Say only 10% of humans had any skill whatsoever that would provide value that could not be more cheaply obtained by deploying technology. Say it got further, to 1%.

Imagine a world where only a handful of the most incredibly skilled humans can do anything better (i.e. more cost effectively) than a computer, and the rest are rightfully unemployed, because they have nothing to offer. A world where your 20 years of hard earned programming expertise is utterly worthless in the marketplace because an open source program exists that will read a plain language spec and spit out a good approximation of the same code you would have written, except without buffer overrun bugs and it'll be done in 20 seconds instead of three weeks. Where "lawyer" and "doctor" are professions-that-were because humans aren't fit to compete, and where one by one all the other careers considered unassailable today are swallowed up by increasingly sophisticated automation. Sure, we might be left with a few professions that require physical humanity (people will still want to see people in entertainment, for instance), but in the end, most of the wealth in the world will not be created by humans in any direct way.

You might start to understand why it's not such a terrible idea to move closer towards a socialist society as time goes on - eventually we're all going to be unskilled labor, possibly even unemployable, and we're close enough to that point (I'd say within a couple generations, at the outside, barring some major disruptions that hold back further technological progress) that we should probably start thinking about the consequences.

And who will write the spec? And why will it be any easier to write the spec than the program? "Plain language" is a misnomer. Idealistic people try to write legal contracts in plain language, but it doesn't work.

Sure, there will be people at the top of the command hierarchy dictating what needs are being met. That doesn't mean that spec-writing jobs will provide gainful employment for as many people as are employed fulfilling specs these days.

As for why it will be any easier to write the spec than the program, it already is. "Plain language" is exactly how people spec things out all the time these days, it serves me quite well when I'm getting work assignments ("The Fizzle button is making the back end Furble on Tuesday nights, fix it!", or "Add a coupon code field to our order page and an administration tab to add codes to the database - just make sure the same code can't be used more than once").

Idealistic people try to write legal contracts in plain language, but it doesn't work.

People tell their lawyers in plain everyday language what they care about achieving with a legal contract, and the lawyers fill in the blanks based on what they know. Similarly, normal people tell their programmers what they care about achieving, the programmers fill in the blanks. Sure, we can't automate that filling-in-the-blanks super well yet, and yes, there often has to be some back and forth to figure out what the person really wants (this is, IMO, on of the most important things missing in a lot of machine learning approaches, as human thought - especially language processing - involves a lot of feedback between systems and backtracking), but to me it's not hard at all to imagine all of that being extremely successfully automated once we have thirty years more NLP research under our belts and laptop computers that would put the most expensive Hadoop cluster you could construct today to shame...

Then again, I'm one of the weirdos that thinks that the "problem" of intelligence can be solved by a much less sophisticated algorithm than the one that our brains implement, under the theory that typically evolution only ever seems to do one thing well: create highly complex solutions to medium difficulty problems.

That's a ridiculous argument...and remember, the example was ridiculous, talking about a world where only 10% of people had jobs: People would build -very difficult- things to find employment. (incidentally, we're all of us doing harder(cognitively) things than our ancestors to find employment today.)

I think the sentiment, not explicitly conveyed in the article, is that "there are less jobs for the number of employable people". And there lies the rub. Perhaps net total jobs is still on the rise, but the rate of population growth is that much faster than the rate of jobs created. And it's not flat rates, but changing rates as well. When technology provides efficiency (I'm thinking scale, atm), it would probably push rate of jobs create slightly lower. With more people on the planet, instead of competing with 1 other person for 1 spot, a person might be competing with 10 other people for 2 spots, hence the sense of joblessness, and a culprit to blame.

The rate of population growth is declining, especially in developed nations. The UN estimates that the world population will peak at ~9B in the 2050s and then begin falling. Also, the demographic balance is changing because there will be fewer young people in the job market.

So perhaps the jobs/people ratio will improve, but these same demographic changes (and increasing life expectancy) pose big questions about fewer workers supporting healthcare and government services for more older people.

Countries like China and India a good place to export our jobs, at a minimum, because they're stable enough. Even if we could trade with China in the late 1950's, the Great Leap Forward would've made such trade very unattractive. Why would I want to export my jobs to a country where the average citizen isn't even fed properly, assuming I could?

I posit that as countries such as North Korea and Libya stabilize, the global job growth rate will have to increase in order to main the job-to-employable-people ratio. A positive, but stable, job growth rate would not be enough.

I could be totally wrong about this in that there will always be some minimum political chaos in the world, but I'm an optimist :)

'past predicts future' is not exactly commensurate with a point of inflection. perhaps a real-life example can help - a medical transcription firm in India hires 25 Indians to transcribe medical records. I then sell the CEO a copy of voice recognition software. He is so happy with the results, he fires 24 and keeps 1 Indian to fix the occasional transcription typos! After one month, he finds the transcription error rate is unacceptably high, so he hires one more Indian to fix typos. For the past 1 year, these 2 Indians & that software has chugged along, making money. Now that CEO is happy, making money, and has opened his second office in India, with another copy of the software & just 2 more Indians to babysit and bugfix! True story. Any way you cut it, that's net job loss. First, a bunch of housewives in Atlanta who transcribed for a living lost their livelihood, but you can rationalize that by saying, hey atleast 25 Indians got hired. But 23 out of those 25 Indians lost jobs & 1 piece of software ended up doing the job! Now unless the Atlanta housewives and the 23 Indians end up working for the voice recognition software company ( a very remote possibility ) or end up in a brand new job ( even more remote possibility ) given their limited skills, there is definitely a block of time when there is net job loss. Ofcourse eventually some of these people will learn new skills and/or are absorbed into new jobs etc...atleast, that's the hope. But the intervening reality can't be wished away. I still think this story will have a happy ending, but so far I don't see it.

You don't see the happy ending yet because this was the worst financial crisis since the Great Depression. We love to trash finance types, but banks facilitate a lot of transactions--including home purchases and capital expenditures. When they hurt, our economy hurts in some non-obvious ways.

If you assume that for a portion of the people in the transcribing job, that the transcribing job was the best marginal fit, then some of them don't re-enter the job market. Others have to take a worse job. If they could have been in a better job, they likely would have been. So their employment has gotten worse. May be quality, may be compensation, may be available time. Something's worse for them.

There are a number of reasons this is not the case. I'd wager that a main reason is "re-training takes too much time." If a worker gets laid off, gets his Associate's, and starts something new, well, that's a net positive scenario. He or she produces output in a new field, while the transcription software replaces their old output. Net positive outcome.

You call it "handwaving," but the above happens to many people. This is not some accounting trick economists have developed.