The AI wiki, minus the jargon
Twenty-one concepts that explain 90% of every conversation about artificial intelligence. Each with an analogy, the technical detail, and an example.
-
LLM (large language model)
The engine behind ChatGPT, Claude and Gemini.
Read more -
Tokens
The pieces AI breaks all text into.
Read more -
Context window
The model's working memory: if it doesn't fit, it doesn't exist.
Read more -
Temperature
The dial that controls how daring the model gets.
Read more -
MCP (Model Context Protocol)
The USB-C of AI: one standard connector for tools and data.
Read more -
RAG (retrieval-augmented generation)
The model stops memorizing and starts looking things up.
Read more -
Fine-tuning
Turning a generalist model into your specialist.
Read more -
AI agents
From answering questions to completing missions.
Read more -
Hallucinations
When AI answers with total confidence… and is making it up.
Read more -
System prompt
The invisible instructions the AI receives before yours.
Read more -
Chain of thought
Asking the AI to think out loud before answering.
Read more -
Embeddings
Turning meaning into coordinates a machine can compare.
Read more -
Multimodal AI
Models that don't just read: they also see, listen and draw.
Read more -
Parameters
The millions of internal “dials” where everything learned lives.
Read more -
Open vs. closed models
Who gets to download, modify and run the model?
Read more -
Prompt injection
When someone hides commands inside the text the AI is about to read.
Read more -
Transformer
The architecture that made everything else possible.
Read more -
Training
How a model goes from knowing nothing to talking about everything.
Read more -
Inference
The moment the model works for you.
Read more -
Benchmarks
The exams used to compare models.
Read more -
AGI (artificial general intelligence)
The hypothetical AI capable of any human intellectual task.
Read more