"Monotonous jobs will get replaced and more interesting jobs that need human creativity will come up."

This is a pretty common misconception with regard to contemporary AI. Creative jobs are already being replaced, to a degree, by generative AI—writing, illustration, design, etc. And the more significant aspect of this is that the systems that are replacing human jobs have been trained on prior human work, collected without explicit consent. Of course, consent wasn't required, as there's no legislation that covers the case of AI model pre-training, so the AI companies aren't legally "guilty" of anything. This is clearly an area where the author's perspective is 100% on-point. We need to be thinking about these things deeply before problems arise, so we're not relegated to some totally insufficient backpedaling situation down the road.

Further, although the author mentions that there are pedagogical benefits to this kind of thinking-ahead, I suspect there could also be technological benefits. Simply put, researchers engaged in broader, more holistic thinking about the technological challenges will be integrating more information into their thought, and such integration is a known hallmark of creativity and innovation.

Win-win.