I have been telling all my colleagues about Cowen’s latest book: The average is over. I really enjoyed the book.
Cowen goes on in his new book to explain that regular folks are doomed. The economy will only need the top 10% of us.
The argument is essentially technological. Computers and robots are getting much better. Soon enough, your local McDonald’s will be entirely automated, except maybe for the manager. Self-driving cars and trucks will make truckers and taxi drivers obsolete.
Cowen reiterate the well-known arguments: the best teachers will produce electronic material that will be used worldwide. There will be little need for most teachers. Maybe Cowen will get to teach economics 101 to 80% of all students worldwide, thus making obsolete thousands of economics professors.
Yes, only a small minority will be needed, but what Cowen failed to consider is that most of us haven’t been needed in a long time. Maybe economists can’t even consider such a possibility… but let us be honest. Most of us don’t have real jobs. Yes, we are very good at maintaining the illusion that what we do is essential… and it is crucial for our self-esteem… but, speaking for myself, if I were to disappear along with my job… what would happen? Some students may graduate later due to my disappearance. My research program would disappear in the ether… maybe some crucial discovery that I am about to publish would not see the light of day… at least until someone else discovers it independently. But if I stopped working, nobody would go hungry, nobody would die, nobody would be out in the street. Things would mostly be fine. The same can be said for most jobs. Yes, we do need the garbage to be taken away, we do need food on our tables… but we need fewer and fewer people to get these results.
What are the two fastest growing jobs right now? Software programmers and accountants. These are abstract, constructed jobs. If we had to, we could make do with far fewer software programmers… and we could certainly have far fewer accountants. We could make things simpler without any child having to go hungry.
Being needed, being useful… are concepts that are constantly redefined. What was once a hobby becomes a serious job (e.g., scientist). The truth is that as the need for work disappears, jobs appear out of thin air. When these jobs don’t appear fast enough, we ask that people get even more education (whether they need it or not).
So, no, regular folks aren’t doomed. Maybe in the near future, we will all spend 20 or 30 years perfecting our education (using software instructors). When we graduate, we will become manager of the local McDonald’s (even though it is fully automated) or supervisor of the local park (and said park will be kept clean by autonomous robots). I am sure tens of thousands will soon make a living creating silly videos for YouTube (assuming that’s not already true).
Whereas jobs kept us fed and warm, jobs are increasingly meant to give us a social status. Essentially, Veblen’s leisure class is eating the world.
People like Cowen would like us to believe that as robots learn to grow and bring us all the food we can eat, learn to drive us around, and entertain us… a large fraction of the population will suffer. Nah.
Further reading: See my posts Jobless recovery, the Luddite fallacy and the 4-hour workweek and Automation will make you obsolete, no matter who you are.
I think that you’re wildly underestimating the necessity for most people to keep doing exactly what they’re doing to keep society humming along.
American corporation (to take just one example) don’t suffer unnecessary employees lightly. If they could continue to be profitable and eliminate 90% of the work force they would. As we’ve been going through the recession, corporations have basically trimmed about as much as they could without suffering debilitating hindrances that preclude them from making as much money as they can.
Yes, many of our jobs will be obsolete in the not too distant future, but to suggest that we’ve already gotten there is just folly. Of course you can try and argue that society doesn’t “need” fancy shoes or as many varieties of cheese on our store shelves or the latest improvements to our Android phones, but that’s not for you to decide… society has decreed that it *does need exactly those things and so society pays for the creation and distribution of those things. To suggest that we can reduce our workforce by more than a % or two and have everything basically remain the same is simply absurd and you only need to look to places like Spain or Russia to confirm said absurdity.
I think we won’t see any real AI any time soon. Not in the next 30 years.
I agree with you that most of us do not have real jobs. However, I am not sure that robots will take our jobs. The reason is economical. Where does profit come from? Marx says that profit comes from the exploitation of “living” workers. If there are no workers, then any business won’t be profitable. Hence, the the scenario you describe will cause a major economical crisis. I do not know if Marx was right on this point, but I think his argument should be considered and I am interested to know what people on this blog think about it.
@Leonid
You don’t need “real AI” (in the sense of human-level general intelligence) to wipe out jobs, the same way you did not need something like a horse to make horse carriage obsolete.
These aren’t particularly new findings. 150 years ago, some wise man claimed that capitalism will destroy itself by creating an increasing demand (and supply) of automation in search for ever increasing productivity and ever decreasing margins. That wise man was called Karl Marx, and I have recently mused about whether Free Open Source in software development is already the final means of unsurpassable productivity (many many work forces all working for free):
http://blog.jooq.org/2013/11/12/oracle-glassfish-or-why-you-should-think-about-open-source-again/
Anyway, interesting choice of professions: software engineers and accountants. I would have thought of other engineering jobs first :-). Why accountants? You didn’t elaborate on that in your article…
I think arguments like Cowen’s neglect to take violence into consideration. Does he actually think that 90% of the population is going to passively sit around and suffer?
At some point, people start throwing bricks through windows.
I have always been fascinated by the backlash against technology that plays such a prominent role in the back story to Frank Herbert’s “Dune.” Before 90% of people are unemployed, we will see some sort of ban on certain technologies.
This means that there is evil in automation.
I’m not very good at predictions. I’ll come back here in a decade or to see how well everyone here has predicted the future :-).
@Daniel, what we have now extremely expensive and untested prototypes. Remember that electric cars appeared > 100 years ago and we still don’t have on the roads. Thermonuclear fusion was discovered 50 years ago and we still cannot use it. Statistical machine learning is a very old field, but only now we can use it. All current AI depends on massive human effort. In that, we are nowhere close to replacing these humans. We would probably need to boost the number of programmers (and people who train machines directly & indirectly) by orders of magnitude.
@Lukas
Yes, engineering typically does well, but not as well as software. Software has grown to be huge. People often fail to realize just how many people have software-related jobs… it is enormous and growing all the time.
Finance and accounting are also huge in terms of jobs. A strong driver for this are ever more complicated regulations coming either from the government or from industry standards. In some ways, it is counter-intuitive as you might have hoped that accounting would have been automated by now… but it is not so simple.
@Daniel: I know what you mean by Software having grown huge. But as I said, 150 years ago in Karl Marx’s days, automation was mostly mechanic and mechanical engineering and factory working was *huge*. “Regular folks” might have been people in manufactures, back then.
Of course, it isn’t exactly the same, but I’m saying that steam power has equally transformed the society as a whole, just like software / the Internet. This situation isn’t that new.
@Lukas
But that’s the whole question… right… is what is happening new or not.
Cowen and others will tell you that “this time it is different”. I used to think so.
@Leonid
I don’t deny that we are long way from achieving human-level general intelligence in software.
> jobs appear out of thin air.
Could you elaborate on the kind of jobs that would appear out of thin air? Can you be more specific than “Big Government will hire lots of people”?
Recent stats seem to say that most newly created jobs are low-skill, minimum wage service jobs.
If the newly created job is as meaningful as digging a hole and filling it up again, will people still want to get that job out of “status” reasons?
@Mark
It is impossible to predict the employment picture…
It seems entirely possible that governments will grow bigger and fund more and more people. That is not what I personally desire… but it seems possible.
I don’t think the purpose of jobs is to stop other people from going hungry; I think it’s to raise the quality of life for those people. I agree that widespread slavery where people are the slave owners (aka AI) would help raise that quality. I am a bit worried about the slaves – if not because of Skynet then because of hackers messing with the AI. Oh well.
One of the reasons we see the jobs disappear is outsourcing. The jobs in manufacturing are not gone, they just moved to low wage countries. And yes, robots can do a job faster and better but not cheaper.