I think there's a difference in using claude code at work to resolve issues or user stories which are patching existing software and already define what is trying to be solved and what the acceptance criteria is versus using claude code to build something from scratch, where you are acting as an architect.

It leaves more room for skill expression when you're making architectural decisions, defining scope, and designing the application.

I've found that adapting my thinking to how LLMs work best is a real friction point. If you're not doing that, it spits out junk. Your job just has low standards, get used to it.