Unplugged in the age of AI

• Audience any audience, 20 to 100 people
• Duration 20 minutes
• Format sheets

A talk, not a workshop: the only thing the audience does is hold a search sheet and put a hand up. The room runs a bigram model Bigram model A model that predicts the next word from one previous word. It's what you build in the fundamental modules: each row of your grid says what can follow a single word. View in glossary of The Cat in the Hat for five minutes, maps what it did onto the real vocabulary (tokens Token A single unit of text that the model works with. In our activities each word and punctuation mark is a token. Modern LLMs use subword tokens, so a long or unusual word may be split into several. View in glossary , parameters Parameters The numbers a model stores to encode what it has learned. Each cell of your grid, holding the count for one word pair, is a parameter. A modern LLM has billions of them, and a frontier model on the order of a trillion. View in glossary , distribution Probability distribution A set of options with a likelihood attached to each. The counts in one row of your grid, the matching cutouts on a spread, or the show of hands when the room is asked "who has `the`?" are all probability distributions over the next word. View in glossary , sampling Weighted random sampling Choosing the next token with probability proportional to its count. A d10 roll against the counts does this, and so does letting your eye land on one of the matching cutouts, because common words appear on more of them. View in glossary ), sees the same model in grid Grid A sheet of grid paper with the vocabulary written down the side and across the top. Training fills each cell with tally marks for how often the row's word was followed by the column's word; generation reads a row and rolls against it. It's the original LLMs Unplugged apparatus. View in glossary form, and then watches three scales grow—context, training text, learned numbers—from the paper model to a frontier LLM LLM (Large Language Model) A language model trained on a very large amount of text, with billions of parameters. The hand-built models in these lessons are tiny language models; ChatGPT, Claude and Gemini are large ones. The core principles are identical. The difference is scale. View in glossary , without pretending the two are the same kind of machine. It ends on the claim that you can’t build Claude, but twenty minutes of paper and hands in the air makes you harder to fool.

We gave it at ANU in August 2026, straight after Tim Bell’s introduction to CS Unplugged; the Bandura quote it opens with had been on his screen fifteen minutes earlier. The deck below keeps the quote as its own opener and drops the event-specific close.

Running order#

You will need#

• one search sheet Search sheets A whole-room format: each person holds a sheet carrying a shuffled handful of word pairs from the text, and the room together is the model. Whoever is running it calls the current word, everyone holding it raises a hand, and one is picked. The show of hands is the probability distribution. View in glossary per attendee, sized to the head count: the ready-to-print Cat in the Hat set holds 36 sheets (or A5 two-up), or deal your own with llms_unplugged sheets -i data/the-cat-in-the-hat.txt -n 2 --sheets <count>
• the brief (page 1 of the PDF) on the lectern; it carries the worked example
• the deck’s annotation layer (⇧W on the board slide) for the generated text, and a scribe if you’d rather not write and talk
• a projector for the slides

Before you deliver#

• don’t hand the sheets out early; they get read, and the “nobody here is holding a story” beat depends on the shuffle
• the numbers on the “between us we have the whole book” and scale slides are for The Cat in the Hat at n=2; if you swap the corpus, rebuild them with llms_unplugged build and read the metadata block
• the as-delivered deck closes on “come to tomorrow’s workshop”, which was for that event; the deck above ends on the QR instead

Using the slides

Open the slides in your browser — there's nothing to install or download.

The slides run in your browser, and every slide carries full presenter notes—what to say, what to watch for, and how long each part should take. A few keys worth knowing:

• f toggles fullscreen—do this before you start presenting
• s opens the speaker view in a separate window, with the presenter notes, a timer, and a preview of the next slide (drag it to your laptop screen and put the slides on the projector)
• o shows an overview of all slides, handy for skipping ahead if you're running short on time