与Sequence Holdings联合创始人兼首席执行官Michael Lee一起，重新定义AI时代的在任者

世界上每家公司都有一个备受推崇的代表人物。

Every company on the planet has a celebrated persona.

在一个你相信阿尔法来自工程和人工智能的世界里，你需要创造一种文化，让工程师成为备受推崇的人物，这就是所需要的。

In a world where you believe that alpha comes from engineering and AI, you need to create a culture whereby the celebrated persona is the engineer, and that's what's required.

想想你对基金的典型投资，你通常试图弄清楚如何开始包装，以便在三年内出售这个东西。这与我们如何看待技术转型以及你愿意在这里投入哪些投资是非常不同的框架。

You think about your typical investment in a fund, and you're typically trying to figure out how do we start to package to sell this thing in three That is just a very different framing for how we think about technological transformation and what what are the investments you're willing to bring to bear here.

我们所做的事情是稀缺的。

What we do is scarce.

我们试图每年做一笔交易。

We're trying to do one deal per year.

就这样。

That's it.

我们的工作就是完成一笔。

Our job is to do one.

嗨，听众朋友们。

Hi, listeners.

欢迎回到《No Priors》。

Welcome back to No Priors.

今天，我请到了迈克尔·李，他是Sequence Holdings的联合创始人兼首席执行官。

Today, I'm here with Michael Lee, cofounder and CEO of Sequence Holdings.

他们刚刚宣布了迄今为止最大的一笔AI私有化交易，与戴尔家族办公室Baldwin合作，金额高达77亿美元。

They just announced the largest AI take private to date with the Dell Family Office of Baldwin for $7,700,000,000.

我们聊了聊这个，Sequence的故事，AI转型将如何真正渗透到经济中，以及从公开市场投资到私募股权再到运营的转变。

We talk about this, the Sequence story, how AI transformation is actually going to permeate the economy and what transfers from public investing to private equity to operating.

欢迎你，迈克尔。

Welcome, Michael.

迈克尔，感谢你来做客。

Michael, thanks for doing this.

谢谢你的邀请。

Thanks for having me.

你创立了Sequence Holdings，这是一家永久控股公司，与管理团队合作，收购并重塑他们的业务，使其成为AI领域的市场领导者。

So you started Sequence Holdings, which is a permanent holding company that works with management teams to buy and refound their businesses to be market leaders with AI.

我们得聊聊这个，这个模式确实很令人兴奋。

We should talk about that, and it's a very exciting model.

但你还宣布了和Baldwin的那笔大交易。

But you also announced this big transaction in Baldwin.

给我们讲讲吧。

Tell us about it.

你知道，我们二十个月前创办这家公司，当时就是想看看能不能和世界级的企业及管理团队合作，和我们的前沿工程团队紧密配合，打造出市场领导者。

You know, we started the company twenty months ago really with an eye towards can we partner with world class businesses and management teams, working closely with our frontier engineering team to forge market leaders.

经纪业务这个领域，从公司成立以来我们就一直在花很多时间研究。

The brokerage space is an area that we've been spending a lot of time on since the founding of the business.

我们见过几十个保险经纪公司，但Baldwin真的是我们见过的所有保险经纪公司里独一无二的一个。

We've met dozens of insurance brokers, and Baldwin was truly N of one of all the insurance brokers that we've met.

过去几个月，能和Trevor Baldwin及其管理团队共度那么多时间，真的是一种荣幸。我们对未来我们、Baldwin团队以及支持这笔交易的Dell家族办公室之间的合作旅程，感到无比兴奋。

It was it's been a privilege over the last few months to spend a lot of time with Trevor Baldwin and the management team, and we could not be more excited about the journey that lies ahead between us, the Baldwin team, and the Dell Family Office that supported us in this transaction.

我想把话题拉远一点，聊聊你的背景，你和联合创始人Alex是怎么想到创办Sequence的，因为你之前是投资人，现在算是投资人和运营者的混合体。

I want to zoom all the way out to a little bit about, you know, your background, you and your co founder, Alex, how this idea for Sequence came about, because you've been, you know, an investor, now, you know, a hybrid investor operator.

跟我说说，你最初是怎么想到这个主意的？

Like, tell me about how you arrived at the idea to begin with.

你知道，2017年我刚加入Lone Pine的时候，最早让我负责的领域之一就是人工智能。

You know, when I first joined Lone Pine in 2017, one of the first areas that I was asked to cover was AI.

那其实挺早的。

That's actually quite early.

确实，我记得那时候正好是AlphaGo出来的时候。

It was, you know, as I'm sure you recall, it's kind of when AlphaGo came out.

也是第一篇transformer论文发表的时候，当时大家对扩大这些架构规模后能有什么商业价值非常兴奋。

It's actually when the first transformer paper came out, and there was a lot of excitement at the time around what would be commercially viable if you scale these architectures.

那时候主要是一些卷积神经网络、GAN、LSTM之类的。

And at the time, these were convolutional neural nets, GANs, LSTMs, and the like.

而且，我觉得大家都知道后来怎么样了，显然我们看到了很多令人兴奋的成果。

And, you know, I think we all know how the story ends, which is we obviously saw a lot of exciting things come out of it.

但是当你扩大这些架构的规模时，随着时间推移，规模扩大也会带来一些负面效应。

But as you scaled these architectures, there were degrading effects over time to scale.

我之所以提这些背景，是因为当模型或者ChatGPT在2022年底出来的时候，至少对我来说，很明显世界已经永远改变了。

You know, I share that context because, you know, when the models came out or ChatGPT came out at the end of twenty two, you know, what became obvious to me at least is that the world had forever changed.

我们终于有了一个我们知道可以无限扩展的架构。

You know, we finally had an architecture that we knew could scale infinitely.

至于那是否可能，以及会带来什么，当时并不清楚，但我们确实知道，只要有算力，我们就能扩展，架构也会改进。

Now, whether or not that was possible, what would come from that was not clear, but we knew we did know that we could scale with compute, that the architecture would improve.

作为投资者，我开始思考这对世界的影响，我强烈感觉到AI对经济的影响会是不均衡的。

And, you know, like an investor, I started to think about what were the impacts to the world, and I kind of felt strongly that AI would have an uneven impact on the economy.

我觉得经济中有些部分AI根本不会触及。

There are certain parts of the economy that I felt that AI just wouldn't impact at all.

比如餐馆、高尔夫球场之类的。

Take restaurants, golf courses, things like that.

有些行业我觉得初创公司会赢。

There are certain industries that I just felt like a startup would win.

比如编程。

Take coding, for example.

像Conviction和Sequence去买一个外包编程服务的公司，听起来就是个糟糕的主意，你会把钱给Cognition或者Anthropic。

Like Conviction and Sequence going to go buy an outsourced coding services business feels like a terrible idea, which you will give money to Cognition or toanthropic.

但作为研究商业的人，我一直觉得有些行业里，现有企业拥有所有优势，可能是品牌、规模、网络效应或者监管。

But there, as a student of business, there have always been certain industries where I felt like the incumbent had all the advantages, and that could be brand, that could be scale, network effects, regulatory.

面对我们一生中最大的技术变革，我强烈地感觉到，如果你能收购合适的现有企业，继承它的优势，你能打造出市场领导者吗？

And in light of what is going to be the biggest technological shift of lifetime, I felt very strongly that if you could acquire the right incumbent, inherit the advantages of being that incumbent, could you forge the market leader?

这个想法的核心，正是Sequence创立的根基。

And the kernel of that idea is the one that really is upon which Sequence was founded upon.

我在2023年初就有了这个想法，当时有几个人想让我离开Lone Pine去做这件事。

Came up with the idea in early twenty three, had a handful of people try to get me to leave Lone Pine to go do it then.

但那时对我来说时机不对。

Just wasn't the right time for me.

然后我记得在2024年和我妻子散步时，我告诉她，嘿，人生中很少能赶上我们这一生最重要的技术变革。

And then I remember going for a walk with my wife in '24, letting her know that, hey, it's pretty rare in life to be in the middle of the most important technological change of our lifetime.

有了一个我深信不疑的想法，并且我坚信我能组建团队去实现它。

Having the kernel of an idea, I felt very strongly about, and feeling very strongly that I could pull together the team to go and pull it off.

所以这就是我们创办Sequence的原因，我们对目前取得的成就以及未来的前景感到无比兴奋。

And so that's why we started Sequence and couldn't be more excited about kind of what we've been able to achieve to date and kind of what we see ahead of us.

这是一种非常从第一性原理出发的思考方式，你知道，就是这项技术发生的宏观链条是什么？

It's a super first principles approach to, you know, what is the macro chain to this happening technology?

它将如何影响经济？

How's it going to impact the economy?

但是，你知道，很多投资者，至少是风险投资者，可能会想，哦不，现有企业有一大堆优势。

But, you know, many investors, at least venture investors, would be like, oh, no, incumbents have a bunch of advantages.

我觉得这不太适合我们。

I guess it's just not for us.

你是怎么考虑规模以及能够与现有企业合作甚至收购它们的想法的？

How did you think about, you know, scale and the idea of being able to engage with or even own incumbents?

这确实，我想你可以直接去做，但这可是一个雄心勃勃的前提。

That's a, I guess you can just do things, but that's an ambitious premise.

我认为这更像是适应当前机遇的商业模式。

I think it's more just it's the business model that meets the opportunity today.

所以如果我们把视角拉得很大，你知道，至少在Sequence，当我们思考今天AI的发展时，我们认为我们正在经历相当于下一次工业革命的时期。

So like if we take a giant step back, you know, and at least at Sequence, when we think about what's happening with AI today, we think we're living through the equivalence of the next industrial revolution.

而且，你知道，我们在大多数企业中普遍看到的，以及我从今天许多生产AI代理的软件公司那里通常看到的，是这种巨大的推动力，要给人类流水线上的每个工人配发小机器，以加快工作速度。

And, you know, what we generally see in most enterprises, what I typically see from a lot the software companies that are producing AI agents today, is this giant push to give small machines to every human in the human assembly line and to speed up work.

这没什么不对的。

And like, there's nothing wrong with that.

这很好。

That's great.

总比没有强。

It's better than nothing.

但你知道，当你有可以全天候工作、可以靠电力扩展、可以完成任何个人或群体都无法完成的任务的机器时，正确的做法是思考如何开始重新组织一个组织，以抓住技术带来的机遇。

But as you know, when you have machines that can work 20 fourseven, that can scale with electricity, that can accomplish what no individual human or group of humans can do, the right answer is to think about how do you start to reorganize what an organization should be to meet the opportunity of what's available to you with technology.

这就是我们所处的背景。

And that's the context upon which we're sitting in.

而且我认为，真的第一次，我们有一个真正的机会与优秀的公司、优秀的领导团队合作，与我们的平台和工程团队携手，认真思考如何重新组织这个组织。

And I think for the, really in the first time, we have a real opportunity here to partner with amazing companies, with great leadership teams, partnering with our platform and our engineering team, and really thinking critically around how do we reorganize the organization?

我们怎么才能充分利用技术能做到的最好的部分，人类能做什么，然后重新思考我们怎么在这些行业里竞争？

How do we lean into what's best in terms of what the technology can do, what can humans do, and rethink how we can compete in these various industries?

我觉得这更多是一个机会的问题，而不是雄心的问题，因为我觉得这件事必须发生，而且我看不到有什么自然的方式能让它发生，除非通过我们的方法来做。

And I think it's just really more of an opportunity piece more so than an ambition piece in that, like, I just think this has to happen, and I don't see it a natural way for it to happen unless you do it through our approach.

你能谈谈公司的结构吗？

Can you talk a little bit about the structure of the company?

为什么是永久控股公司？

Like why a permanent holding company?

你知道，拥有所有权能让你做到哪些作为供应商或合作伙伴做不到的事情？

You know, what does ownership enable you to do that being a, you know, a vendor or partner doesn't?

是的，我觉得这个问题有两部分。

Yeah, I think there's two parts to that question.

关于控股公司结构本身，你知道，当我思考要完成Sequence Holdings想做的事情需要什么，也就是我们怎么和世界级的组织合作？

So in terms of like the holding company structure itself, you know, when I thought about what is required to accomplish what Sequence Holdings wants to do, which is how do we partner with world class organizations?

我们怎么把前沿工程发挥出来？

How do we bring frontier engineering to bear?

我们怎么建立一个可以在所有投资组合公司之间共享的平台？

How do we build a platform that we can share across all of our portfolio companies?

要做到这一点，唯一自然的方式就是通过控股公司。

The only natural way to do this was through a holding company.

我们怎么打造一个能够锻造市场领导者的业务？

How do we build a business whereby we can forge market leaders?

我们怎么让资本的期限与推动整个转型所需的持续投资和运营承诺相匹配？

How do we align the duration of capital with the sustained investment and operational commitments that are going to be required to drive the transformation end to end?

我们怎么创建一个在文化上以锻造市场领导者为导向的组织，而不是以部署资本为导向？

How do we create an organization that culturally is oriented around forging market leaders, not deploying capital?

最后，我们怎么建立正确的结构，以便利用企业中的留存收益，重新思考如何支持现有公司和新投资。

And lastly, like how do you create the right structure upon which you can take the retained earnings in your business to rethink about how do you support your existing companies and net new investments.

所以当我看到面前的机会时，唯一自然的做法就是通过控股公司来实现。

And so when I think about the opportunities in front of us, the only natural way to go do this was through a holding company.

关于你的第二个问题，也就是所有权的问题，我们不妨通过一个例子来说明。

In terms of your second question around, like, ownership, you know, let's maybe just play this out through an example.

假设你是一家非常出色的公司。

Let's say that you're an incredible business.

我不会点名批评哪一家。

You know, I won't pick on one.

假设我们是一家财富500强公司。

Let's say we're a Fortune 500 company.

你拥有无限的资源。

You have unlimited resources.

如果我挑战你说，嘿，你今天是一家财富500强公司的CEO，你的任务是重新创立自己，你只有三个选择。

The if I challenge you to say, hey, you're the CEO of this Fortune 500 company today, and your job is to refound yourself, you're stuck with three options.

第一个选择是我自己来做。

One is I'm gonna go do it myself.

这样做在结构上的挑战，虽然不算坏，但就是很难招募和留住重新创业所需的人才。

The structural challenge with that, and it's not a bad one, is just it's impossible to recruit and retain the talent required to go and do this refounding.

其实答案比人们想的要明显得多，那就是地球上每家公司都有一个备受推崇的人物形象。

And the the answer to that is actually much more obvious than people think, which is every company on the planet has a celebrated persona.

比如说黑石集团吧，那是个非常厉害的组织，我非常敬佩。

So let's take Blackstone, for example, incredible organization, one that I admire a lot.

实际上，我让每个人都经历一个过程，让你能成为资本之王。

And in fact, I make every person a sequence where you can king of capital.

但黑石最受推崇的角色是投资人。

But the celebrated persona of Blackstone is the investor.

这就是为什么他们能聚集全球最顶尖的投资人。

And that's why they're able to aggregate the greatest investors in the world.

如果你相信阿尔法收益来自工程和人工智能，那你就需要创造一种文化，让工程师成为最受推崇的角色。

In a world where you believe that alpha comes from engineering and AI, you need to create a culture whereby the celebrated persona is the engineer.

这就是关键所在。

And that's what's required.

这也是Palantir存在的原因。

It's the reason why does Palantir exist?

我知道亚历克斯·卡普肯定不喜欢我这么说，但Palantir本质上就是一个聚集了世界级工程师、并且推崇工程师的组织，而他们的核心业务就是把这些能力包装起来，卖给那些自己招不到这样人才的公司。

You know, I'm sure Alex Karp would hate how I would describe this, but it is foundationally an organization that has aggregated world class engineers that celebrates the engineer, and foundationally, they do is it's a wrapper to sell it to other organizations that are not able to get that talent.

所以这就引出了我的第二点。

So that leads me to my second point.

如果你自己招不到这样的人才，那就有一个巨大的、价值数万亿美元的行业，我称之为服务业。这些公司都很棒，像埃森哲、麦肯锡、Palantir，它们确实重视工程师，也确实重视技术和变革。

If can't get the talent to do it yourself, then there's this huge multi trillion dollar industry that I got formed called the services industry, and these are great companies, Accenture, McKinsey, Palantir, that have do celebrate the engineer, that do celebrate change in technology.

但是，当你和这样的服务提供商合作，想要重新思考和重塑你的业务时，你面临的最大挑战是激励机制的问题。

But the challenge that you have when you partner with a service provider in terms of rethinking and refounding your business is an incentives problem.

服务公司通常追求三件事：进入你的钱包，留在你的钱包里，以及扩大在你钱包里的份额。

Services companies optimize for kind of three things: getting in your wallet, staying in your wallet, growing the share of your wallet.

这是一条通往渐进主义的道路。

It is a path towards incrementalism.

更糟糕的是，服务提供商实际上可以改变那些你真正需要的东西，以便真正应对当下的挑战。

To compound that, services providers can actually change the things that you really need to really kind of meet the moment.

你如何改变人员的组织方式、谁在这些组织中工作，以及激励机制？

How do you change how the people are organized, who's working in these organizations, and the incentives?

所以，如果你自己做不到，也不能用供应商，那你还有第三个选择，这也是很多人会做的，那就是去买软件。

And so if you can't do it yourself, you can't use vendors, and you have the third option, which is what a lot of people do, which is they go buy software.

买软件的问题有两方面。

The challenge with buying software is twofold.

显而易见的一点是，它还不成熟。

The obvious one is that it's beta.

如果谁都能买到，那它就没有什么独特性了。

If it's available to everybody, it's available to everybody.

我觉得更微妙、也常被人忽略的一点是，如果咱们俩要创办一家软件公司，我们会优先考虑的是：什么样的工作流程在很多地方都存在且相对统一，然后我们怎么才能既快速又深入地部署它，这样既能大量销售，又能让客户离不开我们？

I think the part that's more nuanced that I think people often miss is that if you and I were starting a software company, the thing we would optimize for is what is a workflow that's reasonably homogenous that exists in a lot of places, And then how do we install it fast enough but deeply enough such that we can sell a lot, but it's sticky?

这些就是

Those are the

基因层面的，结构上确实如此。

genetics Structurally, yes.

听起来很有吸引力。

That sounds very attractive.

那就是我们会寻找的基因。

That would the genetics of what we'd look for.

但这意味着，你总是要针对今天的设计流程来销售，也就是按照今天人类流水线的设置来卖。

What that means, though, is that you're always going to sell to a workflow as it's designed today, as humans are set up in that human assembly line today.

这是唯一能卖出产品的方式。

It's the only way you can sell a product.

你不能围绕一个今天还不存在的新人类流水线来卖产品，因为那只是你自己的想法。

You can't sell a product around a new human assembly line that doesn't exist today because that's how you think it is.

所以当我们思考为什么我们的模式有效时，关键在于如何拥有所有权，从而保持一致。

And so when we think about why does our motion work, it's how do you have ownership so therefore you're aligned?

如何创建一个像Sequence Holdings这样的组织，来鼓励工程师，并发挥前沿工程的力量？

How do you bring create an organization like Sequence Holdings that celebrates to engineer and bring frontier engineering to bear?

第三，如何推动长期视野和经济模式，使我们在如何利用现有技术打造公司最佳版本方面与管理层保持一致。

And then three, how do you drive a long term horizon and economic model such that we're aligned with management around how do we think about how do we forge the best version of this company possible based on the technology that's available today.

这就是我们思考商业模式的方式。

And so that's how we think about our business model.

这也是我们认为它有效的原因。

It's why we think it works.

而且，从BankSouth的早期情况以及我们在实地观察到的现象来看，这一点已经得到了验证。

And the early evidence from kind of BankSouth and kind of what we've observed in the field today have proven that out today.

我也想聊聊BankSouth。

I want to talk a little bit about BankSouth as well.

不过，你知道，你的团队里有很多来自Scale和Palantir背景的人。

But in, you know, you have a team that includes a bunch of like scale and Palantir DNA.

你一直在努力保持很高的质量标准。

You are trying to keep the quality bar really high.

那对于在Sequence工作的工程师们来说，什么最吸引他们呢？

What is compelling to the team, the engineers that work at Sequence?

我觉得有很多因素。

I think it's a bunch of things.

首先，这真的是一个机会，可以参与到那些真正与人们日常生活息息相关的公司中去。

I think the first is really it's an opportunity to kind of work on the companies that actually are the capillaries upon which people interact with.

你知道，那些AI模型公司确实很出色。

You know, the AI model companies are phenomenal.

如果没有Anthropic、OpenAI、xAI这些公司，我们今天也不会站在这里。

We would not be here today if it weren't for Anthropic, OpenAI, xAI, and the like.

但实际上，真正影响大家日常生活的，还是那些每天为他们服务的公司。

But the reality is, is like the impact to everybody's day to day lives really are going to be impacted by the companies that serve them today.

所以对我们来说，这是一个机会，能和那些在经济中扮演重要角色的公司合作，真正思考如何把世界级的技术用起来，为客户提供独特的体验。

And so for us, it's the opportunity to partner with really important companies that play an important role in the economy, and really thinking about how do we bring world class technology to bear to deliver differentiated experiences for their customers.

我觉得这个使命挺有力量的。

And I think that mission is like quite powerful.

我觉得，当我们广泛地考虑我们通常招聘的两种工程人才时，一种是部署工程师，吸引他们的地方在于，他们能直接看到自己创造的价值。

I think, you know, when I think broadly about the two engineering personas that we typically hire from, I think, you know, call it the four deployed engineers, the, I think that the thing that appeals to them is the opportunity to be really aligned with the value that they create.

我觉得这是一种非常 rewarding 的体验，而且能够推动那种作为服务提供商不可能实现的变革，这真的很有吸引力。

And I think that's an extremely rewarding experience, and I think the ability to drive the type of change that would not be possible if you were a service provider is quite appealing.

另一方面，可能对风险投资人来说不太有趣的是，当我跟应用软件工程师聊天时，我会说，嘿，我们是世界上少数几个完全不受模型影响的商业模式之一。

I think the other hand, which is probably not as fun for venture capitalists to think about is like, you know, when I talk to application software engineers, it's, hey, we're one of the few business models in the world that's completely immune to what the model's going to do.

随着模型性能的提升，我们在Sequence这里庆祝，因为这给了我们更多工具，可以用于我们合作的伙伴公司。

As the model performances improve, we celebrate here at Sequence, and that like, that gives us more tools to bring to bear in the companies that we partner with.

而且，我们的业务会不会一直存在，其实没有什么真正的生存风险。

And like, there's no real existential risk to whether or not our business will be around.

我觉得这些就是比较吸引人的地方。

And I think those were kind of things that have appealed.

我觉得最终驱动我们的还是使命，但确实，当我们和候选人交流的时候，有一些实实在在的东西对我们想做的事情是很有吸引力的。

I think the mission is probably the thing that drives it at the end the day, but I do think there are tangibles when we talk with candidates that have been quite appealing for kind of what we're trying to build.

我记得你第一次跟我说，嘿，我们打算买一家银行的一部分，和BankSouth团队合作。

I remember when you first told me, hey, we think we're gonna go buy a piece of a bank and partner with the BankSouth team.

我当时可能有点震惊，觉得这事儿难度大，还有监管方面的顾虑。

I might've reacted with some mild horror about the difficulty and regulatory concerns around that.

我当时就想，迈克尔，你得知道，银行业传统上不是私募股权热门领域，这是有原因的。

And I was like, Michael, you must know that there's a reason that banking is not a popular private equity sector traditionally.

跟我们说说，你是怎么选择金融服务的，还有到目前为止，你在和BankSouth的合作中学到了什么？

Tell us a little bit about how you, you know, chose financial services and then what you've learned in the BankSouth partnership so far.

所以，我觉得值得退一步，想想这笔投资是怎么来的，然后也许再谈谈为什么以及怎么就成了我们创业的完美起点。

So, you know, I think it's worthwhile to take a step back and thinking about how did this investment come to be, and then kind of maybe to tackle kind of how and why it's been kind of the perfect place for us to start our business.

所以，你知道，当我们去年三月开始这项业务时，你第一次写下第一笔投资，Sequence，我们面临一个冷启动的问题，对吧？

So, you know, when we started this business, call it March of last year, when you first kind of wrote the first investment, the Sequence, you know, we had a cold start problem, right?

控股公司模式是个奇怪的业务。

So the holding company model's a weird business.

没有钱，你就做不了交易。

Like, without money, you can't do a deal.

没有交易，你就招不到工程师。

If you don't have a deal, you can't hire engineers.

招不到工程师，他们就会要求我们给你钱。

If you can't hire engineers, they want us to give you money.

幸运的是，你愿意支持我们。

Thankfully, you you were willing to support us.

你知道，我们能够招到一些工程师来和我们一起工作，但我们仍然被冷启动问题困住，没什么真正可做的。

You know, we were able to get a handful of engineers to come work with us, but yet we were stuck with kind of the cold start problem of nothing to really to do.

当时我们争论的一件事是，我们是直接去买点什么来向世界证明我们能买并改变它，还是等待合适的资产？

And one of the things that we were debating at the time was do we just go buy something to prove to the world that we can buy something and change it, or do we wait for the right asset?

而且我想你也知道，我们一直倾向于做大规模和企业级的业务。

And as I think you know, we've always had a bias towards scale and enterprise.

不幸的是，世界上没有多少投资者愿意给你几亿美元去买一家企业，就为了看看你的技术行不行。

Unfortunately, there not are a lot of investors in the world that want to give you hundreds of millions of dollars to go buy an enterprise to see whether or not your technology works.

所以我们决定先找一个客户，于是联系了一位好朋友，杰米·雷诺兹。

And so what we decided to do is to go and take on a customer, and, you know, called a dear friend, Jamie Reynolds.

他是Avenir的联合创始人之一，我基本上跟他讲，嘿，我遇到了这样一个问题。

He's one of the co founders of Avenir, and basically framed for him, hey, like, here's the problem that I've run into.

非常幸运的是，他家在佐治亚州正好拥有一家银行，就这样，它成了我们的第一个客户。

And very fortunately, he his family happens to own a bank down in Georgia, and that's how it became our first customer.

去年八月开始的。

Started off August of last year.

确切地说是8月4号。

In fact, was August 4.

那天是我的结婚纪念日。

It was my anniversary.

我记得那天我跳过了庆祝。

I remember skipping it.

那真是一次难忘的经历。

And it was an incredible experience.

所以从去年八月到差不多十一月下旬，我们基本上是以服务模式在运作，真正开始处理银行里那些关键的工作流程。

And so from August to about November, like kind of late November of last year, basically operated in a services motion and, like, really started to tackle kind of the key workflows that existed at the bank.

幸运的是，那时候那家人问我们是否愿意成为长期合作伙伴，就这样，这成了我们的第一笔投资。

Fortunate to us, you know, the family asked us at that point in time whether or we'd become a permanent partner, and that's how it became our first investment.

我们最终在三月份完成了这笔投资。

We ended up closing the investment in March.

我们花了一些时间来协商所有细节，并且获得了OCC的批准。

It took us a little while to kind of negotiate everything as well as get fed an OCC approval.

但我分享这些背景是想说，从服务模式到成为投资者，我们看到了天壤之别。

But I think it's I share all that context to say that like operating in a services motion and then being an investor, we saw a night and day difference.

这涉及很多方面。

It's a lot of things.

首先从工程角度来看，你知道你会在这里待很长时间。

One is from an engineering perspective, you know you're going be here for a long time.

所以，你愿意承担的复杂度和深度，与作为服务提供商时——知道自己最终会离开，必须把别人要维护的应用程序留下来——是截然不同的。

So the level of complexity and depth that you're willing to take on versus being a service provider, knowing that you'll eventually leave and have to leave behind applications that other people have to maintain is demonstrably different.

第二点是，员工们非常兴奋，觉得这些人会长期支持我们。

The second thing is just a lot of excitement from the employee base around, hey, these people are here to support us for a long time.

那我们怎么推动这个最雄心勃勃的版本呢？

So how do we push the most ambitious version of this?

第三点，我们真的可以采取长期导向，思考这家银行三年、五年、十年后应该是什么样子，以及如何打下技术基础，让我们始终受益于模型性能。

And then three is, like, really, we can take a long term orientation around, like, what do we want this bank to look like three years, five years, ten years from now, and how do we lay the technological foundation such that we're always benefiting from model performance?

这就是我们能够做到的。

And that's what we've been able to do.

从某种奇怪的角度来说，这家银行是我们验证商业模式所需的完美试点。

In a weird way, the bank was the perfect pilot for what we needed to do to validate our business model.

它让我们能够购买少数股权，也就是以风险投资规模的支票，在真正的企业级环境中运营。

It was an ability to buy into minority stakes, so a venture scale type check to operate in a real enterprise like environment.

而且有意思的是，银行的监管性质反而成了优势，而不是问题。

And in a weird way, the regulated nature of the bank was a feature, not a bug.

受监管机构的一个好处就是，它的运作方式非常明确。

One of the nice things about a regulated institution is that how it operates is well defined.

数据管理也做得非常好。

The data hygiene is excellent.

业务怎么运作都有明确的规定。

There are well defined rules around how your business is supposed to operate.

所以在这种环境下，对智能代理来说其实特别合适。

And so if you think about that context, it actually works extremely well for agents.

这是其中一点。

That's one piece.

另外一点对银行业特别重要，也让我们在投资思路上学到了很多，就是它的集中化特点。

The other piece that's really important for banking, which has been quite a learning for us in terms of how we think about investing, is a centralized nature of it.

所以我们在Sequence讨论要找什么样的公司时，经常提到组织物理这个概念。

So when we think talk at Sequence about companies we're looking for, we talk a lot about organizational physics.

我们喜欢那种结构紧凑、运营集中的组织。

We like organizations that are quite dense, and the operations are centralized.

这样你做的任何东西都能在庞大的基础上摊薄成本。

So anything you build can be amortized over a large base.

比如说银行吧。

So like take a bank, for example.

银行可能有很多分行，但所有的审批都是集中进行的。

A bank may have a bunch of different branches, but all the underwriting happens centrally.

所以我们在总部做的所有东西，都能分摊到所有合作的分行上。

So everything we build at headquarters gets amortized over all the branches that they're working with.

你

You

但对比一下现在市场上很多并购整合的公司，我觉得那是个很棒的战略，但其中有很多复杂性。

contrast that with, you know, I think, you know, a lot of the roll ups that we see out in the marketplace today, which I think is an awesome strategy, but there's a lot of complexity with it.

你得整合不同的系统，派工程师去很多不同的地方，还得统一操作流程，统一企业文化。

You have to integrate different systems, you have to send engineers to a number of different places, you have to standardize operating procedures, you have to standardize cultures.

对我们来说，银行的好处在于，虽然它是一个大型组织，销售额超过一亿美元，但业务的核心结构非常紧凑，所以我们能在短时间内取得很多进展。

And for us, the nice thing about the bank is that while it's a large organization with, you know, dollars 100,000,000 plus of sales, like the physics of the business are quite dense, and therefore we've been able to make a lot of progress in a short period of time.

它有一个中枢神经系统，有一本总账，有一些关键流程，还有一些可以在经济上操作的杠杆。

There's a central nervous system, there's a ledger, there's some key processes, there are levers you can pull economically.

完全正确。

Exactly right.

为什么选择保险经纪？

Why insurance brokerage?

自Sequence成立以来，经纪领域一直是我们投入大量时间的领域。

So the brokerage space has been an area that we've been spending a lot of time in since the founding of Sequence.

它符合我们寻找的很多条件。

It fits a lot of the things that we look for.

那么，作为一家组织，我们广泛寻找的是什么？

And so what do we look for broadly when we're as an organization?

我们考虑的第一点是市场的规模。

We think about, one, is scale of the market.

你知道，我们在Sequence这里做的事情没法规模化。

You know, what we do here at Sequence doesn't scale.

所以如果我们投资，就必须让这笔投资有意义。

So if we're gonna make an investment, we need it to count.

我们希望合作的公司，能打造成千亿美元级别的企业。

We want to work on companies that we think we can build into $100,000,000,000 plus companies.

就是说，你不会一年做十次这样的投资吧？

Like, you're not gonna do this 10 times a year?

不会。

No.

我们不是投资机构。

We're not we're not an investment shop.

我们和优秀公司合作，打造行业领导者。

We partner with great companies to forge leaders.

我们非常在意的第二点是，这个行业里，现有企业是否拥有所有优势。

The second thing that we cared a lot about is, is it an industry where the incumbent has all the advantages?

我很乐意更详细地聊聊这个。

And I'm happy to go into that in greater detail.

第三点，如果我们看看组织现在做的事情，以及AI目前擅长的领域，再想想两者的交集，我们能不能做出点特别的东西？

And then three is if we think about what the organization does and what AI is good at today, and we think about the overlap, can we build something special?

而经纪行业在很多方面都符合这些条件。

And the brokerage industry in a lot of ways fits all of that.

每年有超过两万亿美元的保费流向保险公司。

It's a $2,000,000,000,000 plus of premiums per year going to carriers.

经纪人在其中分一杯羹，这个行业多年来也造就了许多非常成功的大公司。

The brokers take a vague on that, and it's been an industry that has minted dozens of very skilled company, large companies over time.

此外，这个行业一直对初创公司来说很难竞争，而且原因比人们想的要更不显而易见。

In addition to that, it's just been a wonderful industry that's been really hard for startups to compete in, and it's for reasons that are less obvious to people than people think.

想想保险价值链，它有三个部分。

If you think about the insurance value chain, there's three parts to it.

有保险公司，有分销伙伴，然后就是客户。

You have the carriers, you have the distribution partners, and then you have the customer.

保险公司有两种赚钱方式。

The carrier makes money two ways.

一种是承保，就是我给风险定价，希望赔付的比收进来的少，另一种就是投资。

You have underwriting, which is I price risk, I hope to pay out less claims than what I receive, and then the other part of the business is investing.

保险行业从古至今，承保几乎不赚钱，主要靠投资赚钱。

The insurance industry since the beginning of time has almost made no money underwriting and has made the principal amount of their money through investing.

所以这就激励了资产积累。

So what that incentivizes is asset gathering.

所以，获取高质量的毛承保保费才是关键。

So getting high quality gross written premium is the name of the game.

这就是为什么经纪行业这么强大。

And so that's why the brokerage industry is so powerful.

经纪行业还有一个有趣的特点，就是你的客户其实不付钱给你。

Another interesting feature about the brokerage industry is that your customer doesn't actually pay you.

付钱的是保险公司。

The carrier does.

所以如果你想想这个行业的本质，这是一个非常注重关系的行业，客户留存率高达90%，而且经纪人实际上并不真正靠价格竞争，所以对于初创公司来说，进入这个行业非常困难。

So if you think about the nature of the business, which is a very relationship centric business, gross retention rates are 90%, and you think about the fact that a broker doesn't can't doesn't really compete on price, it's a very difficult industry for startups to get into.

所以如果你想想什么对Sequence来说是完美的，那就是一个巨大的市场。

So if you think about what's perfect for Sequence is it is a huge market.

你有一些公司，它们基本上不受初创风险的影响。

You have companies that are largely immune to startup risk.

但如果我们能和对的伙伴合作呢？

But what if we could partner with the right one?

而这就是Baldwin带来的机会，它是一个规模化的资产，拥有集中的技术基础和世界级的领导团队，他们既雄心勃勃，又对前沿技术今天能实现的可能性感到非常兴奋。

And that's what the opportunity at Baldwin presented, which is a scaled asset with a centralized technological foundation and a world class leadership team that is both ambitious, but also extremely excited about what's possible today with frontier technology.

从平台技术的角度来看，你能说说哪些是通用的或共享的，跨越不同的公司？

What can you say about what is generalizable or shared from a, like a platform technology perspective across the different companies?

所以我觉得如果你退一步看，你知道，我们认为我们的共享平台能力可以分为两大块，对吧？

So I think if you take a step back, you know, we think about our shared platform capabilities as too broad, brought in two broad areas, right?

一个是Atlas，那是我们的平台，我稍后会详细告诉你，然后还有我们的工程师，以及我们有一套与公司合作的打法。

One is Atlas, which is our platform, which I'll tell you about here more in a second, and then kind of our engineers and kind of, you know, we have a playbook for how we work with companies.

这更像是一门艺术，而不是一门科学。

Like, this is an art, not a science.

我经常在内部开玩笑说，我们在Sequence有两个问题。

You know, I often joke internally that we have two problems here at Sequence.

一个是工程问题，一个是人的工程问题，而人的工程问题比工程问题难得多。

We have an engineering problem and a human engineering problem, and the human engineering problem is much more difficult than the engineering problem.

但说到工程问题，我们就会想到Atlas。

But if we think about the engineering problem, we think about Atlas.

那么，Atlas现在是什么呢？

So what is Atlas today?

Atlas是我们银行内部搭建的一个平台，我们期望它随着时间推移能推广到各个行业。

Atlas is our platform that we have built at the bank that we think that we expect to generalize over industries over time.

这源于我们很多工程师在之前的公司，比如Scale、Palantir等，积累的一个观察：如果你把业务拆解到最基本的单元，80%的部分是高度相似的，只有20%是特定于某个垂直领域的。

And part of that is an observation that a number of our engineers have had from their prior lives, both at scale, Palantir, and others, which is if you break down the business to its atomic units, 80% of it is largely homogenous and 20% of it is vertically specific.

所以，回到Atlas是什么这个问题。

And so if you think about what is Atlas?

Atlas基本上就是我们的平台，它做了很多事情。

Atlas is basically our platform that both does a number of things.

第一，它帮助我们加快部署速度，提升代理性能，提高我们工程师的构建效率，而且最终它是一个构建平台，让运营公司的工程师可以在上面进行开发。

One, it helps improve the speed to which we can deploy, improves agent performance, improves our build rates for our own engineers, and ultimately it's a builder platform that allows the operating company engineers to build upon.

它分为四层。

There's four layers to it.

第一层是数据本体。

So the first layer is the data ontology.

你可以把它简单理解为，我们如何用代码来定义组织以及业务的运作方式。

So think about that as just simply how do we define the organization, the motion of that business in code?

这一点非常重要。

And that's extremely important.

我们内部常说，就是如何让业务变得清晰易懂，让模型变得清晰，对吧？

The way that we talk about internally is how do we make the business legible, the models, right?

比如，一个代理怎么知道这个客户和那个系统里的客户是同一个人，以及这个客户的属性如何与相关的索赔或贷款政策相互关联。

How does a business how does an agent know that this customer is the same as the customer in this system here, and how the properties of that customer interrelate with the claims or the loan policies that are associated with it.

再往上一层是我们的智能体构建器，基本上就是怎么基于真实数据来打造高性能的智能体，这些内容我觉得我们经常听到和读到。

The layer on top of that is our agent builder, which is basically how do we build high performance agents grounded in ground truth, kind of kind of the things that we, I think, we hear and read a lot about.

第三个部分是我们称之为Lattice的编排引擎，就是怎么用我们构建的智能体来实际编排工作流程。

The third component to it is what we call Lattice, which is our orchestration engine, which is like how do we actually instrument workflows using the agents that we've built.

最上面一层是我们称之为artifacts的应用构建器，它构建在我们所有成果之上。

And on top of that, at the top is what we call artifacts, which is basically our application builder that sits on top of everything that we've built.

所以如果你想想我们用Atlas构建的东西，以及我们在银行这里构建的东西，从基础上看，我们构建的核心基础设施在Baldwin以及我们未来做的任何公司里都是可以复用的。

And so if you think about what we've built in terms of Atlas and what we've built here at the bank, foundationally, the core infrastructure we've built is reusable at Baldwin and any future company that we do.

这就是我们迄今为止构建的核心部分之一。

And so that's like a core component of what we've built to date.

从工程角度来看，在playbook方面，是的，我觉得这部分更像是一门艺术，而不是科学。

On the playbook side from an engineering perspective, yeah, like, I think this part is is an art, not a science.

你知道，我觉得拥有来自Scale AI和Palantir这样地方的优秀工程师的一个好处是，他们已经制定了playbook，但你知道，我们做的事情是不同的。

You know, I think one of the nice things about having great talented engineers from places like Scale AI and Palantir that they have developed the playbook, But, you know, what we do is different.

你知道，我想说，我们在投资后从银行学到的一个经验是，我们需要更加关注和敏感于员工的感受，因为尽管这很令人兴奋，但它确实引发了很多焦虑，比如这对我们组织意味着什么？

You know, I would say that, you know, one of the learnings that we've had with the bank post the investment is how much more hyper attuned we need to be and sensitive to how employees feel, which is like there's, you know, as exciting as it is, it does draw a lot of anxiety around like, what does this mean for us as an organization?

对我们Sequence来说，关键的是如何让人们觉得我们所做的和带来的东西，能提升你在工作中的表现？

And the key thing for us as Sequence is how do we make people feel like what we do and what we bring to bear here is how do we make elevate what you do here at the job?

我们怎么去实践？

How do we exercise?

作为人类最棒的部分是什么？

What is the best parts of being a human?

我们怎么让工作更有趣，把那些枯燥重复的任务从你的工作流程中拿掉？

How do we think about making your job more fun, taking away really the rote kind of repetitive work out of your workflow?

我们怎么让你觉得你的组织在赢？

And how do we make feel make you feel like your organization's winning?

这个方法论我们会继续打磨。

And like that playbook is something that we're going to continue to refine.

我觉得我们在银行这里学到了很多。

I think we've learned a lot here at the bank.

我们期待在Baldwin这里也能学到很多。

We expect to learn a lot here at Baldwin.

而且，这将是我们在Sequence长期积累的关键要素。

And like, that's going be a key component of what compounds over time for us to sequence.

你觉得你们在这里长期下来，有什么特别之处是那些传统的大型收购型基金很难复制的？

What do you think is going to be special for you guys over time here that's very hard for, let's say, more traditional large buyout shops to replicate, right?

因为据我所知，现在没有哪家不在考虑AI转型。

Because I don't know a single one that doesn't, I think, at least think about AI transformation today.

而且从回报的角度来看，他们做的事情大多还是围绕承销、平台整合、估值倍数提升，还有一些金融工程。

And like much of what they do from a returns perspective is probably dominated by, you know, underwriting, consolidation to platforms, multiple expansion, some financial engineering.

不过我觉得，他们确实也有像投资组合价值提升团队这样的设置。

But I do think, you know, people have like, let's say, portfolio value teams.

那你觉得怎么才能更明确地划清界限，说明我们不只是投资机构，而且有些事情没有我们，别人确实很难做到？

So how do you draw the distinction more strongly between we're not investing shop and like, here's what we think it's still hard for other people to do without us?

我对所有大型私募股权公司都非常敬佩。

So I have a lot of admiration for all the big private equity firms.

你之前在阿波罗工作过。

You used to work at Apollo.

我以前在阿波罗工作过。

I used to work at Apollo.

我有很多朋友还在那些不同的公司工作。

I have a lot of friends who still work at all these various firms.

我觉得他们会继续做得很好。

I think they will continue to do great.

但我深信我们的业务非常不同。

But I deeply believe that we have a very different business.

如果你想想典型的私募股权公司，它们是以基金结构运作的。

If you think about your typical private equity firm, they are set in a fund structure.

基金结构的经济激励机制促使你部署资本。

The fund structure economics are such that it incentivizes you to deploy capital.

他们的业务是寻找优质资产，合理定价，建立合适的资本结构，进行足够的价值创造，以产生满足有限合伙人要求的回报率。

They're in the business of finding great assets, pricing them attractively, putting the right capital structure in place, doing sufficient amount of value creation to generate the required rate of return to satisfy their LP base.

我们的业务非常不同。

Our business is very different.

我们的业务在于，如何找到由杰出人才领导的顶级组织，并以极其长远的眼光，去打造市场领导者。

Our business is around how do we find world class organizations run by exceptional people and taking an extremely long term horizon around how do you forge the market leader.

正因为如此，我们的做法与众不同。

And what we do is is different by virtue of that.

现在，我觉得你真正想问的是，私募股权公司能否通过他们的运作，逐步实现彻底的AI转型？

Now, I think there's a the pointed question I think you're getting to is can private equity firms over time do radical AI transformation through their motions?

我认为很多公司在这方面肯定会取得不少进展。

And I think many will definitely make a lot of progress on this front.

比如，我确实认为私募股权公司有动力去做这件事。

Like, I I certainly think private equity firms are incentivized.

他们有资源。

They have the resources.

他们也有能力开始在这方面取得实质性的进展。

They have the capabilities to start to go and make meaningful progress on this front.

不过，我认为他们确实会遇到一些限制，而且我觉得很多公司会找到办法解决其中一部分问题。

I do, however, think that there are a number of limitations that they do run into, and I think many will find ways to solve parts of these.

一个问题是，你怎么招募世界级的工程师来你的公司工作？

One is like, how do you recruit world class engineers to come work at your firm?

这真的很难做到。

It's really difficult to do.

典型的私募股权公司设计上就是围绕投资者来打造的。

Your typical PE firm has designed itself to celebrate the investor.

而我看我那些优秀的工程师，他们希望在决策时能参与其中。

And when I look at my great engineers, they want to be in the room when decisions are made.

他们希望对于这是不是一门好生意有自己的发言权。

They want to have a say in terms of like, hey, like whether or not this is a good or bad business.

我猜如果你去一家典型的大型收购公司，很难找到一个25岁的工程师对投资组合有任何发言权。

My guess is that if you go to your typical large cap buyout firm, it'd be highly unusual to find a 25 year old engineer have any say in terms of that investment profile.

所以我觉得这会是个挑战。

So I think that will be challenging.

我觉得第二个挑战就是时间跨度，对吧？

I think the second part that's challenging is just the time horizon, right?

现实情况是，如果你想想你在基金里的典型投资，你知道，他们会告诉你时间跨度很长。

The reality is, is if you think about your typical investment in a fund, it's, you know, they'll tell you it's a long time horizon.

但你通常是在琢磨怎么在三年内开始打包卖掉这个东西。

You're typically trying to figure out how do we start to package to sell this thing in three years.

这跟我们对技术转型的思考方式，以及你愿意在这里投入哪些资源，完全是两码事。

That is just a very different framing for how we think about technological transformation and what what are the investments that you're willing to bring to bear here.

而且我真的觉得最后一点就是，你知道，对我们来说，我们极度极度关注业务和团队的质量。

And I really think the last thing is just, you know, for us, we just have an extreme, extreme focus on like quality of business and team.

这并不是说其他私募股权公司不这样，但像我们做的这种事确实很少见。

And it's not to say that other private equity firms do not, but, like, we what we do is scarce.

我们试图每年只做一笔交易。

We're trying to do one deal per year.

就这样。

That's it.

我们没有固定的投资节奏。

We don't have a deployment cadence.

我并没有一群投资者和有限合伙人催着我们说，嘿，你们今年投资不够啊。

I don't have a group of investors and LPs who are telling us, hey, like, you guys were underinvested in the year.

我们的任务就是做成一笔交易。

Our job is to do one.

如果这意味着今年一笔都不做，那也挺好的。

And if that means that we don't do any this year, that's fantastic.

我觉得这是一种非常专注、非常审慎的做法，跟别人很不一样。

And like, it's just like a very, I think, a very focused and deliberate approach that's just very different.

我们只会在世界上很小的一块领域里出手，然后我期待其他私募股权公司继续做得很好。

We're gonna touch a very small surface area of the world, and then I expect all these other private equity firms to continue to do great.

但我确实认为，在我们建立的公司基因上，存在核心差异。

But I do think there are core differences in terms of the genetics of the companies that we've built.

我希望能够一直持有Sequence Equity，直到我生命的尽头。

I hope to hold Sequence Equity for, you know, the rest of my life.

所以，继续前进吧。

So, you know, go forth.

再问一个关于你们自己的承销和稀缺性的问题。

Just one more question on your own underwriting and the scarcity.

过去几年，每个CEO在财报电话会议上肯定都谈到了人工智能。

It must be true that every CEO in their earnings calls over the last few years has talked about AI.

我觉得大家都意识到了这种变化。

I think people recognize the change.

我觉得大家确实想为他们的公司投资这些能力。

I think people genuinely want to invest in these capabilities for their companies.

你在管理团队中寻找什么，让你觉得你们能帮助他们成为领导者，甚至更占主导地位的领导者？

What are you looking for in management teams where you say like, we can help you become a leader, or even a more dominant leader?

你知道吗，我觉得你可以从管理团队在他们成长起来的行业里如何应对竞争中学到很多。

You know, I think you can learn a lot about a management team just in terms of like, how do they play the game in the industry upon which they grew up in?

这是我们寻找的核心部分，归根结底，Sequence拥有优秀的工程师。

And like, that's like just a core part of what we look for, which is at the end of the day, Sequence has wonderful engineers.

我们有一个非常有才华的投资团队。

We have a very talented investing team.