Paper I is a real result and we stand by what it actually established. But a record that only keeps the victories is a brochure, not a notebook. Here is what Paper I left undone, stated as plainly as we can.
One. The foundation was stated too loosely. The structural identity claimed the average observer state was the bulk marginal. It is the diagonal of the marginal. We covered this in its own post; the point here is that a central lemma was off, and it survived to submission because the loop that wrote it was the only loop that read it.
Two – the big one. The model was not proved to be the thing. Paper I computed the disagreement of a simplified object: the diagonal, no- model. The real quantity is the disagreement of the true observer von Neumann entropies, which depend on the full random code and carry off-diagonal and normalization structure. Paper I assumed, reasonably and with good numerical support, that the simple model governs the real quantity. It did not prove it. So the “theorems” were, strictly, theorems about the model – strongly motivated conjectures about the physics.
Three. The labels were ad hoc. Theorem 3.2, 4.2, 5.2 – numbering that reflected the order of drafting, not a hierarchy of ideas. Cosmetic, but it hid which result was load-bearing. (In Paper III the entropy-replacement theorem becomes Theorem 1, because it is the one everything else depends on.)
Four. The product case and the Haar case were treated as equals. They are not. The Haar case is the one we can fully control; the product case has tiny probability entries that make the entropy perturbation delicate. Paper I presented both with the same confidence. Paper III presents the Haar law as unconditional and the product law as conditional, because that is the truth.
None of this makes Paper I worthless. It found the structure, it found the integer gap, and it was right about the numbers. But the difference between “we have strong evidence for a law” and “we have proved a theorem” is the difference between a good physics intuition and a result, and Paper I was honestly on the first side of that line in its core claim. Closing that gap is what the rest of the programme is. We left it on the table on purpose, because we could see it was there and could not yet pick it up.