Human slop code existed way before we had auto-complete in our IDEs, way before Copilot start writing small functions and thus bad code is not because of LLM. AI slop is no different than Human slop, it uses a shit ton of ressource to generate it and is faster at getting a working output but bad codebase is not new.