Essential AI terms decoded and demystified for apprentices and archmages alike.
The core technique that allows AI to focus on relevant parts of input text.
A way for software to communicate with AI services programmatically.
An AI system that can autonomously plan and execute multi-step tasks.
AI that predicts and suggests code as you type in an editor.
AI that creates original music compositions from text descriptions or parameters.
Ensuring AI systems behave according to human values and intentions.
Systematic errors in AI outputs that reflect prejudices in training data or design.
The field of research focused on making AI systems safe and beneficial for humanity.
The maximum amount of text an AI can consider at once during a conversation.
A prompting technique that makes AI show its reasoning step by step.
A technique for guiding AI image generation with structural inputs like poses or edges.
The AI technique behind most modern image and video generation tools.
A subset of AI using neural networks with many layers to learn complex patterns.
A numerical representation of text that captures its meaning in a format AI can process.
Customizing a pre-trained AI model with your own data for specific tasks.
Teaching AI to perform a task by showing it just a few examples in the prompt.
OpenAI's series of language models that power ChatGPT.
Connecting AI responses to verified, factual sources to reduce hallucination.
When an AI generates false or fabricated information that sounds convincing.
The process of an AI model generating outputs from new inputs after training.
An AI model trained on vast text data to understand and generate human language.
An efficient method for customizing AI models without retraining the entire model.
A compressed mathematical representation where AI models store and manipulate concepts.
AI that can understand and generate multiple types of content — text, images, audio, video.
An open standard for connecting AI models to external data sources and tools.
An AI architecture that uses specialized sub-networks to handle different types of tasks efficiently.
A computing system inspired by the human brain that learns patterns from data.
AI models and tools whose code and weights are freely available for anyone to use and modify.
The text input you give to an AI model to get a response.
The practice of crafting effective prompts to get better AI outputs.
A security attack where malicious instructions are hidden in prompts to manipulate AI behavior.
Compressing AI models to use less memory and run faster with minimal quality loss.
A technique that gives AI access to external knowledge to improve accuracy.
A training technique that uses human preferences to make AI more helpful and safe.
An open-source AI image generation model that runs locally or in the cloud.
Search that understands meaning and intent rather than just matching keywords.
Artificially generated data used to train or evaluate AI models.
The neural network architecture behind modern AI language models.
The basic unit of text that AI models process — roughly 3/4 of a word.
The dataset used to teach an AI model how to perform its tasks.
AI technology that converts written text into natural-sounding spoken audio.
AI that generates images from written text descriptions.
AI that generates video clips from text descriptions.
AI's ability to use external tools like web search, calculators, and APIs.
A setting that controls how creative or predictable an AI's responses are.
A specialized database that stores and searches AI embeddings for similarity matching.
A development approach where you describe what you want in natural language and AI builds it for you.
AI technology that creates a digital copy of someone's voice from audio samples.
A browser-based runtime that can execute Node.js and build tools entirely in the browser.
An AI's ability to perform tasks it wasn't specifically trained for.
