← all field notes
MODEL · REVIEWNeill's Vibe · Field Notes

Three other AIs found the bug mine swore wasn't there.

A model reviewing its own work brings its own blind spots to the inspection.

✓ what worked
01 · the learning

What I set out to add

When one model writes the code and the same model reviews it, you're asking the same mind to catch its own mistake — and it shares its own blind spots. I learned that self-review feels like review but isn't. The confidence is real; the coverage isn't.

02 · the test

The thing I actually tried

I tested a cross-provider review — pipe the same diff to models from different families (a Google model, a coding-tuned model) and let them argue. A small runner sends the change out, collects the verdicts, and flags anything two of them independently distrust.

cross-review.mjsREVIEW-DOCTRINE.md
03 · the frustrations

Where it bit

"Looks good. Tests pass. Ship it." It would have shipped a bug.
04 · the verdict

What worked, what didn't

✓ worked

Cross-family review. A different lineage catches what a shared lineage rationalizes. The bug that 'passed' got named by a reviewer that hadn't written it.

✗ didn't

Trusting a green test run plus one model's thumbs-up. Same author, same blind spot — the inspection was theater.

the durable rule

The reviewer that didn't write the code is the only one whose 'looks good' is worth anything.

🔒 this became a classT3·01 — Three Other AIs Found the Bug members