I had a similar though a while ago[1]:
> the most salient quality of language models is their ability to be infinitely patient > humans have low tolerance when dealing with someone who changes their mind often, or needs something explained twenty different ways
Something I enjoy when working with language models is the ability to completely abandon days of work because I realize I am on the wrong track. This is difficult to do with humans because of the social element of sunk cost fallacy — it can be hard to let go of the work we invested our own time into.