It would be cool if #ytdlp could add a feature that treats #YouTube pages as already downloaded if the v= portion of the URL is found in already-downloaded files, as sort of an adversarial hack against using video titles for A/B testing.