Vibe coding
Prompt-led exploration with lightweight structure.
Useful when
- Fast to test an uncertain idea
- Low setup cost for disposable work
Watch for
- Requirements and checks can remain implicit
01 · Comparison
Five criteria compare prompt-led exploration with structured coding. Code becomes maintainable through requirements, review, verification, and ownership, regardless of who typed the first draft.
5 decision criteria · first-person editorial guidance
02 · The two modes
Prompt-led exploration with lightweight structure.
Explicit requirements, review, verification, and ownership—whether drafted by a person or AI.
03 · Decision criteria
| Criterion | Vibe coding | Structured coding | Result |
|---|---|---|---|
| Time to a first shape | Fast because discovery and implementation happen together. | Slower because intent and checks are made explicit first. | Favours Vibe coding |
| Exploration cost | Low for a question that may be discarded. | Structure may be wasteful before the problem is understood. | Favours Vibe coding |
| Reviewability | Intent can remain inside the conversation that produced the code. | Requirements, changes, and acceptance evidence can be reviewed independently. | Favours Structured coding |
| Maintainability | Depends on whether structure is added after the first draft. | Ownership, naming, tests, and operating constraints are part of the work. | Favours Structured coding |
| Production responsibility | The mode does not define release, monitoring, or recovery ownership. | A responsible delivery process names release, monitoring, and recovery owners. | Favours Structured coding |
My recommendation · reviewed Jul 13, 2026
Calling one side ‘real’ hides the important variable. Human-written code can be unreviewed and fragile; AI-drafted code can be carefully specified, tested, reviewed, and owned. The 3–2 result favors structured coding for anything expected to remain in service.