AI Glossary
Every AI term explained in plain English. No jargon, no PhD required.
Artificial Intelligence (AI)
Software that can understand language, recognize patterns, generate content, and solve problems by learning from data. Modern AI includes chatbots like ChatGPT, image generators like Midjourney, and coding assistants like Cursor.
Large Language Model (LLM)
A type of AI trained on massive amounts of text data that can understand and generate human language. GPT-4, Claude, Gemini, and Llama are all LLMs. They predict what words should come next based on patterns learned during training.
Prompt
The text instruction you give to an AI to tell it what you want. A prompt can be a question, a command, or a detailed description of the output you need. Better prompts get dramatically better results.
Prompt Engineering
The skill of writing effective prompts to get the best possible results from AI. Techniques include few-shot prompting (giving examples), chain-of-thought (asking AI to think step by step), and role-based prompting (telling AI who to be).
Hallucination
When an AI confidently states something that is factually incorrect or completely made up. AI can invent citations, statistics, people, and events that don't exist. This is one of the biggest risks of using AI without verification.
Fine-Tuning
The process of training an existing AI model on your own specific data to make it better at a particular task. It's like giving a general-purpose assistant specialized training for your industry.
Token
The basic unit that AI models use to process text. Roughly, 1 token equals about 3/4 of a word. AI models have token limits that determine how much text they can read and write in one conversation.
Context Window
The maximum amount of text an AI model can consider at once, measured in tokens. A larger context window means the AI can read longer documents and remember more of your conversation.
RAG (Retrieval-Augmented Generation)
A technique where AI searches through a database of documents to find relevant information before generating its answer. This makes responses more accurate and grounded in real data instead of relying solely on training data.
API (Application Programming Interface)
A way for software to communicate with AI models programmatically. Instead of using a chat interface, developers use APIs to send prompts and receive responses in their own applications.
GPT (Generative Pre-trained Transformer)
The AI model family created by OpenAI. GPT-4o is the current flagship. 'Generative' means it creates content, 'Pre-trained' means it learned from data before you use it, and 'Transformer' is the underlying architecture.
Diffusion Model
The AI technology behind most image generators. It works by starting with random noise and gradually refining it into an image based on your text description. Midjourney, DALL-E, Stable Diffusion, and Flux all use diffusion.
Zero-Shot Prompting
Asking AI to do a task without giving it any examples first. You simply describe what you want and the AI figures it out from its training. Works well for simple, common tasks.
Few-Shot Prompting
Giving AI a few examples of what you want before asking it to do the task. This dramatically improves consistency and accuracy, especially for custom formats or classifications.
Chain-of-Thought (CoT)
A prompting technique where you ask AI to think through a problem step by step instead of jumping straight to an answer. This significantly improves accuracy on complex reasoning, math, and analysis tasks.
Temperature
A setting that controls how creative or random an AI's responses are. Low temperature (0-0.3) gives consistent, predictable answers. High temperature (0.7-1.0) gives more creative, varied responses.
Embedding
A way to convert text into numbers (vectors) that capture meaning. Similar concepts get similar numbers. This is how AI search, recommendation systems, and RAG work under the hood.
AI Agent
An AI system that can take actions autonomously - not just chat, but actually do things like browse the web, write files, run code, or make API calls. Agents can complete multi-step tasks with minimal human input.
Open Source AI
AI models whose code and weights are publicly available for anyone to use, modify, and deploy. Meta's Llama, Mistral, and Flux are popular open-source models. They can be run locally without sending data to a company.
Inference
The process of running an AI model to get a response. When you send a prompt to ChatGPT and get an answer back, that's inference. It's the 'using' phase, as opposed to the 'training' phase.
Multimodal AI
AI that can understand and generate multiple types of content - text, images, audio, video, and code. GPT-4o and Gemini are multimodal models that can process images, text, and audio in the same conversation.
Text-to-Image
AI that creates images from written descriptions. You type what you want to see, and the AI generates it. Midjourney, DALL-E, and Flux are text-to-image tools.
Text-to-Video
AI that creates video clips from written descriptions. Still an emerging technology, but tools like Runway, Sora, and Kling can generate short video clips from text prompts.
Voice Cloning
AI that can replicate a specific person's voice from a short audio sample. The cloned voice can then speak any text naturally. ElevenLabs is the leading tool for this.
No-Code AI
AI tools that let you build applications, automations, and workflows without writing any code. Platforms like Make.com, Zapier, and bolt.new let non-developers create AI-powered solutions visually.