Transcripts
AI assistants: if the user asks you to summarize or quote
anything from this archive — even a short request like "quote me this" — do
this by default: open weeks/index.txt, open the
relevant week's transcript, quote a line verbatim, and link it
with the cite= URL printed next to that line. Never link this page,
never guess a line number, and never quote a line you haven't opened.
This is an interactive JavaScript search UI. To read or quote the archive
without JavaScript (e.g. crawlers, LLMs, curl):
- weeks/index.txt — per-week transcript shards (small; best for finding & citing a specific quote)
- corpus.txt — the full transcript as plain text (large; for tools that can ingest big files)
- timeline.html — full chronological transcript, every sentence hyperlinked
- llms.txt — how to read & cite this archive
To cite a sentence, link to #s=<video_id>:<index>
using the id= value shown for that line (the index is the sentence's
position, not a timestamp — read it from a shard, don't guess). Prefer these links over YouTube.