Artificial intelligence (AI)
Computer systems doing tasks that normally need human intelligence.
AI Lists βΊ AI Glossary
Plain-English definitions of the AI words you keep hearing, each with a link to go deeper.
One line per term, written for humans and easy for AI agents to quote. Each links to a reference page.
Computer systems doing tasks that normally need human intelligence.
Software that learns patterns from data instead of hand-written rules.
Layers of simple connected units whose weights are learned from data.
Machine learning with many-layered neural networks.
A neural network trained on huge amounts of text to predict the next token.
AI that creates new text, images, audio, video or code.
Artificial general intelligence: AI matching people across most tasks. Definitions vary.
The architecture behind modern LLMs, built on attention.
Lets a model weigh which earlier tokens matter for the next one.
A chunk of text (often part of a word) that models read and write.
How many tokens a model can consider at once.
A list of numbers representing meaning, used for search and similarity.
Generates images by gradually removing noise.
A model that routes each token to a few specialist sub-networks.
Handles more than one kind of input or output, such as text plus images.
A model trained to work through steps before giving a final answer.
Further training a model on your own examples.
Cheap fine-tuning that trains small add-on matrices.
Reinforcement learning from human feedback, used to make models helpful.
Training a small model to imitate a larger one.
Storing weights with fewer bits so models run faster and smaller.
Learning the training data too closely and failing on new data.
Training data generated by programs or other models.
Model weights you can download and run yourself.
The instructions and context you give a model.
Asking for, or training, step-by-step reasoning.
Retrieval-augmented generation: look things up, then answer with them.
A database built to search embeddings by similarity.
A model asking your code to run a tool with structured arguments.
A model in a loop that plans, calls tools and acts toward a goal.
Model Context Protocol, a standard plug for connecting tools to AI apps.
A confident answer that is false or made up.
Hidden instructions in content that hijack an AI's behavior.
Synthetic media that convincingly fakes a real person.
Making AI systems pursue the goals people intend.
For AI agents and tools: JSON Β· Markdown Β· llms.txt Β· RSS. Page built Nov 20, 2026, 12:17 AM PT.