来源：https://dev.to/jugeni/vibe-coding-is-not-a-level-its-an-axis-12gb

Karpathy gave us vibe coding: "see stuff, say stuff, run stuff, copy and paste stuff, and it mostly works." Since then, the industry has kept trying to turn it into a tidy autonomy ladder — Level 0, Level 1, all the way up to fully autonomous development. That ladder is useful. It is also incomplete. It measures one thing: how much of the building you delegate to AI.

But two people can delegate the same amount and get radically different outcomes. One compounds. The other accumulates entropy. Same autonomy level. Different operating system. That's the missing axis: operator discipline.

By operator discipline I mean one thing: how much of your work survives the session boundary as inspectable state.

Here's the question the vertical can't answer: Two developers are both at Level 4. One ships features that compound — the codebase gets cleaner, their operating context gets sharper. The other ships features that decay — the codebase grows entropy, every new prompt is a fresh negotiation. Same vibe coding level. Different outcomes.

For about three months I kept re-explaining the same architecture decision to the model every few sessions. Each time it would respectfully suggest the alternative I'd already rejected. Then I started writing those decisions down in a separate store, with a status field: proposed → accepted → locked. Once a decision is locked, the model is told not to relitigate it. The relitigation stopped. The work got calmer. Nothing about my vibe coding level changed. What changed was that a decision became a piece of state instead of a thing I had to defend live.

That's the axis. Not "are you good at prompting" — how much of your context is a state machine, vs. how much is reconstructed from scratch each session.

If autonomy is L0–L5 and operator discipline is Low/High, you get twelve cells. The diagonal that matters isn't "low everything → high everything." It's the cross-axis claim: L1 + High operator discipline > L5 + Low operator discipline over any time horizon longer than a sprint.

What operator discipline actually is: A persona file the model loads each session. Three append-only stores: decisions with lifecycle, threads for active workstreams, notes as atomic facts with source-anchoring. A capture habit — decisions go into the store the same turn they happen. Locked decisions stop the death-by-second-guessing loop.

Discipline doesn't beat fluency. They multiply. But if "more AI" hasn't translated into "more leverage" for you, the answer might not be a smarter model. It might be the axis you weren't measuring.

（此帖无评论）