If you stick with the same software ecosystem long enough you will collect (and improve upon) ways of solving classes of problems. These are things you can more or less reproduce without thinking too much or else build libraries around. An LLM may or may not become superior at this sort of exercise at some point, and might or might not be able to reliably save me some time typing. But these are already the boring things about programming.

So much of it is exploratory, deciding how to solve a problem from a high level, in an understandable way that actually helps the person who it’s intended to help and fits within their constraints. Will an LLM one day be able to do all of that? And how much will it cost to compute? These are the questions we don’t know the answer to yet.