Why academics should self publish textbooks

I work as a production editor for a uni press--mainly academics writing for other academics, not so much 'textbooks' in the sense of being mainly for student use in classrooms. I like to think the books are better after we work on them, but it is also true that schedules mean sometimes not as much as should be. Some of these guys cannot write their way out of a paper bag. One key difference between our market and the textbook market is that most of our books will stay in print for a long time without changes. A book like Data Science for Crime Analysis with Python may need to be updated frequently, a biography of a long-dead literary figure, no so much. Sales of such a book are probably less dependent on reviews and award notices. The more specialized your book is, the more frequently it needs to be revised, the more benefit there is to self-publishing.

But I would say, don't use AI for copyediting. Hire a pro with some topical knowledge.

We wrote a book for Calculus, that follows the sylabus of our course here in Argentina in the first year of the university. If people use LaTeX and have more formulas than images, then it's an easy. When people use Word, it's harder because someone has to reapply all the styles to get a consistent look. Images are very hard, both in Word and LaTeX. IIUC the author wrote a book for programming, so I think it's fine, but a book for biology would be a nightmare.

By all accounts I've heard, traditional textbooks have become a profit-maximizing racket. So "yes", obviously...

OTOH, self-publishing requires some serious time and a skill set, and most modern academics are already working horrible hours for crap job security and pay.

Not mentioned in the article - co-authors (spread the load), departmental support ("wouldn't it be nice to have a textbook that actually fit our syllabus for CS 473?"), and other tricks.