How I Supercharge the Feynman Technique and AI for Better Learning

Share with a fellow dilettante!

Worried AI will rot your brain? Learn how to turn ChatGPT into a learning coach with the Feynman technique and AI + spaced repetition for rock-solid retention.

Not going to lie, I love AI. At the same time, I worry about its impacts.

The promises? Exciting. The threats? Bloody scary.

In this article, I zoom in on one specific concern – that AI might become a mind-rotting crutch, dulling our ability to think for ourselves.

Here’s how I’m using AI to hopefully help me learn, using AI as both a student and mentor. I blend the classic Feynman technique of learning, so I’m still doing the heavy cognitive lifting while using ChatGPT to point out my blind spots. 

AI and Learning – A Cautionary Tale

There’s been a lot of chatter lately about how AI is dumbing us down. Students are using AI to write their papers rather than wrestling with the process – and ideas – themselves. And as for the rest of us, we’ve already started outsourcing our thinking to AI.

There’s a genuine concern about AI and how this new technology will impact our thinking. Will relying on AI lead to cognitive decline in a way that social media has eroded our attention?

Will it rot our brains?

Cognitive-offloading research (i.e, getting AI to do the thinking for us) does suggest that habitual reliance on AI correlates with weaker independent-reasoning scores, but causation is not settled yet.

On the other hand, we’ve always catastrophised what emerging technologies will do to our intelligence, from the printing press to radio, TV, the internet, and now AI.

After all, how many times have you been told that TV will rot your brain?

That doesn’t mean we should ignore AI, however. The genie is out of the bottle, and it’s important to learn how to use it as a tool to enhance our thinking rather than a crutch.

I’m still dabbling with AI. I’m no expert. As I tell my kids, I’m trying to make all the mistakes so they don’t have to. Here’s one way I’ve found AI might be able to enhance my learning, not undermine it.

The Feynman Technique for Better Learning – What is It?

Named after the Nobel-prize-winning physicist Richard Feynman, the Feynman technique is basically to teach someone what you’ve learned to solidify your learning.

So, in a nutshell: 

• teach someone what you’ve just learned in simplified language (not using jargon)
• as if they are 12 years old and have no knowledge of the subject

The idea is, if you can’t explain something in a simple way to someone else, you don’t understand the concept well enough yourself. 

When you try to teach someone, you can see the gaps in your understanding because it’s hard to explain how something works when you don’t understand or remember all the details.

So the next steps, after explaining a concept, are to:

• Review your explanation (written or spoken) and identify gaps
• Go back to your source material and learn what you need to
• Re-explain the concept, and test your understanding
• Add analogies if you can that help explain a topic in concrete ways

If it’s a maths or science concept, you might also want to work on examples or problems to check your understanding.

How I Use AI to Help Me Learn, Not Do the Work For Me

We’re all used to asking AI for answers. 

Here, we’re going to flip the script with the Feynman Technique – we’re going to tell AI what we know, and ask it to help us identify our knowledge gaps.

We don’t want answers. We want it to raise questions.

AI isn’t going to do the work for us. We’ve got to put in the work to get the results. Instead, it’s going to fulfil the role of guide or mentor. 

(Caveat: LLMs are designed to be helpful, so they will slip into explanation mode. You will need to redirect them back to working as you need.)

Here’s the prompt I use to get started:

I’m going to explain a concept as best I can. Your job is to listen and then identify gaps, vagueness, or missing pieces in my explanation. Don’t explain the topic to me — just tell me where my explanation seems thin, unclear, or incomplete so I know what to study further.

I’m currently studying a course through Coursera, and this is my process for learning the material.

First, I watch each video lecture and take handwritten notes as I watch. Research shows that people learn and retain information when they handwrite notes, although there are some caveats to that. But it’s also that that’s the system I grew up with, and it feels more natural to me anyway.

After the lecture, I put my notes away and then, after giving it the prompt above, I use the voice-to-text function on ChatGPT to tell AI what the lecture was about and everything I’ve learned or can remember from the lecture, the connections I’ve made, and the questions I have. 

I try to connect concepts to concrete things and even visualise them if I can – that helps with my understanding. 

For instance, in a typography lecture I watched recently, I learned that a 12-point font relates to the height of the metal “sorts” used in traditional typesetting. That detail clicked for me — my grandfather was a typesetter, and I could vividly imagine him at work. That kind of personal connection helps me lock a concept in. Today’s digital point is defined as exactly 1/72 inch, so physical variability across fonts is less dramatic than in hot-metal days, but I’ll forever remember this fact, I’m sure.

To be honest, once I start talking, I have a pretty good idea of where my gaps in understanding are. Ums and ahs are a pretty good indication. 

But AI helps by bullet-pointing those gaps so I can then go and rewatch the lecture, reread my notes or do some further reading outside the course to deepen my understanding.

Once I’ve filled the gaps, I rewrite my notes. The writing process itself helps me to understand the information.

Then I take a second pass, telling AI what I now remember on the topic. I do this as many times as I feel necessary until I have a full understanding of the topic.

To recap the process:

1. Watch the lecture/read the chapter
2. Take notes as I watch/read
3. Review my notes
4. Teach the topic to AI using voice mode
5. Identify gaps in my learning
6. Relearn the information, focusing on the gaps
7. Make new notes and review
8. Reteach the topic to AI

This is not the only technique I use to learn things. 

Actually practising the concept – doing maths problems, for instance – is also an important part of solidifying concepts into real-world understanding and memory. 

A combination of learning techniques can really supercharge your understanding and recall.

From Brain Dump to Action Plan: Use ChatGPT to Organise Ideas

Deepening Your Understanding with AI

If you’re following a set curriculum through school or university, they will have a pathway that leads to a deeper understanding of concepts.

However, if you’re self-learning, AI can help you deepen your understanding of a topic. 

The problem with self-directed learning is that you don’t know what you don’t know.

AI can help by asking questions for you to answer.

Here’s a prompt to get you started:

Now ask me thoughtful questions, one by one, that make me think more deeply about the concept I just explained. Assume I already know the basics, but challenge me to clarify, connect, or apply what I know in new ways. Again, don’t explain anything to me – just ask.

Then answer the questions and ask AI to push you further or identify (and fill) your knowledge gaps.

Finished learning a concept and don’t know which direction to go next?

You can also ask AI what you should learn next, based on your learning goals and outcomes.

I’ve just finished learning the basics of [topic], and I’m not sure what to learn next. Based on someone who understands the fundamentals and has [X] learning goals, what would be the next 3–5 logical or valuable directions to explore to deepen my knowledge or improve my skills? Keep it practical and progression-based. 

Spaced Repetition and Learning

Spaced repetition is important for long-term memory retention.

It’s basically where you review what you learned in gradually wider gaps – day 1, day 3, day 7, day 14, and so on, to make sure the information sticks.

You can bring AI into the mix by using the Feynman technique as part of your spaced repetition: teach AI what you know, wait a bit, teach it again, patch the holes in your understanding (yourself), and repeat until it’s rock-solid.

You could even use ChatGPT’s schedule feature as a reminder to do the spaced repetition.

I’m still experimenting with this in my learning, but here’s how I intend to use it.

To run a spaced-repetition loop with AI:

1. Seed the cycle. Tell AI everything you remember right after finishing a lecture, video, or chapter.
2. Set reminders: 1 day after, 3 days after, etc. 
3. Recall first, notes later. When a reminder pings, don’t peek at your notes. Dump a fresh, from-memory explanation into AI, and see how you go.
4. Review. Once you’ve finished, look at the gaps in your memory and understanding and study further.

Use the prompt:

List any parts of my explanation that are vague, missing, or wrong. Don’t explain the topic. Just point out the gaps so I know what to review. 

Optional extra prompt (depending on what you’re studying):

Turn the key points into 6-8 Q/A flash cards I can drop into Anki.

Rinse and repeat at each interval.

Generative AI isn’t going back into the box. It’s up to us to use it as a tool for our benefit while maintaining our capacity for deep, independent thought. 

Here’s one way I’m using it to enhance (hopefully) my learning, not to replace it. 

Do you use AI for learning? What are your tips and tricks?

The Importance of Lifelong Learning and How to Cultivate the Learning Habit

Share with a fellow dilettante!