It's a pretty complacent article. How long until software development has good quality automation - of a kind that drastically reduces the number of jobs? I suggest it's not as far away as the author thinks.
The question is when will the demand for software be low enough that this ongoing process will reduce the number of programmer jobs.