Basics
Core concepts in plain language: tokens, context, agents and prompts — no jargon, no formulas.
What is temperature in a model — the slider between boring and wild
One setting decides whether the AI answers dull and identical or creative and risky. At zero it's by-the-book; turn it up and it writes poetry, but it may also make things up.
What is RAG — letting an AI peek at your cheat sheet
The model never trained on your documents — yet it answers from them. The trick is that it gets handed the right snippets right before answering. That's RAG.
What is MCP — the single port that gives AI its 'hands'
MCP is USB for AI agents. One common port, and an agent can reach your browser, your database and your files — no custom adapter for each.
What is a context window — and why AI forgets the start of a conversation
In a long chat, old messages get pushed out of the model's memory — that's why it loses the thread. Explained with a simple picture.
What is an AI hallucination — and why the model isn't lying when it makes things up
The model confidently spat out nonsense. Feels like a lie. But it can't lie — it has no idea what truth even is. Let's see how it works and how to catch it.
What is a token — and why Russian costs more than English
The same sentence costs a model more in Russian than in English. Tokens are why — let's see what they are and how to pay less.
What is a prompt — and why it's a brief, not a question
A prompt feels like a question you ask the AI. It's really a brief for a worker. Flip that, and the answers stop surprising you.
What are embeddings — and how AI knows that 'cat' and 'kitten' mean the same
AI turns words into coordinates on a map of meaning. Close meanings land near each other — so it searches by meaning, not by exact words.