Intent is a big deal in this one. Your Python book writer intended for people to read it to learn Python.

I really don't see where intent falls into this, still-- but feel free to change the hypothetical to looking at other people's python code to learn how to use python. It still doesn't change the equation. Did I exploit the people who wrote the python code that I learned from? Does my source of learning matter when it comes to what I produce? Do you really believe that artists create new art in a total vacuum, without drawing inspiration from prior art?

Back-breaking jobs that hurt people's health should be improved with technology. A migrant worker might lost his job to a mechanical fruitpicker but he's likely bilingual and eligible for a translator job. Unless that job, which is better for health and longevity, and allows someone to stay in one place, is taken by an AI.

I am somewhat stunned by the obvious bias you seem to have against manual labor. You really think having an active job is less healthy than sitting in a chair, looking at a screen all day? (Please note: 90% of my job is sitting in a chair, staring at a screen all day.)

There was no "promise of automation". Technology was always going to take everyone's jobs-- the only change is the order it has taken it in. It was assumed that human creativity was some special thing that was so difficult to define in software that it would be towards the end when it came to getting replaced, but it turns out that we're a lot more like computers than we believe, and you can train software-- with relative ease-- to figure out how to achieve an end result without explicitly defining how.

Large companies want to reduce overhead, increase productivity, and maximize profit. I assure you there's no bias as to what kind of jobs get replaced when it comes to those goals. It just happens that creative jobs seem to be easily replaced.

Do you really, honestly, think that it's even possible to hold back a technological advance using legislation? You can already host your own LLMs and train them on whatever material you desire, to better tailor their output. That's today. Even if we assume, for sake of argument, that the law does decide that people have a "right" to control how their art is consumed. (again, very unlikely imo), that won't even slow down the people spinning up their own instances, and even if they follow the rules, how much worse do you really think the models would be using only public domain and open source training materials?