Skip to content

AI Knowledge Hub

An open reference on how modern AI systems actually work, from the first explanation of what a language model is through to running an agent in production. Written for readers and builders alike, whatever their technical background.

Everything here is free. The hands-on half runs against a model on your own machine: no account, no API key, no cost.


Where do you want to start?

  • Understand AI, no code


    Ten pages, four to five hours. What these systems are, how they work, where they fail, and how to judge a claim about them.

    The reading path

  • Build an agent


    Thirteen modules, thirteen labs. Start with a model on your laptop and finish with an agent that survives retries, injection and a real budget.

    The build path

  • Look something up


    More than a hundred terms in plain English, a vetted reading list with a reason to trust each item, and official documentation by vendor.

    Glossary


How this site is layered

Most topics appear more than once, at different depths. That is deliberate. Each page tells you which layer it is on and links to the same idea one level up or down, so you can go as deep as you need and stop there.

Layer What it gives you Code
Understand The concept in plain language, with no prerequisites None
Build The same idea as something you run, break, and fix Every page
Go deeper Design-level depth for builders and architects Some
Reference Glossary, vetted reading, official sources None

Retrieval, for example, is one page in Understand, one module with a lab in Build, and six pages in Go deeper. You can browse by tag to see every layer of a single topic at once.


Who this is for

Analysts, product managers and leaders

Follow the reading path. It is written to be read end to end with no code. If you are evaluating a proposal or a vendor, Safety and responsible AI and Enterprise AI patterns are the two pages that will change what you ask for.

Engineers

Skim the reading path for the vocabulary, then take the build path. Thirteen modules, each ending in code that runs locally. Depth on retrieval is in Retrieval in depth; your daily tooling is in Developer tools.

Students and new joiners

Start with AI 101, keep the glossary open, and work through the reading path in order. Then, if you write code, do the build path. Nothing in it costs money.

Architects

Design patterns and Design principles for structure, Frameworks and platforms for the current landscape, and Production for what breaks when real users arrive.


Recently

The site was restructured in September 2026 around the two paths above. See What's new for the detail, and Resources for what is worth reading elsewhere, with an explicit list of what to avoid.