I guess the way to think of it, from my perspective, is to picture a pyramid. At the bottom you have the least skilled jobs, which traditionally has also represented the greatest number of jobs, with specialization and number of jobs narrowing as the pyramid goes up. As robots become more and more capable, thus climbing up the pyramid, consuming more and more jobs, the number of jobs left become far more specialized than what people have traditionally had to be trained for, requiring more money and time for school, and creating more competition because there are fewer of those jobs, which will also mean that the wages will go down - supply and demand. Furthermore, if it can be demonstrated that a robot can do the job they trained for a decade or so after they secure that job, retraining is a huge undertaking, because of how specialized they are.
The problem is that robotic technologies are getting better at doing more than just the basics at a much faster rate, as mentioned in the post I referenced, and the articles therein, which is why it is a question of timeline. The introduction of robots in the workforce has been pretty minimal, and has had a big impact...but they are just getting started, in terms of what technology can accomplish. The danger is as they are able to do more and more, and cost the corporations that employ them less and less, human beings will be less and less desirable as corporate assets, due to their overhead. And since capitalism is king, there is nothing motivating companies to employ people when they can spend a lot less money on a machine.
