It is 3 AM, I am able to cram for an examination that’s precisely 5 hours away, and the very first thing I do is not crack open my textbook or the lecture slides. I do not hop onto YouTube hoping to seek out that one video that magically covers all the things I would like. In basic 2026 vogue, I do not head to ChatGPT both and ask it to clarify all the things to me like I am 5. As a substitute, the primary software I flip to earlier than all else is Google’s NotebookLM. The software is Google’s AI analysis assistant, and what makes it stand out from most instruments is that it is designed that will help you perceive and work with your individual content material.
You feed it your lecture notes, slides, readings, YouTube sources, and primarily all of the content material you have got associated to what you need to check. NotebookLM then works solely inside that materials. You possibly can ask it any questions, and you will not have to fret about hallucinated information pulled from the web or generic explanations that miss the purpose of how your course really teaches the subject. Nonetheless, NotebookLM is barely as helpful as what you ask of it. Prompting stays an enormous a part of the standard of responses you get from any AI software, and NotebookLM is definitely not an exception. I have been counting on it for finding out since its Google Labs experiment days, and listed here are a couple of prompts that flip it right into a full-fledged examine system.
Getting the total image earlier than I begin finding out
Ctrl+F in your complete course
Even in the event you’re sitting in a category the place you perceive completely nothing the professor says, some a part of it at all times sticks someplace inside your unconscious. While you come again to the idea by yourself, you may really feel a faint sense of familiarity that makes studying it the second time round considerably simpler.
Earlier than correctly starting a examine session inside NotebookLM, I exploit a immediate to copy that very same impact. I ask NotebookLM to generate a fast, skimmable rundown of each idea I have to know for my examination. I sometimes add the course define, all lecture slides, and generally, even the textbook. That is the precise immediate I exploit:
Primarily based on the course define, lecture slides, and textbook I’ve uploaded, give me a whole listing of each key idea, idea, and time period I have to know — organized by matter or module. Hold every rationalization to 1 or two sentences max. I do not want deep dives proper now, only a high-level rundown I can skim to see all the things at a look.
Now, the aim right here is not to memorize the whole output it offers and name it a day. As a substitute, it is simply to learn it effectively sufficient and provides my mind a map of what is forward. This fashion, after I dive deeper into every matter, nothing actually catches me utterly off guard.
The 5 important questions immediate
The “show you perceive this” immediate
This can be a immediate I’ve talked about earlier than, however I wished to spotlight it once more since I exploit a barely modified model of it throughout my examine periods. The 5 important query immediate went viral on X and Reddit, and it is designed to flip the standard NotebookLM course of: as a substitute of you asking it questions, it asks you. The immediate nudges NotebookLM to research all of your sources, determine the core concepts throughout the content material you’ve got uploaded, after which strip it down to 5 important questions.
Provided that the top level of quite a lot of examine periods is commonly an examination (and exams encompass, effectively, questions and solely questions), this immediate is ideal for this use case because it mirrors how instructors construction data. It is also a superb solution to acknowledge precisely the place your blind spots are. In the event you breeze by means of 4 questions however utterly clean on the fifth, you’ve got simply recognized what wants extra consideration than the remaining. Whereas I shared the generic model of the immediate in that article, this is the tailored model I exploit for finding out:
Evaluate all uploaded supplies and generate 5 important questions that seize the core which means. Concentrate on: core subjects and definitions, key ideas emphasised, relationships between ideas, and sensible purposes talked about.
The Feynman Approach, NotebookLM version
Flip NotebookLM into your worst scholar
You would possibly’ve seen that it is considerably tougher to neglect an idea you’ve got taught somebody. There is a cause for that. While you’re explaining one thing to a different individual, you are compelled to arrange the data in a means that truly is sensible to them, fill within the gaps you did not understand have been there, and simplify issues all the way down to their core. That you must adapt your rationalization to their understanding stage, and after an individual asks you a bunch of questions you hadn’t thought of, you stroll away understanding the fabric much better than earlier than you began explaining it. That is primarily what the Feynman Approach is constructed on.
Associated
I hooked Obsidian to a neighborhood LLM and it beats NotebookLM at its personal recreation
My notes now speak again and it’s terrifyingly helpful.
The approach is constructed round the concept that instructing is without doubt one of the best methods to be taught. The everyday means to make use of the approach is to seize a good friend (or a stuffed toy, simply as efficient and much much less judgmental) and clarify the idea to them as merely as you possibly can. However the issue with buddies is that they do not at all times ask the appropriate follow-up questions, and the issue with stuffed toys is that they do not ask any. So, why not flip NotebookLM into that scholar as a substitute?
NotebookLM has a function that permits you to set a customized persona for the way it ought to act and reply to you. You will discover this by opening an current pocket book or creating a brand new one and hitting the Configure Pocket book setting within the Chat panel. Then, hit the Customized button below the Outline your dialog type header. Now, you possibly can describe precisely the way you’d just like the software to behave and reply to you. Provided that I would like it to behave as a scholar on this case, I exploit a immediate alongside the traces of:
Flip our roles. You’re the scholar. I’m the instructor. Subject I am instructing you: [topic]. My stage: [undergraduate/master’s/self-taught]. How laborious I would like you to be: [curious beginner / sharp student who asks follow-ups / skeptical professor who tests every claim].
And then you definitely start finding out explaining! Relying on the way you customise the immediate, NotebookLM will reply in another way to your explanations. A curious newbie would possibly ask you to make clear a time period or give an easier instance. A pointy scholar will press you on the main points and ask how two ideas join. A skeptical professor will flat-out problem your reasoning and power you to again up each declare. This is without doubt one of the best prompts I’ve used for my examine periods, and I am unable to suggest it sufficient.
These prompts assist you really use your notes
I do know quite a lot of college students who check out NotebookLM for finding out and stop utilizing it in hours as a result of they deal with it like every other chatbot. They simply ask query after query and stroll away unimpressed. NotebookLM is not the form of software you simply open and hope for the very best. The worth is at all times in the way you direct it. In the event you’ve been sleeping on NotebookLM, or tried it as soon as and gave up, give these prompts a shot earlier than your subsequent examination. Your 3 AM self will thanks, belief me.
