Unlike some other comments, I don't think there will be a dystopian future where humans will become cognitively switched off and everything will be delegated to machines. (More likely: extinction of the human race?)

Anyway, I work as a programmer, and at work I see a lot of people massively overusing LLMs to generate user stories or basically write functional specifications.

This honestly hurts a bit, because I can see the quality level of my environment going down, but in the end I'm often the one fixing the problems that come from it.

I just think this is a transition phase, and in the medium-term future there will only be room for people who, AI-assisted ofc, know how to apply logic in the best possible way.