There was a time (I believe it was somewhere in the “Roaring Twenties” a
century ago), when the profession of being a telephone switchboard operator was
growing so fast that a properly conservative projection of that growth would
conclude that everyone in the country would need to be a telephone operator
within 5 to 7 years.
Obviously that was not sustainable. Technology “had to” get better (and of
course it did). There is a bit of an analogous issue (on a somewhat more
manageable scale) with regard to programming in 2020. We have had chronic
shortages of good coders for decades now. This is partly due to the fact that the
technology of software production has NOT changed materially, and rapidly enough. Productivity tools and aids are lagging. Too much code is still written essentially the
same way that it was when C programming language was created in the Seventies
which was more than 40 years ago (gulp), soon to be 50 years.
Software as an industry is a chronic under-performer as regards productivity.
Most of the actual productivity “gains” in the last 20 years and more have been
due to the inclusion of millions of people from lower wage economies. The dismal
science of economics measures productivity as output against cost, and
programming has been “keeping up” mostly because effective wages (the costs)
for programmers have been going down, as more and more folks from third-world, relatively low-wage economies have joined the ranks of programmers. The effect of Agile and test-first and pair programming, and other new techniques, such as Scrums and sprints, have been positive, but (probably) not positive enough.
The productivity “bandwidth bottleneck” is still with us… where are the robot
coders who represent true automation?
Still just a ways too far in the future, one supposes.