Glossary

Technical terms explained clearly

A

A/B Testing

A method for comparing two or more versions of something by showing each to a slice of your audience and measuring which performs better.

Agent (AI Agent)

An AI system that can independently plan, make decisions, and take actions to accomplish goals — rather than just answering questions.

AGI (Artificial General Intelligence)

A hypothetical form of AI that could perform any intellectual task a human can — learning, reasoning, and adapting across domains without being retrained for each one.

AI (Artificial Intelligence)

Technology that enables machines to perform tasks that typically require human intelligence — from understanding language to recognizing images to making decisions.

API (Application Programming Interface)

A defined way for software components to talk to each other — usually over the network — using requests, responses, and documented rules.

B

Benchmarks

Standardized tests used to measure and compare AI models — like report cards that show how well a model performs on specific tasks.

C

CI/CD (Continuous Integration/Deployment)

A software development practice where code changes are automatically tested, integrated, and deployed to production.

Cloud Computing

Delivery of computing services — servers, storage, databases, networking, and software — over the internet, typically on a pay-as-you-go model.

D

DevOps

A set of practices and cultural norms that bring development and operations together to deliver software faster, safer, and more reliably.

E

Edge Computing

A distributed computing paradigm that processes data closer to the source, reducing latency and bandwidth usage.

The European Union's regulation on artificial intelligence (Regulation (EU) 2024/1689), which classifies AI uses by risk tier and attaches obligations to each — from outright bans to simple disclosure duties.

F

Fine-Tuning

The process of taking a pre-trained AI model and training it further on your own data to adapt it to a specific task, style, or domain.

Full-Stack Development

Building software across both the parts users interact with in the browser or app (front end) and the servers, databases, and integrations behind them (back end).

G

Guardrails

The rules, limits, and checks that keep an AI system inside safe, intended behavior — what it may do, what it must not, and when it asks first.

H

Harness (Agentic Harness)

The system that runs AI agents together — the orchestration, tools, memory, permissions, and scheduling layer that turns a language model into something that does real work safely.

L

LLM (Large Language Model)

A deep learning model trained on vast text data that can understand and generate human-like text.

M

MCP (Model Context Protocol)

An open standard, created by Anthropic, for connecting AI assistants to the tools, data, and systems they need to do real work.

Mechanistic Interpretability

The practice of reverse-engineering AI models to understand how they actually arrive at their answers, rather than treating them as black boxes.

Multimodal AI

AI that works across more than one kind of data — text, images, audio, and video — rather than text alone.

O

OAuth

An open standard that lets you grant an app limited access to your accounts without sharing your password.

Open-Source

Software or AI models whose underlying code or weights are publicly available — anyone can inspect, modify, and use them, often for free.

P

Pipelines

Automated, repeatable sequences of steps that move work forward — from building and testing code to ingesting, transforming, and loading data.

PPC (Pay-Per-Click)

An online advertising model where you pay only when someone clicks your ad, used across search engines, social platforms, and display networks.

R

RAG (Retrieval-Augmented Generation)

An AI architecture that enhances LLM responses by retrieving relevant context from external knowledge bases before generating answers.

RL (Reinforcement Learning)

A training method where an AI learns by taking actions and receiving feedback — rewards for good choices, penalties for bad ones — until it figures out how to achieve a goal.

S

SEO (Search Engine Optimization)

The practice of improving how visible and compelling your site is in organic search results — through technical health, content relevance, and authority signals.

SolarPunk

A cultural and design movement that imagines an optimistic future where communities live sustainably, powered by renewable energy and technology in harmony with nature.

T

Tokens

The basic units of text that AI models process — roughly word-sized chunks that can be words, parts of words, or punctuation.

Tool Use (Function Calling)

The mechanism that lets a language model call external tools — APIs, databases, code, search — instead of only generating text.

U

UI (User Interface)

The visual and interactive layer of a digital product — screens, controls, typography, color, and motion that users see and manipulate.

UX (User Experience)

The overall experience someone has when using a product, service, or system — how easy, efficient, and satisfying it feels from their perspective.

V

Vector Database

A database optimized for storing and querying high-dimensional vector embeddings used in similarity search and AI applications.

W

Weights

The learned numbers inside a neural network that encode its knowledge — the 'settings' the model adjusts during training to get better at its task.

Workflow

An AI system in which language models and tools are coordinated through a sequence of steps defined in advance by the developer.