Breaking Math Verification

The company appears to be a single-person project named "Breaking Math Verification" (BMV), self-described as a reusable workflow for transforming reported mathematical claims into formal Lean reconstructions with explicit witnesses and logical consequences.

The description is self-reported and unverified — no third-party corroboration, no financials, no customers, no usage data. The project appears to be a proof-of-concept or hackathon submission with no demonstrated commercial viability.

• The description states that BMV provides a reusable workflow for transforming externally reported mathematical claims into:
  • an independent Lean reconstruction,
  • explicit finite witnesses,
  • a selected summit theorem,
  • logical consequences,
  • a focused axiom audit,
  • a provenance record,
  • and a stable public Demo surface.

• The project formalizes a specific case study: a candidate counterexample to the Jacobian Conjecture, using Lean to verify that a polynomial map is not injective and has no set-theoretic left inverse.

• A generic Lean API was extracted:
  • DkMath.Verification.CollisionCertificate
  • DkMath.Verification.CollisionCertificate.notInjective
  • DkMath.Verification.CollisionCertificate.noLeftInverse

• The project includes reusable templates for:
  • reported claims,
  • theorem pipelines,
  • provenance,
  • scope boundaries,
  • Demo contracts,
  • and axiom-audit targets.

• The system uses GPT-5.6 and Codex, with Git repository reports as an auditable handoff channel between GPT review and Codex implementation.

• The project was submitted to the OpenAI 2026 hackathon.

Not evidenced No information on whether this workflow is used beyond the author’s own work, or if it has been adopted by others. No evidence of product-market fit, customer feedback, or commercial usage.

• The description states that BMV was built to answer a real-time question: whether GPT-5.6, Codex, and Lean could turn complicated reported formulas into a precise, reproducible verification package while broader mathematical review was still beginning.

• It positions itself as a tool for fast, reproducible first verification layer for exact formulas under discussion — not for certifying historical priority, authorship, publication status, or community acceptance.

• The tagline: "Local Core preserved. Global address uniqueness released." suggests an approach that preserves local mathematical integrity while allowing for broader, potentially conflicting interpretations or reuse.

• The project is described as a proof-of-concept, submitted to a hackathon, and not yet commercialized.

Inferred The positioning is narrow — focused on formal verification in math domains, using AI tools. It does not claim to be a general-purpose tool for all mathematical claims, but rather a reusable workflow for specific cases like the Jacobian Conjecture.

Not evidenced No evidence of evolution from prototype to product, or any commercial positioning beyond the hackathon submission.

• The description states that BMV is intended for mathematical researchers and formal verification practitioners, particularly those working with complex claims in algebraic geometry or related fields.

• It targets users who need:
  • Independent Lean reconstruction,
  • Explicit finite witnesses,
  • Logical consequences,
  • Averaging of provenance records,
  • And a stable public Demo surface.

• The project is built around Lean, a theorem prover, and uses GPT-5.6 and Codex to automate parts of the process.

Not evidenced No evidence of actual users or customers beyond the author. No indication of whether the tool has been used by others, or if there is a defined customer segment.

• The description does not state any business model or pricing structure.

• It describes a public repository, with demo video and testing instructions available on GitHub.

Not evidenced No evidence of monetization, licensing, or revenue streams. No pricing information, subscription models, or commercial partnerships are mentioned.

• The project is built using:
  • GPT-5.6
  • Codex
  • Lean
  • ffmpeg
  • Python
  • GitHub
  • TTS

• It uses a Git repository as an auditable handoff channel between GPT and Codex.

• The system includes:
  • A generic Lean API for collision certificates,
  • Reusable templates for theorem pipelines, provenance, scope boundaries, Demo contracts, and axiom-audit targets.

• The workflow was developed through six checkpoints, including architecture audit, generic collision certificate, Jacobian adapter, verification contracts, cross-domain validation, and public integration.

Not evidenced No evidence of scalability, deployment infrastructure, or delivery mechanisms beyond the author’s own use. No information on how this would be offered to others.

• The project was submitted to the OpenAI 2026 hackathon, suggesting it is a prototype or proof-of-concept.

• It includes:
  • A public demo video,
  • A GitHub repository with testing instructions,
  • And a formalized case study of a candidate counterexample to the Jacobian Conjecture.

• The author states that the project formalizes an explicit polynomial map over both rational and complex numbers, and verifies specific inputs and outputs using Lean.

Not evidenced No evidence of adoption, usage metrics, or feedback from others. No indication of whether this has moved beyond a hackathon submission into a product or service.

• The description does not mention any direct competitors.

• It is positioned within the formal verification space, particularly for mathematical claims, using tools like Lean and AI assistants (GPT-5.6, Codex).

• It is not clear whether there are existing tools for:
  • Automating formal verification of mathematical claims,
  • Reusing templates or APIs in theorem-proving environments,
  • Integrating AI with Lean-based workflows.

Not evidenced No competitive analysis, no mention of similar tools or platforms, and no indication of market positioning relative to others.

• The project is self-reported, unverified, and submitted as a hackathon entry — no evidence of traction, revenue, or adoption.

• It is built by a single person (D. Deskuma), with no team or organizational structure described.

• The use of GPT-5.6 and Codex raises questions about reproducibility, scalability, and dependency on proprietary AI tools.

• There is no evidence of product-market fit, customer feedback, or commercial viability.

• The project’s scope is narrow, focused on mathematical claims in a specific domain (Jacobian Conjecture), limiting its potential for broader application.

1. What is the intended user base beyond the author’s own work?
2. Has this workflow been used or tested by others outside of the hackathon?
3. Are there plans to commercialize or scale this beyond a proof-of-concept?
4. How does the project handle edge cases or failures in AI-assisted formal verification?
5. What are the long-term dependencies on GPT-5.6 and Codex, and how would they be managed if these tools change or become unavailable?
6. Is there any feedback from the mathematical community on this approach?
7. How is the project intended to generate revenue or value for users?