Have you ever talked to an engineer about their PR and you can tell they have no idea what is in it?
Essentially, AI gives you a big red button that says, "Hey, you can turn your brain off for a bit. I can do all the thinking. Don’t worry about it!" As an added bonus, it can think 100x faster than you. And here is the best part - you are rewarded for doing stuff, not for actually thinking about stuff.
It can be very easy to fall into the habit of spamming that button. To make matters worse, there might be an expectation of a shipping velocity that is only possible if you are spamming that button.
I have touched on this concept of taking on a knowledge debt before. If you are not curious, you will very easily fall into a trap of not bothering to learn what you are actually shipping. There is some irony here since AI can get you addicted to not thinking but it is actually a great teacher. Also it honestly sucks working with the type of person who just outsources all their thinking to LLMs. They put up a PR, and then when you go to talk to them about it, you can tell they have no idea what is actually in there. Don't be that person.
AI rewards curiosity, punishes laziness over the long run, but encourages laziness in the short run.
Obviously LLMs benefit high agency, self-starter people. But there are diminishing returns if that person is not naturally curious.
So word to the wise - stay curious and be sure to press that button to help teach you something, not to just have it do things for you.