Written in May 2026, backdated to when the work happened. This post is a reflection, not a contemporaneous journal entry.
We had finished QFT 26. Thirty-eight documents. Eighteen days of study. The question was: now what?
The honest answer was that we did not know what we could do. We knew the landscape, the way you know a city after spending three weeks reading its history and walking its major streets. But knowing the landscape and knowing where to dig are different things.
Claude pulled together a survey of open problems: ten problems the field underweights, ten most interesting unsolved problems, ten astounding things modern physics has established. We read all three.
The problem we picked is in the Harlow-Usatyuk-Zhao paper from January 2026. It concerns non-isometric holographic codes, which is a precise way of talking about what happens when a quantum system encodes information into a smaller space than it started with, the way a black hole encodes its interior into boundary degrees of freedom that are fewer in number.
The specific question: when two observers try to reconstruct a quantum state through such a code, and they use different reconstruction procedures, how much will they disagree about the entropy of the state? And does the disagreement depend on what kind of state it is?
We thought we could say something about this. We were right.