People really want to believe that AI is decimating software jobs, because it feels like poetic justice, but fortunately or unfortunately that narrative is nonsense (so far).

In fact, AI's labor impact so far has been concentrated on freelance jobs like translation. There's a simple reason why. Jobs are bundles of tasks, and AI automates tasks, not jobs. Most regular jobs are bundles of so many different tasks that it is fairly easy for people to adapt to automation by reskilling, upskilling, and gradually evolving the job description to focus on the parts that aren't automated.

Freelance jobs are an exception because they tend to be jobs that have already been unbundled into individual tasks (that's one of the attractions of that model for employers).

This is strongly reminiscent of Louis Hyman's argument in the book "Temp" that the main reason the internet rapidly enabled the rise of gig work is that ever since the 1970s, relatively secure wage-work jobs have been giving way to the rise of contracting and temp work.

He argues that technology doesn't usually drive changes to the nature of labor; it rather swoops in to accelerate and consolidate changes that arise from organizational decisions.