Reference

AI Glossary

A short, opinionated glossary of the AI terms you'll run into most often.

Agent

An autonomous AI system that can plan, take actions, and use tools to accomplish a goal.

Diffusion Model

A type of generative model that produces images by progressively denoising a random signal — the engine behind tools like Midjourney and Stable Diffusion.

Embedding

A numerical representation of data (text, images, audio) that captures its semantic meaning — used for search, clustering, and retrieval.

Fine-tuning

Training an existing model on a smaller, specific dataset to adapt it to a particular task or style.

LLM

Large Language Model — a model trained on massive text corpora that can generate, summarize, and reason about text.

Multimodal

A model that handles multiple input/output types — e.g., text, image, audio — in a unified way.

Prompt

The natural-language instruction or context given to a model to elicit a desired response.

RAG

Retrieval-Augmented Generation — combining an LLM with a search step over private/custom data so answers are grounded in real sources.

Token

The unit a language model reads and writes — roughly a word or a fragment of a word.

Tool Use

When an LLM can call out to functions or external APIs (search, code, calendars) instead of only generating text.