No, it is really useful to know how a technology works. LLMs work by predicting next tokens.

It is _amazing_ the utility they have given that that is what they are and they are highly useful but suggesting solutions that ignore they are spicy auto-complete is counterproductive on many different levels.