The fastest way to learn something with AI is to stop asking a chatbot questions and start building a grounded notebook around one source. AI Central's How To Learn Anything With AI walks through the method in NotebookLM: upload the book, video, article or transcript you actually care about, optionally let the tool research around it, then generate the same material as an audio overview, a mind map, flashcards, a quiz and a slide deck.
Start with the source, not the prompt
Most people meet AI as a blank box and a blinking cursor. AI Central's How To Learn Anything With AI inverts that. The first move is not a question, it is an upload, so the tool works from the thing you want to learn rather than from whatever it happens to already know.
The ceiling is generous. AI Central notes that NotebookLM handles up to 300 inputs, and that a source can be a YouTube video, a book, a transcript, an article or a whitepaper. The upload screen also takes PDFs, plain text, Markdown, Word files, audio such as MP3s, images, Google Drive files, links, websites and text you paste in.
That constraint is the entire point, and NotebookLM says so on its own upload screen, which AI Central reproduces.
Sources let NotebookLM base its responses on the information that matters most to you.
Grounding is what separates this from asking a general chatbot to explain a book it half remembers. Every summary, every flashcard, every quiz question traces back to the file you handed over, and AI Central's example screens label each generated study guide and quiz with the number of sources behind it.
The research step is optional, until it is not
AI Central marks the second stage optional, and it is the one most people skip without thinking. NotebookLM expands what you upload by searching the web for the context your source is missing. Fast Research is described as great for quick results. Deep Research returns an in depth report and pulls new sources alongside it.
Judge it by how complete your source already is. Upload an entire book and the gaps are small, so extra material mostly adds noise. Upload one talk or one article and the missing context is usually why you were confused in the first place, which makes the deeper setting the cheapest fix available.
Eight ways to read the same material
The middle of AI Central's guide is a catalogue. Once a source is in, NotebookLM's studio can render it in eight formats, and the instruction attached to that page is a single line: match the mode to your moment. In the order AI Central presents them, the eight options are these.
- Video overview, a structured explanation built for focused sessions.
- Audio overview, a two host conversation you listen to like a podcast, so you can learn while walking, commuting or training.
- Mind map, which shows the full picture fast and makes structure legible in seconds.
- Report, which returns short answer quizzes, essay questions and key terms as clean summaries.
- Flashcards, for testing your memory instantly instead of rereading.
- Quiz, which gives immediate feedback and finds gaps before they grow.
- Infographic, for scanning the key ideas and refreshing quickly.
- Slide deck, one slide per concept, on the theory that teaching something is how you learn it.
One book, run through every mode
AI Central does not describe the formats in the abstract. It runs one book through all of them, James Clear's Atomic Habits, and shows what each mode returns.
The mind map breaks the book into central concepts, including the definition of an atomic habit, the idea that habits are the compound interest of self improvement, and the instruction to focus on systems rather than goals. It then branches into the four laws of behaviour change and the three layers of behaviour change.
The report comes back as a study guide of short answer questions, each to be answered in two or three sentences drawing exclusively from the source. The examples AI Central shows cover the aggregation of marginal gains, the mathematical difference between getting one percent better and one percent worse every day for a year, the Plateau of Latent Potential, goals versus systems, and the four steps of the habit loop.
The quiz becomes fifteen multiple choice questions with immediate feedback and a hint on each. The infographic compresses further, laying out the habit loop as cue, craving, response and reward, then pairing each law with its inversion: obvious or invisible, attractive or unattractive, easy or difficult, satisfying or unsatisfying, with habit stacking, temptation bundling and the two minute rule as tactics.
Read across those outputs and the pattern appears. It is one book at four resolutions. The mind map is the skeleton, the report is the exam, the infographic is the revision card, and the audio, a little over seven minutes in the example, is the version you absorb with your hands full.
Why the order is the actual technique
AI Central closes with a workflow rather than a feature, and the workflow is where the value sits. Start with the source upload. Pick the right mode for the moment. Reinforce with flashcards and quizzes. Finish with visuals.
The first half of that sequence produces the pleasant sensation of understanding. The second half checks whether the sensation is real. AI Central states the reason under the flashcards without hedging.
Active recall beats passive reading
That is the load bearing sentence. An audio overview on a commute is comfortable, and comfort is the failure mode of self directed study, because familiarity with an idea is easy to mistake for command of it. A quiz is uncomfortable by design, and the discomfort is the part that moves anything into memory.
The visuals landing last is deliberate too, and AI Central says why in six words.
Finish with visuals to lock everything in
A mind map read before you know anything is decoration. The same mind map read after you have been tested on the material is a memory index, because every branch finally has something attached to it.
What to do with this
The concrete version, for anyone opening NotebookLM tonight, follows AI Central's sequence rather than its feature list.
- Upload one source, not thirty. Three hundred inputs is a capacity, not a target, and a notebook built on one book answers more precisely than one built on a shelf.
- Generate the mind map first and skim it before you study, so you know the shape of what is coming.
- Send the audio overview to your phone and treat the commute as a first pass, not as the study session.
- Take the quiz before you feel ready, because the gaps you cannot see are the ones that cost you.
- Use the report's essay questions as writing prompts, answer them from memory, then check yourself against the source.
- Build the slide deck last, one slide per concept, and explain it out loud to someone who has not read the book.
One detail in AI Central's own workspace is worth copying. Its notebooks are not one per subject, they are several built from the same book, each holding a single source and each aimed at a different angle. Cheap to make, and none drifts into a junk drawer.
How many sources can I add to one notebook?
AI Central puts the ceiling at 300 inputs. In practice you will get sharper answers well below that, because a notebook grounded in one book or one talk has nothing to blur across.
Does this only work with documents, or can I use a video?
Videos work. AI Central lists YouTube videos, books, transcripts, articles and whitepapers as valid starting points, and the upload screen also accepts audio files, images, Word documents, PDFs, Markdown, plain text, links, websites and Google Drive files.
Do I have to use the research step?
No, AI Central marks it optional. It expands what you uploaded by searching the web for missing context, with a fast setting for quick results and a deep setting that returns a fuller report along with new sources. Use it when your source is a fragment, skip it when your source is complete.
Which format should I start with?
Match the mode to your moment, which is AI Central's own phrasing. Mind map when you need structure in seconds, audio overview when your hands are busy, video overview when you can sit and concentrate, flashcards and quiz when you want to know what you have retained, infographic and slide deck when you are refreshing or teaching.
Will it invent things about my book?
The design point is that it works from what you give it. The study guide in AI Central's example instructs answers to be drawn exclusively from the provided source material, and the generated reports and quizzes are labelled with the number of sources behind them. Checking a surprising answer against your own source still takes ten seconds and is still worth it.