I would say any programmer has found themselves so deep in the metaphorical woods that they forgot what they were actually trying to achieve in the first place. For example, you start because your boss told you to implement a new feature, but you realized there was no module you could place the feature in, then by looking at the existing module structure you realized it had become completely unsuitable in general, then you started trying to split some and rename others, then you noticed a circular dependency would appear if you didn’t add a third module between two you just split, and now all you can remember is that you’re fixing the circular dependency. The feature you were supposed to work on is a vague memory now.

I think this is a sort of brainrot which explains why a lot of times things take unjustifiably long to implement. LLMs can help so much with stuff like this if you know how to yield one. But it can also let you more quickly hang yourself if you don’t even care about “which modules even exist” ( to continue with my example).