Skip to main content
Learning Science

Our Methodology

Every concept explained at three levels of depth. Choose your layer based on your goals and background.

1

Simple Layer

For: Complete beginners, non-technical audiences, quick understanding

Uses everyday analogies and plain language. Focuses on what AI does and why it matters. No math, no code—just clear mental models.

"Think of a neural network like your brain learning to recognize your grandmother's face..."

2

Visual Layer

For: Visual learners, those with some technical background, deeper understanding

Interactive diagrams, animated visualizations, and step-by-step walkthroughs. See how AI works with your own eyes. Includes simplified math when helpful.

Neural network architecture diagrams
+ Interactive playgrounds to experiment
3

Scientific Layer

For: Developers, researchers, those who want the full technical details

Rigorous explanations with actual math, code snippets, and citations to research papers. Understand exactly how algorithms work and why they're designed that way.

softmax(x_i) = exp(x_i) / Σ exp(x_j)
Attention(Q,K,V) = softmax(QK^T/√d_k)V

+ Implementation details & research citations

Core Principles

No Hand-Waving

We don't skip the hard parts or use misleading analogies. Every explanation is grounded in how AI actually works.

Progressive Disclosure

Start simple, go deeper when you're ready. Each layer builds on the previous one without contradicting it.

Interactive First

Learning by doing beats passive reading. Every article includes interactive elements you can manipulate and explore.

Embrace the Unknowns

We're honest about what we don't know. The "frayed edges" of AI research are just as important as established facts.

Citations & Sources

Every claim backed by research. All sources linked and verified. Science, not speculation.

Why Three Layers?

Research in cognitive science shows that learning depth should match your goals. If you just need to use ChatGPT effectively, you don't need transformer architecture details. But if you're building the next LLM, surface analogies won't cut it.

Our approach is inspired by Feynman's teaching philosophy: explain at multiple levels, but never lie or oversimplify to the point of inaccuracy. Each layer is complete on its own—you're never left with misconceptions you'll have to unlearn later.

You choose your depth. We provide the clarity.

See It In Action

Experience our three-layer approach on any article

Explore Learning Paths