Content
- A brief history of AI Seventy-five years of artificial intelligence, built live with an agent: a group collects the milestones it knows, and an AI agent researches, checks and turns them into a timeline with a quiz. AI agentsLanguage models
- A map of meaning A model turns every text into a list of numbers, an embedding, which marks its place on a map where texts of similar meaning lie close together, and this lets a computer search for meaning instead of words. Language models
- AI agents A language model that uses tools in a loop to reach a goal: what sets an agent apart from a chatbot, what it is made of, when a fixed workflow is better, and what can go wrong when it acts with your permissions. AI agents
- An AI on your laptop Switch off the internet and still get answers from a language model on your own laptop, then look at the file that holds the whole model and compare a small and a larger version of it. Language models
- Be the agent Your group plays an AI agent with cards instead of a computer: one of you is the language model, one carries the messages, four are the tools, and one keeps the log. AI agents
- LM Studio A free program that downloads language models with open weights and runs them on your own laptop, without the internet; its developer mode shows the settings that decide how a model writes its answer. Language models
- LM Studio Bionic An AI agent by the makers of LM Studio, like OpenCode or Claude Code, that works on code and documents with open models, on your own computer or in the maker's cloud. AI agents
- Next-token prediction A language model writes one token at a time, and every token it writes becomes part of what it reads next: tokens, sampling, the context and its limits, and the transformer from far away. Language modelsRepresenting meaning
- Open models A language model is a file full of numbers, and some of these files you can download: what is inside, why its size decides how good, fast and hungry it is, and when it is worth running one yourself. Language models
- Quiz: A brief history of AI Six questions on the lesson A brief history of AI, each answer with its reason. The questions come from the pool of the module and are in the format of its exam. AI agentsLanguage models
- Quiz: AI agents Twelve questions on the lesson AI agents, each answer with its reason. The questions come from the pool of the module and are in the format of its exam. AI agents
- Quiz: Open models Six questions on the lesson Open models, each answer with its reason. The questions come from the pool of the module and are in the format of its exam. Language models
- Slides: A brief history of AI The slides to the lesson: how a group collects what it knows about seventy-five years of AI on a pinboard, and how an agent researches, checks and turns it into a timeline with a quiz. AI agentsLanguage models
- Slides: A map of meaning The slides that go with the experiment step by step: a map by hand, the numbers of a token, computing with meaning, from tokens to a text, the model's map and its search, what the map does not know, and the sum behind every similarity. Language models
- Slides: AI agents The slides to the lesson: what sets an agent apart from a chatbot, what it is made of, how it works with its models, and why whoever gives it access hands over their own keys. AI agents
- Slides: Be the agent The slides that lead through the card game: the roles, the cards, the rules, the three tasks, and what you have just played. AI agents
- Slides: Next-token prediction The slides to the lesson: how a language model writes one token at a time, how the next token is chosen, what the context is and why it is limited. Language modelsRepresenting meaning
- Slides: Open models The slides to the lesson: a language model is a file full of numbers: what is inside it, why its size decides speed and memory, and what sets open models apart from closed ones. Language models
- The Bitter Lesson Rich Sutton looks back on seventy years of AI research and draws one conclusion: general methods that make use of more and more computation, search and learning, have always beaten the knowledge that researchers built in by hand. Language models
- Word distributions Counting which word follows which in 22 sentences gives a probability distribution of the next word, and drawing from it one word at a time writes new text. Language models
Nothing matches. Fewer filters or another word will help.