论文研究
x.com ·
@a16z (@a16z)
a16z访谈：AI如何改变数学研究激励
多伦多大学数学家Daniel Litt与a16z的Lisha Li讨论AI对数学的影响，指出AI擅长计算但缺乏理论构建，并分享用Codex一小时内生成三篇正确但质量差的论文的实验。
AI推荐理由 a16z发布数学家Daniel Litt访谈，深入探讨AI对数学研究的影响，观点深刻，但为视频内容，需观看获取细节。
80
浏览 35,257
点赞 100
评论 16
转发 19
正文
多伦多大学数学家 Daniel Litt 谈 AI 如何改变了学术数学的奖励激励机制，以及他多么容易地证明了这一点：“数学的目标不是产生数学论文。而是产生某种理解。”“也许有些理解存在于模型权重中或其他地方。对我来说，这相当不令人满意。”“我做数学的动力是我想满足自己的个人好奇心，我认为人们应该有能力做到这一点。这需要相当庞大的设备。”“现在，如果你是一个在市场上的博士后……找到工作的最好方法是发表大量论文，这些论文可能证明旧的猜想。你可以通过玩老虎机直到模型产生一个希望是正确的证明来做到这一点。”“你甚至不必提前选择定理。”“这里有一个你可以做的实验。你可以拿 Codex，对它说：‘上网，找到代数几何中五个最近的猜想并证明它们。’我做过这个实验，经过一些来回，我能够在一个小时内得到大约三篇相当糟糕的论文。但是正确的论文。” @littmath@lishali88 您的浏览器不支持视频标签。🔗 在 Twitter 上查看 a16z@a16z 多伦多大学数学家 Daniel Litt 和 a16z 的 Lisha Li 谈论 AI 对数学的影响：模型在数学的较窄领域表现良好，比头条新闻所暗示的要窄。它们进行长时间的计算，从比任何人类能阅读的更多论文中提取技术思想，并且比几乎任何人都更好地应用每种已知技术。它们不会做的是构建理论，或者保持一种模糊的哲学足够长的时间使其精确，而这正是 Daniel 说他实际上谋生的大部分工作。在这次对话中，他和 Lisha 深入探讨了数学家如何从 AI 证明中窃取部分内容并用它们解决其他几个问题，为什么一千个 AI 数学家可能最终都是同一个数学家，以及为什么模型交给 Daniel 的证明是正确的但仍然毫无价值。 00:00 介绍 02:10 AI 反驳的 Erdős 问题 06:20 AI 的推理看起来像人类可识别的 07:55 为什么英语胜过形式证明 10:00 为什么模型不能构建理论 14:50 开放问题衡量你的无知 17:45 一个图如何成为千禧年大奖问题 18:58 AI 在哪些方面帮不到 Daniel 21:15 为什么丑陋的证明值得做 23:42 真正的猜想比错误的更难 29:32 10 页计算，零洞察 34:55 数学的目标不是产生论文 36:25 5 个猜想，3 篇糟糕的论文，1 小时 38:05 一个数学家复制了 1000 倍 40:48 为什么即使模型获胜人类仍然重要 46:30 当更便宜和更差胜过更好 49:22 为什么最新的 AI 结果不是什么大事 57:05 数学家如何实际检查一个长证明 59:38 Daniel 三岁的孩子已经在做数学了 YouTube：youtube.com/watch?v=tQI35C…@littmath@lishali88 您的浏览器不支持视频标签。🔗 在 Twitter 上查看 🔗 查看引用推文 💬16🔄19❤️100👀36123📊24 ⚡ 由 xgo.ing 提供支持
查看原文（中文正文为 AI 翻译）
University of Toronto mathematician Daniel Litt on how AI has changed academic mathematics' reward incentives, and how easily he proved it: "The goal of mathematics is not to produce mathematics papers. It's to produce some kind of understanding." "Maybe some of that understanding resides in model weights or something. To me, that's pretty unsatisfying." "My motivation for doing mathematics is that I would like to satisfy my own personal curiosity, and I think people should have the capability to do that. That requires a pretty substantial apparatus." "Right now, if you're a post-doc on the market... the best way to get a job is to produce a lot of papers which maybe prove old conjectures. And you can do that by playing the slot machine until the model produces a hopefully correct proof." "You don't even have to pick the theorem in advance." "Here's an experiment you can do. You can take Codex, you can say, 'Go online, find five recent conjectures in algebraic geometry and prove them.' I've run this experiment, and with some back and forth I was able to in an hour get like three quite bad papers. But correct papers." @littmath@lishali88 Your browser does not support the video tag. 🔗 View on Twitter a16z@a16z University of Toronto mathematician Daniel Litt and a16z's Lisha Li on AI's impact on mathematics: The models are good at a narrower slice of math than the headlines suggest. They grind long computations, pull technical ideas from more papers than any human could read, and apply every known technique better than almost anyone. What they don't do is build theory, or hold a vague philosophy long enough to make it precise, which is most of what Daniel says he actually does for a living. In this conversation, he and Lisha get into how mathematicians raided an AI proof for parts and broke several other problems with them, why a thousand AI mathematicians might all turn out to be the same mathematician, and why the proof a model handed Daniel was correct but still worth nothing. 00:00 Intro 02:10 The Erdős problem AI disproved 06:20 AI's reasoning looks recognizably human 07:55 Why English beat formal proofs 10:00 Why models can't build theory 14:50 Open problems measure your ignorance 17:45 How a graph became a Millennium Prize problem 18:58 Where AI doesn't help Daniel 21:15 Why ugly proofs are worth doing 23:42 True conjectures are harder than false ones 29:32 10 pages of calculation, zero insight 34:55 The goal of math is not to produce papers 36:25 5 conjectures, 3 bad papers, 1 hour 38:05 One mathematician duplicated 1000x 40:48 Why humans matter even if models win 46:30 When cheaper and worse beats better 49:22 Why the newest AI result isn't a big deal 57:05 How mathematicians actually check a long proof 59:38 Daniel's 3-year-old is already doing math YouTube: youtube.com/watch?v=tQI35C…@littmath@lishali88 Your browser does not support the video tag. 🔗 View on Twitter 🔗 View Quoted Tweet 💬16🔄19❤️100👀36123📊24 ⚡ Powered by xgo.ing