How to use AI to learn anything faster

Not sure about you, but I learn much better when I can interact with material that I’m learning - instead of just consuming it. On top of that, I learn better in certain formats (e.g. reading instead of listening).

Problem is - depending on what we want to learn, we don’t always have a say about these factors. Maybe the material you’re trying to learn only exists as a written book, or a complicated scientific study, or a super long YouTube video.

Here’s the thing though - with AI, that really is no longer a limiting factor.

Let’s look at how we can use AI to help us learn better? I’ll go over two of my favorite tools today:

ChatGPT (shocker, I know):

Ever wanted to learn how to code? I can’t tell you how many times I’ve tried (and failed) over the years to learn basic code. I’m not trying to become a full stack developer, or know a million coding languages - but I have always felt it is important to understand the language that pretty much every tool we use is built on.

But for some reason, I could not wrap my head around the material in the way the courses presented them. It just felt so foreign to my brain that even the intro lessons were prohibitively confusing.

It wasn’t until I was messing around with some ChatGPT prompts earlier this year that I figured out why - it came down to the lack of a constant feedback loop. Having a “tutor” that can adapt to your learning style, give you constant feedback, and go at your pace is a GAME CHANGER.

So this time, when I felt the itch to learn some basic code, I turned to ChatGPT instead of the millions of courses online.

Here’s how that went: (paraphrased)
Me: Can you teach me how to code starting with Python?
ChatGPT: Sure! Let me know how much time you have, how you learn best (written tutorials, challenges and exercises, etc), your goals for learning, and what prior coding experience you have.

I responded with my answers, and it gave me a lesson plan outline. Because I had mentioned that challenges and exercises were the way I learned best, it started by giving me very basic exercises to complete to help me learn the fundamentals of the language (e.g. write xyz thing with the proper notation).

All I had to do was put my answer as a response, and it would let me know what I had right, and what still needed work. This right here was the secret sauce. Real-time, live feedback to show me exactly what I was missing.

I’d try again until I had done each component correctly, asking for clarity on why certain rules applied, and alternative ways to approach the problem.

Once we started diving into external platforms to actually deploy the code I was writing, I was able to work through a user interface I didn’t understand by screenshotting the window I was confused about, pasting it into ChatGPT, and asking exactly how to use it. This can also be done for error messages or anything else that you’re unable to navigate on your own.

Instead of being stuck banging my head against a wall because I couldn’t figure out where to find a file, or how to understand some syntax, I literally just asked - and got an answer. This frees up headspace for the actual critical thinking needed to learn.

Without boring you with too many details, this is the key point: treat it like a tutor and ask all the questions you want. Ask for more examples. Ask for clarification. Ask for a slower pace. Ask for external resources to reference. Ask for a podcast on a specific topic you’re struggling with if you’re more of an auditory listener.

Note that ChatGPT is fantastic for learning things like code - but if you have a specific resource (textbook, YouTube video, scientific study) that you’re trying to learn from, it might not be your best choice.

Instead, try: Google’s NotebookLM

Instead of pulling from a traditional LLM of general internet data, NotebookLM specializes in synthesizing information from materials you upload and converting it into the most useful formats for you to learn from.

E.g. You are taking a course that has a PDF textbook, recorded lectures, and powerpoint presentations. With NotebookLM, you could upload all of those resources and then interact with that information through a chatbot.

You might want to ask it to create a study guide, or an FAQ reference sheet, or practice questions for a test.

You can even have it generate a “podcast” episode about your content, where two AI (but remarkably human-sounding) hosts will discuss the material you have uploaded in a casual, easy to understand format. Even more bananas is the fact that you can INTERACT WITH THE AI PODCAST HOSTS and ask them questions, as if you were calling into a radio show.

And here’s the best part - unlike other LLMs, NotebookLM will not hallucinate. It will only give you answers from the sources your provided (with citations for you to reference), and if it doesn’t know an answer, it’ll just tell you.

You just gotta try it - this is an absolute cheat code to learning faster and more efficiently.

Another super cool use case I’ve seen for NotebookLM (the possibilities are truly endless) - try uploading all of your household manuals (for appliances, etc) into a notebook and asking the chatbot how to fix things as they break (it will reference the manuals you uploaded to give you an accurate answer).

So there you have it - two easy ways to start incorporating AI into your learning process. Getting instant feedback, conversion into formats that work best for you, and synthesizing large amounts of information are just some of the ways it can change the game.

Let me know what you think - have you ever tried either of these approaches for AI learning?