AI is currently a very effective army of junior/senior engineers. It requires the oversight of a staff level engineer with some product understanding. When properly applied it can speed up development pace ~100-200x. Things that help achieve this outcome: staff engineer reviewing, typed language (it is so much better at Go compared to Python for instance), well structured project/code, solid testing best practices.

It sounds like the author's company doesn't have this setup well.

If you use AI wrong you get AI slop :)