine economically.

This reminds me of a conversation with a well-meaning adult when I was in High School who advised me to stay away from software because of some forecast he'd seen indicating that software jobs were going to decline. That was thirty years ago. So glad I ignored him.

Decades ago Eliza was an extremely popular AI counseling program among the graduate students at a competitve graduate program. Discrete, available, free. It wasn't sophisticated. But, it didn't have to be be an assist.

How are you feeling?

I feel depressed.

Why do you feel that way?

[40 year old memory]..

I want to add here that this freelancing idea is just another disaster waiting to happen. You know I see these people on here whining about how impossible it is to hire someone who's good at some specific skill - well the reason for that is all these companies want to do is hire from the exact same pool of people with 10-15 years experience and what they don't want to do is any kind of training. Unfortunately you aren't going to actually learn the skills if you don't get the experience of working in the field so how do you just jump into freelance right out of school? You don't by-and-large and it's going to be a rude wake up call if this is how businesses expect to hire. They'll just clamor for more innovation and Tyler Cowen will write blog posts about how we need more STEM majors - even as the pool of unemployed graduates swells

Companies will pay for training if it produces value.

For that matter, so will students who have enough real-world experience to understand what value is. (My theory is that most higher ed does a terrible job of teaching this to students. I think every major should have a co-op program of some sort.)

Tyler believes improvements in society's welfare is driven by STEM. In this I agree with you that it would be useless to increase next year's STEM graduating pool by 10,000 idiots.

I'm sure Tyler would agree with this: Society should increase investment in STEM ventures, to do so requires more and better STEM human capital.

Surely you agree innovation is driven by STEM and not by psychology and english majors? Regardless of what the wages are for these people, our economic pie will be larger?

But what I've said is for every 10 STEM graduates 1-2 actually get decent careers related to their major. Are we to burn through 8 people and destroy their lives just to get the tiny payoff of 1-2 new scientists/engineers?
Sitting around chiding people for majoring in the humanities when the prospects of a career in STEM is extremely low is what Tyler loves to do and it's completely misguided.

What's the alternative? I majored in music and now work in STEM. And where do you get this 2 in 10 number?

And does it really "destroy their lives" to get a degree they don't end up using? I'd say that describe a huge percentage of people I know and work with.

Where did I get 2 in 10? Well like most numbers I cite I made it up.

It destroyed my life - and I'm sure MANY others.

I'd like to give you a free link to your site on one of our client sites, in order to increase your search engine visibility. You can write a guest post and embed the link, or just let me know what key phrase you want us to link to and I'll take it from there. It's free, no strings attached. I'm going to suggest you give us a link too, but even if you don't provide us a link, we'll give you one from one of our client’s tech themed sites.
We are in process of building links for one of our well known reputation management client. Of course, if you think our client site is a good match, we'd appreciate a link from you too. Our client site is and they provide Reputation Management. In fact, if you'll give us a link, we'll share the love and upgrade you to a really good link from a page on a much more powerful site (not because I am sure we'd both like to avoid reciprocal links).
If you've some other text advertising program setup, I'd love to hear about it. We may be able to help you there too.
Just e-mail me back with your preference and I'll get things rolling. I hope to hear from you soon.
Patrick Thomas
SEO Specialist
G2C e-Reach
Quantum Reach with G2C e-Reach

Just a reminder that all of those folks who dropped out of high school, or just barely finished high school, or have a few semesters of college, are not going away. They don't just disappear when their jobs disappear. Either we have an economy that makes jobs for those folks, and they are pulling along with us, or they don't have jobs, and they just vote that we all pull a little harder. Plus, if they don't have jobs, who is going to buy all of your crap high tech stuff?

How about jobs for university graduates? I know a few people who didn't finish university or never went and they're doing better then me because they have parents with a lot of money. That's really how you become successful these days - make sure you have parents there who will just throw cash at you.

You could marry someone rich...

No I can't live with that I need to have the power

I don't know about STE, but as a math grad student who has not learned anything of even remote practical/economic value from classes since my first year of undergrad, I really wish people would stop including the M in these claims.

Comments for this post are closed