Persona Drift Detector

4/11/2026 seed

Preamble

Persona Drift Detector is a private tool for comparing my past and present writing. It maps changes in voice, model influence, beliefs, decision style, emotional language, and taste, then leaves the meaning of those changes with me.


It Reads The Record You Approve

The tool reads approved material: journals, drafts, decision notes, model chats, critique notes, and taste references. Each piece keeps its source, date, audience, and context attached because a private note, a public essay, and a model-assisted draft do different work.

There Is No Pure Starting Point

Past writing is an earlier arrangement of pressure: borrowed language, old audiences, admired writers, professional habits, private rituals, and the emotional weather of that time.

A Voice Is An Arrangement

Originality lives in the collision of influences inside one life. A sentence can carry Chekhov, internet cadence, family shame, professional training, and model phrasing at once. The question is whether the arrangement still feels necessary.

The Model Changes The Rate

LLMs make borrowing faster and smoother. They can flatten a risky sentence into competent prose. They can also surface a phrase I recognize as already waiting underneath. The tool can show that a pattern appeared after model contact, then leave the harder question open: did the model deposit it or discover it?

The Map Should Provoke

The first useful output is a pattern map worth arguing with. It might show that a phrase spread, a decision habit weakened, a voice became more generic, or a visual preference moved. I decide whether the signal is growth, contamination, exhaustion, adaptation, or noise.

The Line Stays Human

The system can show examples, dates, comparisons, confidence, and missing data. I choose which borrowings to keep, which to refuse, and which to watch next time.