如果AI模型商品化，谁会胜出？——对话Mistral CEO Arthur Mensch

如果所有领先的AI模型都表现得一样，它们实际上已经如此，那么AI商业会是什么样子？

What does the AI business look if all the leading models perform the same, which they kind of are?

我们将在接下来的内容中，听取Mistral公司的首席执行官对此的见解。

We'll find out with the CEO of Mistral right after this.

AI最有价值的应用是否在于工业场景？

Can AI's most valuable use be in the industrial setting?

在参观了IFS在纽约市举办的‘Industrial X Unleashed’活动，并与IFS首席执行官马克·穆菲特交谈后，我越来越深入地思考这个问题。

I've been thinking about this question more and more after visiting IFS' Industrial X Unleashed event in New York City and getting a chance to speak with IFS CEO, Mark Muffett.

举个明确的例子，穆菲特告诉我，IFS正在派遣波士顿动力公司的Spot机器人进行巡检，将收集到的数据传回IFS的神经中枢，再借助大型语言模型，为需要检修的区域指派合适的技术人员。

To give a clear example, Muffett told me that IFS is sending Boston Dynamics spot robots out for inspection, bringing that data back to the IFS nerve center, which then with the assistance of large language models, can assign the right technician to examine areas that need attending.

这是这项技术的一个令人着迷的前沿领域，我很感谢IFS的合作伙伴让我看到了这一点。

It's a fascinating frontier of the technology, and I'm thankful to my partners at IFS for opening my eyes to it.

如需了解更多，请访问 ifs.com。

To learn more, go to ifs.com.

那就是 ifs.com。

That's IFS dot com.

财政负责、金融天才、货币魔术师。

Fiscally responsible, financial geniuses, monetary magicians.

这些是人们在说那些将车险转投Progressive并节省数百美元的司机时会用的词。

These are things people say about drivers who switch their car insurance to Progressive and save hundreds.

因为Progressive为一次性付清保费、拥有房产等提供折扣。

Because Progressive offers discounts for paying in full, owning a home, and more.

此外，当您需要帮助时，您可以依赖他们出色的服务，让您的每一分钱都花得更值。

Plus, you can count on their great customer service to help when you need it, so your dollar goes a long way.

访问progressive.com，看看您是否能节省车险费用。

Visit progressive.com to see if you could save on car insurance.

Progressive意外伤害保险公司及其关联公司，潜在节省金额因情况而异，并非在所有州或情况下都适用。

Progressive Casualty Insurance Company and affiliates, potential savings will vary, not available in all states or situations.

欢迎收听《大型科技》播客，这是一档致力于对科技世界及其更广泛领域进行冷静而细致对话的节目。

Welcome to Big Technology Podcast, a show for cool headed and nuanced conversation of the tech world and beyond.

今天我们为您准备了一场精彩的节目。

We have a great show for you today.

我们将深入探讨AI产业和技术竞赛的现状，随着一些领先的基础模型开始变得越来越相似，这如何改变了行业内的权力平衡。

We're gonna talk all about what's happening to the AI business and technology race as some of the leading foundational models start to look the same how that changes the balance of power in the industry.

我们非常荣幸邀请到一位完美的嘉宾来讨论这个话题，Arthur Mench 现在与我们同在。

We're joined by the perfect guest to do it, Arthur Mench is here with us.

他是Mistral公司的首席执行官兼联合创始人。

He is the CEO and co founder of Mistral.

Arthur，欢迎你。

Arthur, welcome.

很高兴能来到这里。

I'm happy to be here.

Mistral这个名字在AI圈内广为人知，但对我们的一些听众和观众来说可能还比较陌生。

Mistral is a name that those who are deep in the AI world know very well, but might be new to some of our listeners and viewers.

所以，对于刚接触Mistral的朋友，让我先分享几个数据。

So for folks who are new to Mystrol, let me give you a couple of stats.

Mistral是一家AI模型构建公司。

Mystrol is an AI model builder.

它还做了一些其他事情，我们稍后会谈到。

Does some other things which we're going to get to.

公司位于法国。

It's based in France.

该公司自2023年4月成立以来，估值已达140亿美元。

Company is valued at $14,000,000,000 after starting in April 2023.

不到三年，或者说两年半时间，就打造了一个140亿美元的生意。

So a little under three years or two and a half years to make a $14,000,000,000 business.

不错。

Not bad.

公司有500名员工。

There's 500 people at the company.

亚瑟，你在进入学术界一段时间后，又在DeepMind工作了两年半，现在领导着这家公司。

And Arthur, you are leading it after spending some time in the academy and two and a half years at DeepMind.

没错。

Exactly.

我们总部设在巴黎，但我们的员工中有大约四分之一在美国。

We're headquartered in Paris, but we have around the fourth of our workforce, which is actually in The US.

我们的许多活动实际上都在这里进行，因此我也花了很多时间在这里。

A lot of our activities are actually here, so that's why I'm spending a lot of time as well.

这就是我们为什么在纽约的原因。

And that's why we are here in New York.

好的。

All right.

很高兴你来到演播室。

Well, great to have you in studio.

让我们直接进入我认为当今人工智能最紧迫的问题。

Let's just go right to what I think is the most pressing issue for AI today.

关于谷歌在2025年如何赶上OpenAI的模型，以及OpenAI的模型如何与其他模型相当，已经有很多讨论。

There's been so much talk about how Google at the 2025 started to equal OpenAI's models and how OpenAI's models were somewhat on par with others.

在我看来，我们正以比预期快得多的速度迎来基础模型的同质化。

And to me, it seems like we're just hitting commoditization of the foundational model much faster than I thought it would be.

我原本以为会有一场竞赛，一些公司会遥遥领先，而其他公司需要一段时间才能赶上。

I thought that there was going to be a race where some companies would leap out further ahead and would take others some time to catch up.

但看起来现在，许多模型构建者推出的前沿模型表现非常相似，几乎难以分辨哪个最好。

But it looks like right now, you have lots of model builders with their frontier models exhibiting performance that's so similar is difficult to tell which is the best.

那你对此怎么看？

So what do you make of that?

我认为，这种技术本质上注定会被商品化。

I would say that inherently this is a technology that is going to get commoditized.

原因在于，构建这种技术实际上并不难。

The reason for that is that it's actually not hard to build.

全球大约有十家实验室掌握这项技术，它们能接触到相似的数据，遵循相同的配方和算法，而这些知识实际上非常简短。

You have around 10 labs in the world that know how to build that technology, that get access to similar data, that follows the same recipes and algorithms, which very short actually.

训练模型所需的知识相当有限。

The knowledge you need to actually train a model is fairly short.

正因为如此，这些知识才会迅速传播。

So because it's short, it actually circulates.

因此，你无法建立知识产权上的差异化优势，很难真正实现超越并大幅领先竞争对手，因为知识的扩散使得所有人都在做同样的事情。

So there's no IP differentiation gap that you can create, so it's very hard to actually leapfrog and to be way ahead of the competition, because there's some diffusion of knowledge that is just making everybody do the same things.

因此，问题在于价值究竟体现在哪里，你应该采取什么样的商业模式，才能确保最终实现盈利。

And so the question there is therefore where is the value accruing, and what kind of business model should you pursue to actually make sure that in the end you're turning profitable.

而我们看到一些竞争对手面临的挑战是，他们投入了数十亿甚至数百亿美元来创建资产，但这些资产却在迅速贬值，因为它们已经变成了商品。

And then the challenge that we see with some of our competitors is that they're investing billions or hundreds of billions into creating assets that are deprecating fairly fast, because those are commodities.

因此，对于我们Mistral来说，这个行业一直以来最大的问题之一就是：你需要投入足够的资金来为企业带来价值，但同时也需要合理投入，以便在这样一个模型创造（资本密集型）实际上只带来处于激烈竞争中的资产的世界里，建立起合理的经济模型。

So for us, at Mistral, it has always been one of the biggest questions of the industry, is that you need to invest enough to actually bring value to enterprises, but you also need to invest reasonably so that you can build unity economics that make sense in a world where the creation of model, which is capital intensive, is actually just bringing you assets that are just in a community competition.

那么，让我们来谈谈构建最佳模型的这场竞赛。

So let's talk a little bit then about this race to build the best possible model.

我的意思是，正如你提到的，这非常昂贵。

I mean, like you mentioned, it's very expensive.

OpenAI计划投入1.4万亿美元用于为其模型建设基础设施，至少他们是这么说的。

OpenAI is going to put $1,400,000,000,000 into building infrastructure for its models, or at least it says so.

如果这些模型实际上已经处于同等水平，那么公司会不会说：等等，也许我们没有必要投入这么多钱去开发下一代更好的模型，因为别人很快就能追上来？

If the models are effectively at par, are companies going to say, hey, wait a second, maybe it doesn't make sense for us to invest all this money into building the next evolution of a better model because people can catch up?

我的意思是，从战略上讲，确实需要设定一些方向。

I mean, strategically, I think it's definitely there's some cursor to be set.

你应该投入多少资金来创建足够有价值的产品，让一家科技公司能够为企业或消费者带来价值？

How much do you invest in creating assets that are valuable enough for one company to bring, for one technology company to bring value to an enterprise or to bring value to a consumer.

最终，所有这些投资都需要由下游产生的自由现金流和价值创造来支持。

At the end of the day, all of these investments will need to be funded by the free cash flow and value creation that is being made downstream.

因此，我们公司关注的重点，也是我认为合理的重点，是更多地聚焦于下游应用，弄清楚企业面临哪些障碍，并努力消除这些障碍。

And so the focus that we have as a company, but that I think is the reasonable focus, is to be more on the downstream applications and to figure out what is the friction that enterprises are running into and try to lift these frictions.

因为归根结底，我认为当今行业面临的主要挑战之一是，AI在三四年前许下了许多承诺。

Because at the end of the day, I think one of the major challenges that the industry is facing today is that AI brought a lot of promises, like three, four years ago.

但如果你问一家企业：你真的从中赚到钱了吗？

But if you ask an enterprise, did you actually make money out of it?

他们通常会回答：没有。

They will, in general, say no.

原因在于，他们没有足够地定制化解决方案，也没有从想要解决的问题出发进行逆向思考。

And the reason for that is that they are not customizing things enough, and they are not thinking backward from the problem they want to solve.

所以他们只想着解决方案，却不去思考问题本身。

So they think about the solution, but they don't think about the problem.

因此，我们需要帮助他们找到正确的应用场景，并进行恰当的定制，比如当一个由20人组成的团队在运营某个供应链流程时，你突然能用两个人就搞定。

And so trying to help them to actually go for the right use cases and actually do the right amount of customization, so that when it has a team of 20 people actually operating some supply chain workflow, suddenly you can actually operate that with two people.

类似这样的例子还有很多。

And there's a lot of examples like this.

但行业面临的挑战是，我们需要让企业尽快认识到价值，以证明所有这些集体投入的合理性。

But the challenge that the industry will face is that we need to get enterprises to value fast enough to justify all of the investments that is collectively being made.

这非常有趣，因为长期以来，你总会听到这些公司一味强调模型、模型、模型。

That is very interesting because for a long time you would hear these companies focus model, model, model.

对吧？

Right?

比如，当你想到OpenAI时，最大的新闻曾经是下一个GPT-5是什么。

The next what's GPT-five was, let's say, when you think about OpenAI, the biggest news.

但现在，他们开始更多地谈论如何利用已有的智能来构建真正可用的应用。

Now they're starting to talk more about how do you take the intelligence that you have and build the applications that work.

我只能分享一点最近几周的报道：我曾在纽约市与萨姆·阿尔特曼和一群新闻界领袖共进午餐时听到了这个故事。

Just one bit of reporting that I can share a couple of weeks ago, I had this story basically inside a lunch with Sam Altman and a bunch of news leaders in New York City.

阿尔特曼告诉他们，公司最重要的优先事项之一就是为企业构建应用程序。

And Altman told him the company's one of their biggest priorities was building applications for enterprises.

这基本上将成为2026年的重大优先事项。

Basically, it's going be a major priority in 2026.

这种说法与之前‘我们想打造通用人工智能’相比，已经略有转变，现在变成了‘我们想为企业构建应用’。

And it's a little bit of a shift in rhetoric from we want to build AGI to we want to build applications for business.

那么，为什么会出现这种变化呢？

So talk about why is that happening?

这是商品化问题的衍生结果吗？

Is that an offshoot of this commoditization issue?

我认为问题在于，首先，通用人工智能这个概念太简单了，对企业来说可能过于简单。

Well, think the issue is well, first of all, AGI is a very simple concept, so probably too simple for enterprises.

根本不存在一个系统能解决世界上所有问题。

There's no such thing as like one system that is going to be solving all of the problems of the world.

所以归根结底

And so at the end of the day

还是你根本不相信这个概念

yet, or you just don't believe in that concept

它永远不会存在。

It's at never going to exist.

我的意思是，问题多种多样，就像没有任何一个人能解决世界上所有任务一样。

I mean, have a wealth of problems, just like you don't have any human that is able to solve every task in the world.

当然，你需要一定程度的专业化才能真正解决问题。

You of course need to have some amount of specialization to actually solve problems.

因此，我们从魔法思维回到了系统思维。

And so we are back from magical thinking to system thinking.

我们需要弄清楚，哪些数据将用于提升模型在特定任务上的表现。

We need to figure out what is the data that is going to be used to make the model better at a specific task.

我们需要建立什么样的飞轮，才能从人类与系统的互动中积累更多信号，从而使应用变得越来越好。

What is the flywheel that we need to set so that we accrue more signal from humans interacting with the system so that eventually the application becomes better and better.

因此在现实生活中，企业只是复杂的系统，你无法用单一的抽象概念——即通用人工智能——来解决这些问题。

And so in real life, enterprises are just complex systems, and you can't solve that with a single abstraction, which is AGI.

通用人工智能在很大程度上是我们未能实现的目标，它原本是作为让系统随时间不断改进的北极星目标。

AGI, to a large extent, is what we were not able to achieve, which is basically the north star of just going to make the system better over time.

但正如你所说，现在很难向投资者解释你所构建的技术将永远不会被竞争对手超越，因此叙事自然发生了转变。

But because, as you said, it's hard to explain now to investors that the technology you're building is never going to be matched by your competitors, then there's of course a shift in the narrative.

公司不再致力于构建一个能解决所有问题的单一北极星系统，而是需要深入企业内部，解决它们的实际问题。

Companies are not building a North Star single system that is going to be solving all problems, but that will need to go into the weeds of enterprises and solving their actual problems.

我认为在中期阶段，我们在思考这个问题时已经走在了时代前列。

And I think at mid start, we've been ahead of time in thinking about this.

这塑造了我们的故事。

That set us our story.

我们的故事基于这样一个前提：最终人工智能将更加去中心化，需要更多的定制化，因为我们遇到了数据积累量的极限以及扩展这些数据的极限，因此我们基于这一事实创立了公司，旨在为企业提供更强的定制能力。

Our story has been to assume that eventually AI will be more decentralized, that more customization will be needed, because we were running into the limits of the amount of data we could accrue and the limits of scaling those, and because of that, we created the company on that premise, on the fact that we'll bring more customization ability to enterprises.

是的，我们稍后会谈到那位先生。

Yeah, and we'll get to the Mr.

我们稍后再讲这个故事，但还有一个关于这个问题的问题。

All story in a little bit, but one more question about this.

在我看来，我想知道你是否认为这已经发生了一种转变，你确实领先于这一趋势，但在我看来，人工智能行业似乎出现了一种转变：过去的想法是让模型变得更聪明，它们就能自己解决这些问题。

It seemed to me, and I wonder if you think this has been a shift, you were ahead of this for sure, But it seems to me like there's been a shift in the AI industry, where the idea was effectively make the models smarter, and they'll try to figure out these they'll be able to figure out these problems on their own.

我举个具体的例子。

For instance, I'll just make it concrete.

让模型更聪明，它就能胜任初级助理的工作，或者为多个系统录入数据并自动生成报告。

Make the model smarter, and it will be able to do a lower level associate's job or maybe do data entry for multiple systems and be able to file reports.

但现在看来，趋势已经从‘让模型更聪明’转向了构建基础设施——模型只是其中一个组成部分，而基础设施至关重要。

And now it seems like there's been a shift from do that to actually build out the infrastructure, that the models are just one component, that the infrastructure is super important.

编排以及构建在模型之上的应用程序，才是价值所在。

And things like orchestration and working through the applications that are built on top of the models is going to be where the value is found.

这很有趣。

It's interesting.

是的，我认为从系统角度来看，你有两个组成部分，而这两个部分将永远存在。

Yes, I think if you look at it from a system perspective, you have two components, and we'll always have these two components.

第一部分是工作流应该如何运行以及系统应该如何行为的静态定义。

The first components are like static definitions of what the workflow should be and how a system should behave.

这些静态定义由人类设定，用于规定系统的行为方式。

And those static definitions are set by humans that are defining how the system should behave.

因此，这对应于你用来定义系统的手动信息。

And so this corresponds to the manual information that you're using to define the system.

然后还有一个动态部分，即你将模型与工具连接起来，向模型发出指令，模型可以自行调用这些工具。

And then there's a dynamic component where you're connecting a model to tools and you're giving instructions to the model and the model can go and call the tools itself.

因此，它可以决定将遵循的执行图谱。

And so it can decide on the graph of execution that it's going to follow.

这部分是动态的，而静态部分则是你设置护栏或有时决定决策树的地方。

And so that part is dynamic, and there's a static part where you're setting up guardrails or you're deciding you have a tree of decisions sometimes.

我认为，认为可以在没有人类指导的情况下，仅靠动态系统解决所有问题是不切实际的。

And I think it's a bit and unrealistic to think that you can solve everything with a dynamic system without guidance from humans.

过去三年中，行业发生的变化是，动态部分显著增长，因为模型能够进行更长时间的思考，可以调用多个工具，甚至能够编写代码，但静态部分仍然极为重要。

And what has happened in the industry in the last three years is that effectively the dynamic part has grown because models can think for longer, because they can call multiple tools, because they can code, but the static part remains extremely important.

即使动态部分在增长，静态部分仍能帮助你构建出更好、更有趣、并能解决以往无法解决的问题的系统。

Even if the dynamic part grows, then the static part allows you to create systems that are even better and more interesting, and you can solve problems that you were not able to solve before.

因此，将这些静态系统（如果你愿意，可以称之为编排）与动态系统（可以称之为智能体）相结合，将始终至关重要，因为这两者正在共同进步，使我们能够应对越来越复杂的问题。

So the combination of these static systems, which you can call orchestration if you want, and the dynamic systems that you can call agents, is going to stay super important because the two things are moving up together so that we can tackle problems that are more and more complex.

好的。

Okay.

既然这一点已经明确，我现在正在思考这个商业模式。

And so now, with that established, I'm thinking through what the business is.

假设模型已经商品化了。

Let's say the model has commoditized.

那么人工智能领域的商业模式会是什么？

So what are the businesses going to be in AI?

我认为，会是一些类似聊天机器人的消费类产品，你可以把OpenAI归入这一类。

It will be, I imagine, some form of consumer products like chatbots where you could put OpenAI in that bucket.

还会是那种让你现有产品变得更优秀的业务，比如与微软Excel进行对话。

Will be a business where you could make your existing products better, like for instance, maybe chatting with Microsoft Excel.

这可能是现有公司改进其产品的一种方式。

That could be one way that current companies can make their products better.

但还有另一个重要的领域，我们之前已经稍微讨论过，那就是企业方面。

But then there is this other big bucket, which we've talked about a little bit already, which is the enterprise side of things.

那么，你会如何评估这三个领域的商业机会？

So how would you rank the business opportunity in those three buckets?

嗯，是的，我认为在消费端，因为人工智能正逐渐成为人们获取信息的方式。

Well, yes, I think on the consumer side, because AI is starting to be well, it's becoming the way you access information.

你基本上可以构建一个广告业务，这显然会得到发展。

You basically have an ad business to be built, and that's pretty clearly going to be built.

但这并不是我们公司的重点。

It's not the focus of our company.

而如果你看一下企业端，我们实际上是在重新构建所有企业软件。

Then if you look at the enterprise side, we're basically replatforming all enterprise software.

在企业中，你有人员、数据和流程。

In enterprises, you have people, you have data, and then you have processes.

历史上，运行多个流程、多个数据系统和多个记录系统的工具是分散的，团队也无法同时访问所有信息。

Historically, there was a fragmentation of the tools to run multiple processes, multiple data systems, multiple system of records, and there was a fragmentation in teams that were not able to access all information at the same time.

本质上，AI在企业中的作用是，你可以从统一的数据开始，甚至可以从分散的数据源开始，因为AI能够导航这些数据源。

Essentially, what AI allows you to do in an enterprise is to start with a unified data, or even you can start with fragmented data sources because the AI is able to navigate them.

你在上面部署一个AI，它能构建适当的智能，理解企业中正在发生的事情，然后AI系统能够生成对每个员工都有用的界面。

You put an AI on top that is building the right amount of intelligence, understanding what's going on in the enterprise, and then the AI system is able to generate the interfaces that are useful for every human to actually work.

因此，这种对企业软件栈的重新平台化，是企业中能够创造大量价值的唯一关键点。

And so that part, that replatforming of the entire enterprise software stack, is the one thing where a lot of value can be created in the enterprise.

掌控上下文引擎，即那个持续运行、监控发生的事情并自动生成相关文档的系统。

Owning the context engine, so the system that is constantly running, that is looking at what's happening and figuring out, creating documentation for what's happening.

同时掌控前端界面，这些界面正越来越多地按需生成。

Owning the front end as well, that are more and more getting generated on demand.

比如说，我是一名律师，我想解决一个问题，需要进行一项特定的审查，我只需上传我的文件，系统就会自动演化，向我展示正确的工具和所需的信息。

So let's say I'm a lawyer, I want to fix one of my problems, and I have a very specific review to make, I just bring my documents, and then the system actually evolves in showing me the right widgets and showing me the right information I need.

因此，这是在持续更新企业动态的上下文引擎之上，构建生成式界面，而底层的记录系统本质上将只是纯粹的数据库。

So the generative interfaces on top of a context engine that is constantly updating its representation of what's happening in the enterprise, on top of system of records that are essentially going to be just pure databases.

你不再需要之前所有那些堆叠在上面的东西。

You don't need everything that was sitting on top before.

这就是未来的发展方向，这种重新平台化可能需要十年时间，因为企业采纳这些技术需要一段时间，但其中蕴含着巨大的价值，因为突然间你可以围绕这样一个事实重组公司：许多原本需要大量人力的流程，现在可以快得多地运行。

This is where this is going, and that replatforming is going to be I think it's going to take a decade, because it takes a while to get enterprises to adopt these things, but there's just immense value to be created, because suddenly you can reorganize your company around the fact that for many of the processes where you had a lot of people, you can actually run those very much faster.

所以，一方面，是效率。

So that's on one side, efficiency.

另一方面，我认为这是企业的一种商业模式。

And the other thing, is the most, that's, I'd say that's one of the business modalities of the enterprise.

企业的第二个方面是帮助企业利用其高度专有的数据，比如制造业中由机器产生的资产，并将其转化为他人无法复制的智能。

The second one in the enterprise is about working with enterprises to help them take their really proprietary data, the assets being produced by their machines, if it's in the manufacturing industry for instance, and turning that into intelligence that nobody else can reproduce.

因此，当我们与一家制造飞机的公司合作，或与ASML合作时，打造专门擅长其机器操作的模型，这具有巨大的价值，因为突然间你不再只是提升公司内部的效率，而是解锁了因缺乏AI而被封锁的技术进步。

And so making models specifically good at a certain kind of physics when we're working with a company doing planes for instance, or when we're working with ASML, making models that are specifically good at operating their machines, that's huge value, because suddenly you're not building efficiency within the company, but you're effectively unlocking technological progress that was locked by the absence of AI.

因此，新系统所提供的这种解锁，带来了巨大的增长。

So that unlock that the new systems are providing, that's immense amount of growth.

这实际上更难衡量，因为第一种是短期的。

It's actually harder to measure, because the first one is shorter term.

你可以看到一家公司在五年后的样子，因为你已经精简了公司的某些部分。

You can look at what a company will look like in five years because you've reduced certain parts of the company.

你重新调整了其他人的工作方向，让他们专注于创造增长。

You've reoriented other people to be creating growth.

你可以为此建立模型。

You can create models of that.

在技术层面，我认为这稍微困难一些，因为我们知道像核聚变或更精细的半导体刻蚀之类的技术存在。

On the technological side, I think it's a little harder because we know there are things like nuclear fusion or sharper engraving of semiconductors, for instance.

这些领域我们正开始遇到物理限制，而人工智能实际上可以帮助突破这些物理限制。

These are things where we are starting to run into physical constraints, and artificial intelligence can actually help to lift those physical constraints.

因此，技术进步的加速，我认为将是价值创造的主要来源。

And so the acceleration of technological progress is, I think, where most of the value creation will be.

这需要一些时间，并且其效果将不如人工智能带来的效率提升那样容易衡量和预测。

It will take a little bit of time, and it will be less measurable, less predictable than the efficiency gains that AI is going to produce.

但这两方面同样重要。

But the two things are as important.

好的。

Okay.

所以让我试着在这里稍微推演一下。

So let me see if I can sort of game this out here a little bit.

如果这将成为人工智能世界中的关键价值驱动因素，那么有两种方式可以实现。

So if that is going to be the key driver of value in the AI world, there's two ways to do it.

一种是构建一个比所有人都更好的模型，并以高价出售。

One is to build a model that's better than everybody else and sell it for a premium.

但我们已经讨论过，这种情况似乎不会永远持续下去。

But we've already talked about the fact that that doesn't seem like it's going to be emote forever.

另一种方式是，模型本身并不是价值所在。

And the other way is the model is actually not the value.

真正有价值的是专业知识和实施层面。

It's the know how and the implementation side of things.

你可以将模型开源，但同时为企业提供服务，帮助它们理解如何将该模型付诸实践并真正取得成果。

So you can make the model open source, but then provide a service to businesses to be able to figure out how to take that model and put it into action and actually get results.

这两种选择吗？

Are those the two choices?

是的，这正是行业中模糊不清的地方。

Yeah, that's kind of the fog that we see in the industry.

我们的观点是选择第二种，

And our view there has been to be on the second one,

专注于开源

to really The open source

开源。

The open source.

以及实施层面。

And implementation side.

这带来了定制化，同时也带来了去中心化，因为如果你假设整个经济都将运行在AI系统上，企业自然会希望确保没有人能关闭他们的系统。

Which brings customization, but it also brings decentralization, in that if you assume that the entire economy is going to run on AI systems, well, enterprises will just want to make sure that nobody can turn off their systems.

就像你有一个工厂，接入电网时，你希望确保没有人会因为不喜欢你而切断电网。

So the same way, if you have a factory, you connect it to the grid, you want to make sure that nobody is going to turn off the grid because they don't like you.

如果人工智能真正成为一个社区，而这正是正在发生的事，并且如果你把智能看作电力，那么你只想确保自己对智能的访问不会被限速。

If AI effectively becomes a community, which is what's happening, and if you treat intelligence as electricity, then you just want to make sure that your access to intelligence cannot be throttled.

因此，这也是开源技术能够带来的一项优势。

And so that's also one of the things that open source technology can bring.

你使用开源技术，就不必担心违反Anthropic的用户条款，从而导致你的操作能力被暂停。

You're using open source, you don't have to worry about going astray of I'm just saying Anthropics user terms, and so then pausing your ability to do what you do.

如果你使用开源技术，基本上就可以按照自己的方式运行它。

If you use open source, you can basically run it on your own terms.

是的。

Yeah.

你可以按照自己的方式运行它。

You run it on your own terms.

你可以建立所需的冗余。

You create the redundancy you need.

你可以提供更高质量的服务。

You can serve with higher quality of service.

你可以确保，无论地缘政治局势如何，只要你愿意，都能继续运行这些系统。

You can make sure that whatever the geopolitical situation may be, you can still run the systems if you want.

然后，这主要涉及IT层面。

And then, so that's really on the IT side.

所以，如果我是一名首席信息官，我会把开源视为创造杠杆效应和独立性的途径。

So if I'm a CIO, I really look at open source as a way to create leverage and independence.

但在科学层面，这也是唯一能够有效利用员工经验知识、你数十年积累的知识的方式。

But more on scientific side, it's also the only way in which you can create systems that are effectively using the folklore knowledge of your employees, the knowledge that you've accrued for decades.

将这些知识转化为无人能获取的资产的唯一方法，就是基于这些开源模型构建你自己的模型。

The only way in which to turn it into an asset that nobody gets access to is to create your own models based on those open source models.

但这很难。

It's But hard.

真正构建这些模型是困难的。

It's hard to actually build those.

因此，你需要合适的工具、专业的技能，而这正是构建开源模型的互补商业模式。

And so that's where you need the right tools, you need the right expertise, and that's the complement business model to building open source models.

但即使是闭源模型的提供商，比如Anthropic公司，也会说他们能够用你的数据定制模型。

But even the closed source model providers, companies like Anthropic, will say they'll be able to customize their models with your data.

你不相信吗？

You don't believe that?

他们会这么说，但随后会在上面加上一些限制措施。

They will say that, but then they will put some guardrails on top of it.

所以你本质上是在信任他们的工程师会给你足够的系统深度访问权限。

So you're basically trusting that their engineers are going to give you enough access to the depth of the system.

你能永远信任这一点吗？

And can you trust that for eternity?

我不确定。

I'm not sure.

因此，这个问题既是控制权的问题，也是定制化的问题。

So the issue there is as much a question of control as a question of customization.

就像供应商会试图把你锁定在他们的系统里。

Like a vendor is going to try to lock you in.

所以，如果你获得了访问权限，并基于开源模型（比如我们的开源模型或其他人的）进行构建，你就不太会被厂商锁定。

So if you get access and if you build on top of open source models, like our open source models or anyone, you're basically less locked into the vendor.

而这项技术如此重要，你绝对不希望被单一厂商锁定。

And this is a technology which is so important that you don't want to be locked into a single vendor.

因此，这也是我们带来的机遇。

So that's also the opportunity we bring.

你知道让我感到震惊的是什么吗？

You know what's stunning to me?

我们已经距离ChatGPT问世三年了，它让很多人意识到了这一点。

We're three years past ChatGPT, which basically brought this into a lot of people's consciousness.

尽管我认为科技圈的资深人士可能早就有所了解，尤其是因为我们曾在ChatGPT发布前就采访过那些认为这类技术具有意识的人——那是另一个话题了。

Although I think big technology listeners would have known about it a little beforehand, especially since we were interviewing the people that thought this stuff was sentient before ChatGPT came out, that's a conversation for another time.

今天我们基本上要说的是，我来总结一下你提出的两个主要观点。

What we're basically saying today, I'm going to sum up two of the main points that you've made.

第一个是，如今的AI模型无法独立完成所有事情。

One is that today's AI models can't do it all themselves.

它们需要协调。

They need orchestration.

你提出的第二个重要观点是，要利用当前的智能进行协调或实施，你需要一项服务，一种托管服务。

And the second big point that you made is to do that orchestration or implementation with the current intelligence, you need a service, a managed service.

因此，令我感到有趣的是，我们已经从曾经希望打造一个能完成一切的‘神级模型’的观点，转变为意识到这可能是我们一生中见过的最强大的技术。

So it is interesting to me that we've gone from this perspective of maybe working towards a god model that could do it all to the fact that this may be the most powerful technology that we've seen come through in our lifetimes.

然而，当你真正想使用它时，它在某种程度上必须变成一项托管服务。

However, when you actually want to use it, you kind of need it becomes a managed service in a way.

是的。

Yes.

这是对的。

That's This is true.

这并不是历史上我们第一次观察到这种情况。

Don't think it's the first time that we observe it in history.

这是一种新技术。

It's a new technology.

这是一个新平台。

It's a new platform.

因此，如何使用它的知识仍然非常稀缺。

And so the knowledge on how to use it is actually still pretty scarce.

所以能够构建出可大规模运行、稳定可靠并真正解决问题的系统的人员并不多。

So there aren't that many people that can build systems that are performing at scale, that can run at scale reliably, that can actually solve an actual issue.

因此，在与企业合作时，由于实施的复杂性，即使对于像数据库这样相对成熟的技术，也需要提供上层服务。

And so when working with enterprises, always need to have some services on top because of the complexity of implementation, even with fairly well understood technology like databases.

但对于人工智能而言，这种需求更为必要，因为它需要改造企业业务，因此你还需帮助团队思考如何围绕系统本身开展工作。

But for artificial intelligence, it's even more necessary in that it requires to transform businesses, so you need to also help in thinking how the team should perform around the system itself.

而且这确实需要定制化，因此你需要懂得如何利用数据并将其转化为智能的数据科学家，而如今这类人才仍然非常稀缺。

And it does require to customize things, so you need data scientists that know how to leverage data and turn it into intelligence, and today this is still a pretty scarce resource.

我认为，这些部署中软件所占的比例将会增加，如今通过微调、强化学习等方式实现定制化的过程，将会被抽象化，不再由企业客户直接处理，因为这太复杂了。

I would say I do expect the part of software in those deployments to increase, so the amount of the way customization occurs today with fine tuning, reinforcement learning, these kind of things, this is going to be abstracted away from the enterprise buyer, because it's too complex.

他们真正应该关注的是拥有能够从经验和与人的互动中学习的自适应系统，而不是纠结于究竟该用微调还是强化学习来将知识注入模型。

They actually should just worry about having adaptive systems that are learning from experience and from deployment with people, instead of thinking about, should I use fine tuning or should I use reinforcement learning to actually put that knowledge into my models?

我们正在做的工作是，将数据科学家所熟悉的底层操作抽象为业务负责人能够实际使用的高级系统。

The work that we are doing is to try and abstract away from lower level routines that data scientists understand, to higher level systems that business owners can actually use.

这将会发生，我们正在为此努力，但服务部分仍然会非常重要。

And so it's going to occur, and we're working on it, but the service part is still going to be quite important.

如今，将这两者结合是企业实现价值的最快途径。

And today, the combination of the two things is the fastest way to value if you're an enterprise.

因此，我们一直将两者结合起来。

So we've been combining the two.

我一开始称你为模型构建者，当时我停顿了一下，说了一些稍后我们会深入探讨的其他事情。

I started our conversation by calling you a model builder, and I kind of paused on it and I said some other things that we're going to get into it later.

而我们现在听到的，基本上是你所说的：Mistral 显然是一位引以为豪的模型构建者，但如果没有服务，无法与企业坐在一起，向他们展示如何使用它，那一切就只是个不完整的拼图。

And here we are being basically what I'm hearing from you is that Mistral obviously proud model builder, but it seems like without the services, without being able to sit with a business and showing them how to use it, it just would be an incomplete puzzle.

那么，你认为自己最重要的工作是构建模型吗？

So do you consider yourself is the most important thing you do building the models?

还是说，你最重要的工作是提供服务？

Or the most important thing you do the service?

你是主要作为一个模型构建者，还是主要作为一个服务提供商？

Or are you primarily a model builder or primarily service provider?

我们的目标是帮助客户实现价值。

We are there to help our customers get to value.

所以是服务。

So service.

我们在这里是为了实现价值，但他们需要优秀的模型。

We're here to but to get to value, they need to have great models.

为了实现价值，他们还需要合适的工具来训练模型。

And to get to value, they need to have the right tools to train the models.

因此，创建这些工具的最佳方式就是训练出最优秀的模型。

And so the best way to create those tools is effectively to train the best models.

所以这两者是紧密相连的。

So the two things are extremely linked together.

创建非常容易定制的模型。

Create models that are very easy to customize.

我们创建带有工具的模型，然后将这些工具导出给客户，让他们能够使用，并帮助客户训练他们自己的模型。

We create models with tools that we then export to our customers so that they can use them, and we help our customers train their own models.

因此，如果你无法向世人证明你确实是开源技术的领导者，你就无法向企业销售服务，声称你能帮助他们创建高度定制的系统。

So you can't go and sell to an enterprise that you're going to help them create very custom systems if you can't show to the world that you're effectively the leader in open source technology.

因此，这两方面同样重要。

And so the two parts are equally important.

一方面促进另一方面，实际上形成了一种飞轮效应，因为我们在模型设计上的选择都是为了支持我们各种各样的客户。

The first is enabling the other, and there's effectively a flywheel there, because we make our choices when it comes to the model design in a way that is enabling the various customers we have.

一个例子是，我们非常重视开发在物理方面表现优异的模型，因为我们与面临物理问题的制造公司合作。

One example is that we've put a lot of emphasis on having models that are great at physics, because we work with manufacturing companies that run into physical problems.

这就是我们设立的飞轮机制：让科学团队和业务团队真正坐在一起工作。

So that's the flywheel that we have set up, by having the science team and the business team actually sit together.

好的。

Okay.

我们今天与亚瑟·门奇对话。

We're here with Arthur Mench.

他是Mistral的首席执行官，也是联合创始人。

He is the CEO of Mistral, also co founder.

广告后回来，我们将讨论开源与闭源之间的区别。

When we come back after the break, we are going to talk about open source, the open source movement versus closed source.

记住，DeepSeek和开源本应超越闭源。

Remember, DeepSeek and open source was supposed to surpass closed source.

它有什么？

What has it?

我还会谈谈地缘政治和监管，以及这些是否会为这家公司带来优势，然后可能深入一些实际案例，因为我们应该谈谈这项技术在实际中的应用。

I'll also talk about the geopolitics and regulation and whether that's going to give this company a leg up and then maybe get into some more practical examples because we should talk about how the technology is being used on the ground.

广告后马上回来。

We'll be back right after this.

我们回到《大科技播客》，与Arthur Mensch对话。

And we're back here on Big Technology Podcast with Arthur Mensch.

他是Mistral的首席执行官。

He's the CEO of Mistral.

亚瑟，我想问问你关于我的NordVPN合作伙伴的事情。

Arthur, I wanna ask you about, you know, the progression of Let me tell you about my partners at NordVPN.

如果你想要观看你所在地区无法提供的体育赛事、电视节目或电影，可以通过NordVPN切换虚拟位置到正在播放这些内容的国家来实现。

If you ever wanna watch sporting events, TV shows, or films that aren't available in your region, you can do it by switching your virtual location to a country which is showing that content with NordVPN.

NordVPN还能在你旅行并使用全球任何地方的公共Wi-Fi时保护你的数据安全。

NordVPN also helps protect your data while you're traveling and using public Wi Fi wherever you are in the world.

它是世界上最快的VPN，流媒体时不会出现缓冲或延迟。

It's the fastest VPN in the world with no buffering or lagging while you stream.

NordVPN拥有超过7400台服务器，覆盖118个国家，轻松切换虚拟位置。

NordVPN has 7,400 plus servers across 118 countries with easy virtual location switching.

它支持最多10台设备，而且速度极快。

It supports up to 10 devices, and it's extremely fast.

要获得NordVPN计划的最优惠折扣，请访问nordvpn.com/bigtech。

To get the best discount off your NordVPN plan, go to nordvpn.com/bigtech.

通过我们的链接，你还能在两年计划基础上额外获得四个月服务。

Our link will also give you four extra months on the two year plan.

NordVPN提供30天无风险退款保证。

There's no risk with Nord's thirty day money back guarantee.

链接也在播客节目描述框中。

Link is in the podcast episode description box as well.

这就是NordVPN。

That's NordVPN.

过去一年，开源领域的发展。

Open source over the past year.

我记得今年一月曾阅读过关于DeepSeek的报道。

I remember reading about DeepSeek, doing reporting on DeepSeek in January.

主导的主题是，DeepSeek为开源领域带来了巨大飞跃，很快，像OpenAI的GPT、Anthropic的Claude，甚至谷歌的Gemini这样的闭源模型都将被开源模型超越，因为开源社区正在协同合作、相互借鉴创新，而闭源社区却各自为战。

And the overriding theme was it was such a leap forward for open source that soon the closed models models like Open Ice GPT and Anthropic Anthropic Claude and maybe Google's Gemini would be surpassed by open source because open source the open source community was working together and building on each other's innovations, where the closed source community was kind of going at it on their own.

我们刚刚经历了这样一个时刻。

We just had this moment.

我们在节目开始时讨论过，也许Gemini使OpenAI的GPT模型变得商品化。

We talked in the beginning of the show about how maybe Gemini commoditized OpenAI's GPT models.

但关于开源能否实现年初预期的这种讨论，并没有发生。

But that conversation was not being had about open source being living up to that expectation from the beginning of the year.

所以是我漏掉了什么吗？

So am I missing something?

还是我理解错了？

Or am I reading it wrong?

或者你认为是什么阻碍了开源的发展？

Or what do you think if something has held back open source, what has it been?

如果你看看2024年的趋势，我认为可能有一个六个月的差距。

Well, if you look at the trends in 2024, I'd say there might have been like a six month gap.

如果你看看2025年的趋势，我认为这个差距大约缩短到了三个月。

If you look at the trend in 2025, I think the gap is more around three months.

所以，明年差距会是多少，就由每个人自己猜测了。

So I guess it's up to anyone to guess what the gap is going to be next year.

但事实上，这个差距已经显著缩小了。

But effectively, this gap has been shrinking quite significantly.

原因在于，当你预训练模型时，在10的26次方FLOPS左右会出现饱和效应。

The reason for that is that basically you have a saturation effect when you pre train models, around ten twenty six FLOPS.

原因在于预训练模型时，你能找到用于压缩的数据量是有限的。

The reason for that is that there's only that much data you can find to compress when you pre train models.

因此，那些起步稍晚的实验室实际上也创造了足够的计算能力来训练这种规模的模型。

So effectively, labs that maybe started a little behind created enough compute capacity to train models at this kind of scale.

效率也提高了，这意味着如今每个人都能在几个月内获得10的26次方FLOPS级别的计算设施。

And efficiency has also increased, and so what it means is that today, everybody has access to 10 to the power of 26 FLOPS facilities over the course of a few months.

这是对计算能力的一种衡量。

And that's a measure of compute.

这是对计算时间的衡量，所以现在任何实验室都能在几个月内实现10到26次浮点运算。

That's a measure of compute times, so you need to 10 to 26 FLOPs is something that any lab today can achieve in a couple of months.

正因如此，饱和效应意味着开源模型已经迎头赶上，因为那些领先起步的闭源模型某种程度上撞上了预训练这堵墙。

And because of that, the saturation effect means that open source models have caught up because closed source models that were started ahead kind of run into that wall of pre training.

这意味着差距只会继续缩小，如果我们看看我们最新的开源发布——Deathstyles two，这是一个编码模型，它的表现，我认为，相当于Entropic大约两三个月前的水平。

So what that means is that this is only going to continue shrinking, and if we look at the latest open release we did, which is Deathstyles two, which is a coding model, well it's performing, I think, the performance of Entropic around two or three months ago.

所以，我认为差距正在缩小。

So yeah, I think the gap is shrinking.

而且again，我认为这个问题可能提得不太对，因为它们提供了两种截然不同的价值主张。

And again, I think the question is probably not posed in the right way that way, because it's also offering very two different distinct value propositions.

因为一方面，这是经过良好管理的，你会依赖于提供商本身。

Because on one side, this is well managed and you will depend on the provider itself.

另一方面，这需要多花一点精力，因为你需要更多地拥有它。

On the other side, well, it takes a little more effort because you will need to own it more.

你需要学习如何自定义它。

You will need to learn about how to customize it.

你需要使用合适的工具来做到这一点。

You will need to use the right tools for doing so.

如果你选择在自己的设施上部署，你还需要维护其部署。

You will need to maintain its deployment if you choose to deploy it on your own facilities.

但最终，这为你对抗闭源提供商创造了所需的杠杆。

But at the end, this is creating the leverage you need against closed source providers.

所以这两个类别本质上是不同的，但如果你只看纯粹的性能方面，它们确实在趋同。

So the two categories are effectively different, but if you look at the pure performance side, they are definitely converging.

你提到存在一种饱和效应。

You mentioned that there's a saturation effect.

所以，不谈太技术性的内容，这些模型是不是已经差不多到头了，不再进步了？

So without getting too technical, are the models sort of done with getting better?

让我这么问吧。

Let me put it this way.

考虑到所有模型似乎都遇到了瓶颈，AI模型还会继续进步吗？

Are AI models going to continue to get better given the fact that they all seem to be hitting saturation?

它们将在越来越多的特定领域变得更好。

They will get better in more and more specific domains.

在这方面，我认为我们 collectively 已经让它们变得非常聪明，能够处理长上下文并调用多个工具，但如果你想要将它们真正投入银行或制造公司的生产环境，这些模型就需要学习公司自身所包含的所有知识。

In that, I think we've really collectively made them very clever and able to read them about long context and able to call multiple tools, But if you go and want to effectively put them into production in a bank or in a manufacturing company, well, the models need to learn about all of the knowledge that is contained into the companies themselves.

因此，这实际上意味着，对于非常精确的方向，比如我想让我的模型在发现新材料方面极其出色，或者在设计飞机方面极其出色，我就需要付出更多努力，获得正确的反馈信号，找到合适的专家，请他们帮助我让模型在这一特定方向上表现得特别出色。

And so what it effectively means is that for very precise directions, let's say I want to make my model extremely good at discovering materials, or extremely good at designing planes, I will need to go and sweat it a little bit, and get the right reward signal, and get the right experts and ask them to make my model specifically good in that very precise direction.

因此，我们绝对还没有完成这项工作，因为我们所有人竞相追求的是为特定能力提供正确的环境和正确的信号源。

And so we are definitely not done doing that because what we are all racing for is the right environment and the right signal provider for specific capabilities.

但就广泛的横向推理能力而言，我们仍会继续提升它们，不过没有人会以造成与竞争对手显著差距的方式去改进它们。

But the broad horizontal reasoning capabilities, we're still going to improve them, but nobody is going to improve them in a way that is creating a strong gap versus its competitors.

因此，真正的显著差距在于与垂直领域专家合作，这些专家确切知道如何设计飞机，并能向模型解释如何做到这一点。

So the strong gap is actually in working with vertical experts that know exactly how they design a plane, and that actually explain to the model how to do it.

你可以选择众多方向，因为你可以从物理学、化学、制药、生物学等领域入手。

And you have a wealth of directions that you can take, because you can do it in physics, you can do it in chemistry, pharmaceutical, in biology.

因此，在我看来，未来两年最令人兴奋的部分将是模型在众多精确方向上取得突破性进步的爆发。

And so to me, the most exciting part of what's going to happen in the next two years is that explosion of very precise directions in which the model are going to get better.

对我们而言，机遇在于打造一个合适的平台，以支持这类垂直化发展，无论是面向企业，还是那些专注于高度专业化能力的AI初创公司，我们都乐于提供帮助。

And for us, the opportunity is to have the right platform for enabling those kinds of verticalization, whether with enterprises or you have AI startups actually that are working on very verticalized capabilities, and we're happy to help them as well.

这就是我对该领域未来走向的看法。

So that's my view of where the field is going to go.

我们一直致力于横向智能的发展，让系统变得越来越聪明，而未来两年将是把模型打造成在特定技能集上极为出色的时代。

We have been about horizontal intelligence growing and things getting more and more clever, and the next two years is going to be about taking model and making them extremely good at a certain skill set.

这实际上更令人兴奋，因为我们正达到这样一个阶段：只要你选择一个领域，就能让它变得超人般强大。

And that's actually more exciting, because we are getting to a point where if you pick a domain, you can just make it superhuman.

但我们不会同时在每个领域都让它变得超人般强大。

But we are not going to make it superhuman in every domain at the same time.

好的，但之前在我们的对话中，你提到你们不会拥有一个能做所有事情的模型。

Okay, but then on that note earlier in our conversation, you said that you're not going to have a model that can do everything.

但如果这些训练在某些垂直领域完成了，为什么不行呢？

But if that training gets done in certain verticals, why not?

事实上，我们正达到这样一个阶段：你所选择的垂直领域并不会真正迁移到其他领域。

Well, we are also getting to a point where the verticals that you choose do not really transfer to the others.

因此，没有必要去打造一个既精通精确生物学又精通精确物理学的模型，因为这两者之间的迁移效果实际上非常不明确。

So there's no point in making a model that is good at very precise biology and very precise physics, because the transfer between those things is actually pretty unclear.

问题是，如果你真的希望你的模型能同时解决所有问题，那它会变得非常庞大、昂贵，且服务成本极高。

The problem is that if you actually want your model to be able to solve every problem at the same time, you're making it very big, very expensive, and very costly to serve.

所以专门化的模型就是你要为生物学专门做一个，为化学专门做一个，为这个特定的物理问题再做一个。

So specialized models is really you're gonna specialize one for bio, one for chemistry, one for this particular physics problem.

实际上这样更合理，因为如果你想大规模运行它，想让它在后台运行，想让它昼夜不停地思考特定问题，你就希望它尽可能小，因为模型的成本实际上与其规模成正比。

Well, it actually makes more sense because if you wanna run it at scale, if you want it to run on the background, if you want it to run day and night thinking about specific problems, you well, want it to be as small as possible because the cost of a model is actually proportional to its size.

如果你为了使模型在多个领域都表现优异而扩大其规模，那么当你希望尽可能广泛地部署和使用它时，这实际上并不高效。

And if you inflate the size by making the model great at multiple modes, well, you're actually not very efficient if you want to deploy it and use it as much as possible.

所以，从经济角度来看，在某些方向上开发专用模型是有道理的。

So if you look at the economies of it, it does make sense to make a specialized model in certain directions.

让我问你一点关于Mistral的事。

Let me ask you a little bit about the Mr.

所有竞争领域。

All competitive area.

我想在美国，我来告诉你美国人怎么说，让你来回应一下，因为这值得讨论。

I think that here in The US, I'll just tell you what people in The US say and let you address it because it's worth talking about.

我认为，有些人——不是所有人——有一种感觉，认为Mistral在欧洲设立，实际上是利用了监管俘获，因为美国公司很难在欧洲竞争，因此Mistral将接管所有的AI业务。

I think there is a feeling among some, not all, but some that Mysteral has been set up in Europe to effectively take advantage of regulatory capture because US companies have a hard time competing in Europe, and therefore Mistral will be there to pick up all the AI business.

你对这个观点怎么看？

What do you think about that argument?