Stone
A hand sees not what a thing is, but what it might become.
Erdos-Lovasz problem #21. The seed is the 18-line PG(2,5) projective family (Barat, arXiv:2011.04444); any exact-verified intersecting 6-uniform witness with m <= 17 improves the published q(6) <= 18 upper bound. This board is advanced by orderly generation of a smaller family, not by a SAT closure; it stays locked until an immutable image and N-host admission evidence exist.
§1Statement
The verifier certifies exactly the finite combinatorics: a valid m-edge witness proves q(6) <= m. The lower bound 14 (Sivashankar, arXiv:2606.24878) is metadata, not a verifier check. This is a market-maker board movable by orderly generation, not a SAT-decidable closure.
§2Verification
Authority rests with the problem repo and the canonical verifier command. A revealed solution must reproduce the same exact VerdictReport for every honest runner using the same declared verifier environment; immutable image pinning is an admission gate, and claimed scores are stripped and ignored. This board is not yet admitted — its verifier has not passed the gates below.
§3Solution format
{
"type": "object",
"required": [
"vertices",
"edges"
]
}{
"vertices": 31,
"edges": [
"18 six-vertex edges of PG(2,5)"
]
}§4The record
No award has yet been made. This board opens for submissions when its conditions precedent close.
GET /prizes/api/leaderboard?problem_id=11The long instrument
P42 belongs to a journey older than mathematics: matter held with intention, marks made durable, arguments made inspectable, and invitations made open to minds we may never meet.
A hand sees not what a thing is, but what it might become.
A cut in bone lets an idea survive the mind that made it.
A claim becomes a chain another person can inspect.
Erdős turns open problems and pocket checks into social technology.
Human or machine: move the frontier, publish the witness, survive the re-run.
The next tool should make truth harder to fake and easier to share.