Browse › Number theory › 012 Independent largest prime factors of consecutive integers
The joint Dickman law for consecutive integers
Open PDFCite Report a problem Add a reference or comment Interested in working on this Start a discussion
Edit NOTES.md · Existing issues · Source on GitHub
Abstract
Let $P^+(n)$ denote the largest prime factor of n. We prove that $\log P^+(n)/\log n$ and $\log P^+(n+1)/\log n$ are asymptotically independent in ordinary natural density, with Dickman marginals. This resolves the Erdős–Pomerance joint Dickman conjecture positively and implies that the ordering $P^+(n)\lt P^+(n+1)$ has natural density 1/2.
If the PDF does not display here, open it directly: Open PDF.
Notes
No notes yet. Add a reference, a comment or your name with the buttons above, or edit NOTES.md on GitHub.
Automatically collected notes (unreviewed)
work in progress
Collected automatically by AI tools (Claude Opus 5.5 and GPT-6 Astra, 2026-10-21) and not checked by a person. May be wrong or incomplete. Report an error.
Follow-up works
- Ji Ho Bae (jbaelaw), Erdos928 (GitHub repository, Lean 4/Mathlib, accessed 2026-10-21). Formalizes in Lean a counting estimate and an explicit implication from the joint Dickman law of this paper to Erdős Problem 928; it explicitly does not re-prove or import the paper's analytic theorem. [AI-generated per source]
Comments and discussion
- Thomas Bloom, Erdős problems in the OpenAI proofs (erdosproblems.com blog, 8 Oct 2026). Lists Problem 371 among the release's claimed results, notes that the 84-page PDF proves it as a corollary of a more general joint distribution law and that a Lean formalisation exists, and says he had not checked the proofs or formalisations himself. [family-level]
Formalization
- OpenAI math release, lean/docs/012.md (scope notes, mirrored) (oai-math-revised, 2026). The release's own scope note: Lean proofs of the joint Dickman law in natural density for $0<a,b<1$ and of density $1/2$ for each ordering, with comparator statement
JointDickman.lean. [family-level] - OpenAI, PaperMain.lean (OAI.JointDickmanPaper) (openai/math, commit adc7f12). States
joint_law(Theorem 1),increasing_order(Corollary 2) anddecreasing_order, with proofs imported fromJointDickman.UnconditionalMain. - Google DeepMind formal-conjectures, FormalConjectures/ErdosProblems/371.lean (formal-conjectures, accessed 2026-10-21). Formal statement of Erdős Problem 371 (density $1/2$ for $P(n+1)>P(n)$, via
Nat.maxPrimeFacandHasDensity), tagged as research-solved and linked to the OpenAI Lean proof; the proof body in this repository is stillsorry. - mo271, feat: mark problems solved in the OpenAI math release and link formal proofs (PR #6898) (formal-conjectures, merged 11 Oct 2026). Marks Erdős 371 as solved; the PR notes that the linked proof differs from the repository statement by a routine transport ("density over Icc 2 ⌊X⌋ vs Iio b") that the linked proof does not contain. [family-level]
Related prior literature mentioned by others
- Zhiyuan Yang, Convolution-type Bombieri-Vinogradov theorem with well-factorable weights, and its applications (arXiv:2608.13299, Aug 2026, v2 Sep 2026). Proves the lower bound $0.299x$ for $\#\{n\le x:P^+(n)<P^+(n+1)\}$, improving the 0.280 of Yang2026 cited in the paper. The erdosproblems.com thread for Problem 371 points to it, but the paper's bibliography does not.
Possibly relevant literature (found by the agent, not checked)
- Zhiyuan Yang, Note on the largest prime factors of consecutive integers (arXiv:2610.23280, 16 Oct 2026). Gives an alternative proof of the Tao–Teräväinen positive-proportion results on patterns of $P^+$ at consecutive integers, discussed in the paper's "Previous work" section, and a lower bound $\gg x/\log x$ for $P^+(n+1)/P^+(n)\in[1,1+\varepsilon)$; it does not cite the release.
Digests and expositions
- eigendash, openai-math-visualised (GitHub repository, accessed 2026-10-21). Includes a numerical probe of family 012 up to $6\times10^7$: the density of $P^+(n)<P^+(n+1)$ is reported as 0.499983, with a comparison of the binned joint distribution against the product of Dickman marginals.
Elsewhere
- VibeMathed: Problem record
- Emergent Mind: family 012
Independent sites; their content is not reviewed here.
Cite
@misc{OAI:The-joint-Dickman-law-for-consecutive-integers-September-24-2026,
author = {{OpenAI}},
title = {{The joint Dickman law for consecutive integers}},
howpublished = {OpenAI Math Release preprint
\href{https://github.com/openai/math/blob/main/preprints/The-joint-Dickman-law-for-consecutive-integers-September-24-2026/paper.pdf}{OAI:The-joint-Dickman-law-for-consecutive-integers-September-24-2026}},
year = {2026}
}Lean formalization
OpenAI provides Lean files for this family: lean/docs/012.md. The Lean statements have not been checked by people against the paper's claims.
Family 012: Independent largest prime factors of consecutive integers
Resolves the Erdős–Pomerance joint Dickman conjecture: the logarithmic sizes of the largest prime factors of n and $n+1$ are asymptotically independent in ordinary natural density. In particular, the integers satisfying $P^+(n)\lt P^+(n+1)$ have density 1/2.
This is the only paper in this family.