Erdős Problem #1196
Editorial depth: Authored explanatory DRAFT — NOT editorially accepted.
Snapshot classification: proved (not automatically the live result).
Subject areas: number theory, primitive sets.
Read this status correctly
The status above is reproduced from GCL's immutable external-catalogue snapshot. It is not a new claim of current mathematical resolution, a source-semantic review, an independent proof check, or GCL certification. A published proof and a GCL replay are separate events.
UNACCEPTED DRAFT — mathematical editorial review pending
This material is a provisional, source-qualified manuscript published for line-by-line scrutiny and real HTML/LaTeX testing. It has NOT received separate GCL editorial acceptance or independent mathematical certification. Its presence on this website is not evidence of a solution, reviewed coverage, or final publication.
Review the underlying draft on GitHub.
The question in plain language
Erdős Problem #1196 is EXTERNALLY PROVED IN 2026 in the source as of 15 October 2026.
A set of integers \(A\subseteq[x,\infty)\) is primitive if no two distinct elements divide one another. Erdős–Sárközy–Szemerédi conjectured that, as \(x\to\infty\), all primitive \(A\) satisfy
The use of an infinite sum over a possibly infinite primitive set and logarithm convention matter; \(a\ge2\) avoids \(\log1=0\). The result is a uniform upper bound over all primitive sets whose smallest element is large; it is not saying every primitive set has a sum tending to one.
A small example
The first five primes \(\{2,3,5,7,11\}\) are pairwise nondividing, hence a primitive set. Direct numerical calculation gives the finite reciprocal-log sum approximately 1.260353667393, which is greater than one and is not a contradiction: the theorem applies asymptotically when every element exceeds a large cutoff \(x\), not to \(A\) containing 2. This is an exact primitive-set structural fixture with rounded numerical evaluation.
What is known
The maintained https://www.erdosproblems.com/1196 marks PROVED and attributes a 2026 GPT-5.4 Pro argument, expounded by Alexeev–Barreto–Li–Lichtman–Price–Shah–Tang–Tao [ABLLPSTT26], yielding even a tail bound \(\sum_{a\in A,a>x}(a\log a)^{-1}\le1+O(1/\log x)\) for arbitrary primitive \(A\). Earlier Lichtman [Li23] achieved constant \(e^\gamma\pi/4\approx1.399\), and GLW24 gives fine near-one behavior of the primitive set of integers with exactly \(k\) prime factors. Original 2026 proof/source, exact hypotheses and potential Lean formalization have NOT been independently reviewed. This is not a novel GCL proof.
GCL's current state
Authenticated original issue-bound research source return. The exact bounded example was independently replayed through ci/erdos_compendium_007_phase23_checks.py; general mathematical results here remain explicitly attributed to primary/external proofs, not independently certified by GCL. This manuscript is a local UNACCEPTED editorial draft.
What remains
Independently source-review [ABLLPSTT26] and the GPT-5.4 Pro proof, verify uniformity in \(A\), the precise \(x\) cutoff and logarithm/rate in \(O(1/\log x)\). Compare Lichtman and GLW24 constructions for asymptotic sharpness and ensure that a claimed formalized statement is not conflated with a kernel proof.
Any critical sample approval, theorem source fidelity, public site rollout and protected V2 admission are separate uncompleted steps.
Sources and claim boundaries
https://www.erdosproblems.com/1196 ; [ESS68b],[Er80,p.101],[Li23],[Li20],[GLW24],[ABLLPSTT26].
Finite evidence: governance/erdos_compendium/batches/EC-BULK-007_FINITE_CHECKS_PHASE23.json. Frozen original source issue identity retained; independent mathematical source approval PENDING.
Catalogue and evidence provenance
- Pinned catalogue source: teorth/erdosproblems at
536408716724. - GCL issue: #2467 — protected issue/assignment identity; its labels and results are live, not reproduced automatically here.
- GCL triage: OUTSIDE_STRICT_OPEN_TRIAGE.
- Formal-statement coverage: A provider statement was included in the pinned snapshot. Presence does not mean the statement or proof has been checked.
- External mathematical status: consult current problem page and its dated bibliography.