I doubt they'll change the general amount of time and energy required to train human programmers all that much, which is something this article also illustrates. E.g.
> "After a day of being focused on getting "the tedious stuff" right (using the right API, checking the right errors, properly calling API X, implementing the unit tests and mocks for Y, writing an API wrapper for Z)..."
Programmers will still have to learn the how, when and why of all of the above steps in that and many other standard programming pipelines, and that will take about the same time to master as before. The list of most-commonly-used-skills might change, but that's nothing new - fewer programmers now regularly use low-level bit-flipping tricks in their daily job, for example.
People doing computation work could become more like carpenters who usually work with power tools and prefabbed structural units, with niches for custom construction in unusual scenarios. Carpenters themselves are unlikely to be replaced by autonomous robots anytime soon.
> "After a day of being focused on getting "the tedious stuff" right (using the right API, checking the right errors, properly calling API X, implementing the unit tests and mocks for Y, writing an API wrapper for Z)..."
Programmers will still have to learn the how, when and why of all of the above steps in that and many other standard programming pipelines, and that will take about the same time to master as before. The list of most-commonly-used-skills might change, but that's nothing new - fewer programmers now regularly use low-level bit-flipping tricks in their daily job, for example.
People doing computation work could become more like carpenters who usually work with power tools and prefabbed structural units, with niches for custom construction in unusual scenarios. Carpenters themselves are unlikely to be replaced by autonomous robots anytime soon.