36 key AI terms, explained in plain language. Each entry links to tools you can try so the concept is not just theory.
A large language model is a neural network trained on massive text data to predict and generate human-like tex…
Read more →GPT is the family of generative pre-trained transformer models behind many chat assistants. 'Generative' means…
Read more →Prompt engineering is the practice of writing clear instructions so an AI model returns the result you want. G…
Read more →RAG is a technique where a model first retrieves relevant documents, then generates an answer grounded in them…
Read more →Fine-tuning adapts a pre-trained model to a specific task or style by training it further on a smaller, labele…
Read more →A token is a chunk of text - roughly a word or a piece of a word - that a model reads and charges by. Token co…
Read more →An AI agent is a system that plans steps and calls tools on its own to reach a goal, rather than answering one…
Read more →Multimodal AI handles more than one type of input or output - text, image, audio, video - in a single model, s…
Read more →An embedding is a list of numbers that represents the meaning of text or an image. Similar meanings get simila…
Read more →A vector database stores embeddings and finds nearest neighbors fast. It is the storage layer behind semantic …
Read more →A hallucination is when a model confidently outputs information that is false or invented. Checking facts and …
Read more →LoRA is a lightweight fine-tuning method that adds small adapters instead of retraining a whole model. It is w…
Read more →The transformer is the neural architecture behind modern LLMs. Its attention mechanism lets a model weigh all …
Read more →Computer vision is the field of teaching machines to understand images and video - detection, segmentation, OC…
Read more →NLP is the branch of AI that works with human language - classification, translation, summarization, and dialo…
Read more →Generative AI refers to models that create new content - text, images, audio, code - rather than only classify…
Read more →A foundation model is a large model trained broadly on diverse data that others build on via fine-tuning or pr…
Read more →Inference is the step where a trained model actually produces an output from your input. Latency and cost are …
Read more →A chatbot is a conversational interface, usually powered by an LLM, that replies in natural language and can c…
Read more →AI art is imagery created by generative models from a text prompt. Style, composition, and subject are steered…
Read more →Text-to-image is the task of turning a written description into a picture. It is the core capability behind AI…
Read more →Text-to-video generates short clips from a prompt or an image. It extends text-to-image into motion, with timi…
Read more →Voice cloning synthesizes a speaker's voice from a short sample so it can read any text. It powers dubbing and…
Read more →Speech synthesis, or text-to-speech, turns written text into spoken audio. Modern neural TTS sounds close to a…
Read more →OCR reads text out of scanned pages and photos so it becomes searchable and editable. It is a building block o…
Read more →Background removal automatically cuts a subject out of its backdrop in a photo, used for product shots, portra…
Read more →Upscaling enlarges a low-resolution image while recovering detail with a model, instead of just stretching pix…
Read more →An AI avatar is a digital presenter driven by a script and a synthetic voice, used for explainer videos and tr…
Read more →No-code AI lets non-developers build workflows and apps by dragging blocks, often with an AI step inside. It l…
Read more →An API is the interface a model exposes so your code can call it. Most AI products are wrappers around a model…
Read more →An open-source model publishes its weights so you can run or fine-tune it yourself. It offers data control and…
Read more →A benchmark is a standardized test set that scores models on reasoning, coding, or safety. Benchmarks help com…
Read more →RLHF tunes a model using ratings from human reviewers so its answers better match what people want - a key ste…
Read more →Synthetic data is artificially generated to stand in for real data - used to train or test models when real da…
Read more →Edge AI runs models on your device or local server instead of the cloud, cutting latency and keeping data on-p…
Read more →A diffusion model generates images by gradually removing noise from random input. It is the core method behind…
Read more →