Much as I agree with the anti-AI stance on AI-generated art, applying it to code doesn't make much sense. Unlike art, there is nothing creative about writing most code. The creativity comes at an earlier stage, when you conceive what you're going to build and how - and no sane engineer would outsource that to AI. And yes, I agree that low-effort, low-skill vibe-coding produces results that stink of LLM smell, but not all AI-assisted coding is like that, and giving them the same label just feels unfair. Out of personal honesty I did mark my open-source library as "AI Assisted, Code" - starting with v0.3.0 it was essentially rebuilt with heavy AI assistance - but look at the sources, including the commit history: would you be able to tell it was AI-assisted if I hadn't disclosed it?