AI Glossary

Agentic Coding

Agentic coding means AI tools can plan, write, and test code autonomously across multiple iterations, not just suggest one-off snippets.

AGI (Artificial General Intelligence)

AGI refers to hypothetical AI that matches or surpasses humans across almost all cognitive tasks, unlike today's specialized systems.

AI Agent

An AI agent is software that observes context, plans, and executes steps toward a goal, often using a large language model and tools.

AI Image Generator

An AI image generator creates images from text prompts or references, often using diffusion models such as Midjourney, Flux, and DALL·E.

AI Model

An AI model is a trained function that maps input to output, from classic ML models to large language and image models.

AI SEO

AI SEO (AEO/GEO) is about getting cited in AI assistant answers while still performing in search through structure, trust, and clear source pages.

AI-bias

AI bias occurs when models reflect skewed data, design choices, or context and produce unfair or inaccurate outcomes.

Anthropic

Anthropic is a U.S. AI safety company that develops the Claude family of language models and related products.

Application Programming Interface (API)

An API is a contract that lets programs communicate, and web APIs power most LLM and automation services today.

Artificial Intelligence

Artificial intelligence (AI) refers to computer-based systems that can solve tasks traditionally requiring human perception, reasoning, or decision-making.

ASI (Artificial Superintelligence)

ASI refers to AI that systematically surpasses human expertise across almost everything, a future scenario discussed in AI safety and ethics.

Automation

Automation means recurring tasks are run automatically using rules or tools, often with AI and workflows that connect systems.

Chatbot

A chatbot is an interface that replies in conversation, from simple rule-based systems to LLM-powered assistants.

ChatGPT

ChatGPT is OpenAI’s chat interface built on GPT models, optimized for dialogue, tool use, and work with code and text.

Claude

Claude is Anthropic’s family of assistants and models for writing, coding, and analysis, with a focus on long context and safety.

Claude Code

Claude Code is Anthropic’s CLI workflow where Claude can work with file trees and terminal commands as an agentic coding tool.

Codex

Codex is OpenAI’s coding agent for reading codebases, proposing changes, and working across terminal, IDE, and developer workflows.

Context Window

A context window is how many tokens a model can consider at once, including your prompt and prior messages.

Crawling

Web crawling means a bot follows links and loads pages, forming the foundation of search engines and some AI training pipelines.

CRM

CRM systems collect customer data, opportunities, and communication, while AI adds transcripts, prioritization, and drafts.

Cron Job

A cron job runs a command or script on a schedule, commonly used for backups, reporting, and automation.

Cursor

Cursor is an AI-driven code editor built on VS Code, with agent modes, inline editing, and full-project context.

Deep Learning

Deep learning is machine learning with many neural-network layers that learn representations directly from data, forming the backbone of modern AI.

Fine-tuning

Fine-tuning is additional training on an existing model using domain-specific data for better tone, format, or accuracy.

GDPR

GDPR is the EU data protection regulation covering rights to access, deletion, and legal basis when personal data is processed, including by AI services.

Gemini

Gemini is Google’s family of multimodal models and assistants across Workspace, Pixel, and cloud products.

Generative AI

Generative AI creates new content, such as text, images, audio, or code, based on training data and prompts.

GPU

A GPU is parallel hardware that accelerates matrix multiplication, which is critical for training and inference in neural networks.

Hallucination

Hallucination means a model produces confident but false claims, often when facts are missing or source context is weak.

Large Language Model (LLM)

An LLM is a large language model trained on text to predict the next token, powering systems like ChatGPT, Claude, and Gemini.

Lovable

Lovable lets you describe an app in natural language and iterate toward code and preview, offering an AI-driven build workflow for web apps.

Model Context Protocol (MCP)

MCP is an open standard, initiated by Anthropic, for how assistants discover and call tools and data sources in a unified way.

Multimodal

Multimodal AI combines multiple data types, such as text and images, in the same model or pipeline.

n8n

n8n is an automation platform with visual workflows, self-hosting, and AI nodes for connecting APIs and databases.

Natural Language Processing (NLP)

NLP is the field that teaches computers to understand and generate human language, now largely driven by transformer-based LLMs.

Neural Networks

Neural networks are connected nodes with weights adjusted through training, and they are the foundation of deep learning.

OpenClaw

OpenClaw is a tool for building and running AI agents that can operate across systems and tasks, related to the Moltbot ecosystem.

OpenRouter

OpenRouter is an API gateway that lets you switch between many model providers using the same client, often with unified billing.

Prompt

A prompt is model input, from a short question to long context, that controls response tone, format, and quality.

Prompt Engineering

Prompt engineering is the practice of designing instructions, examples, and formats so a model returns useful, robust answers.

Singularity

Singularity (technological singularity) refers to a hypothetical turning point where AI progress accelerates beyond what humans can adapt to.

Skills

Skills are packaged instructions, files, or scripts that extend an AI assistant with domain-specific workflows, such as Claude Skills.

Speech-to-Text (STT)

Speech-to-Text converts audio into written text using acoustic and language models, forming the backbone of transcription and voice interfaces.

Text-to-Speech (TTS)

Text-to-Speech synthesizes natural speech from text with neural voices, used in podcasts, accessibility, and voice assistants.

TPU (Tensor Processing Unit)

A TPU is Google’s specialized chip optimized for tensor operations in machine learning, often used for cloud training and inference.

Transformer

A transformer is a neural-network architecture with self-attention that surpassed older sequence models and powers GPT, Claude, and more.

Vibe Coding

Vibe coding is a loosely defined style of building with AI through fast iteration and intuition rather than formal specs, often using agentic tools.

Webhook

A webhook is an HTTP POST sent by a system to your URL when an event happens, enabling event-driven automation.