AI Building Blocks: The Art of Taking Apart to Build Better

The true power of artificial intelligence lies not in treating it as a monolithic oracle, but in breaking down our interactions into their fundamental components, examining each piece, and rebuilding them into something greater than the sum of their parts.

This is part of my special series on AI in Education where I guide you through transformative principles that are reshaping how we interact with artificial intelligence and design learning experiences.

Last month, I shared insights about the importance of viewing AI as a collaborative programmable partner rather than just a tool.

A Late Night Revelation

It was 2 AM on a Tuesday when this idea struck me. I'd spent countless hours working with various LLMs, trying to optimise my prompts and improve my results. But something wasn't clicking. The responses, while good, felt mechanical, and predictable. That's when I realised - we've been approaching this all wrong.

The fundamental question isn't "How do we get better responses from AI?" but rather "How do we break down our interactions into their most basic elements and rebuild them in more meaningful ways?"

This question is incredibly important because it shifts our focus from output to process. When we understand the components of effective AI interaction, we feel more in control, more creative, and more capable of achieving breakthrough results.

But often along the way, we get caught up in treating AI like a magical black box. Or maybe we fall into the trap of using the same patterns repeatedly, expecting different results. Maybe you've found yourself copying and pasting prompts from online guides, hoping they'll work for your specific needs.

And that leaves a lot of passionate educators and innovators feeling frustrated and limited.

Take a moment to think about your last significant interaction with an AI system. What were you trying to achieve? How did you approach it?

As a follow-up, consider how that interaction might have been different if you'd broken it down into smaller, more focused components.

The Fractal Nature of AI Interaction

Every interaction with AI contains multiple layers of complexity. There are moments of clarity when everything flows perfectly. This is especially true for those of you who've spent time learning to craft effective prompts.

But eventually, every person reaches a plateau—whether it's after weeks or months—when they begin to question their approach.

Researchers and practitioners in other fields do this too. Consider how a chef breaks down a complex dish into its component flavors and techniques, or how a musician deconstructs a piece to understand its fundamental elements.

"Am I really getting the most out of these tools?"

"Why do some interactions work brilliantly while others fall flat?"

And one of two things can happen in that moment. Either you continue with your current approach, accepting the limitations...

Or you step back and reconsider everything from first principles.

The Power of Recursive Decomposition

This isn't something a lot of people talk about, and yet it's incredibly important.

And you might be thinking, "Isn't this making things needlessly complicated?"

You're not alone. Let me share something from my own experience. When I first started exploring this approach, I was sceptical. It seemed like overkill. But then I applied it to designing a lesson on photosynthesis.

Instead of just breaking down the concept into its basic components, I broke down each component into ten parts, and then each of those parts into ten more. The results were astounding. I discovered connections and teaching opportunities I'd never considered before.

About 80% of my initial assumptions about what made an effective lesson were challenged and improved through this process.

What do you have if you don't have this level of understanding? Surface-level interactions that barely scratch the surface of what's possible.

You see, you need to embrace the complexity before you can find the simplicity on the other side.

Even if it seems overwhelming at first, you have to trust the process...

When people don't take the time to break things down properly, they end up with fragmented understanding, instead of deep, interconnected knowledge.

Practical Implementation

So, what can you do?

Start with a single interaction or lesson plan. Break it down into ten distinct components. For each component, ask:

1. What is its essential purpose?
2. How does it connect to other components?
3. What assumptions am I making about it?
4. How could it be improved?
5. What would happen if I removed it?
6. How does it affect the emotional experience?
7. What variations are possible?
8. How does it build on previous knowledge?
9. What misconceptions might it address?
10. How does it set up future learning?

This allows you to see patterns you might have missed while developing a deeper understanding of your own thought process. And it often reveals opportunities for innovation you hadn't considered.

The Reconstruction Phase

Another option you might consider is approaching reconstruction as a creative act rather than just reassembly.

This option allows you to maintain rigour but also introduces an element of artistry to the process.

Think of it like a building with recycled bricks. The same pieces can create countless different structures.

This might not be the right choice for everyone, but in my work with educators, I've seen it transform how they approach both lesson planning and AI interaction.

As an educator and researcher, you gain a deeper understanding, create more effective learning experiences, and develop a more intuitive grasp of how AI can support your goals.

We spend so much of our lives trying to optimise our use of technology; we owe it to ourselves to ask the right questions like...

How can we make our interactions with AI more meaningful?

What patterns emerge when we examine our approach at different scales?

How can we rebuild our understanding in ways that serve our goals better?

And whether you're excited about these possibilities or feeling overwhelmed, taking time to break down and rebuild your approach is something everyone should experience.

It's about growth through understanding.

Here's to your journey of discovery! I look forward to sharing the next instalment in this series in my upcoming newsletter!

Phil