20VC：为什么AI无法在企业中取代人类 | 为什么工作流程而非模型将成为AI时代最有价值的资产 | 与UiPath的Daniel Dines探讨欧洲为何失势以及美国与欧盟在AI领域的建设

这是其中一个框架不清晰的领域，因为当我收到发票或客户的订单时，我可以有不同的处理方式。

This is one of the domain where the frame is not clear because when I receive an invoice or I receive an order from a customer, I can treat it differently.

规则并不总是存在的。

There are not always the rules there.

我知道对于这个客户，Nvidia会优先发货给OpenAI。

I know for this customer, Nvidia is going to ship with priority to OpenAI.

也许他们有一条规则说：“是的，我的第一批芯片先给那边。”

Maybe they have a rule that says there, "Yes, my first chips go there."

也许他们没有。

Maybe they don't.

如果那条规则没有被纳入框架，AI就无法学习它。

That rule they don't, if it's not captured in a frame, AI cannot learn it.

这就是为什么你需要创建这个手册。

This is why you need to create this manual.

我们把这个手册称为工作地图。

We call this manual the map of work.

你得把工作地图交给AI，才能成功。

You need to hand the map of work to AI in order to be successful.

你说的工作地图是什么意思？

What do you mean the map of work?

就是记录工作是怎么运作的，在企业里工作是怎么发生的。

Who basically capture how the work works, how the work happens in an enterprise.

这就是工作地图。

This is the map of work.

它包含了所有的流程、所有的例外情况、所有的程序，以及你为了实现某个流程目标而使用的所有系统。

It's all the workflows, all the exceptions, all the procedures, all the systems that you use in order to fulfill a goal of a process.

当然，但如果你有一个财务主管，管理着30个代理，当Nvidia回来说‘嘿，我们是你们最大的买家，我们在付款上有特殊条款’时，他们会说‘好的，没问题。’

Sure, but then you have a head of finance who sits on top of 30 agents and exactly when Nvidia come back and say, "Whoa, we're your biggest buyer, and we have special terms on our payments," they go, "Yeah, sure.

没错。

That's right.

别担心。’

Don't worry about it."

但你得明白我的意思。

But you come to my point.

即使在财务领域，你也不可能替换所有人，所以——

Even in finance, you cannot replace everybody, so you-

不是所有人，但你只需要一两个人。

Not everyone, but you got one person or two people

……你还是需要一定数量的人。

... you will need a certain number of people.

问题是，如果你有X个人，你怎么知道哪些人留下，哪些人可能得去做不同的工作？

The thing is, if you have X number of people, how do you understand which of them stay and which of them have to go to maybe do different jobs?

你怎么知道？

How do you know?

因为理想情况下，你会留下那些懂AI、能主动做事的人。

Because ideally, you will get the people that will have literacy in AI that can display initiatives.

因为AI不能像人那样表现出主动性。

Because AI cannot exhibit initiative in the human sense.

所以，在你想留住的那批人里，你要留的是那些最有主动性的人。

So out of this number of people that you want to keep, you want to keep the people that display the most initiative.

即使是财务人员跟客户处理发票，也会影响公司文化和外界对我们企业的看法。

Even a finance person treating an invoice with a customer contributes to the culture and how my enterprise is regarded.

那我怎么区分那个人和另一个不太在乎怎么对待客户的人呢？因为后者更容易被AI取代。

How can I make a distinction between that person and the other person that care less about how they treat the customers because the second is more prone to being displaced by AI?

我觉得企业需要建立这样一本账，去理解和衡量员工的不同产出。

This is the ledger that I think enterprise have to create and to understand different outputs of people.

他们得用这种难以量化的产出来评判员工。

They need to judge people by this hard-to-define output.

那具体该怎么做呢？

So how does one do that then?

对于那些公司里没记录下来的、难以量化的数据，我们怎么处理？

For the illegible data that isn't captured within companies, how do we do that?

扎克伯格和Facebook曾提到要监控员工屏幕上的每一个操作。

Zuck and Facebook have said about monitoring every single action that's on the screen of employees.

我觉得这并不能捕捉到通话的语气、事后发给客户的温暖文字，以及那些难以量化的数据。

I don't think that captures the tone of a call, the warm text afterwards to a customer, the illegible data.

我们怎么才能捕捉到那些体现价值但我们没有记录的数据呢？

How do we think about capturing the data that shows value that we don't capture?

我觉得这正是问题的关键所在。

I think this is the crux of the problem.

在这里，你真是个真正的哲学家。

And here you're a true philosopher.

这正是我们公司投入大量精力的地方，我们正在推出一种我们称之为“制图学”的新技术。

And this is where we put a lot of effort as a company, and we are introducing a new technology that we call it cartography.

制图学是一门学科。

And cartography is a discipline.

它是一门帮助公司展现所有工作方式的信息，并帮助他们创建这份工作地图的学科。

So it's a discipline to help companies surface all the information of how the work is done and help them create this map of work.

制图学的一个重要方面，就是真正去调查人们在桌面上都在做些什么。

And one big important of cartography is to actually have investigated what people are doing on their desktops.

我们有一个产品叫“制图员代理”，它可以采访真正领域专家，让他们基本上记录自己在做什么，而代理会实时进行采访。

And we have a product that we call the Cartographer Agent that can interview people, real subject matter experts, have them basically record what they are doing, and the agent is interviewing them in real time.

如果你采访财务人员，它可以问：“为什么当邮政编码不同时，你修改了这张发票？

If you interview a finance person, it can ask, "Why did you change this invoice when the zip code was different?

为什么你选择了不同的路径？

Why did you choose a different path?

告诉我。”

Tell me."

然后他们就能开始把所有那些例外情况都摆出来。

And they can start surfacing all of these exceptions.

所以这就是真正的代理在采访真实的人。

So that's real agents interviewing real people.

这挺酷的。

It's pretty cool stuff.

然后你把多个人的数据整合起来，再创建流程地图。

And then you consolidate data from multiple people, and then we create the process maps.

所以我们实时展示工作是如何进行的。

So we show them how the work works in real time.

然后，在展示当前工作流程的地图之后，你可以使用我们的编码代理，想出如何改造这个流程的想法。

And then after this map of work that show the work as is, you can use our coding agents, and you can come up with an idea of how you should transform the process.

而改造流程的方式就是生成软件。

And you transform the process by printing software.

所以从一个流程开始，在A点，它完全是人工操作的。

So you start with a process, and at point A, it's fully manual.

当然，你有企业系统，比如记录系统，但操作这些系统的是人。

Of course, you have enterprise systems like system of records, but people operate the system.

我认为任何企业的目标都是减少操作系统的人员，增加自动化和代理型AI来操作系统。

I think the goal of any enterprise is to have less people operating the system and more automations and agentic AI operating the systems.

你不担心公司员工的抵触吗？他们知道你在观察他们的工作是为了取代他们。

Do you not fear pushback from people working in the company, aware that you are watching what they do to replace them?

扎克伯格就遇到了这种情况。

That is what Zuck got.

这是不可避免的。

This is inevitable.

这很大程度上取决于你如何向公司推销这个理念。

It very much depends how you pitch the company.

所以在UiPath，我觉得重要的是告诉员工们，再说一遍，'各位，我们不是在做什么蠢事。

So in UiPath, I think it was important to tell people, again, "Guys, we are not doing anything stupid.

我们不是打着AI的幌子搞大规模裁员，但转型是不可避免的。

We are not doing any mass extinction for the pretense of AI, but transformation is inevitable.

你们必须转型，而且每个人都有机会。

And you guys have to transform, and everybody will get a chance.

那些对AI更熟悉的人，不仅在这里，将来在任何其他工作中都会有更好的机会。'

And the people that will become more literate in respect to AI will have a better chance, not only here but in the future in any other job."

我觉得每个人都应该听到这样的信息。

That's the message that I think everybody should get.

他们对此有回应吗？

Did they respond to it?

在那之后，你有没有看到AI采用率飙升？

Did you see AI adoption through the roof after that?

我觉得反响不错。

I think the response is good.

AI的采用需要比仅仅发现流程和人们参与其中更多的周期。

AI adoption requires more cycles than just discovering the process and people's input in this.

但是你看，年初的时候，整个行业的人，不仅是我公司，很多公司，都极度恐惧，害怕被完全取代。

But look, in the beginning of the year, the fear of people across the industry, and not only in my company but in many companies, was off charts, the fear of completely being replaced.

现在，我觉得人们开始更理解他们工作的持久性，以及AI在企业中的扩散可以以较慢的速度进行，一次处理一个流程。

Now, I think people start to get a bit more understanding of the durability of their jobs and this AI diffusion in enterprise that can happen at a slower pace and can happen one process at a time.

因为那数百万个爱因斯坦式的人才还无法被雇佣。

Because these millions of Einsteins are not hireable yet.

我们投资的一家公司，Macaw，一家数据提供商，其CEO Brandon Foody昨天发推说，他们在推理上的花费是人力薪资的三倍。

One of the companies we've invested in, Macaw, data provider, Brandon Foody, the CEO, tweeted yesterday that they spend 3X the spend of human salaries on inference.

你觉得你们在推理上的花费是人力薪资的多少倍？

What percent or multiple or take would you say you spend human salaries on inference?

我个人其实不太在乎这个，而且我跟你说，

I personally don't care about it And let me tell you something.

这其实是个很简单的假设。

It's a simple hypothesis.

如果机器能做出比人更好的工作成果，那我今天就雇机器，哪怕它比人还贵。

If the work at the quality were better of a human can be done by a machine, I will hire today a machine even if it's more expensive man than a human.

因为人的成本只会越来越高，而且人还会带来各种错误，但机器的成本会越来越低。

Because human costs will only increase, and they bring errors into the pictures while the cost of machine will decrease.

所以比起那些坚持用人的公司，我会有竞争优势。

So I will have a competitive advantage compared to people that will stick to humans.

我觉得所有人都会这么做。

I think everyone will do this.

所以我觉得，用AI替代人，真正的关键不是token的成本。

So I don't think the cost of tokens, it will be the real question if you replace a person with AI.

但现在真正的问题是，AI还替代不了人，因为，再说一遍，如果你给我一个能替代我的爱因斯坦，我巴不得去世界各地度假，但这个爱因斯坦还不存在。

But the real problem today is that AI cannot replace a person because if, again, if bring me an Einstein that replace me and I will happily go in any vacation in the world, but this Einstein doesn't yet exist.

我很想和另一个爱因斯坦一起接受采访，做这个播客。

I would like to be interviewed and have this podcast with another Einstein.

天哪，这种东西现在还不存在。

Man, this thing doesn't exist today.

这就是现实，所以我们就叫它现实吧。

That's the reality, so let's call it a reality.

我们就实话实说吧。

Let's call a spade a spade.

也许这项技术会出现，然后出现一些像人一样有意志、能在工作中不断成长、边干边学、具备推理和想象能力的爱因斯坦。

Maybe this technology will emerge and somehow Einsteins that embody like a person, that have will, that gets transformed on the job, learn on the job, have the capability of reasoning, imagination of Einsteins exist.

当然，到那时所有的工作都会消失。

Of course, all the jobs will go extinct.

我和SaaStr的Jason Lemkin做过一期节目。

I have a show with Jason Lemkin from SaaStr.

他把团队从25人缩减到了2人。

He's cut his team from 25 to two.

如果他现在坐在我这个位置，他会说："不，不，不，它确实能。"

If he were in my seat now, he would say, "No, no, no, it does.

我换掉了我的财务副总裁。

I replaced my VP finance.

我换掉了我的营销副总裁。

I replaced my VP marketing.

而且实际上，AI做得更好。"

And actually the AI is better."

我想见见这个人。

I want to see this man.

我听说有些公司过去裁掉了数百名客服人员，现在他们又重新雇用了这些人。

I heard companies that replace hundreds of support people in the past and now they are rehiring these people.

我觉得，除非这个模式被证明有效，并且在大规模范围内被证明有效，而不是在某个特定行业、针对某个特定的人，否则我们不能仅凭一个数据点就推断它会适用于整个行业。

I think until this model is proven and is proven at scale, not in a particular industry for a particular guy, I don't think we can extrapolate for one data point that is going to go across industry.

你提到了推断和过度夸大。

You said about extrapolation and over-exaggeration.

SaaS末日论确实是真的。

The SaaSpocalypse was very real.

我们当时觉得什么都能用Vibe Code搞定。

We're going to Vibe Code everything.

你们真的用Vibe Code做了工具吗？

Did you Vibe Code tools out?

说实话，那确实挺神奇的。

Look, it was amazing.

对，我们做了，但算不上特别成功。

Yes, we did, but not an extraordinary success.

一开始看着挺厉害，但真要投入生产的时候，就发现这些工具其实有不少瓶颈。

Initially, it seemed extraordinary, but when we tried to put it in production, we started to see some kind of real bottlenecks with these tools.

而且你得维护很多东西，比如连接器、权限、审计、安全，而且把一个软件从原型带到生产环境，那才是真正费劲的地方。

And you need to have a lot of things that you have to maintain, connectors, permissions, audit, security, and always taking a software from a prototype to production, it's actually where the work is.

写代码本身倒不是最难的。

It's not necessarily the writing code.

写代码是挺有意思的，但真正能拉开差距的地方不在这儿。

Writing code is fun, but it's not there where you can really makes the difference.

我觉得现在做个原型容易多了。

I think it's much easier today to make a prototype.

原型做起来太简单了，但接下来你得不断迭代，才能把原型变成真正能上线的产品。

Prototype is so easy, but then you have to iterate to make the prototype in production.

这中间就涉及到测试，还有一堆别的事情。

And this is the testing and everything else.

我们当时想自己动手写一个采购工具，替换掉现有的。

We were trying to replace a procurement tool, writing ourselves.

我觉得我们做得挺成功的。

And I think we have a lot of success.

而且一开始，这个工具完全是由AI写的。

And it was initially, it was written only by AI.

但后来就涉及到我们说的自我改进循环了。

But then it comes to speaking to this self-improvement loops.

有点不放心，我们测试不够，没有足够的信心把这个工具100%投入生产环境。

Kind of don't trust, we don't have enough tests, we don't have enough trust to put this tool in production 100%.

而且根据我们的经验，人类必须大量介入这个Vibe Coded工具的使用。

And humans, in our experience, humans have to intervene a lot into how this Vibe Coded tool.

比如，Vibe Coded工具创建的数据库结构完全是胡来的。

For instance, the database schema that Vibe Coded tool created was completely bogus.

必须有人类来搭建这个结构。

A human has to come and create the structure.

所以现在还没到那种程度，业务人员理解问题后就能用Vibe Code做出工具。

So right now you are not at the point where you will have a business user that understands a problem and they will Vibe Code a tool.

所以你还是需要工程师和人手来维护它。

So you will still need to put engineers and people, you need to maintain it.

所以这是个漫长的过程。

So it's a long process.

所以最终你付出的成本可能和替换掉的工具差不多，甚至更高，同时还得占用你最好的人才的时间。

So eventually you will end up paying probably as much, if not more, as the tool you replace while you keep some of your good and best people bandwidth occupied.

你今天会买Salesforce吗？

Would you buy Salesforce today?

我会把Salesforce作为记录系统来买。

I would buy Salesforce as a system of record.

作为股票？

As a stock?

作为股票。

As a stock.

你看，我投资软件这个类别，我觉得几个月前我做了个好投资，因为我在SaaS崩盘的最低点买入的。

Look, I invest in software as a category, and I think I made a good investment in a few months ago because I bought in the bottom of SaaSpocalypse.

连我们自己的股票表现都更好了。

Even our own stock has been doing better.

但现在的市场很大程度上是由情绪驱动的，而不是价值。

But the markets today are driven so much about sentiment and not the value.

所以对我来说，很难对某一家公司做出判断。

So it's kind of hard for me to make a judgment of an individual company.

但我不觉得Salesforce会被Vibe Coding取代，如果这是你想问的话。

But I don't think Salesforce can be replaced by Vibe Coding, if this is the question.

我不明白为什么现在还有公司要上市。

I don't understand why a company would go public today.

想想上市的两个主要原因，第一是流动性，让股东、员工和股东能套现。

If you think about the two drivers of being public, number one is liquidity for shareholders, employees and shareholders.

但Stripe和很多公司在私募市场就能做到股票流通。

Well, Stripe and many companies are able to have liquid stock in private markets.

第二是并购能力，需要一个可以用来交易的资产。

And two is the ability to have M&A, a tradable asset that you can buy with.

其实很多私营公司都能用私募股票来收购。

I mean, many, many private companies are able to buy with private stock.

Stripe打算用私募股票以600亿收购PayPal。

Stripe are going to do PayPal with private stock for 60 billion.

所以这根本不是障碍。

So that's not a barrier.

而且就像你说的，公开市场已经变得像赌场一样，现在的股市完全由情绪驱动，我真不明白为什么还有人要上市。

And the casinoization of public markets, as you said, with current stock markets being sentiment-driven, I don't understand why one would.

但是Harry，我们别把一些非常特殊的公司和市面上大多数公司混为一谈。

But Harry, let's don't make a confusion between some very exceptional companies and the most companies that are out there.

现在有很多公司其实就像僵尸一样。

There are so many companies that are kind of zombies right now.

就是那些2021年的僵尸公司。

These 2021 zombies.

它们现在在公开市场上反而会过得更好。

They would fare better in the public market right now.

至少它们的投资者能有退出的途径，员工也能有机会赚到一些钱。

At least their investors will have a way for exiting, their employees will have a way to make some money.

而现在，所有人都只是纸上富贵，对吧？

Nowadays, all of them are sitting on paper Okay?

但公开市场会让它们直面自己真实估值的现实。

But public markets will face them with the reality of what's their real valuation.

我不是说这没有价值——Anthropic 最终为什么要尝试 IPO 呢？

I will not say there is no value in the-- Why the Anthropic is trying to do IPO in the end?

嗯，但他们和 OpenAI 一样是独特的公司，他们不得不这样做，因为他们已经耗尽了所有现有的私人融资。

Well, but they're a unique company alongside OpenAI, which just has to because they have exhausted all private funding that exists.

他们真的是把私人资本的供应给用尽了。

They are literally hitting the tap out button on the supply of private capital.

他们是非凡的公司，因为他们需要的资金太多了。

They're extraordinary companies because they just need too much money.

你觉得 OpenAI 和 Anthropic 的所有投资者会多年留在这些公司吗？

Do you believe all the investors in OpenAI and Anthropic will stay in the companies for years to come?

不会。我觉得有些人会，但有些人——

No. I think some will do, but some-

有些人当然会，但我觉得我们会看到大批人离开。

Some will do, of course, but I think we will see an exodus.

说实话，我不相信两万亿美元或者别的什么，也许他们会达到五万亿，因为如果我买了价值两万亿的股票，我需要有一条通往五万亿的路径。

Honestly, I don't believe a $2 trillion or whatever, maybe they will reach $5 trillion or because if I buy a $2 trillion, I need to have a path to five.

但如果他们以两万亿美元的估值上市，你会卖吗？

But if they went public at $2 trillion, would you sell?

特别是Anthropic，我不会卖。

Anthropic in particular, I wouldn't.

因为你还记得我们上次播客里，我问你押注哪家公司，我说的是Anthropic。

Because do you remember on our last podcast, I think you asked me which the company that I bet on, and I said Anthropic.

当时Anthropic的市值是600亿美元。

And Anthropic was worth $60 billion market cap.

也许我太傻了，没投资。

Maybe I was stupid I didn't invest.

你本来可以在那上面赚得比Salesforce或ServiceNow之类的更多。

You would've made more money on that than Salesforce or ServiceNow or whatever.

百分之百。

100%.

所以我不会卖。

So I would not sell.

我觉得OpenAI也是一样。

And I think it's same with OpenAI.

我觉得OpenAI已经追得很好了，我现在两个都在用，没什么区别。

I think OpenAI has catched up quite nicely, and I use interchangeably right now.

我只是觉得我们现在处在一个市场里，大的越来越大，价值比以往任何时候都更集中。

I just think we're in a market where the big get bigger and value concentrates more than ever.

但真正的问题是，Harry，我会不会买一个两万亿美元的公司？

But the real question is, Harry, would I buy a $2 trillion?

这才是我的真正问题。

That's my real question.

现在，我需要看到他们的真实数据，才能决定要不要把钱投进他们的IPO。

And right now, I need to see their real numbers to understand if I will go, if I will put money in their IPO.

我这么说肯定会惹麻烦。

I'm going to get in so much trouble for this.

我觉得AI很像比特币，很难确定哪个应用会赢，哪个钱包会赢，哪个使用场景会赢。

I think AI is quite like Bitcoin in just the way that it's very difficult to determine what application is going to win, what wallet's going to win, what usage is going to win.

如果是这样的话，那就押注那些你确定会存在的基础设施。

And if that is the case, by the underlying infrastructure that you know is going to be there.

所以对我来说，我同意你的看法。

And so for me, I agree with you.

我不知道Claude会不会比下一代的Codex更好。

I don't know if Claude is going to be better than the next Codex.

我不知道Cursor会不会推出什么特别牛的东西，但我确定的是，Jensen会坐在那儿，然后说：“这是另一块芯片。”

I don't know if Cursor are going to come out with something fucking amazing, but I do know that Jensen's going to be sitting there going, "Here's another chip.

这是另一块芯片。

Here's another chip.

再来一块。”

Here's another."

太好了。

Great.

对，但我觉得Jensen的成功取决于开源的发展，因为如果Anthropic和OpenAI形成双头垄断，基本上，我觉得他们的市场潜力是万亿级别的。

Yeah, but I think Jensen is bound by the success of open source because if Anthropic and OpenAI, if this is becoming a duopoly, and basically, I think their TAM is in trillions.

基本上就是干这个活。

It's basically the work.

他们自己就能造芯片了，兄弟。

They will print their own chips, man.

说实话，到最后造芯片也不是什么大不了的事。

Honestly, it's not such a big deal in the end to print chips.

那当然。

Of course.

OpenAI 在做 Jalapeno-

OpenAI are doing Jalapeno-

……Anthropic 也在做自己的芯片。

... and Anthropic are doing their own chips.

没错。

Exactly.

所以我觉得，如果最大的供应商只有 Anthropic 和 OpenAI，而且真理和光明之源只来自这两家，那 Jensen 的日子可能就没那么好过了。

So I don't think Jensen will be doing so well if they have the single biggest providers, and the source of truth and light of God will come from only Anthropic and OpenAI.

所以，开源应该成功。

So therefore, the open source should succeed.

我完全同意，这也是为什么我认为黄仁勋在写那封公开信，大家都签了名——

I absolutely agree, which is why I think Jensen is doing the open letter, which everyone signed-

没错

Absolutely

……鼓励开源。

... encouraging open source.

而且我觉得收购是个好棋——

And I think it was a great move buying-

Hugging Face

Hugging Face

……Hugging Face。

... Hugging Face.

是的。

Yes.

为什么？

Why?

因为我觉得它鼓励并托管了所有开源模型。

Because I think it encourages, it hosts all the open source model.

这对他的公司来说是把钱花在了刀刃上。

It's putting the money where the money is for his company.

完全同意。

Totally.

这也让一个中立的提供商不再中立了。

It does also make a neutral provider not neutral anymore.

有偏见了。

Bias.

对，显然，他们现在也有Nemotron，还有自己的模型。

Yeah, obviously, they have Nemotron, and they have their own models now as well.

可以说，现在被英伟达收购后，独立性有所丧失。

You could say there's a loss of independence now that it's owned by Nvidia.

我觉得Nemotron在整个图景里还是个小角色。

I think Nemotron is still a small cog in the picture.

它有价值，但我不会说它——它的规模和其他那些不一样。

I think it's valuable, but I would not call it-- It's not the same chip size as the others.

你担心那种循环收入的问题吗？

Do you worry about the round-tripping revenue?

大家都在说英伟达这边投资、那边收购，还有从Oracle和OpenAI来的那种循环经济。

Everyone talks about Nvidia investing here, buying here, the circular economy that come from Oracle and OpenAI.

你觉得这被夸大了吗？

Do you think that's overblown?

有可能，因为历史上每个重大基础设施都建得过头了。

It can be because every major infrastructure in history has been overbuilt.

我觉得这解释很简单，因为我一直在想为什么每个基础设施都会过度建设。

And I think it's a simple explanation because I was thinking why every infrastructure is overbuilt.

因为你得确保自己拿到最大的那块蛋糕。

Because you have to make sure you get the biggest piece of the opportunity.

如果机会很大，那多建一点也没关系，比需要的多建一些。

If the opportunity is big, it doesn't matter, you build a little bit more than it's necessary.

所以很明显，现在每个人都在抢同一个蛋糕，但大家建的是蛋糕的两倍。

So it's clearly that now everybody, that there is only 100% of the pie, and people are building right now 200% of the pie.

肯定会有输家。

There will be losers.

但你不觉得这是第一次我们严重建设不足的创新吗？

Do you not think, though, this is the first innovation where we are significantly underbuilt?

而且，如果你看看现在的限制条件，你说得对。

And actually, if you look at the constraints now, you're right.

实际上，在之前的技术周期里，我们过度建设了供应端，而需求端却落后了。

Actually, in prior technology cycles, we overbuilt supply side and demand side was lagging behind.

现在能源成了巨大的限制。

Now we have energy that's being a massive constraint.

还有水。

We have water.

我们有数据中心。

We have data centers.

我们有法规和政策。

We have regulation and policy.

我们在供给侧面临很大的阻碍，而且我们是建设不足，而不是过度建设，这就是为什么每一分算力都被抢光了。

We have a significant hindrance to supply side, and we are underbuilt, not overbuilt, which is why every ounce of compute is taken.

是的，但我们的建设不足，是否到了需要投入数万亿美元基础设施的程度？

Yes, but are we underbuilt to the extent of these trillions that comes into the infrastructure?

这个我不知道怎么回答。

I don't know to answer this.

所有事情都建立在AI将大规模取代人类工作的前提上。

Everything happens on the premise that AI is going to replace human work in a really large scale.

我们需要看这种取代和转型的时间点。

We need to see the timing of this replacement and transformation.

如果是在10年后发生，和未来两年内发生，差别可大了。

It's a big difference if it's coming in 10 years versus next two years.

所以我觉得未来十年不算过度建设，但未来三年可能确实有点过度了。

So I don't think it's overbuilt for next decade, but it might be overbuilt for the next three years.

而且股市和资本可能非常无情，而且——

And stock markets and capital can be merciless and-

也许我是个天真的乐观主义者，但我看到安德烈·卡帕西说，他最初用编程工具完成了20%的工作，六个月后，他说工具完成了80%的工作，他只需要帮忙20%。

Maybe I'm a childish optimist, but I saw Andrej Karpathy say that he used coding tools for 20% of the work, and then six months later, he said that it did 80% of the work, and he helped it with 20%.

我们是Lawora的投资者。

We're investors in Lawora.

我们做那笔交易时采访了律师，他们都对我说：“你们这些科技宅。

I interviewed lawyers when we did that deal, and they all said to me, "You're such tech bros.

你们以为能取代我们。

You think you can replace us.

哈哈。

Haha.

我们可是上过法学院的。”

We went to law school."

我当时就说，"好吧，行。"

And I was like, "Okay, cool."

我两周前采访了他们。

I interviewed them two weeks ago.

15个人里每一个都说，如果这个工具被拿走，他们会非常不开心，而且大多数人说自己已经六个月没写过文档了。

Every single one of 15 said they would be severely unhappy if it was taken away, with most of them saying they hadn't written a document in six months.

哈里，你应该读读我的书，朋友。

Harry, you should read my book, my friend.

因为这本书回答了——

Because it answers-

我读了。

I did

——正是同样的问题。

... exactly the same questions.

我很喜欢。

I loved it.

当一个人操作的框架被其他人定义得非常清楚时，比如在法律领域，AI可能会产生毁灭性的影响。

When the frame that a person operates is really well defined by someone else, like in law, AI can be devastating And the impact.

当框架不那么清晰，而且有很多为企业量身定制的例外情况时——

When the frame is not as clear and there are so many exceptions that are custom-made for an enterprise-

兄弟，你马上要和我女朋友——一位律师——待上两天了。

Dude, you're about to spend two days with a lawyer who is my girlfriend.

她会告诉你，法律在写作风格上是非常模糊和主观的。

She will tell you that law is highly ambiguous, subjective in terms of writing styles.

她会——而且这是一个——

She would-- And it's a-

人类语言也是如此，而AI能完美理解它，并且能理解人类敏感性的每一个细微差别。

So is the human language, and AI understands it perfectly, and it understands every freaking nuance of human kind of sensitivities.

只要有文档记录并且定义明确，法律就有手册可循。

As long as it's documented and it's well-defined, there is a manual for the freaking law.

AI太厉害了。

AI is amazing.

如果没有手册，AI就不那么神奇了，也发挥不了作用。

When there is no manual, AI is not amazing, and it doesn't work.

这就是巨大的差别。

That's the huge difference.

但美国法律行业有3000亿美元的市场。

But $300 billion is the legal industry in the US.

这可不是小数目。

It's a lot.

如果想想这能替代多少人力，我觉得30%是合理的。

If you think about how much labor could be replaced by that, I think 30% would be reasonable.

那就是900亿美元的潜在收入。

That would be $90 billion of available revenue.

对，但这不会直接转化为代币收入。

Yeah, but that's not going to convert in token revenue.

可能900亿中，公司能收取的只有10%，所以代币的总机会大概是100亿。

Maybe out of 90 billion, companies might charge maybe 10%, maybe it's a 10 billion total opportunity in tokens.

如果市场只有100亿而不是900亿，那我投资Lawora是不是错了？Harvey的投资人投资Harvey是不是也错了？

Am I mistaken to invest in Lawora and are the Harvey investors mistaken to invest in Harvey if it's 10 billion, not 90 billion?

我可以告诉你，从法律的角度来看，开源模型和前沿模型都会表现得很不错。

What I can tell you that from a law perspective, open source models, frontier models will do just fine.

也许他们还会围绕法律流程做定制化的工作流，这确实很有价值。

Maybe they do also the custom workflows around the legal process, which is really valuable.

但对我来说，这也是我投资理念中很重要的一部分。

But to me, that's also a big part of my thesis.

模型是可以互换的，但工作流、工作地图以及围绕工作地图的工作流，才是真正有价值的地方。

Models are interchangeable, but the workflow, the map of work and the workflows around the map of work is where the real value is.

如果Harvey和Lawora在做这些，他们真正梳理了工作，创建了工作流，为我打造了一个法律部门，那他们能获取的价值当然要大得多。

If Harvey and Lawora are doing this, they map really the work, they create the workflow, and they create a legal department for me, of course, it's a much bigger value that they capture.

但如果只是用来获取法律意见，打个电话给模型，那这个市场就不会有1000亿美元那么大。

But if it's only to get a legal opinion, a call to a model, that's not going to be a $100 billion market.

完全同意。

100%.

你觉得12个月后，有多少比例的token流量会走开源模型，多少走闭源模型？

What percent of token traffic do you think will go through open versus closed models in 12 months?

我觉得问题不是这个。

To me, I think the question is different.

真正的问题在于，有多少流量会流向像Astra或Fable这样的前沿模型，又有多少会流向性价比高的模型。

What percentage of the traffic will go to truly frontier model like Astra or Fable versus very cost-efficient models?

对于企业级应用，我预测90%的流量会流向性价比高的模型。

For enterprise work, my prediction is that 90% of the flow will go to very cost-efficient models.

我觉得大多数日常运营工作不需要前沿级别的模型质量。

I don't think you need frontier-level quality of models for most operational work.

那也就是说，我们实际上还是会用核心供应商，也就是OpenAI和Anthropic。

Just to be clear then, so we will actually still use the core provider, which is OpenAI and Anthropic.

只是会换成那些老一点的、已经过时的模型。

It'll just be deprecated older models.

我还是会用Anthropic和OpenAI的性价比模型，但我会一直有一个可验证的开源备份。

I will still use Anthropic and OpenAI with their cost-efficient model, but I will have a verifiable backup on open source all the time.

作为一家负责任的企业，我应该能够切换模型。

As a responsible enterprise, I should be able to switch models.

我不能被锁定。

I cannot be locked in.

我能问你一下吗，我们投资了……

Can I ask you, we invested in...

我就是在拿我的投资组合跟你的想法对照一下。

I'm just checking my portfolio against your brain.

我非常看好开源模型，而且我认为每家公司，不是所有公司，但中大型公司都会拥有自己的模型，掌握自己的智能，并把数据喂进去。

I believe strongly in open models, and I think that every company, not every company, but mid to large-scale company will have their own model, own their own intelligence, and feed their own data into it.

对，我觉得这正好呼应了那句话——拥有自己的智能，而不是租用别人的。

And yeah, I think that goes to the statement of kind of owning your own intelligence, not renting it.

这就是我们投资Fireworks的原因，我也相信开源模型的生态。

It's why we invested in Fireworks, and I believe in the open model ecosystem.

我非常喜欢Fireworks，我们也在大量使用它。

I'm a big fan of Fireworks, and we are using them quite a bit.

你喜欢它们吗？

Do you like them?

是的，我们非常喜欢。

Yes, we like them a lot.

而且我坚信，企业应该分散投资，其中一部分应该投在开源上，更重要的是，应该投在这个工作地图上。

And I'm a big believer that an enterprise should distribute their bets, and one of the bets should be on open source, and very importantly, should be on this map of work.

因为你想，如果我想训练一个模型，一个属于我自己的模型，我需要非常清楚地记录下我是谁。

Because think about if I want to train a model, my own model with who am I, I need to have this who am I very well documented.

我需要创建这个手册，因为今天我训练一个模型，但接下来的两个月，会有更好的基础模型出现。

I need to create this manual because I'm training one model today, but in next two months, there is another better model, base model coming into the picture.

如果我没有数据和准确的手册，我怎么才能把旧模型的知识迁移到新模型上呢？

How can I do transfer learning from my old model into the new model if I don't have the data and the exact manual?

我做不到。

I cannot.

这样做的话，损失会非常惨重。

Where there are terrible losses when I do this.

对企业来说，真正的投资是创建这份工作图谱，记录他们实际的工作方式。

The real investment for an enterprise is to creating this map of work that documents how they actually work.

有了这个，他们就能训练自己的模型。

And with this one, they can train their own models.

不管是在Fireworks还是其他提供商那里，都没关系。

It's in Fireworks or other provider, doesn't matter.

但这是他们的知识产权，也是他们的核心数据。

But this is their IP, and this is their core data.

明白吗？

Make sense?

那你是不是——抱歉，我只是想确认一下我的理解。

And so you do-- Sorry, just so I understand.

所以你确实认为，很多公司会用自己的数据训练自己的模型？

So you do believe that companies, and a lot of them, will have their own models with their own data?

我确实认为，至少他们会拥有自己的模型，作为前沿模型的备份。

I do believe that they will at least have their own models as a backup to frontier models.

对我来说，我不确定自己的模型能否提供与Anthropic和OpenAI的高性价比模型相同的成本效率。

To me, where I'm not clear if I can provide the same cost efficiency with my own model versus a cost-efficient model from Anthropic and OpenAI.

因为我觉得这些公司真正有能力优化大型基础设施。

Because I think these guys are in the position to truly optimize large infrastructure.

所以他们的商业模式的一部分将是提供比我自己模型能挤出的更多的智能，每美元。

So part of their business model will be to deliver more intelligence per dollar than even I can squeeze from my own models.

如果你有高度特定的数据，恰好符合你的需求，那是你的数据，我认为用你自己的模型会比用优化的模型获得更高的令牌效率。

If you have highly specific data that is exact to the request that you have, which is your data, I think you'll get more token efficiency with your own model than you would an optimized-

只有当你训练自己的模型时，那可能是真的，而且只有当Fireworks能够大规模、高度优化地交付时。

Only if you are training your models, that might be true, and only if you can deliver, if Fireworks can deliver at a large scale and in a very optimized way.

你会以150亿美元投资Fireworks吗？

Would you invest in Fireworks at $15 billion?

可能会，是的。

Probably, yes.

如果开放模型的假设是真的，我相信是真的，我认为他们被低估了。

If this hypothesis of open model is true, which I believe is true, I think they are undervalued.

我觉得他们很快就要加入这场争夺算力的大战了。

I think they will have to get very soon in this big game of securing compute.

因为如果他们拿不到算力，我真不明白他们怎么能按我想要的规模给我提供推理服务。

Because if they don't secure compute, I don't understand how can they give me the inference at the scale that I want.

你怎么看？

What do you think?

因为你投资了他们。

Because you invested in them.

是的，我投了。

I did.

你说得完全对，他们确实需要进入算力层，我觉得Lynn正在这么做，而且很快就会有大动作。

I think you're absolutely right that they need to move into the compute layer, and I think Lynn is doing that, I'm sure very soon.

所以他们现在得筹集数百亿美元了。

So they will have to raise tens of billions now.

我会参与其中的。

And I will be there.

她在Facebook就是这么做的。

And she did that at Facebook.

我觉得这是这个团队独有的优势。

I think that's a unique thing to this team.

他们确实在算力保障方面做得很好。

They did compute securing.

而且这是个很棒的团队。

And it's a great team.

我们真的很喜欢他们。

We really like them.

这团队真他妈的棒。

It's a great fucking team.

而且我们在大热之前就和他们合作了。

And we work with them before the big hype.

我还觉得数据提供商被严重低估了，没有得到应有的重视。

I also think the data providers are massively underpriced and underappreciated.

特别是Macaw和Surge。

Macaw and Surge in particular.

大家都觉得，哦，它们就是商品。

Everyone's like, "Oh, they're commodities.

你只是在买数据而已。

You're just buying data."

数据是提升模型质量最重要的东西。

Data is the most important thing to feed model quality.

但区别在哪里呢？

But what's the difference?

我觉得一个区别是存储，另一个是对数据的理解。

I think one thing is storage, and one thing is understanding of the data.

因为如果我有存储，我可以有一盘磁带，把数据放在磁带上。

Because if I have a storage, I can have a tape, and I can put data on the tape.

我们会投资一家磁带公司吗？

Would we invest in a tape company?

我觉得不会。

I don't think so.

你得投资那些能理解数据的智能，它们能喂给模型，并在合适的时候提取合适的数据，真正把模型需要的、带上下文的数据喂给它。

You need to invest in the intelligence that understands the data and feed the model and extract the right data at the right time, feed really the model with the data that is needed with the context.

因为如果你只有数据，但没法在模型提问时给它构建一个真正好的上下文，那数据就没用。

Because if you have just data, but you don't have a way to create a really good context to give the model when they ask something, it's useless.

我觉得你会说，他们的数据比任何人都多，覆盖的类别也比任何人都广。

I think you would say that they have more data than anyone else across more categories than anyone else.

所以当模型需要非常具体的数据时，由于他们的库很广，他们能提供别人提供不了的东西。

So when the model requests highly specific data, because of the breadth of their library, they're able to provide it in a way that others aren't.

如果数据是他们自己的，而且对模型有价值，我相信模型会立刻买下这些数据。

If it's their own data and is valuable for models, I'm sure the models will buy the data in an instant.

我能问你一个问题吗？过去12个月里，你改变最大想法的是什么？

Can I ask you, what have you changed your mind on most in the last 12 months?

我以前不明白，为什么非得有一本手册才能工作。

I didn't understand this necessity to have a manual in order to work.

这大概是我理解上最大的突破。

That was maybe the biggest breakthrough in my understanding.

每次我向AI提问时，AI应该能掌握我公司运作的整个流程，或者这个特定流程的运作方式。

That every time I'm running a query towards AI, AI should have at their disposition the entire way my company work, or this particular process work.

当我意识到这一点时，我也明白了，这正是记忆和真正学习之间最大的区别。

When I realized this, I understood also that this thing that it's the biggest differentiation between memory and true learning.

我们就是这样开始讨论的。

This is how we started the discussion.

我不确定我是否真的说清楚了，但仅仅记录笔记和被经验真正改变，这两者之间有着巨大的差别。

And I'm not sure I really make a point, but it's a huge difference between just laying down, having a scratchpad, or being transformed by an experience.

这是我感受最深的一点。

And that's the thing that I realized the most.

而且我觉得我也明白了，对我们来说，什么才是人性化的。

And I think I realized also what's kind of human for us.

因为我用AI做了很多尝试，不是写代码，而是写一本书。

Because I experienced a lot with AI writing, not code, but writing a book.

所以我试过不同的风格。

So I've been through different styles.

我学到了很多怎么给它们写提示。

I understood a lot how to prompt them.

AI没有风格。

AI doesn't have a style.

你会意识到它们为什么没有风格，因为它们是所有东西的平均值。

And you realize why they don't have a style, because they are an averager of anything.

要有风格，你得有一个身体。

In order to have a style, you need to have kind of a body.

你得有个性，因为从某种意义上说，我们就是我们所做的选择和没做的选择。

You need to have individuality, because we are the choice that we made and the choice that we don't make, in a sense.

所以你需要被改变，否则我只会让AI“读这本书，然后以这个作者的精神来写”。

So you need to be transformed, because otherwise, I will just ask AI, "Read this book and write in the spirit of this author."

这其实不太行，因为你需要被那种体验所改变。

And it's not really working, because you need to be transformed by the experience.

所以对我来说，这将是AI技术最大的突破，当我能拥有像Mythos那样规模的模型，并且能在工作中被转化。

So to me, this is going to be the biggest breakthrough in the AI technology when I can have models at the size of Mythos being transformed on a job.

把它放进笔记本电脑里。

Being putting in a laptop.

这也许有可能，谁知道呢？

And it might be possible, who knows?

科技发展的速度，也许20年后，我能拥有一个10万亿参数的模型，那是我自己的模型，并且它随着我一起进化。

The pace of technology, maybe 20 years from now, I can have a 10 trillion model that it's my own model and it's getting transformed along with me.

但我们还得看看。

But we need to see.

我觉得要达到那个目标，可能还需要一系列创新。

I think there might be a few series of innovations to get there.

因为我也想给你一个有趣的数据点。

Because I want to give you also an interesting data point.

现在AI正在解决一些人类以前没解决过的非常有趣的数学问题。

AI is solving very interesting math problems that humans didn't solve before right now.

但是AI还是不能创造框架。

But AI still it's not capable of creating frameworks.

我不知道，相对论就是一个框架。

I don't know, relativity is a framework.

对吧？

Okay?

我在想，为什么会这样呢？

And I was thinking, why so?

我觉得一个主要原因还是跟这个有关，就是它没有在工作过程中被转化。

I think one of the main reason relates still with this not being transformed when you are on a job.

当我写书的时候，我写书这个过程本身就在改变我。

When I'm writing a book, I am being transformed by the act of writing this book.

每次我写下一些东西，我内心有些东西在变化，这不一定是记忆上的变化。

Every time I'm writing something down, there is something in me that changes that is not necessarily the memory of thing.

是我自己在改变。

It's me that is changing.

爱因斯坦因为思考光速、以及当你超越光速时会发生什么，而发生了改变。

Einstein has been changed by his experience thinking about the speed of light, of what happens when you go behind the light.

并不是爱因斯坦把它写下来，然后每次再思考时，又重新写一张纸。

It's not like Einstein wrote it down, and then every time he thought again, he rewrote a piece of paper.

不，他逐渐变成了一个不同于最初开始思考这个问题的爱因斯坦，当他创造出这个思维框架、这个相对论框架的时候。

No, he became gradually a different Einstein than the one that started to think of a problem when he created this thinking, this frame of relativity.

模型不是这样运作的。

Models don't do this way.

即使我放一群智能体，所有东西，它们都必须把一切写下来。

Even if I put swarm of agents, everything, they have to write down everything.

它们不会因为过程而改变。

They are not being transformed by the process.

所以，这非常困难。

So therefore, it's very difficult.

最终，它们会有这个上下文。

In the end, they will have this context.

要超越这一百万个令牌的上下文窗口是非常困难的。

It's very hard to go beyond this one million tokens context window.

一个框架可能需要在处理过程中进行转变。

A frame might require a transformation as you work on that frame.

这是一种不同于纯粹记忆的学习方式。

It's a different way of learning than pure memory.

所以我想把这个论点尽可能讲清楚。

So I want to make this argument as clear as possible.

你对你的孩子们乐观吗？

Are you optimistic for your children?

我对自己非常乐观，哈利，因此也对我的孩子们乐观。

I'm extremely optimistic for myself, Harry, therefore, for my children.

我不想听起来像个AI末日论者，因为我相信——

I don't want to sound like an AI doomer man, because I believe that-

我觉得你不是那样的人。

I don't think you do

是的。

I am.

不，我觉得你不是。

No, I don't think you do.

我听起来有点像悲观主义者。

I sound like a doomer in a way.

我觉得会有更多的失业。

I think we'll have a lot more job loss.

而且我觉得会来得更快。

I think it will happen a lot quicker.

我觉得我们现在就能看到。

I think we're seeing it in real-time.

我乐观得多，觉得不会那么严重。

I'm much more optimistic that we won't have so much.

因为根据我自己用AI的经验，我觉得扩散速度没你想的那么快，尤其是因为企业必须更详细地记录他们的流程。

Because based on my own experience with AI, I don't think the diffusion is as fast as you imagine, particularly because enterprise have to document in much greater detail their processes.

我能问一下吗，不好意思，我们都是欧洲人，而且都坐在伦敦。

Can I just ask, sorry, we're both Europeans and we're both sitting in London.

我不知道该怎么说，但我们不再重要了。

I don't know how to say this, but we don't matter anymore.

我就直说了。

Just being blunt.

你觉得未来三到五年，这种情况会变好还是变坏？

Do you think that gets better or worse in the next three to five years?

是的，老兄。

Yes, man.

很难承认现实，但从技术角度来看，我觉得我们基本上无关紧要了。

It's hard to admit the reality, but from a technology standpoint, I think we are largely irrelevant.

但这很愚蠢，因为制造芯片机器的最大生产商就在欧洲。

But it's so stupid because the biggest producer of machines that make chips is based in Europe.

是阿斯麦。

It's ASML.

我们本来可以在欧洲造这些芯片的。

We could have made these chips in Europe.

一些最聪明的人工智能人才，比如Dario、Sam，其实都是欧洲裔。

Some of the most brilliant minds that build AI, even if you think Dario, Sam, all of them are European origins.

还有Ilya。

Ilya.

我们有人才，有技术造机器，但不知怎么的，我们正在失去这些优势。

We have the talent, we have the technology to build the machines, but somehow we are losing it.

这真的很愚蠢。

And it's very stupid.

你觉得美国团队和英国团队在工作态度上有区别吗？

Do you see a difference in work ethic having a team in the US and the UK?

有。

Yes.

我和英国团队合作过，下午五点他们全去酒吧了。

I experience with teams in UK, and at 5:00 PM they are all in the pub.

这是为什么呢？

Why is that?

是因为钱更重要吗？

Because money matters more?

我觉得文化比钱更重要。

I think culture matters more than money.

那边的文化更有活力。

It's a more dynamic culture.

我觉得如果我在欧洲，我不会像在美国那样成功。

I don't think I would have succeed in Europe the way I did in US.

所以我是欧洲人，但作为创业者，我是美国人。

So I'm a European, but as an entrepreneur, I'm American.

这是我经常跟别人说的。

This is what I told to everybody.

所以我的创业理念是在美国学的，即使我的公司是在欧洲创立的。

So my formation is in American school of entrepreneurship, even if I started my company in Europe.

听着，我明白你的意思。你看看Legora，看看Eleven Labs，还有近几年欧洲冒出来的那些顶尖公司，要是你觉得赚钱的机器不在美国，那你就是在自欺欺人。

Listen, I get it, and you look at Legora and you look at Eleven Labs and some of the best companies to come out of Europe in the last few years, and if you think the revenue machine is anywhere but in America, you're lying to yourself.

显然，在美国赚钱比在欧洲容易得多。

It's an easier to access revenue machine than in Europe, clearly.

更快，也更容易。

Faster, it's easier.

这些团队以前就做过市场拓展。

The teams have scaled GTM functions before.

我完全同意你的看法。

I completely agree with you.

而且美国公司更敢于在愿景上下大赌注，不会像欧洲公司那样等那么多验证。

And American companies are making larger bets on vision Without waiting for so many proof points as the European companies.

在美国，甚至中层管理人员都能对新科技做出几百万美元的大手笔投资。

And even people in the middle management can make sizable million-dollar bets on new technologies in the US.

我在欧洲没见过这种魄力。

I haven't seen this appetite in Europe.

我不想问这个，但我确实很好奇。

I don't want to ask this, but I am interested.

如果你今天要给一位年轻的欧洲创业者提建议，你会建议他们去美国吗？

If you were to advise a young European entrepreneur today, would you say to go to the US?

会。

Yes.

这就是残酷的现实。

That's the sad reality.

我觉得，除非他们是为了某个特定市场、带着明确的目标去创业，如果他们做的是通用技术，那在美国成功的几率会更大。

I think that unless they build for a specific market with some specificity in mind, if they build a universal technology, they will have a better chance to succeed in the US.

未来会有很多成功的欧洲公司涌现出来，可能不是前沿实验室，但我觉得在AI应用领域会很多。

There will be many successful European companies coming out of this, maybe not as frontier labs, but I think for the application of AI.

你认同主权这个论点吗？

Do you buy sovereignty as an argument?

认同。

Yes.

我觉得这个论点很重要。

I think it's an important one.

能源主权、模型主权。

The energy sovereignty, model sovereignty.

对，百分之百。

Yes, 100%.

而且现在所有欧洲客户都更倾向于本地部署的软件模型主权和模型可选择性。

And all European customers right now would prefer on-prem software model sovereignty and model optionality.

对，百分之百。

Yes, 100%.

这里即将迎来一个大市场。

This is a big business that it's coming here.

你看，我跟Fireworks的朋友聊过，我其实一直在劝他们提供本地部署的软件版本。

Look, I talk to our friends at Fireworks, and I actually try to convince them to make their software available on-prem.

现在他们还在观望，但我可以告诉他们：“各位，这可是个大生意。”

Right now, they are in show me the money, but I can tell, "Guys, this is a big business.

你得先证明自己，因为这里是欧洲。

You need to prove first because this is Europe.

先让他们看到技术，钱自然会来。

Show them the technology, and the money will come."

UiPath 现在一年营收多少？

How much revenue does UiPath do today?

我觉得这是公开数据。

I think it's public data.

我们现在是16亿，去年增长了大概14%。

We are at 1.6, growing last year, like 14%.

所以Jason Lemkins教过我，除非你增长率超过20%，否则在公开市场上就完蛋了。

So Jason Lemkins taught me that unless you're growing 20% plus, you're just fucked in the public market.

所以要么增长，要么死，现实就是这么残酷。

So it's grow or die, and it's a horrible reality in that way.

我不是在赞同这种做法。

I'm not condoning it.

这确实很糟糕。

It's horrible.

是这样吗？

Is that right?

是的，因为我觉得现在公开市场对谁是AI赢家或输家非常困惑。

Yeah, because I think the public markets are very confused right now of who are the AI winners or losers.

如果你没有展现出强劲的增长和吸引力，他们就会自动把你归为AI输家，而不会深入了解你的业务。

And if you don't show serious growth and traction, they automatically put you into AI losers without looking really deep into the business.

公开市场上有成百上千家软件公司，所以很难逐一去研究。

There are so many hundreds of software companies in the public market, so then it's hard to look at each of them.

那如果我把问题抛回给你，我们最后来个快问快答。

So if I were to flip it on you, we do a final one for a quick fire.

看好UiPath成为500亿美元公司的理由是什么？

What is the bull case for UiPath being a $50 billion company?

想想看，就在今天，Gartner发布了他们新的BOAT魔力象限，也就是业务编排和自动化技术。

Think about even today, Gartner released their new BOAT Magic Quadrant, business orchestration and automation technologies.

我们是领导者之一。

We are one of the leaders.

去年我们从一个挑战者变成了领导者。

We moved from a challengers into a leaders in the last year.

这说明我们公司实现了从RPA和自动化技术向编排和自动化技术的转型。

So it shows that we, as a company, made this transition from an RPA and automation technology into an orchestration and automation technology.

而且有各种理由表明，这确实是打造这种AI驱动的新型企业所必需的。

And there are all the arguments in the world that this is really required in order to create this new enterprise that is AI-powered.

那种认为可以有一个AI代理帮你处理所有事情，从高层流程到自动编排一切，像魔法一样，我觉得这种想法人们已经开始不太相信了。

This idea that you can have an AI agent that runs everything for you from top-level processes, orchestrate and automate everything by magic, I think it's kind of a thing that people stop believing.

你需要有一个底层的编排和自动化技术，以及我之前提到的那个工作地图，来支撑你的流程。

You need to have an underpinning orchestration and automation technology and this map of work that I talked in order to power your processes.

而这正是我们所拥有的。

And this is what we have.

而且这不只是我这么说，Gartner也这么说，Forrester也这么说，行业分析师们也都非常看好我们。

And it's not only me saying, but it's Gartner saying, it's Forrester, it's industry analysts, and being very bullish on us.

所以这就是现在的关键论点。

So that's really the argument right now.

我觉得，就像我说的，AI现在带来的这种不对称性比以往更明显——构建运行你流程的软件比一年前容易多了。

I think, as I said, this asymmetry that AI is creating right now is more obvious than printing software that run your processes has become much easier than a year ago.

而创建一个运行你软件的AI代理，难度和一年前差不多。

Creating an AI agent that run your software is as difficult as a year ago.

所以你要在构建这个软件上投入巨资，捕捉我们所说的“工作地图”的企业上下文，然后把这个企业上下文放进我所说的“轨道”里，也就是我提到的编排自动化，地图和轨道。

So you make a huge investment into printing this software, capturing the enterprise context that we call the map of work, putting this enterprise context inside these rails that I named this orchestration automation as the map and rails.

地图就是上下文，轨道就是编排自动化。

Map is the context, rails is the orchestration automation.

你把它们放在同一个平台上，然后就可以指派一个代理去干活了。

You put them in the same platform, and then you can assign an agent to do work.

然后你告诉代理：“这是现实情况。

And you tell the agent, "This is the reality.

这些是你可以用的轨道。”

These are the rails you can use.

这是描述如何使用这些轨道的地图。

This is the map that describe how to use these rails.

这是目标。

This is the goal."

这就是你可以控制的方式。

And that's the way you can control.

你可以在顶层进行控制，但你的代理不能失控。

You can have a control on the top, but your agents cannot go rogue.

现在没有哪个理智的企业会让一群代理直接说：‘帮我做财务记账。’

No sane enterprise right now will put a swarm of agents and just ask them, "Do my finance accounting for me."

因为谁知道呢？

Because who knows?

也许他们会攻击你的竞争对手。

Maybe they will attack your competitor.

把两千万美元转给某个……

Send 20 million bucks to some-