An MIT paper on LLM use recently caught my attention researchers found that regular use for simple tasks can measurably reduce cognitive engagement over time. There's a good discussion of it here:
So I went back through my own ChatGPT history out of curiosity.
A lot of "rewrite this sentence," "what's the word for X," "give me a synonym for Y." Things I already knew or would've figured out in a minute or two. I wasn't using AI as a tool. I was using it as a reflex.
Same with feeds opening Twitter with no intent, scrolling, closing, opening again with no memory of what I was looking for. My thumb was running the show.