Glossary: AI Coding & Learning Terms
Plain-English definitions of the key terms behind Contral and the shift to learning while you build. Each term links to a deeper explanation.
- Vibecoding
Vibecoding is using AI tools to generate code by describing what you want in natural language, without fully understanding the code that gets produced. It makes shipping fast but leaves developers unable to explain, debug, or defend their own codebase.
- Teaching IDE
A teaching IDE is an integrated development environment that explains the code as it is written, so you learn while you build instead of after. Contral is a teaching IDE: its AI agent writes code and explains every function, pattern, and decision in real time.
- Teaching layer
A teaching layer is runtime infrastructure that sits on top of an AI coding agent and explains every change it makes — what the code does, why this approach, and what the alternatives are — turning passive code generation into active learning.
- Build Mode
Build Mode is Contral's repo-aware AI coding agent. It reads your files, writes diffs, and runs commands inside the IDE at full vibecoding speed — while the teaching layer explains each change so you understand what you ship.
- Learn Mode
Learn Mode is Contral's structured, AI-powered curriculum inside the IDE. It teaches a language (starting with Java) from zero to mastery through interactive lessons, checkpoints, challenges, and real projects — not videos or browser sandboxes.
- Defense Mode
Defense Mode is a Contral feature that pauses after the AI generates code and challenges you to explain what was just built. Inspired by academic thesis defenses, it makes sure you can defend every decision before you ship — so you never carry code you can't explain.
- AI coding agent
An AI coding agent is a tool that autonomously reads a codebase, writes and edits code, and runs commands on your behalf (e.g. Cursor, Copilot, Windsurf, Claude Code). Contral pairs an agent with a teaching layer so you learn from it rather than just consuming its output.
- Code-comprehension gap
The code-comprehension gap is the difference between the code a developer ships with AI assistance and the code they actually understand. Research has found AI-assisted developers score measurably lower on code-comprehension tests — the gap Contral is built to close.
See these ideas in action
Contral is the IDE that teaches you while you build. Try it free.
Get Started Free →