Technology

I published an article about a vulnerability. Then I ran a review on it and watched it fire the vulnerability — in its own closing paragraph, with the rule against it three screens up.

Equal care, equal drift — that was the bet. The audit found the drift stacked in one file: the one with the most amendments.

My content-strategist agent started rejecting good work after the model changed. Two words fixed the routing: shape and fidelity. The vocabulary that turns cross-model review from judgment calls into rule application.

I ran two AI reviewers on the same code review. Zero overlap in findings, five of six rounds. Here's the asymmetric-coverage property—and how to run Dual-Rival correctly.

I was three quarters of the way through writing a rule about vocabulary drift when I hit save, ran a review, and found I'd done it again. Third time that session.

I had two versions of the same rule on my screen. Same rule. Different shape. One of them broke when the model on the other end changed how it reads.

I built an AI-assisted operating system on Claude. Then I packaged 55 files and asked Codex to audit it cold. The diagnosis was sharp. The prescription was 6x what the patient needed. Here's what happened — and what it teaches about cross-model review.

I built a security hook to prevent my AI from editing governance files. It worked so well it blocked the AI from fixing the hook itself. The pattern has a name now: F-GUARDIAN.

You don't need a team of 50 to compete with companies that have them. You need infrastructure that amplifies—and the clarity to direct it.