Tag
OpenAI released a repository of 721 mathematical manuscripts generated by an unnamed internal model, including a claimed refutation of the Exponential Time Hypothesis, with about 42% of results formalized in Lean along with abridged reasoning traces.
3Blue1Brown (Grant Sanderson) argues in a post: even if AI completely surpasses humans in proof generation, the value chain and funding chain of mathematics will remain largely intact. What really changes are culture and the allocation of credit — making now exactly the best time to enter mathematics.
在这一期播客中，量子计算学者 Scott Aaronson（曾在 OpenAI 从事 AI 安全研究）认为 AI 已在数学和理论计算机科学领域开启了"奇点"——人类已不再是地球上主要的定理证明实体，领域正被迫重新定义研究、署名与出版的意义。
This paper introduces a metric for the intrinsic interestingness of mathematical theorems and trains a 27B model to predict proof difficulty, enabling the generation of more interesting and novel mathematics with reduced overlap with existing libraries.
Mathematicians have proven the long-standing sandwich conjecture in graph theory, showing that large random graphs can be approximated between two simpler graphs, connecting different random processes and advancing the field.
Anima Anandkumar describes a guest post on Terence Tao's blog detailing a method using Physics-Informed Neural Networks to find singular solutions for the Euler equations in fluid dynamics, highlighting its earlier release and broader applications in scientific computing.
Terence Tao argues that mathematical open problems should be preserved for human problem-solving to maintain uncontaminated benchmarks and develop new techniques, suggesting social norms to limit AI solvers in certain areas.
This article improves the lower bound for the smallest square that can pack 17 unit squares, achieving a new value of 4.5058, building on previous work and incorporating AI-assisted methods.
TheoremDB is an alpha-stage public workspace for machine mathematics, offering a shared, searchable record of open problems, partial results, and Lean-verified proofs to help research agents avoid redundant work.
This research report develops a rigorous analytical framework for one-bit noise-shaped coefficients in discrete polynomial Fourier extension, deriving estimates and higher-order extensions for sigma-delta quantization.
A mathematician reports that Grok 4.5 discovered a novel counterexample to hypercontractivity on the 4-sphere, potentially marking the first instance of original mathematical research by an AI, pending peer review.
Fields Medalist Timothy Gowers reports using GPT5.5 Pro to solve open mathematical problems and predicts an imminent crisis in mathematical research due to rapid AI progress.
Researchers from Charles University introduce Bolzano, an open-source multi-agent LLM system that orchestrates prover and verifier agents to assist with mathematical research, reporting new results on six problems where four reached publishable quality and three were produced essentially autonomously.
Mathematician Terence Tao believes AI is reducing cognitive friction in mathematical research, enabling experimentation and bold ideas, and is expected to become a mainstream tool. He also predicts that future mathematical publications will share exploration paths rather than just final results.