For anyone who wants to finally understand AI

Artificial intelligence, explained honestly and without tech jargon.

What is a Large Language Model? What does a harness do? What do you need an SDK for? You'll get answers here. Step by step, with examples, no prior knowledge required.

16
Chapters, from the first aha moment to practice
0
Prior knowledge needed, really zero
Examples, analogies & mental pictures
Why this page?

AI gets mentioned everywhere today, but rarely explained.

"The LLM communicated with the harness via the API, and the agent called the tool through MCP." If all you understood there was word salad: that's exactly why this page exists. We take every word apart, one at a time, with calm words and pictures that stick in your head.

The topic map

Everything you should understand about AI.

You can go through them in order, or jump straight to the topic you're interested in right now.

Chapter 01
🌱
What even is AI?
Machines that "think", really? We clear up the difference between magic, math, and real capability.
Read →
Chapter 02
🎲
How does an AI work?
Probabilities, tokens, predictions. Why AI is really just "very good guessing", and doing it cleverly.
Read →
Chapter 03
🧠
Neural networks
The brain as a blueprint: why you need layers, neurons and weights, explained simply.
Read →
Chapter 04
📚
Training & data
How does an AI learn? What is training data? And why "garbage in, garbage out" is a hard truth.
Read →
Chapter 05
💬
Large Language Models (LLMs)
ChatGPT, Claude, Gemini, what's actually behind these "language models"? And why are they so big?
Read →
Chapter 06
✍️
Prompts & prompt engineering
How to actually get great results from an AI. The difference between "ok" and "wow" lies in the prompt.
Read →
Chapter 07
🔌
API & SDK
Two words programmers use constantly. We explain them with a pizza delivery service and a toolbox.
Read →
Chapter 08
🦾
What is an AI harness?
The "harness" that actually makes an AI useful. We explain why raw models alone can do very little.
Read →
Chapter 09
🤖
AI agents
When the AI suddenly operates tools itself, reads files, and solves tasks in multiple steps.
Read →
Chapter 10
🔗
MCP. Model Context Protocol
The "USB port for AIs": a common format that lets AIs plug into all kinds of things.
Read →
Chapter 11
📖
RAG. AI with a library
How an AI accesses your own knowledge without needing to be retrained.
Read →
Chapter 12
💭
Why AI hallucinates
When AI confidently makes stuff up: where that comes from, and how to guard against it.
Read →
Chapter 13
🛡️
Safety, alignment & bias
How do you make AI safe? How do you stop it from learning prejudice? A short tour through the topic.
Read →
Chapter 14
📊
Understanding benchmarks
MMLU, SWE-bench, GPQA, what do these numbers actually mean? An overview of all the important AI benchmarks.
Read →
Chapter 15
Prompting perfectly
The practical guide: how to write really good prompts in 2026. For code, building, writing, analysis, brainstorming.
Read →
Chapter 16
🔤
Glossary A–Z
All the terms once more, compact and ready to look up, from "agent" to "zero-shot".
Read →
What you'll be able to do afterward

Join the conversation. Understand. Judge for yourself.

01

Place terms correctly

API, SDK, harness, agent, MCP, RAG, you'll know not just what it is, but also what it is not.

02

Spot nonsense

You'll recognize when an AI is inventing a plausible-sounding story, and when it really knows what it's talking about.

03

Write better prompts

You'll understand why small changes to a prompt have huge effects, and how to use that on purpose.

04

Judge what's feasible

You'll be able to assess which problems are a good fit for AI, and which are better left alone.

Our promise

No black magic. No buzzwords. Just clarity.

Every chapter starts with an everyday analogy. Every unfamiliar term is explained the first time it appears. Every chapter ends with a three-point summary. If a concept is hard, we slow down, not speed up.