I did finish the article, but for me, it was missing a discussion and review, probably manual improvement, of the code that came out of the LLM. Reverse-engineering means understanding a system that you didn't previously understand, which is still (with some degradation) possible while using an LLM.