1) That for every low-skill job lost, a high-skill job replaces it. That's not the case. When robotic arms replace 100k factory workers, the robotic arm companies do not need 100k new programmers, nor are the factory owners going to use the reduced labor costs to turn into software shops. It's not like the industrial revolution all over again; computers and robotics are producing such enormous productivty gains that eventually there will be fewer people required by the labor force than the working-age population.
2) That people whose skills are no longer in-demand will somehow acquire other skills. Some will, but large swaths of the unemployed won't, especially among people 50+. That's why so many millions, especially 50+ women, have now been unemployed for years. It's not such an easy transition as you make it out to be -- people won't simply adjust, no, a large portion will essentially drop from middle-class to poverty and stay that way until they die. It's not they who will take the new jobs, it's the younger next generation that never had the obsolete skills.
A 55 year old with only a HS diploma for whom a computer has never been more than a magical toaster that delivers e-mail, only existing in the past few years of his/her life, is simply not going to become a professional programmer even if he/she wanted to.
This is entirely true, but was also true in the industrial revolution. With the increasing efficiency of manufacturing we've seen a tremendous growth in service industries that are absorbing a large portion of the displaced workforce.
Whether or not that ratio is 1:1 is hard to guess, though I suspect not quite.
The real problem here is not so much lack of jobs, but like you mentioned, it's nigh-impossible to actually transition people from one field to another. We've outright eliminated most manual and unskilled labor, and what remains is disappearing quickly. The jobs that people had, vs. the jobs that they must now pursue are incredibly different, and most people aren't going to make that jump successfully.
We are dealing with an entire lost generation, if not more.
(Bear in mind that not all "assembly lines" or "programmers" are the same -- there are very likely line workers at Boeing who are multiples more productive than programmers at [big Fortune 500 company of your choice].)
Workers can improve productivity by taking jobs that are otherwise less desirable. Maybe they have to commute further, or work shifts they don't prefer. There is a small segment of the workforce that can improve its productivity just by showing up on time. Some folks will have to make more drastic changes, such as moving. Many people moved to Florida to build houses that, as it turns out, no one wants. Those people won't find construction work in Florida but might well find it in North Dakota, which is currently going bonkers.
And they can seek retrainings that might seem marginal. Maybe they learn to weld, or operate a crane. A semi-literate adult who buckles down achieve actual literacy probably opens very large possiblities for productivity gains.
So moving up the productivity curve doesn't mean learning 55 year olds learning Python. It can mean finding a way to do 5%, 10% 20% better.
That said, I have to agree that improving productivity is a lot easier for people with sound education and training to begin with. One of our great problems is the failure of our schooling culture to properly prepare people to understand and learn new things.
EDITED TO ADD:
Also, I don't want to sound ignorant of all the difficulties of this. Some gains are easily achieved, some are hard won but worthwhile for non-economic reasons. And some of the sacrifices required are pretty hard. It is one thing for economists to talk about "labor mobility" as a factor driving productivity, and another for those who find they have to move half-way across the country.
I disagree, if he really wanted to, he could. I do agree that most don't/won't. At that age, most people have low motivation to start over.
What you're discounting is that people are born with vastly different mental capabilities - I'm quite certain everyone on HN is at least in the top 20% of the country, and it's easy for us to assume that, just because we've encountered few hard caps on our mental abilities, that others are the same.
What changed my mind was working in a factory - where I had months to hang out with line workers, machinists, technicians, and just shoot the shit with them. My takeaway from it is: a lot of them simply do not have the intelligence to operate in many highly technical jobs. This isn't a slight on them - they're amazing, worthwhile human beings, but it's unreasonable to expect them to be able to write code, even given an arbitrarily large amount of training. For every IQ 120, there is an IQ 80. Remember that.
The other side of it is that our school system performs appallingly poorly, and fails a lot of the population. Even if someone had the natural intelligence and aptitude, for a lot of them their educational background is so poor that there is no meaningful way for them to catch up, ever. A person of solid intellect who, for whatever reasons, barely made his way out of high school math, is not going to be able to pick up a CS curriculum. The foundation we're working from is not very solid.
The third component is of course the system - if someone wanted to retrain as a professional programmer from a factory job, and let's assume magically that he/she has the intelligence to do it, and unexpectedly did really well in school a long time ago. They're about as ideally positioned as one can be... except who's going to pay for the retraining? Who's going to hire the 40 year-old who just got into programming? Even when you have created the ideal state, the person's odds are still horrible.
I wonder what we're all going to do when humans doing app development by writing code becomes obsolete. Yes, it will happen.
> "But it's not that hard."
It wasn't for you, evidently, and it wasn't really for me either, but for a lot of people it strains at the edge of their ability to comprehend and mentally model, even with practice. That's the point I'm making: we've been born into the portion of the population that can, with some elbow grease, grok it. Not everyone is as fortunate. Even some of the ones who are, lack the educational background - in math, in logic, in whatever - to make that leap.
It took working a wildly different job, with a wildly different demographic, to really comprehend just how wide that gap is.
Even if we could successfully retrain, say, 20% of the industrial work force into knowledge-based jobs, that's still a lot of people out in the cold, and even then - who would pay for the retraining, which for a lot of people would take years? Who will bankroll their education, or even their living costs, while this transition occurs?
There are so many systemic problems with this it's mind-boggling. Suffice it to say, the notion that displaced workers from now-collapsed industries can retrain and reapply elsewhere is dramatically oversimplified. The reality is that the vast majority will never make the leap anywhere else.
Unfortunately, the rate of technological acceleration is increasing, and society is not able to keep up with even the current rate, viz. the huge numbers of unemployed workers with outdated skills. They suggest that the disruptions we have experienced over the last ten years are just a hint of the disruptions coming over the next ten years, with increasing rates of change. It is the elephant in the room for the presidential campaign, but is pretty much being ignored by both sides. They even suggest that in ten years or so, all that manufacturing in China could move back to US soil as factories become completely robotic.
With all the productivity improvements couldn't workweek hours decrease causing a rise in the number of people employed (maybe at a cost of less take home pay for workers overall)?