The sum-product conjecture is false for real numbers, disproved via constructions of sets A with max(|A+A|, |AA|) ≤ |A|^{2-c} for absolute c > 0.
hub
Remarks on the disproof of the unit distance conjecture
17 Pith papers cite this work. Polarity classification is still indexing.
17
Pith papers citing it
abstract
We present a short, digested, human-verified version of the recent OpenAI-generated counterexample to the Erd\H{o}s unit distance conjecture, and a sequence of reflections on it. The argument relies crucially on ideas that may, at least in retrospect, be attributed to Ellenberg-Venkatesh, Golod-Shafarevich, and Hajir-Maire-Ramakrishna.
hub tools
years
2026 17representative citing papers
Constructs sets A subset R with |{x+y+(x-y)^2 : x,y in A}| <= |A|^{2-c} for some c>0, giving a counterexample to the Elekes-Rónyai problem via prime-splitting amplification.
There exist n-point planar sets where every subset A has a distance occurring ≥|A|²/n^{1−δ} times, confirming Erdős's 1980 isosceles-triangle conjecture and answering a repeated-distance question negatively.
Constructs lattice point sets with many rectangles and few isosceles triangles to produce explicit counterexamples to the Mizohata-Takeuchi conjecture for the paraboloid via transference principles.
Proves R(B_8, B_10) = 37 via an AI-assisted short proof with a Lean formalization of the upper bound.
Introduces an auditable four-stage diagnostic for LLM physics reasoning in novel frameworks and applies it to three parallel worlds, yielding pass rates of 6/15, 6/15, and 0/15 on frontier models with noted qualitative-quantitative asymmetry.
A framework incentivizes inter-coalition collaboration via asymmetric Shapley values and expander graphs to achieve verifiable decentralization in blockchains.
Embedding models fail to capture mathematical equivalence due to lexical bias, but contrastive training on informal-formal pairs improves performance on MELD and retrieval tasks.
LRMs show a large production-evaluation gap on the VAIR dataset with valid answers but invalid reasoning, driven by answer confirmation bias as evidenced by CoT analysis, linear probes, and causal patching.
The US is dismantling its mathematical training pipeline precisely as AI begins producing genuine mathematical discoveries, creating a strategic vulnerability that requires policy intervention and formal verification mandates.
A counterexample is provided to the pairwise independent correlation gap conjecture using AI assistance.
Derives weighted recursions for coefficients of mock theta functions f and ω via holomorphic projection on vector-valued Rankin-Cohen brackets, exploiting vanishing cusp form spaces.
Proves lattice sphere packings in R^N achieve density at least c N² loglog N 2^{-N} for infinitely many N, improving Klartag's bound via stochastic ellipsoid evolution with cyclotomic symmetries.
Adapts known construction to prove existence of c>0 and large finite A subset R with |AA+A+A| << |A|^{2-c}, plus corollaries for other sum-product expressions.
Optimizes integer multiplicities and a real parameter in Sawin's unit-distance certificate via an evolution strategy to obtain improved explicit lower bounds u(n) > n^{1.0152} and u(n) > n^{1.031}.
AI models discovered the worst-case complexity of the Kaczmarz algorithm for solving linear systems.
An incomplete attempt to show that the O(n^{4/3}) upper bound for unit distances among n points in the plane is not sharp, plus remarks on Szemerédi-Trotter incidences.
citing papers explorer
-
The sum-product conjecture is false for real numbers
The sum-product conjecture is false for real numbers, disproved via constructions of sets A with max(|A+A|, |AA|) ≤ |A|^{2-c} for absolute c > 0.
-
Split primes and the Elekes-R\'onyai problem
Constructs sets A subset R with |{x+y+(x-y)^2 : x,y in A}| <= |A|^{2-c} for some c>0, giving a counterexample to the Elekes-Rónyai problem via prime-splitting amplification.
-
The Minkowski grid has robustly many repeated distances
There exist n-point planar sets where every subset A has a distance occurring ≥|A|²/n^{1−δ} times, confirming Erdős's 1980 isosceles-triangle conjecture and answering a repeated-distance question negatively.
-
Rectangles, triangles and Schr\"{o}dinger waves
Constructs lattice point sets with many rectangles and few isosceles triangles to produce explicit counterexamples to the Mizohata-Takeuchi conjecture for the paraboloid via transference principles.
-
An automated proof that R(B_8,B_10)=37
Proves R(B_8, B_10) = 37 via an AI-assisted short proof with a Lean formalization of the upper bound.
-
Testing Frontier Large Language Models' Physics Literacy in Parallel Physical Worlds
Introduces an auditable four-stage diagnostic for LLM physics reasoning in novel frameworks and applies it to three parallel worlds, yielding pass rates of 6/15, 6/15, and 0/15 on frontier models with noted qualitative-quantitative asymmetry.
-
Rethinking Collaborative Trust for Verifiably Decentralized Blockchain Systems
A framework incentivizes inter-coalition collaboration via asymmetric Shapley values and expander graphs to achieve verifiable decentralization in blockchains.
-
Does My Embedding Reflect That $A = B$? Evaluating Mathematical Equivalence in Embedding Models
Embedding models fail to capture mathematical equivalence due to lexical bias, but contrastive training on informal-formal pairs improves performance on MELD and retrieval tasks.
-
An Enigma of Artificial Reason: Investigating the Production-Evaluation Gap in Large Reasoning Models
LRMs show a large production-evaluation gap on the VAIR dataset with valid answers but invalid reasoning, driven by answer confirmation bias as evidenced by CoT analysis, linear probes, and causal patching.
-
Automation Without Understanding
The US is dismantling its mathematical training pipeline precisely as AI begins producing genuine mathematical discoveries, creating a strategic vulnerability that requires policy intervention and formal verification mandates.
-
Counterexample to a conjecture on the pairwise independent correlation gap using AI
A counterexample is provided to the pairwise independent correlation gap conjecture using AI assistance.
-
Recursions for Mock Theta Functions
Derives weighted recursions for coefficients of mock theta functions f and ω via holomorphic projection on vector-valued Rankin-Cohen brackets, exploiting vanishing cusp form spaces.
-
Stochastically evolving ellipsoids with symmetries
Proves lattice sphere packings in R^N achieve density at least c N² loglog N 2^{-N} for infinitely many N, improving Klartag's bound via stochastic ellipsoid evolution with cyclotomic symmetries.
-
More sum-product type counterexamples: products with shifts and $AA+A$
Adapts known construction to prove existence of c>0 and large finite A subset R with |AA+A+A| << |A|^{2-c}, plus corollaries for other sum-product expressions.
-
Optimizing Explicit Unit-Distance Lower-Bound Certificates
Optimizes integer multiplicities and a real parameter in Sawin's unit-distance certificate via an evolution strategy to obtain improved explicit lower bounds u(n) > n^{1.0152} and u(n) > n^{1.031}.
-
How AI settled the complexity of the oldest SGD algorithm
AI models discovered the worst-case complexity of the Kaczmarz algorithm for solving linear systems.
-
An incomplete attack on the upper bound of the unit distance problem
An incomplete attempt to show that the O(n^{4/3}) upper bound for unit distances among n points in the plane is not sharp, plus remarks on Szemerédi-Trotter incidences.