Until you project a piece of yourself — an image, a voice, your words, anything with you in it — it is one locus, here, yours, in a single state. The moment you do, it exists in many states at once. Each copy takes on its own life. None of them collapse back.
The light at the center is you, before you project — undivided, in your hands. Each time you project, a copy is released into the field and begins a life you no longer hold. It doesn't come back. Tap the field, or the button.
Public photos have been built into searchable face indexes. Clearview AI assembled one from billions of images scraped from the open web.
Open image sets scraped from the web — LAION-5B holds about 5.9 billion image–text pairs — became the training material behind generative models.
Once a face has been seen, it can be re-rendered — moved, aged, given a voice it never spoke — by a model, without you present.
What you post is assembled into a profile that moves with you across the network — read by parties you will never meet, whether or not you post again.
“When you project a piece of yourself — an image, a sound, a poem — you offer it to a field where it exists in many states at once.”
An image is only the most legible vector; the law is the same for anything with you in it. Until you project it, it is potential — one locus, here, yours. The act makes it actual, and plural: many copies, each in its own state, each in someone else's hands. That is the oldest law there is — potential becoming actual the instant something acts on it. The new part is only the scale.
It is an offering — given freely, in hope of return, to powers that receive and dispose of it beyond your sight. That it is given, not taken, is why it never felt like theft. The only question that matters is the one the offering raises: does the field return anything close to what you gave?
What this instrument shows is documented, not asserted: the Clearview AI facial-recognition index (restricted under Illinois' Biometric Information Privacy Act, 2022); the open LAION-5B image dataset behind diffusion models; routine likeness synthesis. It presents what is. You read what it means.