If you’re like me, you probably like having your own personal set of notes where you can place everything you learnt — like a centralized repository of knowledge that you could refer back to whenever needed. It’s extremely helpful and allows you to refer back to anything you’ve missed or forgotten.

Leveraging AI

Recently, I had to be absent from school for an entire week. Normally, for me to best retain the information, I would jot down notes during discussion into my own personal note-taking application. However (especially considering the fact that I’m not in school), this process would be very time consuming to do asynchronously.

So, I had an idea: why not just submit the file to GPT, and allow it to create notes for me to study on?

By leveraging the capabilities of GPT-4 Vision, I could upload class slides directly to the AI, and it would analyze the content — thus creating comprehensive notes for me.

This approach has the potential to do two things.

1. Saves you time: I don’t have to manually type everything out in my note-taking application. Instead, I can just copy + paste and add personal touches (like highlighting, bolding text)

2. Become your own personal teacher: Especially due to the fact that I was not currently in school during that time, this approach allowed me to take charge of my learning process. The AI model can then explain each concept to you — where you can continue to ask unlimited follow-up questions and refinements to ensure you get the concept.

Ask and Refine

The process here is mainly what it says: ask and refine.

Simply start by asking the model to explain to you a particular topic. In this query, it is very important to add context. For example, context could refer to slides uploaded by the teacher, or you could say “explain it to me like I’m a 8th grader” to request for a simplified, more broken down approach.

Once you receive your initial response, continue refining the response by sending out more specific queries. If the AI model’s response may be too general, ask even more specific questions and provide feedback, like: “Take a deep breath first, and try again, it seems like you’re doing it wrong” or “Can you delve deeper into the implications of [topic]?”

On the surface, this is a very simplified approach, and is probably something the majority of us have already been doing! However, oftentimes, we tend to be lazy: providing a query without much context, and thus we receive an answer from ChatGPT that doesn’t make much sense. In this case, it’s always better to guide the AI through specific prompts and verbalize what it is you want it to achieve.

On a side note, NotebookLM!

Recently, I’ve discovered a whole new tool, NotebookLM. This tool seemed to go unnoticed to the general public, without Google making much press about it.

(This got me thinking: what if, the next step to efficient comprehension and learning from documents and notes are from tools like these?)

Looking on the internet, I’ve found a ton of positive reviews on Reddit (which somewhat contrast the general public opinion on Gemini, Google’s AI model). Compared to native models like ChatGPT or Claude, Google’s NotebookLM seems to excel when analyzing large documents.

And overall, this is a tool which I would want to keep exploring and would recommend you do as well. This tool can analyze long lecture recordings and research papers, and simplify and break it down for you. There’s even an option to produce a custom-made podcast from the data given!

I’ve been looking at Google’s blog, where they describe different expert tips on how to get started with NotebookLM; describing it mainly as a tool for understanding things.

It recommends uploading your recent documents to first experiment, and see its capabilities. From there, the blog recommends to start by creating a “main” notebook, which contains documents from everything you do in your daily life (like, your school curriculum, your brainstormed ideas, your past papers, etc), as well as specific project-based notebooks as needed (for example, based on a work project you might have — which allows that notebook to be personalized).

Conclusion

To conclude, AI is significantly changing the way we take down notes, and in a deeper level, the way we retain and learn information. I do believe AI tools can help you significantly with teaching and retention, especially with tools like NotebookLM simplifying large documents or long videos into a short summary — or a podcast for you to listen to.