陶哲轩博客写数学问题时，通常先把对象定义清楚，再给直觉、反例和证明轮廓。把「Online reading seminar for Zhang’s “bounded gaps between primes”」改写成可阅读的中文笔记，重点是：问题在问什么、已知到哪一步、下一步最容易走偏在哪。原站广告、分享条和导航已去掉。

问题在问什么

In a recent paper , Yitang Zhang has proven the following theorem:

Theorem 1 (Bounded gaps between primes) There exists a natural number such that there are infinitely many pairs of distinct primes with .

已知结果和反例

Zhang obtained the explicit value of for . A polymath project has been proposed to lower this value and also to improve the understanding of Zhang’s results; as of this time of writing, the current “world record” is (and the link given should stay updated with the most recent progress.

Zhang’s argument naturally divides into three steps, which we describe in reverse order. The last step, which is the most elementary, is to deduce the above theorem from the following weak version of the Dickson-Hardy-Littlewood (DHL) conjecture for some :

证明或构造的主线

Theorem 2 ( ) Let be an admissible -tuple, that is to say a tuple of distinct integers which avoids at least one residue class mod for every prime . Then there are infinitely many translates of that contain at least two primes.

Zhang obtained for . The current best value of is , as discussed in this previous blog post . To get from to Theorem 1 , one has to exhibit an admissible -tuple of diameter at most . For instance, with , the narrowest admissible -tuple that we can construct has diameter , which explains the current world record. There is an active discussion on trying to improve the constructions of admissible tuples at this blog post ; it is conceivable that some combination of computer sear

阅读时建议盯住的点

The second step in Zhang’s argument, which is somewhat less elementary (relying primarily on the sieve theory of Goldston, Yildirim, Pintz, and Motohashi), is to deduce from a certain conjecture for some . Here is one formulation of the conjecture, more or less as (implicitly) stated in Zhang’s paper:

Conjecture 3 ( ) Let be an admissible tuple, let be an element of , let be a large parameter, and define

值得单独记下的条目

• Going through Zhang’s argument in order to improve the value of (perhaps by decreasing ); and
• Gaining a more holistic understanding of Zhang’s argument (and perhaps to find some more “global” improvements to that argument), as well as related arguments such as the prior work of Bombieri, Fouvry, Friedlander, and Iwaniec that Zhang’s
• A recent blog post of Emmanuel Kowalski on the technical details of Zhang’s argument.
• Scanned notes from a talk related to the above blog post .
• A recent expository note by Fouvry, Kowalski, and Michel on a Friedlander-Iwaniec character sum relevant to this argument.

阅读和落地时建议先做的 5 件事

1. 用自己的语言重写定义和结论，不看原文能不能说清对象是什么。
2. 找一个最小反例或边界情形，确认假设少一条会怎样。
3. 把证明拆成可独立检验的引理，每步只保留一个新想法。
4. 若涉及计算或形式化，先写可复现的小例子，再谈一般情形。
5. 记下尚未解决的缺口：缺估计、缺构造，还是缺正确的范畴。

和智能体、形式化工具怎么接

龙虾PRO做 OpenClaw 落地时，数学笔记最有用的部分往往是「可检验的步骤」：定义、反例、引理边界。智能体适合帮忙展开计算和检索，不适合代替你决定哪条假设能扔。

本文侧重全链路风控方法论。落地时请用自身业务单据做回放验证，不要把示例阈值直接当生产策略。 相关：风控体检 · 方案资源