29 prompts that beat “summarize”

Run this before anything else on messy material — transcripts, mixed PDFs, scattered notes. The index becomes the map for every later question.

The follow-up to the index. “Explain” keeps structure and nuance that “summarize” throws away; going topic-by-topic forces full coverage.

The heavyweight. A research-grade answer with explicit rules, so the model can't paper over gaps or drift outside the sources.

Sources rarely agree as much as a summary implies. This surfaces the disagreements a polite answer smooths over.

Finds what the material takes for granted — the fastest way to spot where an argument is weakest.

One answer invites confirmation bias. Three deliberately different readings of the same question show you where the conclusion is actually solid.

The simple explanation is the test; the list of what it had to leave out is where the learning happens.

Twenty questions across three difficulty tiers, with answers kept at the end so you can actually test yourself.

One pass that extracts every term a newcomer would stumble on. Save the output as a note — or as a new source.

Write what you remember, then let the sources grade you. Harsher and more useful than re-reading the chapter.

Kills the warm-up minute and the both-sides framing. The single most repeated community fix for Audio Overviews.

Makes the hosts engage the strongest objections instead of narrating agreement. Good for anything you have to defend later.

The episode gets dramatically more useful when it's for one person in one situation instead of a general audience.

Turns the podcast into retrieval practice: one host quizzes the other on the most testable points.

Typography-only output: no stock illustrations, no icon noise. Works for technical and business material alike.

For anything with a chronology — history material, project retrospectives, case timelines.

When the sources compare options, force the output into a grid instead of paragraphs-as-pictures.

Set once per notebook. Every answer arrives pre-wired for verification, and “not in the sources” replaces confident guessing.

Generate a one-page glossary of your project's acronyms and names, then add the output back as its own source. Cross-references stop misfiring.

Belt-and-braces with the source checkboxes: names the allowed sources inside the prompt so the answer can't quietly borrow from the rest.

Nudges the model's defaults from quick-and-shallow to long-and-enumerated for the whole notebook.

The after-meeting artifact that meetings never produce themselves: decisions, owners, deadlines, contradictions.

One row per paper, blanks left blank. The starting table for any lit review or state-of-the-art section.

Rewrites findings for a non-specialist without losing the citations you'll need when they ask “says who?”

A full academic-review harness: rules, process and a seven-part output structure. The most complete research prompt the community has shared.

Doesn't summarize — it red-teams your notes and designs five custom prompts to expose blind spots. The author runs it in Gemini with the notebook attached; it also works pasted into NotebookLM chat.

Three lenses on the same question — analytical, creative, skeptical — to break confirmation bias before an important decision.

One line in the notebook's custom settings that reliably makes answers longer, deeper and more carefully reasoned.

Fill a notebook with prompt-engineering material first, then use this to have it write high-quality prompts for any goal.