The Geometry of Intelligence: How AI Finally Closed a 47-Year Mathematical Cold Case

The Unlikely Heroes: OpenAI's Astra and Anthropic's Claude

For nearly half a century, the n=11 square packing problem remained a stubborn outlier in geometry, defying formal proof since Walter Trump first proposed his arrangement in 1979. Today, that silence has been broken by a high-stakes collaboration between OpenAI's Astra and Anthropic's Claude, which together navigated the complex logical landscape required to confirm the optimality of Trump's configuration. This achievement builds upon OpenAI's 722-result deluge, signaling a new era in human-AI collaboration.

"This isn't a case of a chatbot explaining a known proof back to a student. It's a formal, machine-checked argument that two competing labs' models helped construct from scratch, on a geometry problem that had sat unresolved since Martin Gardner first wrote about it."

The significance of this milestone cannot be overstated. By leveraging the reasoning capabilities of Astra and Claude, human researchers were able to bridge the gap between heuristic intuition and rigorous, machine-verifiable truth. The project, spearheaded by the GitHub user Queuingtheorydotcom, proves that frontier models are no longer just assistants—they are active participants in the formalization of mathematical knowledge.

The Lean Proof Checker: A Crucial Tool in Mathematical Verification

At the heart of this breakthrough lies the Lean proof checker, a sophisticated environment that transforms abstract mathematical reasoning into machine-executable code. Unlike traditional peer review, which relies on human fallibility, Lean requires every logical step to be validated by the kernel, leaving no room for ambiguity or unproven assumptions.

WORKFLOW_TIMELINE: The Path to Verification

• 1979: Walter Trump identifies the optimal packing arrangement for 11 squares, though it remains unproven.
• 2024-2025: Development of advanced AI-assisted formalization techniques using Lean 4.
• October 6, 2026: The 11SquaresFormalized repository is published, featuring a complete, verified proof.
• Post-Verification: The community audit confirms zero admissions across 7,920 local Lean modules.

This verification process was not merely a computational brute-force exercise. It required the careful assembly of geometry, checker soundness, and proof logic, all while maintaining the integrity of the Lean kernel. By pinning the toolchain to specific versions, the researchers ensured that the proof remains reproducible and robust against future software updates.

The Community Reacts: A Hacker News Discussion on AI-assisted Proof

The mathematical community, often skeptical of AI-generated claims, has responded with a mix of cautious optimism and genuine excitement. Discussions on Hacker News have highlighted that this is not just about solving a puzzle; it is about the changing nature of mathematical labor. This achievement underscores the potential of AI in accelerating mathematical discovery, as discussed in our previous article, AI is breaking the speed limit of mathematical discovery.

Key Takeaways from the Community:

• Formalization as the New Standard: Participants noted that AI-assisted formalization is becoming the gold standard for verifying complex proofs that are too large for human review alone.
• The End of 'Best Known' Conjectures: The community emphasized that the transition from 'best known' to 'proven' is the most critical shift, effectively closing the door on decades of speculation.
• Collaborative Transparency: The open-source nature of the repository allowed for immediate community scrutiny, which is essential for building trust in AI-generated mathematical outputs.

Ultimately, the successful verification of the 11-square packing problem serves as a proof-of-concept for future endeavors. As AI models continue to integrate with formal verification tools, we are likely to see a cascade of long-standing mathematical problems falling to this hybrid approach. The era of the 'AI-mathematician' has officially arrived, and it is already rewriting the rules of discovery.