This Has Never Happened in Math Before | OpenAI Just Dropped BOMBSHELL

This Has Never Happened in Math Before | OpenAI Just Dropped BOMBSHELL

Formal & Physical Sciences Mathematics PBMathematics
🎙 Animated Math 👥 26K 📅 October 10, 2026 ⏱ 20 min 👁 18 📄 science communication 🧭 2026-10-10
Available in: English (current) Français

Keywords

omega exponentmatrix multiplicationStrassen's algorithmAI-generated proofsLean theorem prover

Summary

On October 6, 2026, OpenAI released 722 AI-generated mathematics papers. Among them, a 13-page paper claims to prove the matrix multiplication exponent over complex numbers is at most 9/4, a drop from 2.371177, larger than all progress since 1990 combined. The video explains the historical context: Strassen’s 1969 algorithm, the definition of omega, and the long-standing records. It then details the new proof’s approach via polynomial multiplication and ‘profiles’. Critically, it highlights two underreported points: the result only holds over complex numbers, and the proof is nonconstructive, giving no practical finite size. The video also discusses the use of Lean for formal verification, noting that while many papers have Lean files, not every claim is verified—the pi paper’s Lean check does not cover the Flint Hills series consequence. Furthermore, within a day, OpenAI withdrew three papers due to a sign error in a foundational paper. The video frames this as a Bob Beamon-like leap, but emphasizes that human verification is still pending, comparing it to measuring the jump with a tape measure.

172 words

Critical Evaluation

Value of the Information & Strength of the Argument

The video provides substantial value by going beyond headline coverage, explaining both the mathematical significance and the caveats. The argumentation is solid: it systematically builds from basic matrix multiplication to the new proof, then deconstructs the claims. It presents multiple perspectives, including expert comments (e.g., Strogatz, Anj Daw, Thomas Carl) and explicitly addresses potential weaknesses. The reasoning is clear and avoids oversimplification, making it both accessible and rigorous.

Scientific Rigor, Source Quality, Title Accuracy

The video demonstrates scientific rigor by citing the actual sources (OpenAI index, GitHub repository, specific papers) and by scrutinizing the formal verification process. It distinguishes between computer-checked statements and human-reviewed claims, and it notes the retraction of three papers. The title is accurately sensational—it highlights a genuinely unprecedented event in mathematics—but the content does not overstate the certainty. The sources are exactly those provided in the description, and the video aligns well with them.

157 words

Title / Content Match

The title is sensationalist but the content genuinely describes a major potential breakthrough in mathematics, so the title captures the essence without misrepresenting the core facts.

Quality & Reliability

8/10

The video is meticulously researched, explains the mathematical background clearly, and transparently notes the limitations of the claims, including the small print and the scope of formal verification. It correctly distinguishes between what is proven and what remains to be verified, and sources are provided for key claims.

Chapters

Cited Sources

Concurring Sources

  • OpenAI: Sharing AI progress in mathematics — Supports the claim of 722 papers and the existence of the 9/4 paper.

External References

Contribution & Novelties

The video adds unique value by uncovering the ‘small print’ that most coverage missed: the proof only holds over complex numbers, the result is nonconstructive, and the Lean verification covers only specific parts. It also explains the retraction of three related papers due to a sign error, illustrating the fragility of even AI-generated proofs. The video provides a clear, accessible explanation of the mathematical concepts, making this major development understandable to a general audience.

Pour aller plus loin :

126 words

Radar Profile

The profile shows high scores in information quantity and quality, reflecting the video's thorough explanation and careful analysis. Technical level is moderate, suitable for a broad audience while preserving depth. Reliability is slightly lower due to the unresolved nature of the proofs and the need for human verification, but the video's transparent discussion of limitations keeps it credible.

Reliability 8/10