Glossary

Use this page when a word is unfamiliar; you do not need to memorize it. BLI means Big Lever Institute. MC labels mean mini-course; exercise numbers are stable identifiers, not a required sequence.

Basic AI words

Assistant/app

The service you open to work with AI, such as ChatGPT, Claude or Gemini.

Model

The underlying AI producing the response; an app may offer several.

Prompt

The message or instructions you give the assistant.

Chat/conversation

Your messages and the assistant's replies together.

Context

Relevant background, facts or material for this task.

Source

Material you can inspect to check a claim, such as your original notes.

Instructions

Directions about what to do or how to work.

Project/workspace

A place for related continuing work; features and retained context differ between services.

Upload/attach

Add a file to the service so it may use its contents; check that it can read them.

Critique

A check for important weaknesses. The assistant can be wrong about its critique too.

Subscription/plan

Your account's access and usage allowance. A paid AI plan is separate from free BLI classes.

Hallucination

An AI answer that invents or misstates information while sounding plausible. Compare important claims with sources you can inspect.

Memory

A service feature that may retain selected information across conversations. It is different from the current chat history and from files saved in a Project. What is retained depends on settings and the service; check it rather than assuming.

Web search

A tool for retrieving information from the web. An answer from an assistant is not proof that it searched or opened a page. Ask for sources and inspect them.

Token and context limit

A token is a small unit of text processed by an AI model. The context limit is how much material it can work with at once. In a long conversation or large collection of files, check which relevant information it can actually use.

Project Brain

Our name for a method of organizing sources, questions and approved learning around a topic. It is not a separate product or a promise of automatic memory.

Plugin

An optional package that adds a configured guide or capability to a supported AI service. The BLI student plugins are Guide Me and Your Learning Path; their current installation route uses ChatGPT.

Learning path and Exercise Library

A learning path connects useful exercises in an order that fits a goal. The Exercise Library holds each exercise once, so different paths can link to the same lesson.

Shippers and Prod-ject

Shippers means people who build and release something useful. Prod-ject is the intentional name of a future course combining product and project management.

Terms you may meet in practical lessons

Hypothesis

A claim you can examine using evidence that might support or weaken it.

WWMCAD

What Would My Customer Archetype Do? Chris Myers’s customer-learning approach: begin with real experiences, current behavior and workarounds before proposing a solution. AI can help prepare questions; it does not supply customer evidence.

Customer discovery

Learning about real customer problems and behavior through observation and conversations.

Synthetic interview

An AI-generated rehearsal. It can suggest questions or assumptions; it is not customer evidence.

MVP

Minimum viable product: a limited version used to learn about an important assumption.

PMF

Product-market fit: evidence that an offering meets meaningful demand in a particular market.

BMC

Business Model Canvas: a framework for describing nine connected parts of a business model.

TAM / SAM / SOM

Total addressable market, serviceable available market, and serviceable obtainable market: progressively narrower market scopes whose assumptions need support.

LTV / CAC

Customer lifetime value and customer acquisition cost: estimates used to examine customer economics.

ICP

Ideal customer profile: a description of the customer a business intends to serve.

CTA

Call to action: the next action a message or page asks someone to take.

RICE

Reach, Impact, Confidence, Effort: a prioritization framework whose inputs still require judgment.

MoSCoW

Must have, Should have, Could have, Won’t have for now: categories for negotiating scope.

4R

Results, Roles, Rules, Rhythms: a framework for organizing team work.

ADR

Architecture Decision Record: a short record of a decision, its context, alternatives, and consequences.

Repository / branch / commit

A versioned project, a line of work within it, and a recorded change.

Pull request

A proposed change presented for review before integration.

Instructions / context / input

How the assistant should work, the facts it should use, and what changes for this particular run.

This is alpha material. Report unclear instructions or errors through a public issue or email help@bigleverinstitute.org. Share only redacted or fictional material.