19 concepts · a path, in order
AI concepts, explained.
One idea per page. A plain definition, an everyday analogy, and a picture of how it works. Start at the top, or see how it all connects on the AI map.
See the AI map →01 — Foundations
LLM The technology behind ChatGPT, Claude and Gemini, explained without the jargon. Read → Neural network Follow one email through a neural network, step by step: how it turns into numbers, gets a spam score, and how the network learns from its mistakes. Read → Token AI models don't read words. They read tokens. Here's what that means, and why it changes how you use them. Read → Parameters What the B in '8B' or '70B' means, and why a bigger number isn't always a better model. Read →
02 — Talking to models
Prompt Everything you type to an AI is a prompt. Here's how a few small changes get much better answers. Read → Context window Why long AI chats forget what you said at the start, what a context window is, and three habits that help. Read → Temperature The setting that decides whether an AI plays it safe or gets creative. Read → Hallucination Why AI sometimes makes things up with total confidence, and how to protect yourself. Read →
03 — Giving models knowledge
Embeddings How computers turn meaning into numbers, so they can tell that 'refund' and 'money back' mean the same thing. Read → RAG How AI can answer questions about your own documents, and be more accurate, without being retrained. Read → Fine-tuning Teaching an existing AI model your style or task with extra training, and when you shouldn't bother. Read →
04 — Giving models hands
Tool calling How a model asks your app to run a function, such as checking an order, and then answers with the result. Read → MCP The 'USB-C for AI': one standard way to plug AI apps into tools and data. Read → AI agent The difference between an AI that talks and an AI that gets things done. Read → Harness The code wrapped around an AI model that turns it from something that writes text into something that can do work. Read →
05 — Running it for real
Guardrails The checks around an AI model that stop bad requests going in and bad answers coming out. Read → Prompt injection When text the AI reads, from a user or a document, sneaks in instructions that override the app's own. Read → Evals Tests for AI: a fixed set of questions with known good answers, run every time something changes. Read → Semantic caching How an AI product can skip the model and reuse an earlier answer when a new question means the same thing. Read →