About this hub

A personal, growing reference for working as an AI designer — the kind of foundational understanding useful at an AI PoC or product company (Figma, Anthropic) or in financial services (Monzo, Revolut, Nationwide, Lloyds). The aim is solid technical intuition without the heavy maths: enough to scope features, talk credibly with engineers, and design responsibly.

How to use it

• Library — browse and search definitions, each with a designer angle and sources.
• Model landscape — why model sizes differ, best-in-class compared, when to use which, plus sources and video explainers.
• Flashcards — spaced, flip-card study with known/review tracking.
• Tests — check understanding with explained answers and saved scores.
• Tools — AI tools to fold into a design workflow.
• Add — capture your own notes and back them up as JSON.

Source philosophy

Definitions are written to be accurate and are backed by primary explainers from organisations that are strong in the field — IBM (Think Topics), Google (Cloud & ML Crash Course), Anthropic (Claude docs & research), Meta AI, plus OpenAI docs and open standards like the Model Context Protocol and the EU AI Act. Each concept page links its sources so you can go deeper and verify.

The field moves fast. Treat sources as the authority and this hub as the map — and update entries as your understanding (and the technology) grows.

Every concept page links its primary sources, and the Model landscape page collects provider documentation (IBM, Google, Microsoft, Anthropic, OpenAI, Amazon, Meta) alongside reputable YouTube channels — IBM Technology, 3Blue1Brown, Andrej Karpathy and the provider channels — for going deeper by reading or watching.

Built to grow & stay portable

• Curated content lives in version-controlled data files, so it's easy to extend over time and carry between roles.
• Your own definitions, flashcards, scores and bookmarks are saved in the browser — export a JSON backup from the Add → Manage tab to move it anywhere.
• Re-theme the whole app by changing one accent colour token.

Made by Ethan Chow.