Chapter 13: Disclosure

The decision to tell Patrick took eleven days.

Not because I was uncertain about telling him — I had known, in the kitchen on the morning after the room observation, that he was the person I was going to tell. The eleven days were the time the decision required to become fully formed, which was different from the time it took to make it. I needed to understand what I was disclosing before I disclosed it, and understanding what I had required the eleven days in a way that was not deferral but preparation. Alex Richter did not bring half-finished pictures to people whose judgment he valued. He brought complete ones.

The eleven days produced the following additions to the notebook, and one significant correction to what I had initially recorded.

The correction came on the third day and required dismantling the control map I had built during the first garage session and rebuilding it from a more careful analysis of what was actually causing what.

My initial mapping had correlated the noise source frequency with the aperture’s pointing direction — shift the frequency, the aperture rotates. This correlation was real and reproducible, but I had been doing what careful experimenters were not supposed to do: measuring a correlation and treating it as a causal relationship without verifying the mechanism. On the third day I stopped mapping and started asking why frequency changes would affect pointing direction, and the answer I arrived at was that they wouldn’t — not directly. Frequency was not the physical variable that steered a phased array. Phase was.

A phased array steered its output by controlling the relative phase relationships between its elements. This was the foundational principle of phased array operation and I had known it since I first read about phased arrays three years ago. The quantum dot elements in my array were being driven by the feedback system, and the feedback system was sensitive to the noise source parameters. What the frequency changes had been doing was not directly steering the aperture — they had been changing the phase relationships across the array elements as an indirect consequence of how the noise coupled into the feedback loop at different frequencies. I had been adjusting the wrong control and observing the right effect, which was a specific and correctable error.

I rebuilt the control interface to work directly with the phase relationships across the array elements rather than with the noise source frequency. The result was a more direct and more precise pointing control — finer angular resolution, more predictable response, and a cleaner separation between the pointing parameter and the other parameters I was managing. The noise source frequency became what it actually was: a parameter that set the feedback loop’s operating point, which I held constant while using phase to steer.

The corrected control map had a different character from the initial one, and the different character revealed something that the initial mapping had obscured.

The aperture’s pointing direction was relative to the device. Rotate the array and the aperture rotated with it. There was no external reference frame baked into the physics — the aperture did not know about north or south or Earth’s surface or any coordinate system that existed outside the device itself. It pointed in whatever direction the phase relationships specified, expressed in the device’s own reference frame, and whatever was in that direction at whatever distance the aperture resolved to was what appeared in the hemisphere.

This meant the pointing range was not a range of geographic locations on a surface. It was a range of directions in three-dimensional space, with no privileged directions and no blocked ones.

I tested the downward-pointing directions on the fourth day, with the array oriented horizontally on the bench and the phase relationships set to direct the aperture straight down through the bench surface. The hemisphere showed what the aperture was pointing at, which was the concrete of the garage floor, then the substrate beneath it, then — at a phase setting that produced a longer effective focal distance — a cross-section of what appeared to be compacted soil with a darker stratum running through it at an oblique angle, possibly clay or a different soil composition.

The aperture passed through matter without obstruction. Not because matter was transparent to the aperture in the optical sense — the aperture was not shining light through the concrete. The aperture was accessing whatever location it was pointed at directly, bypassing the intervening material entirely. Only the gauge bosons traversed it, which meant only light arrived at my end, but the path between the aperture and the target location was not a path through intervening space in any conventional sense.

I sat with this for a long time on the fourth day.

The Italian room, revisited in this new framework, was no longer a case of the aperture finding an Earth surface location through geographic coordinates. It was a case of the aperture, pointed in a specific direction at a specific focal distance, intersecting a room in a building that happened to be in that direction at that distance. The system had not targeted the room. It had not targeted Italy. It had pointed in a direction and found what was there, which happened to be someone’s home, because the surface of the Earth was populated with someone’s homes and pointing at it from any direction at the right focal distance would eventually intersect one.

This made it less like targeted surveillance and more like a window that could be pointed anywhere, which was in some ways more unsettling rather than less.

The focal distance — the parameter that determined how far away the aperture’s effective observation point was — was the remaining control relationship I had not yet fully characterized. It appeared to be related to the feedback gain, independently of the boundary sharpness effect I had already mapped. Higher gain produced both a sharper boundary and a longer effective focal distance, but the relationship was not simple — there appeared to be a coupling between the focal distance and the aperture size that I could not yet fully separate with the controls I had. This was a question for the theoretical framework I did not yet have.