This article resonates with me like no other has in years. I very recently retired after 40 years writing software because my role had evolved into a production-driven limbo. For the past decade I have scavenged and copied other peoples' code into bland cookie cutter utilities that fed, trained, ran, and summarized data mining ops. It has required not one whit of creative expression or 'flow', making my life's work as dis-engaging as that of... well... the most bland job you can imagine.
AI had nothing to do with my own loss of engagement, though certainly it won't cure what ailed me. In fact, AI promises to do to all of software development what the mechanized data mining process did to my sense of creative self-expression. It will squeeze all the fun out of it, reducing the joy of coding (and its design) to plug-and-chug, rinse, repeat.
IMHO the threat of AI to computer programming is not the loss of jobs. It's the loss of personal passionate engagement in the craft.
I've just finished a PhD in (specialised) topics around computer performance. And I have a similar feeling about the stagnant state of innovation in computer and software engineering over the last 15 years or so.
The vast majority of research is funded or incentivized in some way by the big internet companies. The big internet companies have a very narrow scope of problems that they are commercially interested in. They also have so much power and money that getting people to listen to diverse ideas and opinions is incredibly difficult, both commercially and academically because everyone somehow has to cater what they do to be in line with internet company practices.
And of course, the internet companies have found ways to industrialize their core competencies of data warehousing and analytics so that every year, fewer inputs (staff, hardware, software, data) are needed to achieve the same outputs.
I think that people are experiencing a loss of independence, and creative thinking. Not a loss of passion for the craft.