Compilation··3:28·Combinatorics
Borsuk's Conjecture Fails in Dimension 9, Sidorenko's Conjecture and 8 More
Watch on YouTubeAbout the problem →
About this video
10 results from OpenAI's public mathematics catalogue, one per chapter. Each chapter names the problem, shows the real objects, and states what OpenAI claims; none of it is peer reviewed yet.
Chapters:
0:00 Borsuk's Conjecture Fails in Dimension 9, OpenAI Claims
0:21 Sidorenko's Conjecture
0:42 Ryser's Conjecture Disproved?
1:02 Hindman's Finite Sums and Products Conjecture Solved
1:24 Erdős Unit Distance Problem
1:45 Combinatorial Invariance Conjecture
2:06 Shareshian–Wachs Conjecture
2:27 Seymour's Second Neighbourhood Conjecture
2:47 Talagrand's Expectation Threshold
3:08 Second Kahn–Kalai Conjecture
Borsuk's Conjecture Fails in Dimension 9, OpenAI Claims: A ball can be cut into 4 pieces that are each narrower than the ball.
Explained: gradientdamage.com/p/156
Paper: github.com/openai/math/blob/main/preprints/A-nine-dimensional…
Sidorenko's Conjecture: Sidorenko's conjecture says that for every bipartite pattern H, every graph G contains at least as many copies of H as a random graph with the same edge density: t(H, G) ≥ p(G)^e(H).
Explained: gradientdamage.com/p/161
Paper: github.com/openai/math/blob/main/preprints/A-counterexample-t…
Ryser's Conjecture Disproved?: Ryser's conjecture (1971) says that if a hypergraph's vertices split into r parts and every edge takes exactly one vertex from each part, then r − 1 vertices are always enough to touch every edge; in the intersecting…
Explained: gradientdamage.com/p/162
Paper: github.com/openai/math/blob/main/preprints/Balanced-Counterex…
Hindman's Finite Sums and Products Conjecture Solved: Colour every whole number with finitely many colours.
Explained: gradientdamage.com/p/164
Paper: github.com/openai/math/blob/main/preprints/Monochromatic-fini…
Erdős Unit Distance Problem: Erdős asked in 1946 how many pairs among n points in the plane can be exactly distance 1 apart; since Spencer, Szemerédi and Trotter (1984) the best upper bound has been of order n^(4/3).
Explained: gradientdamage.com/p/167
Paper: github.com/openai/math/blob/main/preprints/A-power-saving-for…
Combinatorial Invariance Conjecture: Kazhdan–Lusztig polynomials attach a polynomial to every interval of the Bruhat order of a Coxeter group, and they are defined through the Hecke algebra, not through the interval's shape.
Explained: gradientdamage.com/p/168
Paper: github.com/openai/math/blob/main/preprints/Combinatorial-Inva…
Shareshian–Wachs Conjecture: Line up intervals of length 1, join the ones that overlap, and colour the graph so neighbours differ.
Explained: gradientdamage.com/p/169
Paper: github.com/openai/math/blob/main/preprints/Elementary-Positiv…
Seymour's Second Neighbourhood Conjecture: Draw arrows between dots, one way only, never both ways.
Explained: gradientdamage.com/p/173
Paper: github.com/openai/math/blob/main/preprints/A-proof-of-Seymour…
Talagrand's Expectation Threshold: Add random links to a network and patterns appear suddenly: a triangle shows up right around when you would expect one, but pairing up every dot takes longer.
Explained: gradientdamage.com/p/175
Paper: github.com/openai/math/blob/main/preprints/Graph-Decompositio…
Second Kahn–Kalai Conjecture: In a random graph G(n, p) every link appears independently with chance p.
Explained: gradientdamage.com/p/176
Paper: github.com/openai/math/blob/main/preprints/The-second-Kahn-Ka…
The animations illustrate the results; they are not proofs.
Gradient Damage: mathematics you can see, equations you can hear.
youtube.com/@gradientdamage
Chapters
One problem per chapter. Each name links to the result's own page, each time to that moment on YouTube.
- Borsuk's Conjecture Fails in Dimension 9, OpenAI Claims Combinatorics
- Sidorenko's Conjecture Combinatorics
- Ryser's Conjecture Disproved? Combinatorics
- Hindman's Finite Sums and Products Conjecture Solved Combinatorics
- Erdős Unit Distance Problem Combinatorics
- Combinatorial Invariance Conjecture Combinatorics
- Shareshian–Wachs Conjecture Combinatorics
- Seymour's Second Neighbourhood Conjecture Combinatorics
- Talagrand's Expectation Threshold Combinatorics
- Second Kahn–Kalai Conjecture Combinatorics
Chapter 1: about the problem
Combinatorics·OpenAI Math·Result 156
Borsuk's conjecture fails in dimension nine
Constructs a compact subset of ℝ9 that cannot be covered by ten sets of strictly smaller diameter, disproving Borsuk's covering assertion already in dimension nine. The example consists of rank-one orthogonal projectors onto lines in ℝ4, with the Frobenius metric. (Lean)
Chapter 2: about the problem
Combinatorics·OpenAI Math·Result 161·from 0:21
Counterexamples to Sidorenko’s conjecture and the forcing conjecture
Disproves Sidorenko's conjecture with a connected bipartite pattern on 35 vertices and 66 edges that occurs less frequently than in a random graph of the same edge density. The same pattern disproves the forcing conjecture of Skokan and Thoma: matching its density and the edge density of a constant graphon need not force quasirandomness. (Lean)
Chapter 3: about the problem
Combinatorics·OpenAI Math·Result 162·from 0:42
Counterexamples to Ryser’s covering conjecture
Disproves Ryser's covering conjecture by constructing intersecting (q+1)-partite, (q+1)-uniform hypergraphs with covering number q+1, rather than the predicted bound q, for every sufficiently large prime q. A separate construction over extension fields also disproves Gyárfás's monochromatic tree-cover conjecture. (Lean)
Chapter 4: about the problem
Combinatorics·OpenAI Math·Result 164·from 1:02
Hindman’s finite sums and products conjecture
Proves Hindman's finite sums and products conjecture: every finite coloring of the positive integers contains sets of any prescribed finite size whose nonempty subset sums and nonempty subset products all have one common color.
Chapter 5: about the problem
Combinatorics·OpenAI Math·Result 167·from 1:24
Planar distinct distances and unit-distance bounds
Proves the weak pinned Erdős distance conjecture: for every fixed ε > 0, all but o(n) points of any n-point planar set determine at least n1-ε distinct nonzero distances. A complementary theorem bounds the number of unit-distance pairs by O(n4/3-δ) for an absolute δ > 0. (Lean)
Chapter 6: about the problem
Combinatorics·OpenAI Math·Result 168·from 1:45
Combinatorial invariance of Kazhdan–Lusztig polynomials
Resolves the full combinatorial invariance conjecture: isomorphic Bruhat intervals in arbitrary Coxeter systems have identical equal-parameter Kazhdan–Lusztig polynomials. Thus the abstract order of the interval determines the polynomial, even across different Coxeter systems. (Lean)
Chapter 7: about the problem
Combinatorics·OpenAI Math·Result 169·from 2:06
Shareshian–Wachs elementary positivity
Resolves the elementary-positivity part of the Shareshian–Wachs conjecture: the chromatic quasisymmetric function of every natural unit interval graph has elementary-basis coefficients in ℕ[q]. The coefficients count explicitly described permutations, giving a combinatorial explanation of positivity. (Lean)
Chapter 8: about the problem
Combinatorics·OpenAI Math·Result 173·from 2:27
Seymour’s second-neighborhood conjecture
Proves Seymour's second-neighborhood conjecture: every nonempty finite oriented graph has a vertex with at least as many vertices at directed distance exactly two as at directed distance one. Oriented graphs may be arbitrary apart from the exclusion of loops and oppositely directed edge pairs. (Lean)
Chapter 9: about the problem
Combinatorics·OpenAI Math·Result 175·from 2:47
Talagrand’s expectation thresholds, discrete convexity, and graph decompositions
Proves that integral and fractional expectation thresholds differ by at most a universal factor, and resolves Talagrand's discrete-convexity conjecture. An application proves the Ascoli–He–Park–Talagrand graph-decomposition conjecture: every graph's edges split into a universally bounded number of fixed pieces, each with containment threshold at most a universal constant times the original graph's integral expectation threshold. The pieces' embeddings need not agree on shared vertices. (Lean)
Chapter 10: about the problem
Combinatorics·OpenAI Math·Result 176·from 3:08
The second Kahn–Kalai conjecture with an edge-count bound
Proves the second Kahn–Kalai conjecture: for every finite simple graph H with h ≥ 1 edges and at most n vertices, its appearance threshold in G(n,p) is at most C pE(n,H)(1+log2 h), with universal C. Here pE is the least density at which every subgraph of H has expected copy count at least 1/2. (Lean)
Topics
Borsuk conjectureSidorenko's conjectureRyser's conjectureHindman's conjectureErdős unit distance problemKazhdan-Lusztig polynomialsShareshian-Wachs conjectureSeymour's second neighbourhood conjectureTalagrand conjectureKahn-Kalai conjectureCombinatoricsOpenAI mathAI proofAI mathematicsmathsGradient Damage