TL;DR
What did @connerdelights use NotebookLM for?
A reader puts books into NotebookLM's podcast feature to get primed and decide if a book is worth reading; one PDF failed to parse beyond the cover pages, yet the AI hosts filled 26 minutes anyway, a useful warning sign to watch for. It works as a primer, but with one PDF the parser only picked up the front cover, copyright page, dedication and back cover. The hosts didn't notice and produced a 26-minute episode reading far too much into four pages.
What they did
They upload a book's PDF to NotebookLM and generate an audio overview, not as a replacement summary but to see whether the book grabs them before committing to reading it.
How it went
It works as a primer, but with one PDF the parser only picked up the front cover, copyright page, dedication and back cover. The hosts didn't notice and produced a 26-minute episode reading far too much into four pages.
Worth knowing
- Check that the source actually loaded: open it in NotebookLM and scroll to confirm the full text is there. The hosts won't tell you when they're working from almost nothing.
Try it yourself
Paste this into a chatbot, then add your own details underneath.
Before you summarize: tell me how many pages or chapters of this book you can actually read in the source, and list the chapter titles. Then, in a short overview, tell me who this book is for, its main argument, and 3 reasons someone might not enjoy it.
Original post
See @connerdelights's original post on X → (Aug 5, 2026)
Sources
- @connerdelights on X, Aug 5, 2026 — the original post