关于“Lonely Runner Conjecture”进展的文章中，关键思想是将问题转化为一类几何-数论等价表述：在平面格点中心放置小方块后，一条非水平、非垂直直线会在何时必然碰到方块。随着该问题在数学中的多个版本扩散，研究者分别动用数论、几何和图论等方法，但缺乏统一框架：许多证明只能解决具体跑者数目。到2000年代中期，数学家已证明当跑者数不超过7时，每位跑者最终都会“孤独”。他们还成功简化了问题，把对所有实数速度的证明归约为整数速度情形（即只需考虑whole-number speeds），但这仍涉及无限多种速度组合。

2015年，Terence Tao提出“速度上界”思想：若猜想在某一范围内成立，则可推出更高速度下也成立；换言之，对给定跑者数只需检验有限集合。尽管如此，实际可检验组合数仍“天文级且完全不切实际”。Rosenfeld在此基础上用计算机结合数论思路推进：他研究反例所需速度结构，证明反例中速度乘积需被一系列指定素数整除。若猜想为假，可在某阈值以下找到反例；而该阈值下的乘积必须同时满足素数可整除条件，这会迫使乘积异常巨大，进而超过阈值。由此推出8跑者情形无反例，因此该情形成立（即第1次把已知范围推进到n=8）。

随后，Kravitz将其方法推荐给本科生Paul Trakulthongchai（Tanupat），他改进了筛选必需素数因子的计算技术，快速排除了9跑者和10跑者的反例；Rosenfeld也独立证明了9跑者情形。至此同一思路一次解决了两个新案例。若扩展到11跑者，现有方法计算成本过高，需要“全新视角”。尽管如此，这种统一化路线被视为关键突破，因为此前每增加一个人数通常需全新证明。Schymura等人已据此在罗斯托克召开研讨会（10月），跨领域协作；Wills预计该猜想可能在20到30年后解决。

In this article on the Lonely Runner Conjecture, the key idea is an equivalent geometric reformulation: on infinite graph paper with tiny squares at grid centers, start from a grid corner and draw a non-axis-aligned line, then ask how large the squares can be before the line must hit one. As versions of the problem spread, proofs came from number theory, geometry, and graph theory, but there was no unified strategy, only ad hoc arguments for specific runner counts. By the mid-2000s, the conjecture had been proved for at most 7 runners. The problem was also reduced from infinitely many real speeds to testing whole-number speeds, which was much easier but still infinite.

In 2015, Terence Tao introduced a speed-threshold idea: if the conjecture holds for relatively low speeds, it then holds for all higher speeds, so each runner count requires only finitely many cases. Yet the practical number of combinations remained astronomically large. Matthieu Rosenfeld exploited this framework with computer-assisted, number-theoretic reasoning about counterexamples: he showed any counterexample would force the product of speeds to be divisible by certain primes. Recasting Tao’s threshold in product terms, any counterexample would have to lie below a threshold, but the required prime divisibility made such a product so huge that it exceeds that threshold. Hence no counterexample exists for 8 runners, and the conjecture is true there.

Paul Trakulthongchai then optimized the computation, using Rosenfeld’s method to identify mandatory prime divisors more efficiently. He ruled out counterexamples for both 9 and 10 runners, and Rosenfeld independently confirmed the 9-runner case. The same approach is still too computationally expensive for 11 runners, so a new viewpoint seems required. Still, researchers view this as major progress because one method solved multiple cases instead of each case needing a separate proof. A workshop in Rostock is being organized to unite communities across disciplines, and Wills estimates a likely resolution may still be 20 to 30 years away.