20VC：Nebius 联合创始人谈 AI 基础设施泡沫 | 开源对 OpenAI 和 Anthropic 的真实影响 | 算力需求的价格弹性如何 | 若拥有更多算力，Nebius 能否将算力销量提升 10 倍？以及更多与 Roman Chernin 的对话

顺便说一下，他们对待封闭生态的方式，比如Anthropic和OpenAI那样，也是类似的。

By the way, the same way like they do with closed ecosystems, like Anthropic's OpenAI's.

你提到了‘差异化’这个词。

You mentioned the word differentiation.

我之前和合伙人讨论过一件事，关于我们在这几层面上的定位，正好想和你聊聊。

And one thing that I was discussing with my partner before that we have is a thing we have to discuss, and it's within these layers.

你已经详细谈过产品开发，以及在能力基础上构建产品的关键性。

But you've spoken extensively about product build out and the importance of building the product underneath capacity.

当人们拿你们和其他新兴云服务商比较时，比如和CoreWeave对比，你们都用GPU，都有英伟达的合作关系，也都服务像Matter这样的客户。

When people look at you versus other neo clouds, we look at you versus a core weave, you both run GPUs, you both have NVIDIA relationships, you both have matter as a customer.

那你们的区别到底在哪里？

What's the difference?

我一般不拿自己和别人比较，但我们坚持的原则是全栈式架构。

I don't, like, compare with others, but the principles we build are full stack.

我们称之为全栈整合。

We call it full stack integration.

你可以把它理解为从底层到顶层的全栈布局。

And you you can think about it like full stack down and full stack up.

全栈向下，就是我们深入到了物理世界。

Full stack down is we are really deep in physical world.

我们自己建数据中心、机架和服务器，也打造整个平台。

We build data centers, we build racks and servers, we build the platform.

一旦掌握了这些下游环节，就能跑得更快，降低成本，为客户提供更经济的解决方案。

And when you control these things downstream, you can move faster and you can squeeze more cost and provide more economically viable solutions for the customers.

而上游的垂直整合，其实就是我们之前提到的产品层面——如何紧跟客户的需求和细分市场，不再局限于只服务那些单纯需要基础设施的人，而是真正走进企业与产品公司，到他们需要的地方去提供支持。

And then your vertical integration upstream is actually what we spoke about, like product and how can you follow the customer's needs and customer segments and not be limited by the small population of the people that just need infrastructure, but really serve kind of enterprises and product companies with, like meet them where they need us.

我觉得，这正是我们的不同之处。

And this is like, I think what we different.

具体表现就是，我们的业务分布更分散，客户群体也更多元。

And then how it showing up, would say, is again, less concentration in the business, more diversified customer portfolio.

从长远来看，这种布局让我们在进军企业市场时更具优势，因为未来大部分需求最终都会来自这里。

We believe long term better positioning for going to enterprises where we believe eventually a lot of demand will come from.

现在，我们的大部分客户群体都已经在使用我们的服务了。

Again, now most of our segment is working.

主要是做AI的团队在和做AI的团队合作。

It's AI natives working with AI natives.

但我们面对的市场里，还有大量传统企业，他们也需要服务。

But we have a huge market of enterprises, existing companies, and someone needs to serve them.

他们不会直接买原始的计算资源。

They will not buy raw compute.

他们需要的是平台。

They will need platforms.

他们需要的是工具。

They will need tools.

他们希望我们能尊重他们的既有系统，适应他们更复杂的环境。

They will need us to respect their legacy and being able to work with their more complex environment.

他们没那么灵活。

They are not nimble.

他们需要迁移数据。

They have data to migrate.

他们需要整合现有系统。

They have systems to integrate.

这才是关键所在。

And that's the big game.

我觉得这对我们来说，是主要的发展方向。

And I think that for us, it's kind of the main direction to move.

你提到四层架构中的第三层是托管推理。

You mentioned the third layer of the four pillared stack being managed inference.

对于不了解的人，你是怎么理解这一层的？会怎么向他们解释？

For people that don't understand, how do you think about this layer, and how would you explain it to them?

是的。

Yeah.

很简单。

Very simple.

你当初是随便选了个平台，就拿它来开发你的产品。

You built your product on whatever, call your where you wipe code.

我其实是OpenAI和代码方面的专家。

I'm actually an OpenAI and a code expert.

明白了。

Okay.

够用了。

Good enough.

你用OpenAI做出了很棒的产品。

You built your great products with OpenAI.

你找到了关键的应用场景，开始快速增长，用户增长势头非常好。

You cracked the use case, and you started growing and you have amazing traction.

唯一的问题是，你的利润空间可能不够，或者你想更主动地利用数据、调整模型的行为，但在封闭生态里你做不到。

The only problem may be you don't have enough margin or you want to start applying more aggressively the data and tune the the behavior of the model, you cannot do it in the closed ecosystem.

于是你上网去查，发现有很多优秀的开源模型，在性能测试上已经接近OpenAI了。

So you go to Internet and you read, there are a lot of great open source models that on the benchmarks are closed to OpenAI.

你心想，太好了，推理成本能便宜十倍。

And you think, oh, great, it will be 10 times cheaper, inference is cheaper.

我可以调整这些模型。

I can tune those models.

我可以用自己的数据，让产品变得更好。

I can apply my data, and my product will be better.

我的增长速度也会加快。

My growth will accelerate.

于是你去Hugging Face下载模型权重。

So you go, you you take the weights from Hugging Face.

再找一个运行引擎，比如VLLM或者SGLANG之类的，结果根本跑不起来，因为你根本没法兑现预期的性能。

You take some engine to run it, like VLLM, SGLANK, something, and then it doesn't work because you need to to really extract the value you expect.

你得做各种优化。

You need to do, like, optimizations.

还得正确地部署它。

You need to deploy it in the proper way.

你不能只靠一块GPU做令牌提取，或简单调整一台服务器的配置，你做的是大规模产品。

You need not just one GPU tokens extraction or one host setting, but you you are the large product.

你已经在使用数十万张GPU了。

You run on hundreds of thousands GPUs already.

你需要完整的调度系统。

You need all the orchestration.

你需要缓存机制。

You need the caching.

你需要可观测性监控。

You need the observability.

你的客户会问你，这东西到底是怎么运行的？

Like, your customers ask you, like, how does it work?

诸如此类的问题还有很多。

And so on and so forth.

顺便说一句，你之前在OpenAI那边早就有了这些全套方案，因为那本来就是你的生产级服务。

And by the way, you had all of that on OpenAI because this is, the production service for you.

你用 OpenAI 的时候，根本不会去想底层基础设施的事。

You don't think about infrastructure when you work with OpenAI.

只需要选个合适的套餐，用多少付多少就行。

Just subscribe for the plan you need, and you pay for whatever.

结果就是这么简单。

End result.

所以这时候，你就需要 Token Factory 这样的产品。

And so that's where you need the product like Token Factory.

Token Factory 能帮你托管推理服务，支持开源模型或专用模型。

Token Factory gives you the managed inference with the open source or specialized models.

你可以直接用现有的开源模型，也可以自己微调并部署专属的权重。

You can run existing open source, vanilla open source model, or you can tune the model and deploy your own, like, weights.

剩下的事，我们来搞定。

And then we'll take care about all the rest.

我们会把所有优化技术都给你安排上。

We'll apply all the optimization techniques.

我们会帮你优化成本结构。

We'll manage the better economics for you.

系统会非常稳定。

It will be reliable.

你不用操心接下来一百张显卡该从哪儿找。

You don't need to think about the next 100 GPUs where you will find them.

这就像是托管服务。

It's like managed service.

使用 Token Factory，你可以运行六十种开源模型。

With Token Factory, you run on 60 open source models.

你之前提到过，通过优化能把推理成本降低最多百分之七十。

And you said before about cutting inference cost by up to 70% through optimization.

我能问个傻问题吗？你们到底是怎么让每个令牌变得更便宜的？

Can I ask a dumb question, which is how do you actually make a token cheaper?

是的。

Yeah.

这可不是什么魔法把戏。

So it's not a magic game.

你拿一个基础模型，然后根据你具体的使用场景进行优化。

You take the model, the like, some baseline model, and then you can optimize it for particular scenarios that you have.

你可以对模型进行蒸馏。

So you can actually, you can distill the model.

打造一个更小、但效果同样好的模型。

You can make the like, the the smaller model that works with the same quality.

你可以用推测解码。

You can do spec decoding.

还可以优化缓存之类的各种方法。

You can optimize caching and so on and forth.

你拿一个模型，然后基于它构建一套系统，专门适配你的需求，实现更优的成本效益。

So you you take the model, and out of this model, you actually build the system that in your particular case works with your your requirements with optimized economics.

顺便说一句，对于客户来说，使用像 Token Factory 这样的托管平台，还有一个重要原因是：模型本身一直在更新。

And by the way, one of the things that also I think important for customers to use managed platforms like Token Factory, the models are changing.

每周、每月，甚至今天，MiniMax 3 刚刚发布，Nebatron Ultra 也刚宣布了。

Every week, every month, like right today, MiniMax three was released, and there is Nebatron Ultra that was announced.

这种事每隔几周就会发生一次。

This happens every few weeks.

每次新模型发布，可能在某些基准测试上表现更好，但在其他测试上就不一定了，诸如此类。

And every time the new model released, it may work better on some benchmarks and maybe not on other benchmarks and so on.

你得有灵活性。

And you want to have flexibility.

你希望有人能帮你做实验，每次有新的最佳模型上线，都能帮你快速接入。

You want someone to support you on experimenting and actually adopting the new best models for your use case every time they come online.

像我们这样的平台，会帮你把换模型、做基准测试这些麻烦事都屏蔽掉。

And then like the platforms like ours, abstracts from you all the work that you need to do to actually change from one model to another, to benchmark all of them and so on.

你可以放心，永远站在技术最前沿。

You can be sure that you will be on the frontier.

只要一有新东西出来，平台立马就能用上。

Every time something new is happening, it will be in the platform.

你可以去试用一下。

You will be able to test it.

如果发现它更适合你的场景，你可以轻松切换，整个过程对你来说都会很顺畅、无感。

You if it works better for your use case, you will be able to switch, and it all will be kind of smooth and transparent for you.

模型发展的速度能跟得上吗？

Does the pace of model development sustain?

你说的是每几周一次，我倒是觉得， respectfully 地说，应该是每几天就有一回。

Like you said that, I would argue respectfully, you said every couple of weeks.

我觉得是每几天就出一个新版本。

I'd say every couple of days.

这种速度能持续得下去吗？

Does that sustain?

五年后，我们还能看到这么快的迭代节奏吗？

In five years' time, are we seeing that level of iteration?

这个嘛，我也不确定。

Well, I don't know.

我觉得很有希望看到更多垂直领域的模型不断涌现并得到改进。

I think that it's a good chances that we will continue to see a lot of niche models show up and improved.

我不确定。

I don't know.

我坚信，我们离极限还远得很，未来还会看到大量模型持续优化。

Again, I'm a believer that we're quite far from the wall and we will see a lot of models improvement happening.

我觉得我们还看到了越来越多新模态和专用模型的出现。

I think that what we also see is much more new modalities and specialized models coming in game.

我们总在谈前沿大模型，但还有整个生命科学、机器人、世界模型、视频模型、图像模型的世界。

So we speak about this frontier LLMs, but there is entire world of life science models, robotics, world models, video models, image models.

这些模型各自都有自己的应用场景。

They all have their own use cases as well.

我们正看到越来越多的小型专用模型，针对特定用途做了高度优化。

And we see more and more small specialized models, particular use cases coming with, like, very much optimized.

就在今天早上，我跟以色列这边一个团队聊了聊，他们正在开发一种网络安全基础模型，专门用来构建网络安全代理。

Just this morning, I spoke with a team here in Israel that develops cyber defense foundational model, like the model that optimized for to build cyber defense agents.

而且他们也不是从零开始的。

And again, they don't start from the scratch.

他们会拿一些开源的基础模型，再针对特定场景做训练，优化质量和延迟，比如在网络安全这种应用里。

They take some of the open source foundational model, but then they they train it for the particular case, optimized for the quality and the latency that needed in this, like, cyber defense use cases.

我觉得这种趋势还会继续下去。

And I think we'll continue to see it.

我们会看到越来越多专门训练的模型，但这些模型还需要优化推理过程，以及配套的基础设施，才能让客户顺利使用。

We'll see a lot of specialized, both both trained models that still need optimized inference and optimized, like, infrastructure around them to let customer use them.

关于令牌成本和使用量，你注意到哪些别人还没怎么谈的事？

On token costs and token usage, what are you seeing that you don't think other people are talking about enough?

最近有什么事让你特别惊讶吗？

What has shocked you recently?

我觉得大家说的其实都差不多，就是增长速度太快了。

Again, I I think everybody is speaking the same thing, like how fast it's growing.

看到像Anthropic、Cortex、Cognition这些公司的发展轨迹已经够惊人了，现在我们又在医疗、金融这些领域看到类似的应用，真的挺不可思议的。

When we see these trajectories of some companies like Anthropic and Corsor and Cognition and coding, and now we see start seeing in other verticals as well, some health care examples, financial use cases, I think I think it's, like, quite amazing.

有趣的是，看看那些非AI初创公司是怎么行动的。

What's interesting is to see how, like, non AI startups are moving.

我可以举个例子。

So we I I can give an example.

我们有个客户是Revolut。

We we have the customer of Revolut.

刚开始合作的时候，他们99%的推理预算都花在了闭源模型上，主要是OpenAI。

And when we started working with them, I think 99% of their budget, inference budget, was in closed models, in OpenAI.

后来他们尝试了一些应用场景，但发现有些在经济上根本行不通。

And they started to crack some of the use cases, and some of them didn't work for them economically.

结果就是，他们没法用AI替代人工，也没法有效提升人工效率。

So they practically couldn't replace the humans or cannot enhance the humans in the use cases they wanted to address.

于是他们开始转向开源模型，但进展没那么快，因为他们得从头搭建整个内部系统。

And they started moving to open source models, but it didn't move fast for them because they had to spend time on building the entire engine internally in the company.

首先，他们花了很多时间在评估上。

And first of all, they were focusing on evaluations.

我认为很多人低估了这一点：当你清楚地知道团队和公司哪些地方做得好时，建立一套用于改进和实验的基础体系有多重要。

And I think this is something that people underestimate how important to build kind of the foundation for improvements and experimentation engine when you understand as a company, as a team, what is good you.

因为，再怎么说，某个应用场景你搞定了，确实能跑通。

Because, again, like, you close some use case, it works.

但接下来你想换模型，该怎么办？

But then you want to change the model.

你怎么确定不会把质量搞砸了？

How do you know you you don't ruin the quality?

你得有明确的指标。

You need to have, like, metrics.

你得有个类似阀门的控制机制。

You need to have a valve mechanism.

AI开发也得有成熟的持续集成和持续交付流程。

You have you need to have this CICD process established for AI development.

我们看到很多客户，像Revolut这样的公司，都是这样。

And I think that what we see a lot of customers are like Revolut.

他们很早就投入资源，搞清楚了如何迭代模型，以及如何安全地把模型集成到生产流程中。

They have these foundational investments that need to do in the understanding of how to evolve the models, how to actually safely integrate them in their production processes.

但一旦解决了这些基础问题，他们的增长就会呈指数级爆发。

But when they solved these foundational problems, they start growing exponentially.

你千万别小看，一旦建立起能快速交付的系统，这些客户的发展速度有多快。

And I wouldn't underestimate how fast those customers can grow when they build the system that led them ship fast.

所谓快速交付，就是他们懂得如何迭代，也知道怎么做决策。

And ship fast means they know how to evolve, they know how to make decisions.

这种现象在很多客户身上都能看到。

And this is something that we see across a lot of customers.

他们一开始都会遇到这个难题——不管是叫基础投入，还是冷启动问题，就是怎么才能真正把产品推出来。

They have this, you can call it foundational investments or cold start problem, how to start shipping.

只要跨过这道坎，增长就会立刻加速。

When they solve it, they start to grow exponentially.

而且他们还能灵活切换不同的模型。

And they can use different models.

他们可以在公司内部开发出更多产品，诸如此类。

They can build much more products inside the company and so on and so forth.

我觉得，从外部看，你可能会觉得他们没什么增长。

And I think this is this is something that when you look from outside, you kind of, oh, they are not growing.

他们起步很慢。

They start small.

需要花时间，等等。

They take time and so on.

但如果公司团队强大，建立了这套基础体系，之后就会迎来爆发式增长。

But this is if the company has a strong team, they build this foundation and then they start growing exponentially.

我觉得，像Revolut、Shopify、Prosus、Booking.com这类云原生企业，未来会看到大量爆发式增长。

And I think we'll see a lot of explosive growth in enterprises, the digital, like in the cloud native companies like Revolut, Shopify, Prosus, booking.com.

一旦解决了起步难题，建立起快速交付的体系，他们在AI应用上的增速就会快得惊人。

When they solve this cold start problem, they build the system how to ship, and then they will grow, like, in their AI adoption like crazy.

你觉得三年后，Revolut会给你多付多少钱？

How much more do you think Revolut will pay you in three years' time?

我不知道。

I don't know.

我真的不知道。

I don't know.

不行。

No.

我不想谈这个。

I don't wanna speak about it.

不行。

No.

但我可以这么说，他们整体的增长模式大概是这样的。

But I I can say that they in total, I think they, like, they grow like this.

我们都看到这些科技公司报告的收入在持续增长。

We all see the CI companies reporting IRR growth.

对吧？

Right?

对他们来说，根本不是什么IRR。

For them, it's not IRR.

而是他们的预算。

It's like their budget.

但我认为最前沿的公司，他们的AI预算根本不是那种虚张声势的竞赛。

But I think that the most advanced companies, their AI budget, and it's not like this fake all this stocky maxing race.

我们看到的是，他们怎么在实际生产中应用的。

We see it, like, how they do it in the production workload.

所以他们的增长节奏和那些AI原生公司一样，AI消耗量的增长和IRR持平。

So they they grow the same pace like this AI native companies reporting they are growing their AI consumption equal to their IRR.

像Revolut这样的公司，增长轨迹也是呈指数级上升的。

So the companies like Revolut, they're growing the same exponential trajectory.

我总是会反驳那些说开源会对大型模型提供商构成真正威胁的人，我说，听好了，大企业最看重的是稳定性。

I always push back on people who claim that open source would be a credible threat to the largest model providers because I said, listen, the biggest enterprises want reliability.

他们需要的是安全性。

They want security.

最重要的是，他们想要简单易用。

And most of all, want ease.

他们不想整天去折腾底层那些复杂的架构和乱七八糟的东西。

They don't wanna be tinkering around with all the architecture and shit beneath the surface.

你的意思是，你能做到这一切，让他们摆脱那些供应商，获得更便宜、更好的体验，因为你把那些繁琐的底层工作都给解决了。

What you're telling me is you're able to be all of that to allow them to pipe away from those providers and have a cheaper, better experience because you take away the plumbing.

对。

Yes.

但我想说的是，闭源模型和开源模型之间的区别，根本不是可靠不可靠的问题。

But my point is closed models with open source models, it's not about, like, reliable or not reliable.

像Nebius这样的公司，正是为了让用户用其他模型时，根本不用操心底层那些事，才做了这些工作。

Again, the work of the companies like Nebius to make possible, like, as you say, not think about plumbing if you want to use alternative models.

但我觉得关键在于能力。

But I think it's about capabilities.

我觉得像前沿模型这样的闭源模型非常出色，而且还会变得更好。

I think that closed source models like frontier models are great, and they will become even better.

它们将解决许多我们目前还解决不了的问题。

And they will solve so many problems that we don't solve yet.

我们要解决的使用场景多种多样，因此市场上会同时存在最智能的模型、最快的模型，以及那些够聪明又够便宜的中间型号。

And we have such a diversity of the use cases we wanna solve that there will be market for the smartest models of the world, the fastest models of the world, the in between models of the world, smart enough but cheap enough.

作为用户，你可以为每个具体任务自由选择最适合的模型服务。

And you as a customer will be able to just pick the right the right source of token for each particular task.

回到智能代理层这一点，也许最终决定该调用哪个模型，本身就该由用户来决定。

And back to the agentic layer point, maybe it's even want to be the customer task to choose the which model to call now.

引擎会掌握所有底层模型的能力，了如指掌。

It will be the engine that knows all the capabilities, like all the models underneath.

当你去研究OpenAI的时候，你根本不会去想它要循环多少次。

And then when you go to OpenAI and you do the research, you don't think in terms of how many loops you want it to make.

你也不会去纠结什么时候该用大模型，什么时候该用搜索。

You don't think in terms when it should go to LLM and when it should go to search.

你根本不用操心现在该调用哪个提示词。

You don't think should it now call which prompt to call.

对吧？

Right?

这事儿已经在发生了。

It's it's happening.

你给个任务，有个推理引擎负责决定怎么执行，然后你就拿到结果了。

You give a task, there is an engine, the reasoning engine that decides how to run this task and you got the result.

我觉得很多企业场景、那些特别复杂的任务，都会用同样的方式解决。

So I think that a lot of enterprise cases, a lot of these gigantic tasks will be solved in the same way.

当不再是你这个开发者去直接关注客户需求时，我们就得去协调这些模型和调用资源。

When it's not you as a developer that focusing on customer need, we'll need to kind of orchestrate all these tokens and models.

我们需要各种模型——最聪明的处理最复杂的任务，快速的模型负责快速迭代。

And then we will need all the models, the smartest one for the most complex intelligence and the fast models that can do quick iterations.

我们甚至还没谈各种模态，还有物理世界里AI需要的东西等等。

And, we don't speak even about all the modalities and, like, what we'll need in the physical AI world and so on.

所以，我再说一遍，不同模型都有足够的空间。

So I think, again, my point, we will have enough of pie for different models.

作为一家基础设施公司，我们需要做的就是尽可能帮助开发者更轻松地使用这些模型提供的各种能力。

And what we need to do as a infrastructure company is just help for extent we can to make developers comfortable how they use all these capabilities that models provides.

正如你所说，问题根本不在模型本身的能力。

Because as you as you rightly said, it's not about model capabilities.

这不仅仅是模型能力的问题。

It's not only about model capabilities.

关键是别让开发者陷在底层 plumbing 里，得让他们能顺利运行、优化并稳定使用这些系统。

It's about, like, not plumbing, getting them working, getting them optimized, getting them reliable.

当我们看到模型数量激增、分工越来越细，以及不同应用场景下会涌现出多少新模型时，很明显的一点是，欧洲在模型建设上的投入远不如美国和中国。

When we look at the explosion of models and the specialization of models, like you said there, and how many will be built and the depth across different use cases, sadly, the one thing that is quite clear is that Europe does not have anywhere near the model build out that we've seen both in The US and in China.

你认为各国拥有自己的主权模型有多重要？

How important do you think it is that nations have their own sovereign models?

这是个好问题。

It's a good question.

但看起来，世界已经分成了不同的阵营。

But looks like the world is divided.

我们可能不喜欢这样。

We may not like it.

让世界上大部分地区都能用上足够好的基础模型，这一点很重要。

Having good enough foundational models available for the big parts of the world is important.

我觉得在欧洲，我们得好好想想，怎么才能在这里拥有足够的能力。

And I think here in Europe, we should think how we have enough capabilities available here.

过去几年里，我们谈了很多关于主权之类的话题，整个主权AI的议程。

We had a lot of conversations over the last couple of years in, like, about the sovereignty and so on, all of this, like, so Sovereign AI agenda.

但那时候讨论得太集中在电力和能耗上了，反而忽略了开发者这一层的东西。

And I think it was too much concentrated around, like, megawatt and power rather than on what we have on the builders layer.

对吧？

Right?

电力供应迟早会跟上的。

Megawatts will come.

我觉得，这正是我们在Nebius一直强调的：我们要做基础设施。

I think that it's what what we in Nebius always told is we will build infrastructure.

像我们这样的公司，只要有需求就会去搭建基础设施，而需求来自那些开发者。

The companies like us will build infrastructure if we have demand, and demand is coming from the builders.

我们真正该关注的是，多培养一些像 Lovables、Black Forest Labs、Mistras 这样的优秀公司。

What we need to care about here is to have more great companies like Lovables, Black Forest Labs, Mistras of the world.

只要我们有足够的资金投入研究，有足够的资金支持产品开发，需求自然就会起来，飞轮效应也会重新启动，到时候就能打造出足够好的模型。

And we have enough people that invest in research, have enough people that invest in the products, and then they will create enough of demand, and there will be enough of flywheel again to have a good enough models if we need.

所以我觉得，这才是我们真正该重视的事。

So I think this is something that, like, we should care about.

今天最值得投资的领域是什么？

What is the most interesting area to invest today?

明白了。

Okay.

我给你四个选项。

I'm giving you four options.

基础设施、通用模型、垂直模型、应用层。

Infrastructure, horizontal model, vertical model, application layer.

我们确实建了基础设施，所以现在挺满意的。

I mean, we built infrastructure, so we are quite happy here.

我觉得在当今这个环境下，我们处于一个不错的位置，尽管某种程度上，我们做的其实是相对容易的部分。

I think it's a good place to be in the current world, even though we we for some extent, we are building kind of the easiest part.

不是执行有多复杂，而是我们大概知道需要什么，客户也会帮我们明确需求。

Not it's complex execution, but we kinda know what's needed, and our customers help us to understand what's needed.

在我看来，这个行业里最了不起的人，是那些敢于去打造终端产品的人。

I think the most amazing people in this industry are those who take a risk to go and build end user products, in my view.

他们才是真正推动增长的核心力量。

They actually drive the most of growth here.

敢于冒险去打造用户可能需要、也可能不需要的产品的人，我认为才是我们AI旅程中最了不起的英雄。

People who take a risk, like the real risk of building something people would need or not need, I think this is the most the heroes of our AI journey.

说到AI旅程中的英雄，我每次录节目前，都会去和一些人聊聊，我现在挺幸运的。

Speaking of heroes of AI journeys, before I do a show, I go and speak to I'm very fortunate now.

你之前提到过，我确实采访过一些重量级人物。

You mentioned earlier, I've interviewed some some big people.

我会去和这些大人物聊聊。

I go and speak to some of those big people.

我和他们交流时，一个反复提到的主题就是和英伟达的关系。

A theme that did come up when I was speaking to them was the relationship with NVIDIA.

如果一方掌握着绝对主导权，这还能叫合作关系吗？

And is a marriage a marriage if one has more power than the other?

当英伟达拥有如此大的话语权时，你怎么看待这种权力关系？

How do you think about the the power dynamics in a relationship with NVIDIA when they have so much power?

谁能在英伟达面前谈权力？

Who has power against NVIDIA?

我们看这个问题其实很简单。

We look at this in a very simple manner.

我们只需要专注做好自己该做的事。

We just need to build what we build.

我们得把产品做出来。

We need to build our product.

我们需要讲好自己的故事，其他的一切自然会跟上来。

We need to tell our story, and then the rest will complement it.

我觉得NVIDIA最吸引人的一点是，到目前为止，它依然是一家由工程师主导的公司。

I think what is the most fascinating NVIDIA is still, for big extent, is an engineers driven company.

我觉得要想赢得NVIDIA的尊重，最好的办法就是——

And I think the best thing you can do to get respect from NVIDIA it's my read.

他们可能有不同的看法。

They may have a different point of view.

但如果NVIDIA的工程师认可你的团队，那你们的关系就有了坚实的基础。

But if engineers in NVIDIA respect your engineers, you will have the right foundation for relations, let's say.

我们一再证明了，我们清楚自己在做什么，而且拥有强大的工程团队。

We managed to prove again and again, we know what we build, and we have a strong engineering team.

我觉得他们看在眼里，也确实尊重这一点。

And I think that they see it and they respect it.

我们在硬件、软件层，还有推理平台层，都有大量工程师之间的直接合作。

And we have a lot of, like, engineers to engineers relations on physical like, on a on a hardware level, on the software layer, on the inference platform layer.

NVIDIA的工程师越认可你，合作关系就越容易建立起来。

And the better engineers than NVIDIA think about you, the better relations and partnership I think it enables.

我们可能这么想是错的，但这就是我们实际看到的、能做到的。

And, and, again, we may be we may be wrong thinking this way, but that that's what we see, like, we can do.

我们只专注于讲道理，着眼于长期价值。

And we just focus on being reasonable and being kind of focused on the long term value.

听起来有点虚。

It sounds like fluffy.

谁都这么说。

Everybody say it.

但归根结底，你得把活儿干好。

But do your fucking job at the at the end of the day.

对吧？

Right?

我打算给这个起个名字，叫罗马。

I'm gonna I'm gonna title this, Roman.

好好干你的活儿。

Just do your fucking job.

不对。

No.

我们还能做点什么？

What what else we could do?

我的意思是，我们现在正处于一场激烈的竞争中。

I mean, we are in such a race.

我们只能尽力把工作做得更好。

We just can do we we can do the best to do our work better.

好好干你的活儿。

Do your fucking job.

乔，我知道这听起来挺搞笑的。

I Joe, so I know it's funny.

我喜欢这句话。

I I like it.

但话说回来，今天真正最难的部分到底是什么？就是好好把活儿干好。

But, like, what's the hardest part of just doing your fucking job today?

四个方面：扩展规模、开发产品、服务客户。

Four dimensions, build scale, build product, work with customers.

其实我们之前聊过的主要是两个方面：规模和产品。

It's actually like two dimensions we discussed, like scale and product.

客户这一块，才是我们得认真想的。

The thought is customers.

我们做的是现场服务的生意。

We are in the field business.

我们常说，云计算是售后的活儿。

We we like to say that cloud is post sales business.

你卖的时候，卖的只是一个承诺。

When you sell, you sell the promise.

卖完之后，你得让客户真正满意。

And then the customer, you need to satisfy the customer.

和客户打交道，服务客户，拥有一个强大的客户-facing工程团队——也就是FDE团队，这才是第三个维度。

And working with the customers, covering the customers, having this strong customer facing engineering team, FDE team, this is the third dimension.

去和你的客户聊聊吧。

Go talk to your customers.

让他们认识你，你也得了解他们。

Make sure that they know you, that you know them.

这就是第三个维度。

This is the third dimension.

第四个维度，最枯燥但也最让人兴奋的，就是资金。

And the fourth, the most boring, but also the most exciting is the capital.

我们玩的是资本密集型的游戏，对手都是世界上资金最雄厚的公司。

We are in the capital intensive game, and we're competing with the most capitalized companies in the world.

如果你给我无限的预算，你会怎么做不一样？

If if I gave you unlimited budget, what would you do differently?

更快地推进。

Build faster.

这很简单。

That's very easy.

建什么更快？

Build what faster?

建数据中心，再用GPU把它们填满，就是建得更快一点。

Data centers and fulfill them with GPUs, like just build faster.

我们今年的资本支出是2.025万亿美元。

Our CapEx program this year is $2,025,000,000,000 dollars.

我们的竞争对手，那些超大规模云服务商，规模是我们八到十倍。

Our competitors' Hyperscalers have 10 times, like eight times bigger.

如果我有十倍的资金，我就会多建些数据中心，用GPU填满它们，更快地服务更多客户。

If I would have 10 times bigger capital, I would just build more data centers and fulfill them with GPUs faster and serve more customers.

我们当初就是这么起步的。

That's what we started with.

要是我有十倍的供应量，我会怎么做？

Like, what would I do if I had like 10 times more supply?

我会加快进度。

I would move faster.

加文·贝克曾很明智地指出，审批流程和监管以及数据中心建设的延迟，实际上反而有帮助。

Gavin Baker said, I think quite intelligently that permitting and regulation and the delayed build out of data centers has actually helped.

因为如果我现在让你建十倍的数据中心，反而会造成供过于求。

Because if I enabled you to build 10X the data centers today, it would actually create the glut.

是的，这确实是个好问题。

Yeah, it's it's actually a great question.

我们的投资者有时候会问我们，主要的瓶颈是什么？

And and like our investors sometimes ask us, like, what is the main bottleneck?

而主要的瓶颈，说到底，方方面面都是问题。

And the main bottleneck, again, it's everything.

你要从时间投入的角度来看这个问题。

You need to look at this from the time time spent perspective.

在未来六个月里，再多的资金也帮不上忙。

In the next six months, the capital cannot help.

六个月时间太短了。

Six months is too short time.

你现在有什么资源，就得先交付出来。

You have what you have, you need to deliver.

接下来十二个月，你可以加快一些进度，但更多是受制于产能瓶颈。

Then in the next twelve months, you can accelerate something, but it's more like capacity constraints.

在接下来的十二个月里，我们能靠资金或执行效率来提速。

And in the next twelve months, we can accelerate with the capital or with execution something.

但到了二十四个月，你肯定能解锁大量可能性。

But then in twenty four months, you definitely can unlock so many things.

我们不是在建一个数据中心。

We are not building one data center.

这一点也很重要，得明白。

It's also important to understand.

我们是在打造一套完整的产能组合。

We are building the portfolio, the portfolio of capacity.

我们执行力越强，资金越充足，就能并行推进更多事情。

The more execution power we have, the more capital we have, we can do the things in parallel.

这就是我们做事的方式。

That's why we do how we do.

我们先锁定电力和土地，再建数据中心，最后部署GPU。

We secure power and land, then we build data centers, then we fulfill them with GPUs.

每个阶段都需要更多资金，但我们尽量提前做好准备，确保进入下一阶段时，电力资源已经到位。

Every next stage requires more capital, but we do as much as possible in advance to make sure that when we will be on the next stage, we already have power secured.

等我们有足够的资金来部署GPU时，数据中心都已经建好并运行了。

When we will have enough capital to deploy in GPUs, we will have data centers that up and running.

这就像分阶段投入。

So it's like phases of investments.

而且，不同时间维度上的瓶颈也各不相同。

And again, the bottlenecks are different on the different time spent perspective.

显而易见，资金越多，推进速度就越快。

So obviously, if you have more capital, you can move faster.

不是在六个月内，但十八个月到二十四个月肯定是没问题的。

Not in the six months, but in the well, whatever, eighteen, twenty four months for sure.

我想问一下，当你在考虑数据中心建设时，有没有注意到公众对人工智能的抵触情绪越来越明显？

Can I ask you, when you think about the the the data center build out there, we're seeing more and more public angst towards AI?

埃里克·施密特上台被嘘下台，不是因为他说的内容，而是因为大家对AI的反感。

Eric Schmidt's getting booed off stage, not because of the content, but because of the AI inventions.

现在我们看到公众对数据中心的不满也在增加。

And we're seeing like public resentment towards data centers.

我觉得现在有四成的数据中心项目，卡在规划和审批环节，根本建不起来。

I think 40 out of a 100 now are not being built when they go through planning and approvals.

你们内部是怎么看待和应对这个问题的？

How do you think about and reflect on that internally?

这就是我们当前必须面对的环境。

This is the environment we need to work in.

事情有两面性。

There are two sides of the thing.

一方面，我们要从商业角度务实地看待这个问题。

One is how we think pragmatically as a business.

我刚才就是这个意思。

That's what I said.

我们会把项目看作一个投资组合来管理。

We think about it as a portfolio of the projects.

我们必须确保需求超过我们的供给能力。

We need to make sure that we are oversubscribed.

即使某个数据中心延期了，我们依然能为客户提供足够的算力。

And if one data center will be delayed, we will still deliver enough capacity to our customers.

大多数客户并不绑定在某一个具体地点。

And most of the customers, they are not locked in one physical location.

他们用的是云服务。

They just it's a cloud.

我们可以在不同地方建设，然后把工作负载调度到有余量的区域。

We can build in different places and then bring the workloads where we have capacity.

但这是现实层面的考量。

But this is the pragmatical side of the things.

很明显，社区和地方政府都希望我们这样的公司能和他们紧密合作，解释我们的工作，展示我们的做法，并认真回应他们的关切。

Then what we obviously see that communities and the local authorities require the companies like us to work closely with them and explain and show what we do and work with them on their concerns and like address them.

这就是现实。

So this is the reality.

你可以想想优步刚起步的时候。

You can compare it when Uber started growing.

在很多地方，都遇到了抵制。

And in many places, there was the pushback.

对吧？

Right?

那到底发生了什么？

So, oh, what's happening?

这毕竟是个新事物。

It's something new.

我们没想到事情进展得这么快。

We it's moving too fast.

原本没料到会这么快，事情发展得太急了。

We didn't didn't expect it to move so fast and so on.

你得主动去沟通、去解释，这本身就是你的责任——要和那些依赖你的新社区建立联系。

And you go and work and you explain, and it's just a part of your duty to engage and work with the new communities that become dependent on you.

他们确实有顾虑。

They have concerns.

有时候他们的顾虑是因为了解不够。

And sometimes they just they have concerns because they're not educated enough.

有时候他们的担忧是合理的，是可以解决的。

Sometimes they have rational concerns that you can address.

总之，该做什么就做什么。

And the same, do your job.

你觉得到目前为止，你们做得怎么样？

Do you think you've done a good job at it so far?

我们一直觉得自己做得还不够。

We come from the place where we we we always we always think that we didn't do enough.

在我们刚开始建设的地方，现在已经取得了不少进展。

I think that we've got quite a progress in the places where we started building.

过去，我们在欧洲的经验更丰富一些。

Historically, we had more experience in Europe.

现在，我们中期新建的产能中，大概有70%到75%都在美国。

We now probably 7075% of the new capacity that we built midterm is in The US.

所以我们在美国各地都建立了实地存在，积极与当地社区沟通。

So we built a lot of presence on the ground and in like, to communicate with those local communities in US.

我们在这里也尽力做到最好。

And we try to do the best job here.

我们总能做得更好，但至少我们一直在前进。

We need to do better always, but we are moving.

你能再帮我一下别的事吗？

Can you help me on another one?

我们刚才说到太空的时候，都笑出来了。

We laughed earlier when we said about space.

在地球上建数据中心，物流难度已经够大了。

Data centers on planet Earth is a very difficult logistical build out.

在太空建数据中心？

Data centers in space?

我特别喜欢科技。

I love technology.

我是个乐观派。

I'm an optimist.

这难道不疯狂吗？

Is that fucking nuts?

我觉得我们看到的每件事都疯狂得离谱。

I think everything we see is fucking nuts.

我的看法很简单。

My view is very simple.

现在有这么多聪明人在努力让这件事成真。

So many smart people now working to make it happen.

所以，我可能没那么悲观了，觉得我们迟早会看到一些意想不到的东西。

So most likely, I may be less pessimistic that we'll see I I don't know what is there.

三年内，我们在太空建的数据中心会比地球上还多。

We'll build more in space than in on the Earth in three years.

也许三年还不够。

Maybe not in three years.

我觉得自己够谦虚了，既然这么多聪明人都在拼命解决这个问题，要把算力送到太空，我有什么理由不信它会发生呢？

My view, I'm humble enough to say that so many smart people are trying to solve this task and bring compute to the space that why wouldn't I believe it will happen?

当然，还有很多挑战，很多问题等着我们去解决。

And I think there are a lot of challenges still, a lot of things to figure out.

但要是三年前有人跟我们说，将来会建成吉瓦级、大规模互联的计算集群，你会信吗？

But if someone would say us that even three years ago that we will build, like, multi gigawatt data centers, It will be, like, large interconnected compute clusters.

你信吗？

Would you believe?

我以前没这么想过。

I I didn't think like that.

我们现在就在这儿了。

We are here.

这已经成常态了。

It's it's routine.

我想跟你来个快速问答。

I wanna do a quick fire with you.

我先说一句简短的话。

So I say a short statement.

你马上说出你的第一反应。

You give me your immediate thoughts.

现在还没有、但五年后可能会很常见的工作是什么？

What job does not exist today that you think will be very common in five years' time?

太空数据中心的技术员？

Space data centers technicians?

不是。

No.

可能是机器人。

Probably it will be robots.

说正经的，现在明显在发生的一件事，就是我们正在让‘开发者’这个身份变得大众化。

So on a serious note, one thing that's obviously happening is we're democratizing what people called being developer.

对吧？

Right?

现在我们每个人都能成为开发者。

Now each of us can be a developer.

我所说的‘开发者’，指的是把想法变成数字产品。

And, like, what what I mean being developer is to convert the idea in some digital asset.

我希望这种让每个人都能成为创造者的趋势，能打开无数我们目前根本想象不到的机会——让数以百万、甚至千万计的人，轻松把自己的想法变成实际可用的东西。

I hope that this democratizing of building, letting each of us being builder, will open up so many opportunities and, like, that we even don't imagine yet when we will give, like, millions of new people, tens of millions of new people's ability just to convert their idea into something that works very easily.

我们会看到大量新企业、新点子涌现出来，创造出许多我们连想都想不到的新职业。

We will see a lot of new businesses and a lot of new ideas just coming in life, and they will create a lot of new works that we don't even think exist.

所以，这可以说是这种全民创造浪潮的第二波浪潮。

So it's like second second orbital of all this kind of democratizing of the building.

那么，有哪些挑战？又需要做出哪些改变？

Also, what is challenging and what will need to be changed?

我觉得，教育的变革所带来的风险，和它带来的机遇一样大。

And I think it's like as risk as an opportunity as how the education will change.

因为现在每个人都能接触到智能工具，人们到底该学些什么？

Because now when everybody has access to intelligence, what should people learn?

你肯定不需要他们再去死记硬背那些事实了。

You definitely don't need them to learn the facts.

所有信息都触手可及。

Everything is available.

所有的知识几乎都能随时获取。

Like, all the knowledge is kind of available.

当人们不再需要费力思考的时候，我们该怎么培养他们的思维能力？

Like, how do you really, train people to think when they don't need to think so much?

怎么教人们持续适应变化？因为很多职业都不会稳定了。

How to teach people to to continuously change, like many professions will be not stable.

你怎么帮助人们在不断变化的环境中找到自己，学会思考，并持续掌握新概念？

How do you help people to find themselves in the changing environment, think, and learn the new concepts constantly?

这既带来了大量新机会，也带来了不少风险。

This is something that very like, a lot of gives a lot of new opportunities, but also, like, creates a lot of risks.

你提到你有两个十几岁的女儿。

You you mentioned, you know, you have two teenage daughters.

你建议她们未来十年进入职场时该怎么做？

What do you advise them entering the workforce in the next ten years?

你建议她们怎么做？

What what do you advise

我跟她们说的很简单，我觉得有两件事是必须的。

them?

我跟她们说的很简单，我觉得有两件事是必须的。

What I literally tell them is I think two things will be needed.

我不知道具体需要什么，但我确定有两件事是必不可少的。

I don't know what will be needed, but I'm sure that two things will be needed.

第一是能跟人好好沟通，要有同理心，懂得理解别人。

One is, like, being able to communicate with the people within party, with the empathic communications.

也就是真正理解人类，跟人交流，并且有同理心。

So, like, understand humans, communicate with humans and being empathic.

第二是创造力。

And the second is creativity.

所有的艺术性的东西。

All the the art.

我希望这些艺术性的能力能够一直存在下去。

I hope that that art in a way will be will exist.

十年前我以为那些硬技能最重要，觉得孩子最该学的是数学和工程。

So all the hard skills that I thought ten years ago will be needed, I thought that the most important thing they need to learn is math and engineering.

现在我不这么想了，反而很欣慰他们更注重软技能，比我小时候强多了。

Now I'm far from this belief, and I'm quite happy they're much more in the soft skills than I was when I was a kid.

能够与人沟通，理解人，对人有同理心，同时具备创造力，愿意尝试新事物，发挥创意。

Being able to communicate with humans, understand the humans, be empathic to the humans, and have this creativity idea, like being able to try new things and, like, be creative.

我觉得，如果你能帮助孩子培养这两项能力，十年后他们一定会很抢手。

I think these two, if you can help your kids to develop those, I think they will in ten years, they will be in demand.

有一个问题是怎么

There's a question of how do

教人创造力。

you teach creativity.

但我完全同意你的看法。

But I completely agree with you.

你把这句话说完。

Finish this sentence.

对涅玻斯来说，最大的威胁不是竞争，而是……

The biggest threat to Nebius is not competition, but dot dot dot.

而是整合。

But consolidation.

是的。

Yeah.

我觉得Nebius作为一家公司的核心在于，世界会越来越集中。

I think that the main thread for Nebius as a business is the world will be too much consolidated.

我们努力保持多元化。

We try to be diversified.

比如，我们尝试解决不同客户的问题，服务多样化的客户和不同的参与者。

Like, we try to solve, like, problems of different customers and have different customers and different players.

如果未来世界只剩下三五家超级公司、超级巨头掌控全局，那么Nebius这样的公司可能就只能为它们提供底层的物理服务了。

If you'll end up in the world where, I don't know, three, five supermodels, super companies, super empires control the world, then Nebius or companies like Nebius will be needed only to help them maybe serve their needs on physical layer.

所以总的来说，世界越开放、越多元，我们就越有需求。

So I think that the in general, the consolidation is our main the the the the more world democratized, the more world diversified, the more we need as a business.

你觉得这种情况很可能发生吗？

You think that's likely?

我们正看到价值越来越集中在少数几家玩家手里。

We're seeing the concentration of value to fewer and fewer players.

我们看到的恰恰是反向的多元化。

We're seeing the opposite of diversification.

我希望这种情况不要发生。

I I hope it will not happen.

作为一家公司，我觉得这对咱们每个人来说也更好。

As a business, I think that it's better for us as humans as well, for you and me.

世界依然会以各种方式保持多元化，我对这一点很乐观。

The world remain quite diversified in a different manners, and I'm optimistic here.

我觉得有太多人想独立做点自己的东西了。

I think that there are so many people that want to build something independently, let's say.

很多人迫切想尝试新事物、创造新东西，这种需求自然会推动多元化的发展。

Like, there is a lot of people with the need to try things and build new things, that it's organically creates this pressure and organically creates more diversified world.

所以希望这种状态能持续下去。

So hopefully, will remain.

倒数第二个。

Penultimate one.

利奥·阿什恩布伦纳现在是个知名投资人，拥有一大批狂热粉丝。

Leo Ashenbrenner is a famous investor right now, has huge cult following.

他最近公开了自己的一项重大持仓，占这家公司5.3%的股份。

He recently disclosed a very large position for him, 5.3% of the company.

我觉得这占了他投资组合的15%左右。

I think it's 15% of his portfolio.

你们内部是怎么看待这件事的？

How do you guys sit internally?

你们是不是也觉得，这事儿挺靠谱的？

Are you like, yeah.

支持利奥。

Go, Leo.

当然了。

Oh, yeah.

说实话，我们不可能没注意到这件事。

Like, I I would lie if you I I wouldn't say that we didn't notice it.

很明显，大家都注意到了，股价立马就涨了，这事在圈子里闹得挺大。

Obviously, like, everybody noticed it and, like, the the stock jumped and, like, it was a big news around.

我们把它看作是对我们做法的肯定。

We take it as a justification of what we do.

于是你就有了这份认可。

Then you you got this justification.

你心里会想，行，没错。

You say yourself, okay.

那些人依然相信你，觉得你有能力把事情做成。

Those people still you they give you a credit that you will execute.

我一再强调，我们做的其实是售后业务。

I I I come back again and again to what we do is post sale business.

每次签下一个单子，每次有人投资我们，都是在给我们信任和机会，让我们去兑现承诺，然后回到岗位上把事情做好。

Every time we sign a deal, every time someone invests in us, they give us a credit and opportunity to deliver, but to then go back to your job and deliver.

我觉得现在这种情绪化的市场里，更得脚踏实地才行。

And I think that we are in such a market where emotional market as well that you should keep yourself down to the ground.

别忘了，所有这些增长，客户给你的信任，都是让你去兑现承诺、好好干活的机会。

Remember that all this growth, all these credits that customers give you, it's opportunity to deliver and go do your job.

你真是典型的以色列人。

You're such an Israeli.

美国人可能会说，对啊。

Americans would be like, yeah.

走吧。

Go.

你却说，不行。

You're like, no.

我觉得我在这一点上更像是个俄罗斯人。

I I I think I'm Russian in this way.

俄罗斯人总是很务实，凡事都做好最坏的打算，脸上总带着一副‘事情没那么简单’的表情。

Like, Russians always know that you need to look in the very pragmatically, pragmatically and, you know, Russians always with this, like, faces and always expect something will happen.

你得做好准备。

You need to be ready.

你得做好准备。

You need to be ready.

这一点特别重要，也是我们CEO兼创始人凯莉一直强调的。

So it's really important part that comes from our CEO also and founder, Carrie.

每天醒来，面对新客户、新的一天，你都得把活儿干好。

You wake up and it's a new customer, new day, you need to deliver.

没有什么是理所当然的。

Nothing is guaranteed.

你只需要专心把工作做好。

Just you need to concentrate on the work.

我知道团队付出了多少努力，知道每一天的投入有多关键，也知道市场变化有多快。

And I know how much efforts team is putting on things to work and how much depends on everyday's dedication and how fast market is moving.

要想跟上节奏，你就得跟市场保持同样的步伐，甚至更快。

And to stay relevant, you need to continue moving in the same pace or try to move in the same pace with the market.

说点感性的，我觉得我们其实可以多庆祝一下，但就是没时间。

And I think that on a romantic note, I I would say that we could celebrate a little bit more, but we just don't have time.

借此机会，我想向团队致谢，我们确实表扬得不够。

To use the opportunity, actually, to say kudos to the team, I don't think we celebrate enough.

我知道我们不能松懈，但我觉得偶尔也该多庆祝一下。

And I I I think that I think it's right we are not relaxed, but I think we could celebrate a little bit more.

多给团队一些认可，看看他们到底完成了多少工作。

Just give the team, like, more more respect and, like, how much is done.

这从来都不容易，现在也不容易，以后也不会容易。

And it was not easy, and it's still not easy, and it will not be easy.

但无论如何，都不能停。

But, yeah, never stop.

我们绝对不能停下。

We we cannot stop.

就像鲨鱼一样。

Like, it's like a shark.

只有不断前进，才能活下去。

You're alive when you move.

对吧？

Right?

所以这个著名的事儿。

So this famous thing.

所以我们得继续前进。

So we have to move.

借此机会，我由衷感谢你参与这次对话，也感谢你耐心应对我那些东拉西扯的问题。

On that note, I cannot thank you enough for joining me and for putting up with my very meandering questions.

你表现得太棒了，罗马。

You've been fantastic, Roman.

真的非常感谢你。

So really huge thank you.

谢谢你，你太抬举我了。

Thank you, and, too kind to me.

但在结束之前，我想提醒一下，虽然你有想法，但用AI工具时，常常会遇到瓶颈。

But before we leave you today, you have the idea, but often with AI tools, you hit a wall.

Base44 就是让这些麻烦彻底消失的地方，让你说话的方式直接变成构建的方式。

Well, Base forty four is where that friction disappears, turning how you talk into how you build.

全栈的网页和移动应用、网站、自主超级代理，几分钟就能搞定，再也不用花整个周末折腾配置了。

Full stack web and mobile apps, sites, autonomous super agents, all built in minutes, not weekends spent on damn configuration.

Base44 一开箱就自带所有功能。

Base forty four ships it all out of the box.

后端、数据库、身份验证、主机服务，全都有了。

The back end, the database, the authentication, and the hosting.

它把繁重的活都扛了，你只需要专注在创作状态里。

It handles the heavy lifting so you can just stay in the flow.

它不只是取代了那些琐碎的工作。

It doesn't just replace the busy work.

它让你的能力翻倍。

It multiplies you.

它让你变成一个更强大、更高效的自己。

It makes you so much more capable effective version of yourself.

在这个市场里，速度快只是基本要求。

In this market, being fast is the baseline.

但要想赢，你得抢在别人前面。

But to win, you gotta be first.

Base 44 就是你的优势所在。

And Base 44 is that edge.

它让你绕过那些烦人的调试，直接冲向突破。

It's the move that lets you skip the troubleshooting and get straight to the breakthrough.

去 base44.com 发布你的下一个大项目吧。

Launch your next big thing at base44.com.

就是 base44.com。

That's base44.com.

Base 44 帮你上线后，Corgi 会接手后续的一切。

After base forty four helps you launch, Corgi helps you cover what comes next.

天哪。

My word.