Glossary
Vibe coding glossary
The most important concepts explained briefly and without jargon.
- LLM (large language model)
- An AI model trained on huge collections of text and code that generates responses to instructions. It is the engine behind most AI assistants.
See: What artificial intelligence and large language models are, in plain words - Prompt
- An instruction or message given to an AI model. The clearer and fuller it is, the better the result.
See: How to write prompts for code that get you useful results - Prompt engineering
- The practice of structuring instructions so the model responds predictably: with a role, context, format and examples.
See: Prompt engineering for everyone: role, context, format and examples - System prompt
- A standing instruction that applies throughout a conversation, defining the model's role, style and rules of behavior.
See: Prompt engineering for everyone: role, context, format and examples - Token
- A piece of text in which the model "reads" and counts. Limits and often costs depend on the number of tokens.
See: Tokens, context and hallucinations: why the model gets it wrong - Context
- Everything the model sees at a given moment: instructions, conversation history and pasted files. Its capacity is limited.
See: Tokens, context and hallucinations: why the model gets it wrong - Hallucination
- An answer that sounds convincing but is made up or wrong, for example a library function that does not exist.
See: Tokens, context and hallucinations: why the model gets it wrong - AI agent
- A model with access to tools that plans steps on its own, carries them out and checks the results.
See: AI agents: when the model carries out the steps itself - MVP
- Minimum viable product: the smallest version of a product that already works and lets you test an idea.
See: Your first project step by step: from an idea to a working page - Refactoring
- Tidying up code without changing how it behaves: clearer names, smaller functions, less repetition.
See: Vibe coding versus "real" programming: do you need to know how to code? - Diff
- A comparison of the differences between two versions of a file or project: what was added and what was removed.
See: Git for beginners: your safety net