A real bug, a three-line fix, and a one-line rejection: what reaching for 'AI slop' as a reflex gets wrong about verification, scale, and trust in open source. It happened to me a few days ago and I wanted to use the experience to reflect on it and write something that would be more in-depth and (I hope) more useful to the open-source community, which is facing so many difficulties.