Camera Obscura

Kircher, 1646

AI inverts before it shows. You see a projection, not the world.

Vermeer probably used a camera obscura. The almost photographic quality of light in his paintings tells us so. Purists are scandalized: “So he was not a genius, only technique!” But that is exactly the point. Vermeer did not understand optics. He did not know why the image inverted. He did not care. He used that black box to see differently, and from that different vision came art.

AI is our epistemological camera obscura of the 21st century. We do not know exactly how it works, but through its inversion, its distortion, its alien gaze, we can see things that would otherwise escape us.

The Camera Obscura as a lens teaches us that opacity is not always an obstacle to overcome. Sometimes it is an epistemological feature: it forces us to trust an instrument we do not fully understand, and this trust, critical, not blind, but aware of its own limits, is the condition of contemporary vision.

The Set-Up is the studio in which the Camera Obscura is installed. Vermeer could not simply point the device at anything: he had to prepare the space, calibrate the light, position the subject, decide the framing. So with AI: the Set-Up defines the conditions in which opacity can become productive instead of paralyzing.

Without an architecture of governance, the Four Nails: transparency about limits, auditability of decisions, controllability of processes, reversibility of errors, the Camera Obscura does not produce vision, it produces only darkness.

When in the book you meet references to the Camera, think of this: an instrument that inverts to reveal, that darkens to show, that denies transparency in order to produce vision.

→ Full treatment in Chapter 4: A Plea for the Black Box.