Field Guides / The Technician's Read
The Technician's Read
The human operator function that lives at the seam between Layer 1 and Layer 3 of every Atlas instrument protocol. It is not analysis. It is perception — what you see before anything else tells you what to see.
Why It Exists
Analysis models produce fluent, confident readings of data. Those readings are shaped by the framing of the analysis prompt. If the operator reads the model's output before forming their own perception, the model's framing becomes the operator's framing. The independent human read is gone before it ever existed.
The Technician's Read holds that gap open. It is a timestamped record of the operator's perception at two specific moments: before the session begins, and before the session is logged. Everything that happens in between — the model's output, the coded fields, the metrics — sits between those two reads. If the final synthesis contradicts what the operator saw, the contradiction is a finding, not a correction.
Where It Lives in the Protocol
Every Atlas instrument protocol has three layers.
Layer 1 — Investigator Preparation
Determines whether the operator is in a valid condition to run. The Technician's Read #0 is the last step of Layer 1. It is written before the session window opens. Not after. Not during. Before. A Read written after the run is retrospective reconstruction — it does not satisfy Tier A and cannot function as a prediction lock.
Layer 2 — Session Execution
The instrument operation. The operator runs the session, captures fields, computes metrics. The Tech Read Formatter stays open for the duration. Anything that cannot go into a coded field goes here, in real time.
Layer 3 — Investigator Record
Where the science lives. The Technician's Read #1 is written here, before the session is logged or closed. Layer 3 is not optional post-processing. Without it, the run has no interpretive author.
Read #0 — Before the Session
Written before the session window opens. One paragraph minimum. The cold prediction is locked here — it does not get revised after the session begins. The gap between what you predicted and what you observed is data. Closing that gap retroactively destroys it.
Read #1 — Before Logging
Written before the session is logged or closed. Returns to Read #0. This is the only place the operator's interpretive record exists. Models will produce synthesis. The Tech Read is what the operator saw before the synthesis arrived.
The Gap Is the Method
The three-layer structure is designed around one principle: the investigator must see the data before the analysis tells them what the data means.
This is not a check on the models. It is a check on the investigator. Model fluency inflates the reader's confidence in what the model says. A well-formed wrong answer sounds more authoritative than a halting correct one. The Technician's Read is the timestamped record that the operator had their own read before that fluency arrived.
The timestamp is part of the data. The Read is not analysis. It is perception.