Core Viewpoint - The collaboration between mathematicians and AI has led to the resolution of the long-standing Erdős 1026 problem, which had remained unsolved for 50 years, in just 48 hours [1][2][3]. Group 1: Problem Overview - The Erdős 1026 problem was proposed in 1975 and involves determining the minimum possible value of a function related to a game theory scenario involving two players, Alice and Bob [8][10][12]. - The problem's complexity was highlighted by the introduction of a maximum constant c(n) that represents the minimum proportion of coins Bob can guarantee to take, regardless of how Alice distributes them [10][13]. Group 2: AI's Role in the Solution - AI tools played a crucial role in solving the problem quickly, with traditional methods potentially taking weeks or months to reach a conclusion [3][5]. - The use of AI models, such as Harmonic and AlphaEvolve, allowed mathematicians to automate the construction and proof of key inequalities, transforming the original problem into a computational geometry challenge [16][18][22]. Group 3: Collaborative Efforts - The solution involved multiple mathematicians working together, with contributions from Boris Alexeev, Koishi Chan, and Lawrence Wu, showcasing the effectiveness of human-AI collaboration [17][28][32]. - The collaborative approach of combining human insight with AI capabilities is emerging as a new trend in mathematical problem-solving [46]. Group 4: Historical Context and Future Implications - The Erdős problems, proposed by the renowned mathematician Paul Erdős, have been a significant part of mathematical research, with many remaining unsolved [39][41]. - The increasing success of AI in solving these problems suggests a shift in how mathematical research may be conducted in the future, with AI becoming a standard tool for researchers [41][42].

一水 发自 凹非寺 量子位 | 公众号 QbitAI GPT-5又帮陶哲轩解决了一个难题！ 消息来自陶本人最新动态，他衷心提醒大家： AI能够大显身手的场景再+1—— 半自动化文献检索 。 简单来说，陶正在做的事情，其实就是用 AI+数据库比对 来帮忙解决数学里的难题。 结果AI不仅省时省力，成果也十分卓越，正如陶激动所言： 这是Erdos问题/OEIS关联项目的首次概念验证成果。 具体咋回事儿，下面详细来看—— AI在数学难题解决过程中起到"定位器"作用 事情的起因还要追溯到一个关键人物——20世纪著名匈牙利数学家Paul Erdős。 此人一辈子合作了超过500位数学家，毕生发表了约1525篇数学论文，数量之多，至今无人能及。 相应地，他也给后人留下了一大堆至今未解的难题，它们被称为"Erdős问题"。 其中就有一大类问题很"刁钻"—— 它们不是问"算出结果是多少"，而是问"这个结果是不是有理数 （能写成分数的那种） "。 一般而言，准确回答这类问题往往面临两个主要困难： 第一重困难：公式写得简单，但算起来超级复杂，手工几乎算不动。 它不是直接"证明"某个数是不是无理数，而是把这个数列算到很高精度的小数， ...