6 tips I always give NotebookLM users just starting out

Published 1 minute ago

A seasoned mechanical design engineer turned tech reporter and reviewer, Chandraveer brings more than four years of consumer tech journalism experience to the table, with bylines at Android Police and iPhoneHacks. He’s written about everything from UI and UX changes across various apps to emerging software and AI trends. At XDA, he showcases his versatility in the tech reporting space with articles rooted in personal experiences and frustrations.

Chandraveer’s passion for consumer tech spills over into hobbies such as mechanical keyboards, photography, and 3D printing. With an academic background in design and manufacturing, his insatiable curiosity spans beyond the digital domain. Chandraveer’s downtime is an eclectic mix of reading fiction, practicing callig…

Published 1 minute ago

Looking back at the blur that was 2025, it was a banner year for artificial intelligence. We saw models get faster, "agents" become more than just a buzzword, and multimodal capabilities explode. Yet, amidst the noise, Google’s NotebookLM remains largely unrivaled in its specific niche. The second half of the year also brought several important changes for users, such as Gemini 3 support, Infographics, Slide Decks, and Deep Research that scours the web so we don’t need to. The core reason for this tool’s success over all others remains unchanged: an AI fiercely loyal to the supplied documents, and nothing else.

It’s been fascinating to watch NotebookLM seep into mainstream pop culture this year. Millions of people who wouldn’t touch a coding assistant are discovering the joy of using AI with user-defined source data and limitations. It’s the "second brain" concept finally realized without the headache of manual tagging. However, the gentle learning curve still needs smoothing. There are a few things I sorely wish someone had told me when I first dragged my PDFs into the UI. So here are a few tips I give everyone as they start with this powerhouse AI.

AI is scatterbrained if you are

It is painfully easy to treat a new notebook like a digital junk drawer. You see the "Add Source" button, and the temptation to dump every PDF, Google Docs, and random website you’ve bookmarked in the last five years is strong. I did that to one notebook to replace Pocket, but new users don’t have to. Before you create a notebook, define its job. Is it for planning a trip? Learning a new coding language? Managing a project? Your purpose dictates which tools from the Studio sidebar — like the Timeline or the Briefing Doc — will actually be useful. If you just dump random data, those tools will generate confused, disjointed nonsense.

For some users, it might help to organize the digital junk drawer as it is migrated to the new AI tool. I use a Steven Johnson trick also endorsed by my colleague, Mahnoor Faisal. Creating a specific "Everything" notebook on day one makes for a nice sandbox. I can throw in random thoughts, interesting articles, and half-baked ideas for serendipitous discovery as the AI tries to connect recipes to historical factoids. These connections sometimes serve as jumping-off points for new notebooks.

Play your sources smartly

And it keeps your ambition in check, too

NotebookLM is grounded, and only as smart as the information you give it. That’s a big reason why I’d say we should be ruthless with curation. If I’m researching a complex topic, I don’t gamble on the usefulness of discovered sources. I’ll often consult other AIs, like Perplexity or Google Search in AI Mode, to find the highest-quality, most authoritative white papers or PDFs available before I upload them to NotebookLM.

You also need to keep an eye on the source limit. While Google has been generous, the 50-source cap per notebook can creep up on you faster than you expect. If you have twenty tiny text files related to the same sub-topic, do yourself a favor and merge them into a single document before uploading. It saves you a few slots and may help the model understand that those pieces of information are directly related.

The AI doesn’t care, but it keeps you sane

For months after hopping on NotebookLM, I never bothered renaming files like "scan_2025_01_12.jpg" or "Chapter 4_final_v2.pdf" after uploading them to the AI. Sure, Gemini doesn’t care what your files are called, but once you’re three weeks deep into a project and trying to toggle specific sources on or off to narrow down your query, staring at a list of ambiguous filenames is a nightmare.

Give your sources clear, descriptive names so you can quickly glance at the list and toggle off those that are irrelevant to your current question. The AI’s responses become way more useful. Moreover, this method helps identify sources with similar information that you could combine if the free tier’s limits become a hindrance.

Master the Podcast-ification of your notebook

Passive consumption saves time

Reading is an effective mode of learning because you’re actively involved and typically unable to multitask simultaneously. However, some of us learn better through passive, repetitive consumption than through active, focused reading. NotebookLM saves the day here with improved Audio Overviews. It became a genuine productivity tool in 2025 with a prompt box so you can steer the AI hosts’ conversation, control the length, and overall duration or style of the audio stream.

Additionally, I encourage users to interrupt the AI for clarifications as needed—a convenience afforded only to subscribers of human podcasters. This resembles free active tutoring and helps you absorb complex information when you physically can’t sit down and read.

Google Drive sync

Keep files ready, always

I usually create new notebooks while waiting at cafés or taking a cab home, so I maximize the time spent actually using the notebook for learning instead of setting it up. Sure, I add web sources on the go and get home to add downloaded files from my desktop, but it helps to keep research papers and other stuff synced with Google Drive.

Drive integrates directly with NotebookLM, and I’ve configured a folder on my PC to sync files to the cloud periodically. Anything I might need for an existing or new notebook is downloaded straight to this folder by default. The swift change of download location is a tiny detail that saves me the frustration of missing important sources in NotebookLM on the move. Best of all, if you make significant edits to the original doc, you can simply re-sync the source in NotebookLM to get the latest Drive copy, keeping AI in step with real work.

NotebookLM can now search your Google Drive for sources instantly

This feature might just save you from ever digging you through Google Drive.

Just get started already

NotebookLM is not an intimidating tool. The guided workflow is one of the friendliest "onboarding" experiences in the current AI landscape. You can extract value from interactions five minutes into usage, but these small tips add up quickly to separate power users from casual tinkerers. Take it from the collective experience of those of us who have spent way too many hours down this rabbit hole: Small details like custom-named sources and tailored Audio Overviews are worth the effort.

Always have a purpose in mind