袁新意创办了一个数学网站
以下是聊天内容的传播整理：
各位拓扑界同事们好！袁新意牵头办了一个网站 mathhelm.org，专门收集对 AI 生成论文的评论，尤其是那 722 篇。如果各位有任何评论，欢迎发表在上面
Dear Colleagues,
I am writing to bring your attention to a new math forum at
- • https://mathhelm.org/
This website is non-profit, founded by mathematicians, and serves mathematicians. Its major goal is to collect comments on AI-generated papers, especially on the 722 math papers released by OpenAI. For more details about the motivation, please visit https://mathhelm.org/about. For example, you can find the statement:
"No one is obliged to check an AI-generated paper, but if someone happens to read any part of it, their comments are very valuable to the mathematics world."
Some of you may have comments after reading some parts of these papers. I am wondering if you can come to post your comments at mathhelm. You can copy and paste your old comments or write new ones. In order to do that, you need first register on mathhelm.org by clicking the button "Join" on the upper right corner, and then come back to mathhelm.org, navigate on the left of the page to find the math area and the related paper you can post after.
If you find any problem or have any suggestion to the website, please let me know.
Thank you very much!
我们主要目的是把针对AI生成论文的评论收集在一起，包括正面、负面、技术细节、创新点、重要程度、想法来源等等方面的评论，希望由人类数学家来主导对 AI 数学的叙事，而非由 AI 公司主导。
以上是整理内容部分。关于ai产出的论文做解析，数语做过不少，而关于oai的719篇论文的人类解读，数语也转载过几篇，
OpenAI719问题解析之：073 Falconer distance conjecture 的证明思路
OpenAI719问题人工消化之：003 兰道–西格尔零点不存在的证明思路
数学的重要时刻：关于openai一次发布的700多篇手稿
本着是做一系列的，之前有找数学系的人提供稿件，但结果是0篇，看来这事，还是得数学家来联动才行。
转载自久章智能（微信公众号，久章智能于 2026-10-23 转载），内容保持原貌，版权归原作者及原账号所有。阅读原文 ↗