This is the first thing that I used LLMs on. Not code generation, but parser and tooling to gain understanding. Also saves resources in the long run.
One of my favorite uses for Claude Code is to point it at a section of seriously badly written code with undecipherable symbol names, over the top cyclomatic complexity etc and just ask it to make the code readable.
One of my favorite uses for Claude Code is to point it at a section of seriously badly written code with undecipherable symbol names, over the top cyclomatic complexity etc and just ask it to make the code readable.