An explorable guide
AI increasingly gets used more and more every day, in a variety of ways, while the vast majority of us don't fully understand what it does and how it works. It feels like magic—and that's exactly the problem.
This is an interactive guide to how large language models actually work. Not just how to use them. How they actually work.
Every concept has something you can play with. Adjust parameters, see consequences, build intuition. Why does this exist?
The foundational concepts behind large language models—from raw text to reasoning. Start here. Everything else builds on this.
AI models can't read. They work with numbers. The first step is understanding how text gets turned into something a computer can actually process.
Tokens, vocabulary, embeddings 02A model starts as random noise. Through billions of fill-in-the-blank exercises, it learns patterns. Here's what "learning" actually means.
Training, prediction, weights 03The word "bank" means different things in different contexts. Attention is how a model figures out which meaning applies. It's the core of the transformer architecture.
Context windows, attention, memory 04If all a model does is predict the next word, how does it solve multi-step problems and write working code? The answer is more interesting than you'd expect.
Temperature, emergence, chain-of-thoughtThe Fundamentals is just the starting point. Future learning paths will cover how models are improved after training, how to use AI effectively, and the broader questions around AI and society.