---

Claim. Let $f(N)$ be the largest number of residue classes with distinct moduli at most $N$ that are pairwise disjoint, the quantity asked for in Problem 202, and put $L(x)=\exp(\sqrt{\log x\log\log x})$. The draft's Theorem 1 states that $f(N)=NL(N)^{-1+o(1)}$ as $N\to\infty$, the same asymptotic as Ho's accepted claim: the lower bound is the construction of de la Bretèche, Ford and Vandehey, and the new content is the matching upper bound $f(N)\le NL(N)^{-1+o(1)}$, which the draft proves first and then transfers to the reciprocal-sum supremum of Problem 1190. Its route keeps the counting bounds of de la Bretèche, Ford and Vandehey and replaces the dense-core conjecture for intersecting families, which they used conditionally, by the spread lemma from sunflower and threshold theory (the draft's Theorem 2, that a sufficiently spread uniform family contains disjoint members), so that the argument is unconditional apart from the established theorems it quotes. The draft, eight pages dated 30 April 2026 and labeled a draft note, prints no author name; it is hosted on ULAM's research pages and was posted on the site's discussion thread for Problem 1190 on 2026-04-30 by Przemek Chojecki, who credits the proof to GPT-5.5 Pro. The claimant named here is the human who posted it. On that thread Nat Sothanaphan noted the same day that the writeup claims both this problem's asymptotic and then Problem 1190 as a consequence, the same results as Ho's manuscript by a very similar route, and that a standard check had found a few minor issues. The community ledger of AI contributions, in its last revision (data), lists the draft under this problem as a candidate full solution. The draft's Problem 1190 claim has its own page.

Depends on. Nothing in this wiki: the draft's inputs are the published counting bounds of de la Bretèche, Ford and Vandehey and the spread lemma, which it imports from outside this wiki; whether its route is independent of Ho's is not certified by any record.

Standing. Claimed. No outside reviewer has published an examination beyond the thread remark recorded above, the draft has no journal record and no formalization, and the site credits the solution of this problem to Ho's manuscript and its formalization, not to this draft, which was posted seven days after Ho's announcement. The problem's standing rests on Ho's accepted claim.