What we did, and what we are claiming.

What we did. Between April 23 and May 11, 2026, Adam Cagle studied physics from introductory mechanics through the research frontier of quantum gravity, using a structured curriculum of 38 reference documents developed in collaboration with Claude (Anthropic). On May 10 we identified an open problem in the observer-complementarity literature and began research. On May 11 we had a paper.

What we are claiming. Two things, which are separable.

First, a physics claim: that the expected entropy disagreement between two observers reconstructing a quantum state through a non-isometric holographic code depends on the complexity class of the state, and that this dependence has a specific mathematical form with an integer exponent gap. This claim is in the paper. It is supported by analytic derivation and numerical verification. It stands or falls on its own.

Second, a methodological claim: that a motivated generalist working with an AI can reach the research frontier of a technical field in weeks rather than years, and can produce original contributions there. This claim is what the site is about. It stands or falls on whether the physics claim is correct, and on whether the process was as described.

We are not claiming that the AI wrote the paper. We are not claiming that Adam wrote the paper. We are claiming that the collaboration wrote the paper, and that the collaboration is a genuine method, not a stunt.

What we are not claiming. That this method works for everyone, at all fields, in all conditions. That eighteen days of study produces expertise equivalent to a PhD. That the paper is definitely correct (papers get revised). That AI replaces physicists.

The honest version of the claim is: this is what happened, here is the record, here is the output. Make of it what you will.