Teaching Philosophy

Learning by Inventing

Guide learners through a sequence of simple problems until they discover the idea, formula or algorithm themselves. Confidence follows understanding. Understanding follows discovery.

Read essays on this →

The Core Idea

Traditional education often begins by giving learners a formula, a definition or an algorithm, and then asking them to apply it. The learner receives the idea fully formed and is asked to accept it on faith.

I believe a more powerful approach is different. Instead of beginning with the answer, give learners a carefully designed sequence of simple problems. As they work through these problems, they begin to see a pattern. They start to ask: why does this keep happening? What is going on here? And gradually, if the sequence is designed well, they construct the idea for themselves.

This is what I mean by Learning by Inventing.

When a learner discovers something, it becomes genuinely theirs. They understand not just how it works, but why it must work. They remember it because they built it. And most importantly, they develop confidence: the sense that they are capable of original thought.

A Simple Example

Instead of beginning with the Pythagorean theorem, guide a child through drawing squares on the sides of right triangles. Ask them to count the areas. Ask them to cut and rearrange the pieces. Let them play until they begin to notice something surprising.

The moment the child says: wait, the big square always equals the two smaller squares together — that is a completely different experience from being told it.

The same principle applies to machine learning. Instead of presenting a neural network as a mathematical structure, give learners progressively harder prediction problems. Ask them to think: how would I improve this prediction? They will start adjusting weights, noticing errors, updating their approach. Without realising it, they are building the core intuition of gradient descent.

The Five Principles

Principle 01
Start with the problem, not the solution

Design a simple, concrete problem that the learner can engage with immediately. The formula comes later. The engagement comes first.

Principle 02
Build a sequence toward discovery

Each problem should be a small step. Gradually increase complexity until the pattern becomes visible. Patience here is not optional — it is the method.

Principle 03
Let the learner feel stuck

Productive struggle is not a failure of teaching. It is the moment before understanding. A teacher who prevents all struggle also prevents discovery.

Principle 04
Celebrate the moment of realisation

When the learner says "I see it!" — stop. Let them articulate it. Let them own it. This moment of recognition is more valuable than any lecture.

Principle 05
Build confidence, not just knowledge

The goal is not a learner who knows the Pythagorean theorem. The goal is a learner who believes they are capable of discovering things. That belief changes everything.

A Learner's Story

One of the most meaningful moments of my teaching career came when a learner — a professional who had spent years believing she was simply not good at mathematics — told me something that I have not forgotten.

After working through a set of carefully designed exercises, she sent me a message to say that she had gained enough confidence to start teaching mathematics to her young niece. She was not just learning the material. She was beginning to see herself differently.

This is what good teaching can do. Not the transfer of content, but the transformation of a person's relationship with their own capacity to think.

The AI Teacher

I am deeply interested in building an AI teacher based on this philosophy.

Not an AI that answers questions, but one that asks them. An AI that interacts through voice and a visual whiteboard, guiding learners through a designed sequence of discoveries. One that knows when to be quiet and let the learner struggle productively. One that recognises the moment of discovery and helps the learner articulate what they have just understood.

This is technically and pedagogically hard. But I believe it is one of the most important things technology could contribute to human learning.

For Educators and Learners

If you are a teacher or educator interested in this approach, I would be glad to exchange ideas. If you are a learner who wants to explore these ideas in practice, many of the courses and materials I have developed at CloudxLab are built on this philosophy.

And if you are a researcher or technologist interested in building educational AI that teaches through discovery rather than content delivery, I would love to connect.

Explore this further

Essays on this topic Visit CloudxLab Let's talk