Back to The University
Field Guide
Every term, in one place
Plain meaning first, then the real name, then the deeper version when you want it. Nothing here is locked and nothing here is a test.
How your instructions get built
Four stages, in order. Each one turns into the next, and the last one is the text you paste into the AI you use.
How Yu works
The moving parts of Yu itself, so the words on the screen mean something.
- FrameworkA week's worth of work on one topic.
- ActivityOne thing you do inside a Framework.
- AssistantThe living robot assigned to you, who works with the context in your Yu.
- WeekOne Framework, a handful of activities, and one Record.
- Stated FactSomething you add directly, in your own words.
- Drop-inSomething from your real day that you add to Yu.
- MisfireA time your AI got you wrong, kept so future instructions can reflect it.
- FindingOne observation about how you work, pulled out of a week.
- SignalThe raw evidence a Finding was drawn from.
- CityThe map of everything Yu has, laid out as places you can visit.
- CoinsWhat you earn for work you do, and spend in the store.
- StreakHow many days in a row you have done a Yu activity.
AI mechanics
How the machine actually works. Every one of these is taught by a card in AI 101.
- TokenThe chunk of text a model actually reads, instead of a word.
- Training DataThe pile of text a model read before it existed.
- WeightsThe numbers a model kept from its training, frozen in place.
- ModelThe trained pattern engine that writes the answer.
- System PromptThe provider's own hidden instructions, sent before your first word.
- MemoryA product feature that decides which of your saved details to include this time.
- Context WindowEverything the model can see this turn, and it has a budget.
- PromptWhat you type this turn. It lasts exactly one conversation.
- ConnectorA link that lets an AI reach into another system you use.
- HallucinationA confident answer that is plausible rather than true.