The Machine That Finishes Your Sentence
What happens inside an AI response — and why it should not be confused with human thought or verified knowledge.
The learning pathway
Twenty-one lessons in five acts, each 5–8 minutes. We recommend going in order — every lesson builds on the ones before — but nothing is locked, and you can revisit any lesson at any time. All 21 lessons are available now.
Your progress (guest mode)
Progress is saved automatically in this browser, on this device only. It can be lost if you clear browsing data — download a copy to keep or move it.
Act 1 — Remove the mystery
What AI is actually doing: patterns, data, tokens and probability.
A Diet of Words
Discover what AI was fed — and why its diet explains both its brilliance and its blind spots.
Chopping Language into Lego
See your own sentence the way the machine sees it — and discover why AI miscounts letters.
The Dice Inside the Machine
Watch the machine hesitate — and learn the single setting that makes AI creative or careful.
Act 2 — Look inside the model
How a model learns, represents meaning, weighs context and produces an answer. Available now.
Ten Thousand Nudges: How Training Works
How a machine gets better at prediction without anyone programming the answers in — train one yourself.
Layers of Sieves: Neural Networks
Open the machine and find not a brain, but stacked layers of astonishingly simple arithmetic.
The Map Where “King” Lives Near “Queen”: Embeddings
Meaning as a place on a map — how machines find related ideas without matching keywords.
The Spotlight Operators: Attention and Transformers
The mechanism behind the modern AI boom — every word weighing every other word.
Showtime: Inference
Follow your question through the entire machine, token by token, as an answer is built live.
Act 3 — Learn to direct AI effectively
Context, briefing, senses, retrieval, tools, memory and agents. All of Act 3 is available now.
The Desk, Not the Brain: Context Windows
The invisible desk every AI works at — and why long conversations quietly fall off its edge.
Briefing the Machine: Prompting
Why the same request phrased two ways gets a mediocre or brilliant answer — and the briefing habit plus improvement loop that gets the brilliant one.
New Senses: Multimodal AI
How one machine reads photos and scrawled notes — and where its new senses deceive it.
The Library Card: Retrieval
Watch AI stop guessing and start citing — and learn what citations do and don't prove.
Hands for the Machine: Tools and APIs
How a text-only machine books meetings and does perfect arithmetic — by writing work orders.
The Notebook, Not the Mind: Memory
What your AI assistant actually keeps about you between conversations — and who holds the notebook.
Letting Go of the Handlebars: Agents
Watch an AI plan, act, check and try again — and learn when that's brilliant and when it compounds errors.
Act 4 — Use AI prudently
Why AI fails: invention, inherited tilt, and attacks around the model. Available now.
Confidently Wrong: Hallucination and Uncertainty
Why AI invents facts with a straight face — and the three-question test that catches most of them.
The World's Fingerprints: Bias and Blind Spots
How an AI that treats everyone "the same" can still treat people differently — and how to spot the tilt.
Locks, Leaks and Whispered Orders: Security and Privacy
The three ways AI systems get people into trouble — and the five habits that block most of it.
Act 5 — Put AI to work responsibly
Workflow design, judgement and accountability — ending in the capstone: you build your own AI verification & use checklist, the artefact behind the certificate. Available now — the series is complete.
The Relay, Not the Replacement: Human–AI Workflows
Stop asking "can AI do my job?" and start asking which parts each of you should run.
The Anchor Leg: Judgement, Verification and Responsible Use
The final lesson hands you the anchor baton: a durable code for using AI with your name on the result.