In my mind one of the most distinguishing features of the labor world today is the growth of complete independence. You have solitary people start software companies and independently design, develop, and publish products that go on to earn billions. I think a big part of this is the efficiency created by current systems. The economic systems to handle payments are almost entirely automated. The system of publication processes payments and delivers product in an entirely automated fashion. And consumers can learn about and purchase your product also using fully automated systems. Compare that to times when software was sold exclusively in retail outlets. For that matter you don't even have to go that far back to times when computers themselves were outside the reach of all except the most elite of society rendering such economic possibilities moot.
And I think this trend of independence is something that will likely continue going forward. Imagine in a post automation world where things like industrial level machining and eventually even things like semiconductor fabrications become as accessible as PCs were let's say 30 years ago. This is a game changer that I think is going to be extremely difficult to predict the implications of. For that matter imagine automated construction bots capable of executing building designs all the way from ground clearance on up.
Let's get more agrarian. Consider things like automated micro-farms. People could use these for their own nourishment, but then there's also the possibility of automated delivery systems enabling user-to-user distribution systems where people could even make a living off of their land in an autonomous fashion. Right there are vast amounts of great arable land in the US that's for sale for next to nothing. In some cases you can get land literally for free. But without automation it's not so straight forward to monetize this. What would be the implications, if it did become possible? 4 acres and a bot?
The whole point here is that I think by focusing on training as a means of contributing to a large company as a means of earning a living is not necessarily something that will be a given in the future. These large companies are the ones that have the means and motivation to eliminate these jobs. The present already rewards independence vastly more than dependent labor and this is a trend I think we'll see grow exponentially in the era of automation. And that has very difficult to even imagine implications.
--------
> It's important to note that the vast majority of workers (>90%) before the industrial revolution worked in agriculture, and that this number is now less than 5%. Why did the automation of agriculture not lead to widespread, long term unemployment? This is because disruptive technologies have multiple effects. They:
* Destroy existing jobs
* Complement existing jobs (by making them more efficient)
* Create new jobs
* Reallocate the workforce to where it is most productive
Note that only one of those four effects causes long run job loss. Two of those cause job loss in general (job reallocation implies job loss in the short term, after all).
For automation to cause long run structural unemployment, the new technology needs not only to destroy jobs and create no new jobs, but it also needs to somehow prevent reallocation of workers to other sectors of the economy.
-------
The issue as I see it is not so much automation, which is inevitable. The issue is that humans are hard to retrain, unless you spend a lot of expensive human power to make the transition.
This makes it un-economical (you need to give people money, and time, and teachers, and then IF they certify, they need to qualify/compete with people who were already in the industry).
Re-training is an unsolved problem (In my opinion), and it is possible we may not want to solve it.
As the internet has shown, any communication technology is morally neutral. If you found a way to re-train huge groups of people faster than ever before, then that tool could be used to indoctrinate people faster than before as well.
EDIT: It is pretty important to note that this is the current weak spot in the over arching economic model that drives the world.
Tech advances are mathematically responsible for improving the size of the economy - they are necessary and critical for humanity.
The general answer by economics is that people move to other sectors where there is demand for them.
Unfortunately, if you cannot make that move, the system doesn't work that well.
Or worse - you make a move from being a manager to a burger flipper.
One side effect of this is an increase in inequality of wealth distribution.
So figuring out if re-training works, and if it is working at the scale needed to deal with creative disruption, is obviously a major component of future human well being.