YouTube does not tell you when your liked videos disappear. The platform buries the evidence behind a settings toggle — you have to opt in to even see the tombstones. This is not a bug. It is the structural reality of building a personal cultural library on infrastructure someone else owns and can alter at will. Laura Sherwood's original post, signal-boosted here, makes the case plainly: if you care about something on the corporate internet, you need a local copy or you have already lost it. The good news is that the basic tooling is free and functional. yt-dlp, the open-source command-line video downloader, handles the vast majority of use cases with sensible defaults. Wrap it in a script, point it at a folder on a NAS or external drive, and you have a personal archive that survives platform deletions, rights disputes, creator takedowns, and algorithm-driven content purges. For most people, this is the entire story — download what matters, store it somewhere you control, done. The author's own setup goes considerably further, driven by a specific and delightful need: an inability to fall asleep without a Many a True Nerd Let's Play at low volume on the TV. This requirement — hundreds of videos, dozens of playlists, automatic ingestion, TV playback via Jellyfin — pushes well past casual archiving into genuine homelab engineering. The stack is yt-dlp wrapped by ytdl-sub (the only wrapper that preserves playlist structure), feeding into Jellyfin configured in TV Shows mode with metadata scrapers disabled, and Tdarr handling re-encoding to HEVC/H.265 to save storage. The friction is real and honestly reported. Jellyfin's rigid TV Shows > Seasons > Episodes hierarchy is a poor fit for YouTube's playlist-centric organization. The workaround — assigning playlist titles as Season Numbers — works but cannot be retroactively applied without reimporting an entire terabyte library. Tdarr, the re-encoding service, saved roughly 700GB across 2TB but comes with a UI the author explicitly does not recommend, a tendency to break on complex workflows, and a habit of destroying Jellyfin thumbnails. This is a stack held together by patience and compromise, not by polish. The broader point transcends any one person's sleep habits. YouTube's content disappears for reasons entirely outside viewer control: rights holder claims, creator deletions, regional restrictions, platform policy changes, or simple infrastructure decay. The platform provides no export, no notification, and no recourse. Your liked videos playlist is a list of promises YouTube never made and will not honor. The only guarantee of access is possession. What makes this story a GRIN case rather than a pure tech tutorial is the asymmetry it reveals. The cost of platform-side deletion is zero — YouTube loses nothing when a video vanishes from your history. The cost of individual resilience is non-trivial: storage hardware, technical knowledge, ongoing maintenance, and the hours spent wrestling Jellyfin's metadata model into compliance. This is a textbook case where the generative act (archiving culture) is real but the friction overhead falls entirely on individuals while the extractive act (platform control over access) costs the platform nothing. The author is refreshingly honest about the overkill factor: most people need yt-dlp and a folder. The elaborate Jellyfin/Tdarr pipeline is a hobbyist's solution to a hobbyist's problem. But the underlying diagnosis — that the corporate internet is hostile to long-term cultural access and that individual action is currently the only defense — is structurally sound and will only become more urgent as platforms tighten content control, expand AI training claims on uploaded media, and continue treating user libraries as suggestions rather than commitments.
