Let's look at a historical example, coal mining: 1. A coal seam is discovered, mines are struck. Skilled miners can expect high wages, since there is a labour

Let's look at a historical example, coal mining: 1. A coal seam is discovered, mines are struck. Skilled miners can expect high wages, since there is a labour undersupply. 2. Mining companies move to adress this, both using automation, and bringing in more miners. 3. Wages plummet as a result of the above. Meanwhile, the mining company is ever more profitable as it increases its revenue. 4. Coal miners unionize, thereby stabilizing wages. So as it stands, the software industry is between 1 and 2. There was a severe shortage of labour in the 90s, as the computer industry exploded. Now, there still is, but not to the same extent. The tech companes have follwed number 2 to the letter - both in pushing programming training to governments (for much the same reasons that mining companies opened schools), and in investing in better automation - better tooling, better languages, and so on. So, all programmers have to do is to do number 4. If they're successful, programming will become an industry like mining - steady jobs that people enjoy. If they're unsuccessful, programming will become a job like retail - badly paid, unsteady, stressful.