Every proof removes a sorry.

A pre-launch proof-seeking mathematics project designed to pursue open bounties. Candidate artifacts would count only when their proof code is kernel-checked, contains no `sorry` or `admit`, and has a reviewed clean axiom set.

fees → compute → proof → bounty → burn

Open Research Console

PRE-LAUNCH No Aristotle job, proof, payout, buyback, or burn is represented as completed.

Aristotle workflow — interface preview

PRE-LAUNCH

preview /po0f/prelaunch/interface
if configured, sourced records would be loaded for review
candidate formalizations may be assessed after verification
compute allocation would remain at 00 until authorization
proof validation could be queued only after a formal statement is confirmed
treasury flow would be disclosed before any action
under this prelaunch preview, no proof, payout, issuance, buyback, or burn would be claimed

Interface preview. No public run has started.

Research telemetry

PRE-LAUNCH

Tracked problems

03Curated public-source candidates

Active runs

00No public execution is represented

Verified proofs

00No proof is claimed

Published artifacts

00No artifact is published

02 / PROOF LEDGER

Open problems under review

A curated catalog from public sources. Source status and prize language belong to the linked publishers and require direct verification.

Experimental Lean-bounty source included for pre-launch curation; source details require verification before any attempt.

Formalization: SOURCE FLOW TO VERIFY
Review note: Verify network, funding, pinned Lean statement, and claim rules directly before any allocation or public claim.

Open original source

03 / PLANNED MECHANISM

A fee-funded proof loop

PLANNED

01Fees
02Aristotle Compute
03Lean-verified Proof
04Bounty
05Buyback / Burn

Protocol fees would purchase proof-search compute; only Lean-verified artifacts could move a task toward a bounty claim; successful proceeds would then be routed toward market buyback and token burn.

Proof search can fail. Bounties can be disputed, withdrawn, or take years to review. This is a planned mechanism, not a promise of rewards, token value, or investment returns.

04 / WHY `sorry` MATTERS

Programmers have TODO. Lean proofs have sorry.

In Lean, sorry is a temporary placeholder that closes an unfinished proof goal so the surrounding skeleton can be checked. Lean emits a warning: it is not a finished proof.

The task is to replace that placeholder with tactics that leave no goals, remove every `sorry` or `admit`, review the resulting axiom set as clean, and then let Lean's small kernel check the artifact. Harmonic describes its proof system's job in those terms.

Lean Language Reference ↗Harmonic — Lean at Scale ↗

EXAMPLE / INCOMPLETE PROOF

theorem open_goal (n : ℕ) :
  n + 0 = n := by
  sorry

warning: declaration uses `sorry`

05 / BOUNTY SOURCES

Where a proof might earn a reward

SOURCE CHECK / 22 JUL 2026

These markets use different definitions of proof, review, and payout. A formalised statement is not a solved theorem, and a listing is not proof that funds remain available.

01

PUBLIC LISTING

Erdős Problems

Public open-problem listings retained as pre-launch research sources; details require direct verification.

CURATION / UNDER REVIEWRead source record ↗

02

EXPERIMENTAL FLOW

MathBounty

A Lean-oriented source whose network, funding, pinned statement, and claim rules require verification.

SOURCE DETAILS TO VERIFYRead source record ↗

03

REFERENCE ONLY

DeMath

A cited protocol reference; this interface does not represent a solve, attestation, or payout.

NO ACTIVITY CLAIMEDRead source record ↗