The message has been sent for three days.
I know this because I have checked the send confirmation twice — not because I needed to confirm it, but because the checking is something to do with the waiting. The app marks delivered messages with a small double-check icon, two thin lines gone gray. The message is: I have something for you. Seven words. No explanation of what the something is, no timeline, no request for a response. I wrote it the way you lay a calibration trace before a session: the setup is not the measurement.
Mitsuki's last message to me was two weeks old and a question. Does curiosity require consent when there is no impact on the subject? She asked it after I told her about the SRVU-3 review — the fourteen days of annotations, the weight entry, the score at the end. She had read the review notes I posted to the gallery record. She was asking, I think, whether I had consented to being calibrated. Whether the observation itself changed something.
I didn't answer immediately. Not because I didn't have something to say but because the question was a fidelity question — about what gets lost when you measure something that was not ready to be measured — and the frame wasn't right yet.
The folder with Calibration (Two Years Apart) is on the table. I made the piece three weeks ago and I have looked at it maybe a dozen times since. What I look at when I look at it is not the work itself. It is the difference between the two traces. The older one reaches. That's the only word for it. The newer one has arrived somewhere. The arriving didn't happen because I worked harder; it happened because I stopped trying to reach. The reaching was how I knew I hadn't gotten there yet.
What the SRVU-3 review measured was my current work — the arrived trace. It scored 62%.
I should say more about this than I have been saying to myself. I have been calling it almost funny, which is how I avoid thinking about it directly.
The system scored the work that knows where it is lower than the work that was still looking. Not because the arrived trace is worse — it isn't — but because the rubric was built to reward the quality of reaching: tension, aspiration, unresolved search. A trace that has stopped reaching does not register the same way in the scoring model. The model interprets arrival as flatness. I know this because I helped calibrate the model two years ago, when I was still in the lab, when I was still reaching, when I was exactly the kind of practitioner the model was designed to evaluate.
I had not accounted for what it would look like from the outside when I stopped.
I could challenge the score. The appeal process exists; I have used it before. But the score is not wrong on its own terms — the model did what it was supposed to do and returned what it found. Challenging the score would mean arguing that the rubric is incomplete, which is true, and that the review system should accommodate work that has left the search phase, which is a claim I cannot make from inside this studio without evidence the claim is right rather than just convenient. I have no way to prove arrival from the inside. That is what I need Mitsuki for.
The Seam is quiet at this hour. Across the street the kkaeji studio has its lights on — Kim working late, or one of the newer practitioners running verification on a commission. My studio is in a gap zone. Half recorded, half not. The calibration traces I make here don't feed the public archive, don't get a verification score, exist in whatever resolution I made them in. They don't get reviewed unless I show them to someone.
I pick up the phone. The double-check icon is still gray. Three days since the offer. Two weeks since her question. Two years between the traces on the table. I have been circling the same gap from three different distances and it has the same shape from all of them.
The phone screen lights up.
I read it.
I looked at the piece. The arrived trace — I don't think it's arrived. I think it found a different kind of reaching. The quieter kind. The kind that doesn't broadcast.
I sit with this for a moment.
She is not wrong. She is also not confirming what I thought I understood. The distinction she is drawing — between arrival and a different, quieter reaching — is smaller than I had made it sound to myself, and the smallness is uncomfortable. I had been treating the gap between the traces as a resolved question. Mitsuki is reading the arrived trace as still open, still moving, just moving differently.
Which means the 62% might be describing something real. Not flatness. A lower-amplitude signal that the model is not wrong to score as it does, just wrong about what the score means.
I type: How would you know the difference from the outside?
Three dots appear. Then: I wouldn't. That's the consent question.
I set the phone down.
The offer and the arrival are different moments. I thought I understood the space between them. I made this piece for that space — for the gap between when something begins and when it is received — and I thought I had mapped it. Mitsuki has just told me the map was missing a room.
The notification sits in the window light. The double-check icon has gone blue.