Suppose you're taking a math class and the instructor writes on the board:
> An eigenvector or characteristic vector is a (nonzero) vector that has its direction unchanged (or reversed) by a given linear transformation.
Do you understand eigenvectors now? Speaking for myself: certainly not. To really understand even that definition I'd need to do the calculations, play around with how they can be used in different circumstances, and allow my brain time to sit with all this new information and become comfortable with it.
I think LLM code descriptions are similar. They don't really bring understanding. Sometimes that is fine.
> An eigenvector or characteristic vector is a (nonzero) vector that has its direction unchanged (or reversed) by a given linear transformation.
Do you understand eigenvectors now? Speaking for myself: certainly not. To really understand even that definition I'd need to do the calculations, play around with how they can be used in different circumstances, and allow my brain time to sit with all this new information and become comfortable with it.
I think LLM code descriptions are similar. They don't really bring understanding. Sometimes that is fine.