Claude cracks the mystery of the P-256 elliptic curve seed origin in half an hour

To be honest, in the early hours of this morning the tech world saw a miracle worthy of being written into the history of computing, yet most AI tech morning briefs had not even managed to grasp it yet. A top-tier historical mystery that had plagued the global cryptography community for a full 26 years and that countless conspiracy theories suspected was a secret backdoor of the US National Security Agency was completely cracked by Claude in less than half an hour. Brothers, do you know how counterintuitive this truth is? The P-256 elliptic curve, which protects more than half of today's internet traffic worldwide—those mysterious random seeds that nobody knew the origin of back then— were not some high-dimensional mathematical backdoor at all, but the 1,842nd candidate produced in order by an engineer in 1997 calling the standard C language random number function with the timestamps of the time. The first thing shattered was a quarter-century of technical suspicion. When the authorities published these core encryption curves back then, they gave only the final results and always kept silent about the seed generation mechanism, which led scholars worldwide to suspect that privileged constants convenient for specific agencies to crack were hidden inside. Claude worked backward layer by layer along the ancient characteristics of the Unix system of the time and directly pinned down the lrand48 function built into the C language, not only precisely restoring the birth time of the five curves to a specific point in 1997 to 1998, but also fully reproducing the algorithm search chain of that time. Even more crucial is this: everyone must absolutely not panic—this discovery does not mean the encryption system has been breached. Your bank accounts, encrypted communications, and daily browsing need no security changes at all. This is not an attack on mathematical principles, but a perfect archaeology of a lost piece of history. The bounty set up by a well-known security expert has already doubled to $36,864 and been donated in full to an open-source security foundation. Now that the truth is out, it instead proves from the root that this underlying algorithm protecting the global network was not secretly tampered with back then. What really makes my scalp tingle is the dimension of reasoning AI displayed in this incident. It was not simply reciting materials; between the programming habits of decades ago, ancient compiler characteristics, and rigorous mathematical verification, it pieced scattered clues into a closed loop like a top-tier digital detective. People used to think large models could only help write code and do odd jobs. From today on, it is presenting irrefutable evidence on real unsolved scientific mysteries. Truly emotional—when AI has the ability to pierce the fog of history and restore the relics of human engineering, many digital mysteries buried by time may be coming to their own final chapter. Among the unsolved mysteries in the tech world you have come across, which one would you most like AI to investigate and clarify next? Honestly, the details of the first version of the Bitcoin code left by whoever Satoshi Nakamoto really was—I really want to let it dig into that properly now haha https://t.co/CifVqaRlHs

View original

说实话，今天凌晨科技界发生了一件足够载入计算机史册的奇迹，但绝大多数AI科技早报甚至都还没来得及看懂， 困扰全球密码学界整整 26 年、被无数阴谋论猜测是美国国家安全局秘密后门的顶级历史悬案，被 Claude 花了不到半个小时彻底破解了。 兄弟们你知道这个真相有多反直觉吗？
保护着今天全世界大半个互联网流量的 P-256 椭圆曲线，当年那些没人知道怎么来的神秘随机种子， 根本不是什么高维数学后门，而是 1997 年某个工程师调用标准 C 语言的随机数函数，用当时的时间戳按顺序跑出来的第 1842 个候选。 最先被击碎的，是长达四分之一个世纪的技术猜忌。 
当年官方公布这几条核心加密曲线时，只给出了最终结果，却始终对种子的生成机制讳莫如深，导致全球学者一直怀疑里面藏着便于特定机构破解的特权常数，而 Claude 顺着当时 Unix 系统的古老特征层层倒推，直接锁定了 C 语言自带的 lrand48 函数，不仅把五条曲线的诞生时间精确还原到了 1997 到 1998 年的某个具体节点，而且完整复现了当时的算法搜索链条。 更关键的一点是，大家千万不用恐慌，这次发现完全不代表加密系统被击穿。 
你的银行账户、加密通信和日常浏览不需要做任何安全改动，这并不是对数学原理的攻击，而是对一段失落历史的完美考古。 知名安全专家设立的悬赏金已经翻倍到 36,864 美元并全额捐给了开源安全基金会，真相大白之后，反而从根源上证明了这套保护全球网络的底层算法当年并没有被暗中做手脚。 真正让我觉得头皮发麻的，是 AI 在这起事件中展现出的推理维度。 
它不是在单纯地背诵资料，而是在数十年前的编程习惯、古老的编译器特性和严苛的数学验证之间，像一个顶级数字刑警一样把零散的线索拼成了一个闭环。 以前大家以为大模型只能帮人写代码干杂活，从今天起它开始在真实的科学未解之谜上拿出无可辩驳的证据。 真的感慨啊，当 AI 拥有了穿透历史迷雾、还原人类工程遗迹的能力，很多被时间掩埋的数字悬案，也许都要迎来自己的终局。 大家平时遇到过的科技圈未解之谜里，你最希望 AI 下一个去查清的是哪个？ 讲真当年中本聪到底是谁留下的第一版比特币代码细节，我现在真的想让它去好好盘一盘哈哈 https://t.co/CifVqaRlHs