精选7001:16Gary Marcus（RSS）行业动态多源精选 ×9

Gary Marcus与陶哲轩评OpenAI数学成果

Complementary remarks from Gary Marcus and Terence Tao on OpenAI’s giant math drop

原文

发到 X

OpenAI’s massive new math drop:

OpenAI 发布的全新数学成果：

[This essay was written in extreme haste before a very long wifi-less flight; please forgive typos.]

[本文是在一次漫长的无 Wi-Fi 航班前仓促写就的；请原谅其中的拼写错误。]

Part One: My take

第一部分：我的看法

The real news here isn’t the result; it’s not what we were told.

这里真正的新闻并非结果本身；它与我们被告知的不同。

1. AI once tried to be a science. Now we get stuff like the completely vague report from OpenAI below:

1. AI 曾经试图成为一门科学。现在我们得到的却是像 OpenAI 下面这份完全模糊的报告这样的东西：

“Same procedure”? “Using an unreleased model”?

“同样的流程”？“使用未发布的模型”？

This would never pass peer review.

这根本通不过同行评审。

We don’t know what the procedure was.

我们不知道具体的流程是什么。

We know nothing about the architecture. For esxample, were the proofs generated in one shot, and then verified by the symbolic system Lean? Was there an iterative process?)

我们对架构一无所知。例如，证明是一次性生成的，然后由符号系统 Lean 进行验证吗？还是存在一个迭代过程？)

We know nothing about the failure rate. We know nothing about the training/post training/data augmention.

我们对失败率一无所知。我们对训练/后训练/数据增强也一无所知。

2. As a result, we have zero idea of how generalizable the result is outside math.

2. 因此，我们对该结果在数学领域之外的泛化能力毫无概念。

3. A lot of the discussion on social media has been reduced to an ignorant cheering section that applauds without knowing what it is applauding or what it might mean— without ever asking basic scientific questions.

3. 社交媒体上的大量讨论已退化为一种无知的欢呼阵营，他们在不知晓自己为何欢呼、也不知晓其可能意味着什么的情况下鼓掌——甚至从未提出基本的科学问题。

The new system could be a legitimate step toward AGI. Or it could just be a clever leveraging of Lean and synthetic data in a verifiable domain with no generality whatsoever.

这个新系统可能是迈向 AGI（通用人工智能）的合法一步。或者，它仅仅是在一个可验证且毫无泛化能力的领域中，对 Lean 和合成数据的巧妙利用。

From the initial report, we can tell almost nothing.

从初始报告中，我们几乎无法得知任何信息。