These Russian Mathematicians Taught AI Models How to Talk to Each Other Without Using Words
These Russian Mathematicians Taught AI Models How to Talk to Each Other Without Using Words
这些俄罗斯数学家教会了 AI 模型如何在不使用语言的情况下相互交流
I recently met with some brilliant Russian mathematicians who showed me a way for artificial intelligence models to communicate via something akin to machine telepathy. 最近，我见到了一些才华横溢的俄罗斯数学家，他们向我展示了一种让人工智能模型通过类似于“机器心灵感应”的方式进行交流的方法。
The mathematicians work for a startup called Mostik—the Russian word for bridge. It’s a nod to the group’s approach, which allows different models to interact using the mathematical values found in their weights—the things that determine how a prompt gets turned into an output. In practice, this means the capabilities of a larger model can be fed to a smaller model to ramp up its intelligence much more efficiently. 这些数学家就职于一家名为“Mostik”的初创公司——这个词在俄语中意为“桥梁”。这正是该团队方法的体现：他们允许不同的模型利用其权重中的数学值进行交互，而权重正是决定提示词如何转化为输出结果的关键。在实践中，这意味着大型模型的能力可以被“喂”给小型模型，从而更高效地提升后者的智能水平。
The startup used the approach to build a model that has rocketed to the top of ARC-AGI 3, a notoriously difficult competition for AI models. (They wouldn’t tell me more because they want to win the contest.) To demonstrate the idea, however, they also created a bridge between two Chinese open-weight models: the largest version of GLM-5.2, which has 753 billion parameters; and a 4-billion-parameter version of Qwen-3.5 that can run on a mobile device. The resulting hybrid system costs one-twentieth of the full GLM model, and its performance is exactly halfway between the two. 该初创公司利用这种方法构建了一个模型，并迅速冲到了 ARC-AGI 3 的榜首——这是一项对 AI 模型而言极其困难的竞赛。（他们没有透露更多细节，因为他们想赢得比赛。）不过，为了演示这一理念，他们还在两个中国开源权重模型之间架起了一座“桥梁”：一个是拥有 7530 亿参数的 GLM-5.2 最大版本，另一个是可以在移动设备上运行的 40 亿参数版 Qwen-3.5。最终形成的混合系统成本仅为完整版 GLM 模型的二十分之一，而其性能恰好处于两者之间。
“It’s well-known in machine learning that ensembles of models perform better than individual ones,” Sasha Malysheva, Mostik’s CEO, told me over coffee. “在机器学习领域，模型集成比单一模型表现更好是众所周知的，”Mostik 的首席执行官 Sasha Malysheva 在喝咖啡时对我说道。
Malysheva, who developed the approach, shared a running joke inside the company: The future of AI is similar to guessing the weight of a pig. In math circles, it’s well-known that a handful of random people can more accurately estimate a pig’s weight than an expert when their guesses are combined and averaged. 开发了这种方法的 Malysheva 分享了公司内部的一个梗：AI 的未来就像猜猪的体重。在数学界，人们普遍知道，当把一群随机路人的猜测汇总并取平均值时，其结果往往比专家的单次估算更准确。
Much like communally eyeballing porcine heft, combining the outputs of several AI models often nets better results. Typically, this involves feeding the output of one model into another, which takes a good chunk of time and money. The Mostik team, however, figured out a way for AI models to talk to one another without producing text output. If it takes off, it could increase the value of open-weight models, allowing them to better compete with the closed, proprietary models offered by frontier labs like Anthropic and OpenAI. 就像集体目测猪的重量一样，结合多个 AI 模型的输出往往能获得更好的结果。通常，这涉及将一个模型的输出输入到另一个模型中，这会消耗大量的时间和金钱。然而，Mostik 团队找到了一种让 AI 模型在不产生文本输出的情况下相互交流的方法。如果这种方法流行起来，它可能会提升开源权重模型的价值，使它们能够更好地与 Anthropic 和 OpenAI 等前沿实验室提供的封闭式专有模型竞争。
Malysheva says that combining lots of different models may turn out to be a better way to advance AI. “I personally do not think we will have a monolithic model [in the future] or that the capabilities of models will come from scaling,” she told me, referring to the strategy of making models larger and feeding them more data. Malysheva 表示，结合许多不同的模型可能被证明是推进 AI 的更好方式。“我个人认为（未来）我们不会拥有一个单一的庞大模型，也不会认为模型的能力仅仅来自于规模化，”她对我说道，她所指的是那种通过不断扩大模型规模并喂入更多数据的策略。
“If Mostik makes it possible to pair frontier models with domain-specific models—think biology, physics, and so on—many more specialized models would be trained,” says Vladimir Arustamian, the tech lead at the AI software company Lovable, who knows the Mostik team. “This team has been at it for a matter of months and already has something running that I would have guessed was years out.” “如果 Mostik 能够将前沿模型与特定领域的模型（比如生物学、物理学等）配对，那么将会训练出更多专业化的模型，”AI 软件公司 Lovable 的技术负责人 Vladimir Arustamian 说道，他与 Mostik 团队相识。“这个团队才做了几个月，就已经运行出了我原本以为需要几年才能实现的东西。”
The Mostik technique means “you can approach large-model quality without the large model handling the entire loop, giving you substantial improvements with just a smaller model running alongside,” says Karl Tuyls, a former computer scientist at Google DeepMind who is familiar with the company’s tech. The method is a no-brainer for anyone tasked with running models as efficiently as possible, Tuyls says. 曾任 Google DeepMind 计算机科学家、现熟悉该公司技术的 Karl Tuyls 表示，Mostik 的技术意味着“你可以在不需要大型模型处理整个循环的情况下，接近大型模型的质量，只需在旁边运行一个较小的模型，就能获得显著的性能提升。”Tuyls 说，对于任何致力于尽可能高效运行模型的人来说，这种方法是不二之选。
Stanislav Smirnov, a professor at the University of Geneva and a 2010 Fields Medalist, is Mostik’s chief scientist. He says finding common ground between two AI models is surprisingly difficult. “There seems to be no appropriate mathematical language yet,” he says. In the interim, Mostik’s approach is a way to quite literally bridge the gap. 日内瓦大学教授、2010 年菲尔兹奖得主 Stanislav Smirnov 是 Mostik 的首席科学家。他说，在两个 AI 模型之间找到共同点出奇地困难。“目前似乎还没有合适的数学语言，”他说。在此期间，Mostik 的方法实际上就是一种弥合差距的方式。
Smirnov says Mostik’s work could also perhaps reveal new things about how AI models actually function and how this compares to the workings of the human brain. Smirnov says a deeper mathematical analysis may reveal a commonality in the way both AI models and human beings reason over difficult problems. Smirnov 表示，Mostik 的工作或许还能揭示关于 AI 模型实际运作方式的新见解，以及它与人类大脑运作方式的对比。Smirnov 认为，更深入的数学分析可能会揭示 AI 模型和人类在处理难题时推理方式的共同点。
During our coffee meeting, Malysheva told me that she only discovered a talent for math after her older brother told her she wouldn’t be able to solve the Math Olympiad problems he was studying. A few years later, she was studying at one of the top schools in St. Petersburg. More recently, some peers warned her that the bridge approach would be too difficult to pull off. 在我们的咖啡会谈中，Malysheva 告诉我，她是在哥哥告诉她“你解不出我正在研究的奥数题”之后，才发现自己在数学方面的天赋。几年后，她进入了圣彼得堡的一所顶尖学校学习。最近，一些同行警告她，这种“桥梁”方法太难实现了。
“They said it might be too hard for a young girl,” she says. “I decided I need to prove them wrong.” “他们说这对一个年轻女孩来说可能太难了，”她说，“我决定要证明他们错了。”
This is an edition of Will Knight’s AI Lab newsletter. Read previous newsletters here. 这是 Will Knight 的 AI Lab 通讯。点击此处阅读往期通讯。