Rykerr AI Search — TODO / Backlog
Rykerr AI Search — TODO / Backlog
Running list of known-good-enough-for-now shortcuts we want to eventually replace with the real thing. Add to this as new ones come up.
Search indexer
- Auto-detect show-notes-page -> YouTube-video pairing. Right now
(as of 2026-09-27,
link_show_notes.py+curation/show_notes_videos.txt) a new YouTube-notes page needs one line added tocuration/show_notes_videos.txtby hand before it’ll defer to its video. Podcasts already auto-detect fine, via the RSS feed. Auto-detecting the YouTube case the same way needs the RAW page HTML (to find a self-referencing youtube.com/youtu.be link like show-notes-pedi-video.md has) – the already-extracted chunk text strips hrefs down to anchor wording, so it can’t be recovered from what’s already indexed. Likely needs a small hook in crawl.py itself (same shape as podcast_ingest’s own show-notes detection), not just a post-process pass over the finished DB like link_show_notes.py is today. Low priority while there are only ~6 of these pages total.