20VC：Anthropic 对决五角大楼：谁将胜出 | 最佳股票推荐：该买什么 | 数据中心军备竞赛：资本支出战争是否停滞 | 公司增长放缓时代已终结

如果不行，就试试别的方法。

If it doesn't work, try something else.

我想提一下这周发生的一件事，我认为很多人误解了，这对我观点至关重要：当我们运行24小时的AI时，我们需要的算力远超现有水平，而Anthropic刚刚推出了真正的Claude代码审查功能。

I just wanna bring up one thing that happened this week, which I think people misunderstood and is so important to my point of we're not even Like the amount of AI we want to use when we're running 24, we need massively more compute than we have, which is Anthropic launched a true Claude code review.

好吧，太好了，恭喜。

Okay, great hooray.

用于发现错误和问题。

To find bugs, to find issues.

他们说价格是15到25美元，结果互联网炸了。

And they said it's $15 to $25 and the internet blew up.

天啊，这也太贵了吧。

Oh my God, this is so expensive.

我每个月在Mac上只花200美元。

Like I only spend $200 a month for Macs.

我花20美元，现在你却想收20美元来做一次代码审查来检查错误和问题。

I spend $20 and now you want $20 to a code review to check for bugs and issues.

但云代码负责人——至少负责这个功能的人——的回应是：天啊，我们并行启动了十个以上的代理，持续运行二十分钟，只为找出你产品中的每一个漏洞。

But the response back from the head of Cloud Code, at least whoever built the feature was, My God, we're spooling up 10 plus agents in parallel to run for twenty minutes to find every single bug in your product.

这本来可以手动完成，我过去几个月在Replit上一直手动做，但如今只需一键就能完成，这简直太惊人了，人类根本不可能在二十分钟内找出每一个漏洞。

Like, could do this manually, and I've been doing it manually in Replit for months, but to do this in one click, it's profoundly like, humans can't even do this, find every single bug in twenty minutes.

因此，现在他们对算力的需求，甚至包括Claude在内，都大幅增加了。

And so the amount of compute now they need even Claude.

Claude代码要完成这项任务，还需要额外的20美元算力。

Claude code needs an extra $20 of compute to do this.

你知道你真正想做什么吗？

And you know what you'd really like to do?

在每次提交后都运行一次。

Run this after every commit.

而不是偶尔为之。

Not episodically.

不要等到鼓起勇气花20美元的令牌额度时才运行，这可是不少钱，因为这并不 subsidized。

Not when not getting up your nerve to spend $20 worth of tokens, which is a lot because this isn't subsidized.

对吧？

Right?

这是Anthropic。

This is ANTHROPIC.

你肯定希望每天在每次提交后都运行十次这个功能。

You'd love to run this 10 times a day after every commit you make.

与以往相比，这次代码审查所使用的计算量增加了几个数量级。

That is orders of magnitude more compute being used for this code review than we used before.

所以我们现在正处于这个阶段，所有东西都在运行2047，我们才刚刚触及皮毛。

And so we're just between that, everything running 20 fourseven, we've just scratched the code.

因为如果你在Claude Code、Replit、Lovable或任何类似平台上开发过，你就会意识到这仅仅是个开始。

Because if you've built anything in Claude Code or Replit or Lovable or whatever, what you realize is it's just the beginning.

那谁来做所有的质量保证？

Like who's doing all the QA?

谁来做所有的代码审查？

Who's doing all the code review?

我们只是触及了表面而已。

Like we've only scratched the surface.

我们需要在太空中建立数据中心，这样我才能继续编程。

We do need data centers in space so that I can keep coding.

但我认为从新闻角度来看，这个观点很有意思。

But I think the point is interesting from the news.

我认为人们忽略了这一点：花20美元来审查整个代码库并找出所有主要漏洞，这算多吗？

I think people miss this point of this code review thinking, is $20 a lot to review your entire code base and find all your top bugs?

这算多吗？

Is that a lot of money?

20美元？

Is $20?

当然不算。

And of course it's not.

你知道，我们是CodeRabbit的开心投资者，这是一家独立公司，正在多个工具上实现这一功能。

You know, we're happy investors in CodeRabbit, a standalone company doing it across multiple tools.

是的，现在我们有了来自Claude的令人兴奋且有趣的竞争，这虽然不好，但另一方面却验证了这个领域，说明你必须这么做。

And yes, now we have exciting and interesting competition from Claude that's bad, on the other hand, validates the space and says, you need to do this.

我认为你对产品需求的描述完全正确。

I think your description of the product of the need is exactly right.

如果你正在生成海量的代码，使用所有这些代理，认为你不需要进行代码审查的想法是荒谬的。

If you're generating infinite amounts of code, right, using all these agents, the idea that you're not going to kind of do code review is absurd.

所以你会有自动化的代码审查。

So you're going to have automated code review.

你说得完全对。

You're exactly right.

你关于启动所有这些代理的评论恰恰印证了我的观点。

Your comment about spinning up all those agents was exactly the You illustrated my point.

从技术角度来看，有大量的代理需要执行海量的代码审查工作。

From a technology perspective, there's infinite need for large amounts of agents to do a large amount of code review.

从经济角度来看，有人抱怨他们被要求支付20美元。

From an economics perspective, you have people complaining about the fact that they're being asked to pay $20.

你说得完全对，ChatGe。

And you're exactly right, ChatGe.

正确的答案是，克劳德太客气了没说出来，尽管我们知道他们都是好人，不管皮特·赫加特怎么想——天哪，我是在让你自动化整个开发者。

The correct answer, which Claude was too nice to give, because we know they're nice people despite what Pete Hegsart thinks, was, For Jesus' sake, I'm giving you you're automating an entire developer.

我让你每月花200美元生成无限的代码，而你连花20美元在上线前检查一下都不愿意？

We're letting you generate infinite code for $200 a month, and you won't even pay $20 to check it before you put it into production?

看在上帝的份上，老兄，出点钱吧。

For God's sake, man, pony up.

这才是正确的回应。

That's the correct response.

这就是技术与经济交汇的地方。

That's where technology is meeting economics.

你以为所有这些东西都能免费获得，但总有一天会出问题，因为总得有人承担成本。

The idea that you can just have all this shit for free is at some point going to because someone's gonna have to cover their nut.

然后我们会发现真正有趣的问题是：有多少用户愿意为每次代码审查支付20美元，或者为大型代码审查支付200美元，或者其他金额。

And then we'll discover the really interesting question is how many of those users are pay prepared to pay $20 per code review or $200 per big code review or whatever.

对吧？

Right?

所以我是同意你的。

So I do agree with you.

需求是无限的。

The demand is infinite.

你能做的事情简直太惊人了。

The number of things you can do is just freaking amazing.

问题是，人们愿意为哪些东西全额付费？

The question is, what are the things that people are prepared to pay full boat for?

顺便说一句，这很有趣，也稍微验证了我们之前的一个想法。

And it was just it was interesting, by the way, and it kind of validated a little something we were thinking.

我真想成为Claude的定价负责人，你做过定价，杰森。

I would love to be the pricing person for Claude, and you've done pricing, Jason.

你怎么做决定？

How do you decide?

你认为哪些方面比较有意思，因为他们确实有一些很有趣的做法。

What things do you kinda because they've got some really interesting things going on.

API价格昂贵且按使用量计费。

The API is expensive and metered.

另一方面，Claude采用的是订阅制。

On the other hand, Claude called the subscription.

正如你所说，花200美元你能获得远超应得的大量功能。

As you point out, you can get a whole ton more than you should be at $200.

所以他们实际上并没有对这部分进行按量计费。

So they're kinda not metering that.

然后他们做出了决定，在我看来这是明智的：也许代码审查这种功能听起来像是管理层面的需求，或许确实有预算支持。

And then they decided, and in my view wisely, hey, maybe code review is something that kind of feels managerial, and maybe there is budget for that.

你知道，你和我都做过这样的交易：一开始是面向个人用户的低端SaaS产品，属于产品驱动增长（PLG），当时人们都会说：‘你们怎么可能靠这种低价值的小额签约赚钱呢？’

And, you know, you and I both did deals where they were, you know, low end SaasLAN, which were PLG for individual users, and people were like, Oh, how are you ever going to make money on, you know, small, low end signatures?

但当你越往企业级的采购决策层深入，人们愿意付费的倾向就越明显。

But the more you go to management of signatures in the enterprise, the more there is a propensity to pay.

我认为这正是同一种趋势的开端。

And I think this is the beginning of the same thing.

你，先生。

You, Mr.

黑客先生，你花200美元就可以获得无限的免费计算资源。

Hacker, you could have all the free cogen you like for $200.

但如果你要把这个系统部署到美国银行的架构中，你肯定需要的是高级版，而不是25美元的代码审查，而是250美元的安全认证代码审查，你会看到这种盈利模式的推动力。

But if you're going to be rolling this out into Bank of America's systems, you're going to want the deluxe, not the $25 code review, you're going to want the $250 security proof code review, and you're gonna see that kind of push to make money here.

但我觉得这就是整个动态的核心。

But I I think that's the whole dynamic.

你能做的事情多得惊人，但总有一天，有人得为这些数据中心买单。

There's the whole shit ton of stuff you can do, and then at some point, someone's gonna have to pay for those data centers.

是的。

Yep.

而且这钱不可能由战争部来出。

And it ain't gonna be peace at the Department of War.

我认为与此相关的是，我们过去曾讨论过的一些事情确实是真实的，但我们也在主动促成它们的发生。

I think related to that, the thing that there's there's certain things we've talked about in the past that are true, we're also willing into existence.

企业正主动促成其希望用人工智能取代人类的愿望。

The enterprise is willing willing its desire for AI to replace humans.

因此，这种趋势会加速，因为他们希望它加速。

So it's going to accelerate because they want it to accelerate.

我的天，正在加速的是：不要雇佣初级开发人员，不要招任何新人。

I mean, my God, the one that is being accelerated is hire no junior developers, hire no juniors.

这同样是被主动促成的。

This is being willed into existence.

没人愿意培训任何人。

No one wants to train anybody.

这和代码审查有关，听好了，如果你决定不雇佣初级开发人员，当然可以花200美元，或者每月多花500美元，为你的高级开发人员做代码审查。

It ties to this code review because listen, if you're going to hire no juniors, of course you can spend $200 or an extra $500 a month to do code review for your senior developers.

他们本来就已经极其高效了，对吧？

They're already insanely productive, right?

我觉得初级开发者的消亡让我担忧。

And I think the death of the junior, it worries me.

昨晚我和我儿子共进晚餐，他聪明得不得了，但在宾州州立大学读书，那不是前十的常春藤盟校。

And I had dinner last night with my son, who's off the charts smart, but he's at Penn State, which is not a top 10 Ivy League school.

他说计算机科学和数学领域根本没有任何工作机会。

And he said there are just zero jobs for anyone in CS or math.

真的一个都没有。

There's just zero.

连C级公司都没人来招聘他们，但他自己却有工作机会，因为他一直在发布一些关于詹森·费因斯坦理论的东西，我根本看不懂。

There's no one even coming by to hire them for C tier there's just he has job offers because he's like publishing things on Jensen Feinstring theory that I don't understand.

真棒。

So great.

他所在大型州立大学的班级里，只有六个人拿到了科技公司的offer，但不仅在我们的世界里，甚至在他的世界里，也变成了‘我们就是不要初级员工’。

Six people in his class at his large state college have tech offers, but it has become not only in our world, but in his world that has become, we just don't want juniors.

我们不想花三个月或六个月去培训他们。

We don't want to train them for three months or six months.

他们对这个工具一窍不通。

They don't know the tool's cold.

他们根本不懂。

They don't know it.

当你不招聘初级员工时，预算就从这里来，因为我们所处的世界，经验水平都只是中等水平。

And that's where the budget comes from when you hire no juniors, when we're just a world of the middle of the bell curve in terms of experience.

我们不想要那些不愿意学习工具的人，也不想要非专家的人。

We don't want folks that don't want to pick up the tools, and we don't want folks that aren't experts.

这让我对2027年感到非常担忧，因为这种情况在销售领域也会发生。

And this is my big worry for 2027, as it happens, it's going be true in sales.

在客服领域，这种情况已经出现了。

It's already true in support.

你想在客服中保留人类，但又不想招初级员工？

Like, you want humans in support, but you don't want juniors?

天啊。

My god.

我不想要那种半吊子，我想要一个对CodeRabbit了如指掌、能处理紧急情况的人。

I don't wanna like, I I want a human that knows CodeRabbit cold that can handle the escalations.

我不需要初级营销人员。

I don't need junior marketers.

我不需要什么初级小角色。

I don't need junior nothings.

哈里，别扯远了，但我觉得，淘汰初级人员，某种程度上正是我们能为这些数据中心筹到预算的原因。

And Harry, not to get us off track, but I it ties I think getting rid of juniors is where we get budget for these data centers in part.

我们只是发现，他们被淘汰的速度比我们想象的快多了。

We're just getting they're just dying faster than we ever thought.

初级人员的消亡正在我们眼前发生。

It's happening in front of us, the death of the junior.

我觉得这跟我们看到的两次融资有关。

I think it ties to two fundraisers that we saw.

罗里，你作为其中一家的投资人，可能会不同意。

Rory, you may disagree as an investor in one of them.

作为另一个公司的投资者，我同意这个观点，但我们看到Intercom融资了2.5亿美元，Liguora融资了5亿美元。

I I would agree as an investor in another, but we saw Intercom raising $250,000,000, and we saw Liguora raising $500,000,000.

我认为这两者都与你提到的替换初级员工和低级别人员有关。

And I think both tied to your point there after a replacement of juniors and lower ranking

初级人员的消亡。

Death of the juniors.

我为什么想要——我的意思是，我不是Liguora或Harvey的专家，但我确实了解这个过时的领域。

Why would I want why would I mean, I I know I'm not a Legora or Harvey expert, but I I I do know this the dated space.

我的意思是，谁愿意花两年时间等待一个助理成长起来？

I mean, who the hell wants to wait for an associate to scale up for two years?

这曾经是法律行业经典的招聘模式。

Like, that was the classic legal hiring.

你从顶尖学校招聘一个真正优秀的人才，最好是班级前10%，智商在线，然后你补贴他们两年，直到他们达到足够资深的水平，能够服务顶级客户或接手重要案件。

You hire someone real from a top school, ideally that was top 10% of their class, they have the IQ, and you basically subsidize them for two years until they cross the line and become senior enough that they can work with a top client or take a case.

天啊，现在谁还愿意做这两件事？

God, who wants to do those two years now?

除非你不得不。

Unless you have to.

除非你不得不。

Unless you have to.

这一直是我们长期讨论的话题，很高兴我的观点在发生变化，杰伊。

This has been one of our long running discussions, and it's good that I'm evolving my position, Jay.

正如我所说，我越来越认同你的看法，这里显然正在发生一些变化。

As I said this, I've kind of come more to your perspective that there clearly is something going on here.

尽管我不相信大规模失业的灾难性情景，但我觉得你说得完全正确。

Even though I don't believe in the catastrophic mass unemployment scenario, I think you're exactly right.

对于一些非常特定的人群，比如入门级的计算机科学类职位，另一个是实际的客户服务岗位、实际的法律助理岗位，目前确实产生了显著的影响。

For a couple of very targeted demographics, one of which is entry level, comp sci type jobs, Other would be actual customer support jobs, actual legal associate jobs.

目前，失业问题确实受到了实质性影响。

There is a meaningful impact on unemployment right now.

所以我想现在就承认你说得对。

So I just wanna acknowledge you right now.

这很有趣，因为我认为这是对的。

It is interesting because and I think that's true.

另一方面，你们党通常认为，当有新技术出现时，最好的早期采用者是年轻人，因为他们之前的成见较少。

On the other hand, your party says, typically, when there's new technology, the best adopters are young people because they come in with less priors.

有一部分我觉得，这些学校是不是没尽到责任，尤其是现在都2026年了？

And there's a part of me that says, are these schools doing their job, especially now that it's, what, 2026?

如果你培养出来的计算机科学毕业生对AI毫无概念，完全不会使用这些工具，那你们根本就是在失职。

If you're graduating commsci graduates who aren't AI first and totally awash in using these tools, you're just not doing your job.

我觉得这其中还有一部分原因是

And I think there's an element of not

罗里，你根本没法快速调整课程来更新圣克劳德的内容。

Rory, you're you're unable to alter curriculum fast enough to update Saint Claude.

课程的调整需要时间。

Curriculums take time to change.

他们必须改变，因为如果你花每年100美元、四年总计400美元去读一个四年制计算机科学学位，结果却只能找到失业或当咖啡师的工作，你肯定会愤怒。

They need to change that because if you're looking at investing, you know, a $100 a year for four years at a four year college education to get a computer science degree that turns into unemployment and a barista job, you should be pissed.

但我同意你的观点。

But I'm with you.

当你看到安德烈·卡帕蒂在六个月内的转变，从最初仅使用20%到后来达到80%，而安德烈·卡帕蒂就是这样一个人——你真的能期望像大学这样的教育机构，根据克劳德能力的演进，彻底调整整个课程体系吗？

When you look at Andre Kapathi's journey in six months' time from bluntly using it for 20% to 80%, and this is Andre Kapathi, that that I think is unreasonable to expect an educational institution like university to alter the entire curriculum on Claude's ability to progress throughout the kind of skill journey?

这显然并没有发生。

It certainly doesn't appear to be happening.

无论罗里的观点如何，我们是否应该从这些机构中期待这种变化？

Whether Rory's point, should we be expecting it from these institutions?

是的。

Yeah.

它们正变成一种昂贵的看护服务，持续四年。

They're they're becoming just babysitting for expensive babysitting for four years.

而且

And

它们不应该是这样的。

they shouldn't be.

而且要给他应有的认可，我认为十五年前，彼得·蒂尔就提出过一个观点：高等教育的投资回报率正在下降。

And give him credit, I think fifteen years ago, Peter Thiel articulated the perspective that the ROI on higher education is going down.

当时，回报率仍然很高。

And at the time, it was still high.

值得一提的是，直到今天，大学毕业生的失业率仍低于高中毕业生，而高中毕业生的失业率又低于高中辍学者。

And it's worth pointing out, even to this day, unemployment among college graduates is lower than unemployment among high school graduates, which is lower than unemployment among high school dropouts.

所以，如果你不考虑其他因素，上大学仍然算是‘更好’的选择。

So you're still quote unquote better off going to college if doesn't apply.

但关键是，现在你要考虑这40万美元的净现值，除非是最顶尖的学位，否则你根本收不回成本。

But the point is, now you take into account the net present value of that $400, you're not getting your money back on anything other than the best degrees.

这些人并不是地球上最值得被支持的人。

These are not the most deserving people on the planet.

他们的失业率仍然略低，约为3.3%左右，而平均水平是4.4%。

They still have a lower three point something percent versus 4.4 average.

但确实存在一个趋势，如果你是刚毕业的学生，这会让人感到担忧。

But there definitely is a trend here that is would be troubling if you're a recent graduate.

所以我在趋势这个问题上基本同意你的看法。

So I I'm kinda with you on the trend.

假设你并不认为会出现大规模失业的情况，或者觉得这种说法有些夸大。

Say you don't see the mass unemployment case where you don't quite think that or think it's an overreach.

你看法律、计算机科学、客户服务、簿记这些领域，即使其他地方不受影响，仅这些行业本身就会导致大规模失业。

When you look at legal, computer science, customer support, bookkeeping, those alone, if it had no impact anywhere else, would be mass unemployment.

我不确定会不会真的这样。

I don't know if it would be.

再说了，看看数据吧。

Again, look at the numbers.

我今天其实刚看过这些数据。

I actually was looking at it today.

整个科技行业，不包括在非科技公司工作的科技人员。

Entire tech industry, excluding tech people employed in non tech companies.

根据NAICS分类，科技行业有51类公司，美国共有320万人，其中约80万人是软件开发者。

Technology, NAIC classification, 51 tech companies, 3.1, 3,200,000 people in The United States, of whom roughly 800,000 are software developers.

如果减少一半，那就是40万软件开发者。

If you go down by half, that's 400,000 software developers.

这是很多人，但只占失业率的0.2%。

It's a lot of people, but it's 0.2% unemployment.

客服人员有多少？

How many are in customer support?

数量大得多。

Much bigger number.

你说得完全对。

You're exactly right.

但更重要的是，很多低端客服工作其实更具可替代性。

But also way more fungible in the sense of a lot of the low end customer support.

人们还可以从事很多其他服务类工作。

There's lots of other service type jobs folks can do.

我只是相信，哈里，从中期来看，技术的普及速度比硅谷人士想象的要慢，而人们也更具适应性。

I'm just a believer, Harry, that over the again, I wanna be over the medium term, the technology takes longer to diffuse than we in Silicon Valley allow, and people are more adaptable.

所以在中期内，我不认为会出现10%以上的科技驱动型失业，我知道你不同意。

So over the medium, I don't see any kind of 10% plus tech driven unemployment, which is and and I know you're gonna disagree.

关键是，我们没必要为此争论，因为最终我们会知道答案。

The point is we don't have to argue about this because in the end, we'll know.

对吧？

Right?

我只是觉得，我有两百年的事实支持我，而你什么都没有，但也许你是对的。

I just think I have two hundred years of facts on my side and you have shit, but maybe you're right.

话虽如此，我相信你在短期内是对的。

That said, I do believe you are right in the short term.

在某些孤立的领域，一些非常活跃的人确实处境艰难。

In isolated pockets, very vocal people are really struggling.

而且，我有三位刚毕业的大学生，其中两位是过去五六年里毕业的。

And look, I have three recent college graduates, two over the last five or six years.

所以我完全清楚大学毕业生找工作有多难。

So I'm totally aware of how hard it is to get a job as a college graduate.

而且我认为这确实是个问题，我们最好尽快应对。

And I think it's a real issue, and we better get on top of it.

因为法国大革命教会我们的另一件事是，我真的很冷。

Because one of the other things the French Revolution teaches you is because I'm really cold.

你可以长期拥有被剥夺财产的城市贫民，而不会发生任何事。

You can have dispossessed urban poor forever and nothing happens.

但如果你激怒了20岁的中产阶级、过度教育的精英群体，他们往往会惹麻烦。

But if you piss off the 20 year old middle class, the overeducated elites, they tend to cause trouble.

实际上，我认为持续大量培养大学毕业生——尤其是那些主修社会学和英语专业、却要面对计算机科学领域竞争的人——他们二十年来一直按规则行事，结果却遭遇大规模失业，这将带来严重问题。

And I actually think the continued production of masses of college graduates, a fuse with their sociology and English degrees, facing in to and comp science degrees, what they've done played it by the rules for twenty years, and then giving them mass unemployment for that cohort is going to be problematic.

我要说清楚，尽管我认为宏观趋势并不存在，但这并不意味着我不认为这是一个真实的问题。

I want to be clear, just because I think the macro trend is not there doesn't mean I don't think the macro trend is a real issue.

我预测到2026年，这将成为一个政治议题。

And I predict in '26 it's going to be a political issue.

看NPS调查真是有趣极了。

I mean, was so funny to watch the NPS survey.

AI的民意调查非常负面。

AI poll's really negative.

正如有人在推特上说的，你一直告诉我们你会摧毁世界，让我们所有人都失业，现在却惊讶地发现我们不喜欢你。

And as someone tweeted, you've been telling us you're gonna blow up the world and make us all unemployment, and you're surprised to discover we don't like you.

是的。

Yeah.

但与此同时，如果你看一下，我敢肯定Intercom的数据会显示这一点。

But at the same time, if you look I'm sure intercom data says this.

所有在异构人机协作环境中工作的数据都表明，客服代理的客户满意度始终位居前10%。

Everyone's data that works in a heterogeneous human AI agent environment, the CSAT is always top 10% for the agents.

但从不是第一名。

It's never number one.

他们从不说最受欢迎的人总是人类，但代理总是名列前茅。

They never say that the number one most popular person is always a human, but the agent's always in the top.

这一点比大多数人意识到的更有意义，因为现在还处于早期阶段，但趋势正在加速：我们只是更愿意与代理合作。

That is more telling than most people realize because it's early, but it's accelerating, is we'd just rather work with agents.

我们更愿意雇佣代理。

We'd rather hire agents.

我们更愿意购买一个代理。

We'd rather buy an agent.

我们不想说，哈里为他的行政助理职位收到了3000份申请，因为他需要一个真人，但我告诉你，如果代理能胜任，他更愿意用代理。

We don't wanna I mean, Harry's got 3,000 applicants for his EA position because he needs a human, but I tell you he'd rather have an agent if it worked.

他不想面试3000个人。

He don't want to interview 3,000 people.

如果他能用代理，他会选代理，而我们所有人都在经历这一点。

If he could have an agent, he'd pick an agent, and we're all going through that.

即使不划算也没关系。

And it doesn't even matter if it's cost effective.

我们就是更愿意用代理。

We just would rather have an agent.

我一直在思考，但天啊，为什么一些AI B2B初创公司增长得如此迅速？

I've been thinking a lot, but my God, why do some of the AI B2B startups grow so quickly?

简单的答案是产品市场契合、产品驱动增长、代理驱动增长。

The simple answer is product market fit, product led growth, agent led growth.

但随着你深入探究，随着一年的推移，当我们进入2026年底和2027年时，人们只是想购买一个代理。

But as you dig deeper and as the year goes on and as we roll into late twenty six and '27, people just want to buy an agent.

他们不想雇用真人。

They don't want a human.

如果你能提供显著的预期投资回报率，这将极具吸引力。

And if you can deliver perceived massive ROI, that is so appealing.

这并不是裁员。

It is not layoffs.

这也不是别的什么。

It is not this.

而是，天啊，我能够获得客户。

It is, oh my God, I could get customers.

我能够开展营销活动。

I could run campaigns.

我可以不用人类来做客服。

I could do support without humans.

天啊，我想要这个。

My God, I want that.

当我看到许多公共B2B公司推出的平庸产品时，它们根本无法满足这种需求。

And when I look at so many of the mediocre products the public B2B companies have launched, they do not service that demand.

我不想要Atlassian、HubSpot或其他公司的员工，我根本不需要他们。

I don't want humans at Atlassian or HubSpot or other folks with I don't want them.

我不希望用你的AI让我的员工效率只提高8%。

I don't wanna make my humans 8% more efficient with your AI.

我不想要这些员工。

I don't want these humans.

我想要一个代理来完成这项工作。

I want an agent to do this work.

当你做到这一点时，无论Ligora、Harvey，还是节目里提到的成千上万的AI，无论它们是否能完全实现这一潜力，只要你能提供AI代理而不是人类，人们就会排着队来找你。

And when you do that, whether the Ligora, Harvey, whatever, a million on the show, whether they fully can realize that potential, people will line up at your door when you can give them an AI instead of a human, an agent.

他们想要的就是这个。

They that's what they want.

这些才是我们想打造和运营的公司。

These are the companies we wanna build and run.

和往常一样，我同意大方向，但在顺序上可能不太认同，更多是叙事方式的问题。

As is so often the case, I agree with the broad direction and might disagree with the not so much of sequencing as the narrative.

我认为到目前为止，B2B领域的采用，除了客户支持，以及过去一年在编程领域，像Harvey、Ligora和GCAI这样的公司，它们的核心并不是真正取代人类，而只是在边缘上略有涉及。

I think a lot of the B2B adoption to date, with the exception of customer support and in the last year, with the exception of coding, I think folks like Harvey Lago and GCAI, they're all about they're not actually about human replacement except very much at the margin.

在某种模糊的层面上，也许我不需要Linklater的第十七个助理。

At some vague level, maybe I don't need the seventeenth associate of Linklater.

对吧？

So right?

因为到目前为止，它还称不上是代理式的。

Because so far, it hasn't been agentic.

到目前为止，它主要还是：‘好的，让我来帮您，律师先生。’

So far, it's been primarily, yeah, let me help you, mister lawyer.

更聪明一些。

Be smarter.

更高效地工作。

Work more efficiently.

是的。

Yeah.

在某种程度上，不是让实习生去做某事，但这也并非一场全面的替代故事。

At some level, not ask an intern to do something, but it's it hasn't been a wholesale replacement story.

我觉得你说得对。

I think you're right.

这发生在任务层面。

It's been at the task level.

我已经能够增加自己能完成的任务数量，从而减少需要请助理来处理的任务。

I've been able to increase the number of tasks that I can do and therefore decrease the number of tasks I have to pull in an associate to do.

我觉得你说得对，杰森。

I think you're right, Jason.

现在有了智能代理，我认为我们在软件领域首先看到了这一点，过去六到九个月里，这些代理从头到尾编写代码的能力令人震惊。

Now with agents, and I think we've seen it first in software, it's stunning what's happened in the last six, nine months in terms of the ability of these things to just code start to finish.

因此，我觉得你说得对。

And therefore, you know, this is where I think you are right.

我们可能正处在这种‘代理化’的开端，未来十二个月，B2B公司会说：现在我终于能完全自动化这个任务了。

We might be at the start of some of this agentification, and the next twelve months will be B2B companies saying, now I finally can automate this task entirely.

然后我们会发现，我喜欢

And then we'll find out what the I like

去思考一个框架：我的孩子们将来会怎么看我过去的生活，他们会说：天啊，真不敢相信你曾经那样生活。

to think through a frame of what will my children look at me and how I used to live and go, God, I can't believe you used to do that.

我认为他们会提到的一件事是：真不敢相信你花了这么多年培训员工，结果他们却离开了。

And I think one of the things they will say that to is, I can't believe you spent years training people and then they left.

那些年全都浪费了，然后他们走了。

All those years wasted and then they left.

这太疯狂了。

That's so nuts.

我认为这看起来会极其过时。

I think that will seem incredibly archaic.

一个具有代理能力的世界，能够持续地教育和学习，而且它们永远不会离开你。

An agentic world, which compounds education and learning, and they never leave you.

杰森，你会更喜欢这样吗？

Jason, would I prefer it?

会的。

Yes.

对。

Yeah.

我只是觉得，如果你是B2B领域的创始人，你就应该打造这样的东西。

That's what I I just think if your founders in B2B, this is what you should be building.

这会让人排着队来找你，因为他们更愿意和你的代理合作，而不是和人类。

It's something where folks will line up the door because they'd rather work with your agent than a human.

大多数上市公司根本没这么想过，但那些真正爆发的初创公司，正是因为他们做到了让人排长队。

And most of the public companies are not even thinking this way, but the ones, the startups that do blow up, what they line out the door.

请给我一个代理型的例子，区别于自动化任务型的。

Give me an example of the agentic ones, please, as distinct from the automated task ones.

你所说的代理型是指什么级别的？

Give me a what level of agentic are you talking about?

有很多，但我在这类获客代理上做了很多工作，也就是市场推广代理。

There's a bunch, but I've done so much in these GTM agents, these go to market agents.

明白吗？

Okay?

我们最近部署的这个Monaco项目，是Sam Blonde做的，他们在前五天就达成了七位数的销售额，而且已经有六十天的客户在排队等候。

So this last one that we deployed, this Monaco one that Sam Blonde did, I mean, they closed seven figures in their first five days and they have sixty days of people lined up.

而且说实话，我并不是在吹嘘，他们确实都很棒，但为什么客户会排到门外？

And look, I'm not hyping it, they're all good, but why are they lined up out the door?

好吧，这才是更有趣的问题：为什么他们会排到门外？

Okay, this is the more interesting question, why are they lined up out the door?

不管他们是否在整个旅程中都完全成功，对我们来说，这个工具已经奏效了：它能与潜在客户沟通、发短信、推销你的产品，并安排会议，而且会议已经完全准备好，随时可以成交。

Whether they're fully successful on all their journey or not, it's worked for us as a tool, it talks to the prospect, it texts them, it pitches your product, and it sets up the meeting, and the meeting is fully ready to go and close.

如果你能做到这一点，如果你能做到这一点，好吧，这要求很高，我甚至觉得六十天前这根本不可能，而我们用过的所有其他工具，人们都会抢着要这种东西。

If you can deliver that, if you can deliver that, okay, and it's a high bar, and I don't even think it was possible sixty days ago, okay, and all the other tools we use, people will line up out the fucking door for this stuff.

所以，无论你属于哪个领域，你都想打造一个比90%的人类都更出色的代理，这样每个人都会选择它，预算也会莫名其妙地出现。

So that's what you whatever category you're in, you want to build this agent that is sufficiently better than 90% of humans, that everyone will pick it, and the budget magically comes out of nowhere.

像这些产品的总可用市场对我们来说甚至都没有意义。

Like the TAMSA for some of these products don't even make sense to us.

它们与过去的做法不一致，但如果你能不依赖人类做到这一点，人们还是会排长队等待。

Like they're not consistent with the past, but people will line out the door if you can do this without humans.

我的意思是，我们昨天刚自己组建了一位营销副总裁。

I mean, built our own VP of marketing literally yesterday.

这周是它第一次主持我们人类员工的会议。

This week was the first time it led our human staff meeting.

我们的AI营销副总裁主持了团队会议。

Our AI VP of Marketing led the team meeting.

它主持了会议。

It led the meeting.

它总结了我们团队中人类必须做的每一个指标、每一件事。

It summarized every single metric, every single thing, every single thing the humans on our team had to do.

十个关键目标中的八个是由它主导会议的。

10 ks or eight of it led the meeting.

尽管黏土在增强方面很棒，但它确实很棒。

Mean, as great as clay is for enrichment, it's great.

如果它能做到这一点，我明天就给你五万美元。

If it could do that, I'll give you $50,000 tomorrow.

你正在处理Intercom的交易。

You're doing the Intercom deal.

你赌他们会击败Sierra吗？

Are you betting that they beat Sierra?

我告诉你我押注什么。

I'll tell you what I am betting on.

我押注他们能在那个市场中为该产品开辟出重要的空间，并且已经在其中实现了爆炸式增长。

I'm betting on that they can carve out a significant space in that market for that product and that they've had explosive growth in it.

我还认为，这支来自SaaS背景但深谙AI核心需求的团队，能够抢占相当大的市场份额。

And I'm betting that that team coming from the SaaS background, but being grounded in what it takes in AI can take a significant portion of market share.

他们的客户规模平均而言在西拉公司的高端客户群体中略小一些，并且更多地与Decagon竞争。

They tend on average to be slightly smaller in terms of customer size at the high end of Sierra, and they compete more with Decagon.

我认为，在这些已经取得一定市场 traction 的类别中，如果你说我只是部分专家，但如果你想比较 Intercom、Decagon、Sierra 和其他公司，对吧？

One thing that I think in a lot of these categories that have traction, but if you if you wanna say I'm only a partial expert, but if you wanna say Intercom versus Decagon versus Sierra versus others, right?

Zendesk、Salesforce。

Zendesk, Salesforce.

关于赢家，人们稍微忽略了一点，你之前问过Rory，对吧？

One thing that people miss a little bit in terms of the winner, which you kind of asked Rory, right?

目前，全职部署工程师（FTEs）是一个限制因素。

Is FTEs, forward deployed engineers, for the moment are a limiting factor.

让我解释一下我的意思。

And let me explain what I mean.

这意味着，没有人拥有足够多经过充分培训的全职工程师，能够足够快地帮助客户上线。

What it means is no one has enough fully trained FTEs that can get a customer up and running fast enough.

没有人拥有足够的这类资源，明白吗？

No one has enough of these resources, okay?

这意味着几乎每个供应商都得选择自己的赛道。

And what it means is that almost every vendor has to pick lanes.

你不可能像完全靠软件那样做那么多事，你得明确自己擅长的领域，对吧？

Can't do as much as you'd like to do it, if it was all software, you've got to pick where you're strong, right?

如果你在企业数字化转型方面最拿手，也许布雷特·泰勒就是这样，那你就得专注这一点，对吧？

And if you're best at digital transformation for enterprises, maybe that's what Brett Taylor is, you got to pick that, right?

所以，虽然我知道这些公司的销售团队和其他人认为他们是直接竞争对手，但每个人的客户需求都超过了他们能服务的S级现场工程师数量。

And so not that these folks I know that the sales team and folks think they're direct competitors, but everyone has more demand than S tier FDs to service that.

因此，他们现在都在选择各自的赛道，而这是我们尚未跨越的前沿。

So they're all picking lanes today, and that is a frontier we haven't crossed yet.

从这个意义上说，他们都是服务型企业。

And so in that sense, they're all services businesses.

某种程度上，这个风投圈的梗是真的。

In a sense, this VC meme is true.

他们不是长期的服务型公司，但如果你需要人类来训练它、管理它、迭代它，你就面临着与服务型公司相同的限制。

They're not long term services businesses, but if you need humans to train it, if you need humans to manage it, if you need humans to iterate it, you have the same constraints that services businesses have.

这个FDE问题是关键瓶颈。

This FDE thing is a limiter.

这令人兴奋，因为我认为很多失败者甚至都找不到优秀的FDE。

It's exciting because I think the losers, a lot of the losers out there can't even get good FDEs.

如果你看看那些在AI B2B领域挣扎的公司，深入一层就会发现，有时他们的智能代理不够有竞争力，有时速度太慢，对吧？

If you look at folks that are A lot of folks that are struggling in AI B2B, if you peel the layer back, sometimes their agent isn't competitive, sometimes it's too slow, right?

问题很多。

A lot of issues.

但如果你深挖下去，你到底有多少优秀的FDU呢？

But if you really dig down, how many how many great FDUs do you have?

你有多少人能在三十天内完成客户上线？

How many folks do you have that can spool up a customer in thirty days?

我有三个。

Well, I've got three.

好吧，你会输的。

Well, you're gonna lose.

我只是继续关注风险投资和私有领域，报道一些新闻动态，比如我之前说的，Ligure以55亿美元融资。

Just staying on venture in private and just covering, like, the news items so people are up to As I said, Ligure raises 500 at 5,500,000,000.0.

Excel领投了这一轮。

Excel lead that round.

显然，Harvey在最新一轮融资中的估值约为110亿美元。

Obviously, Harvey are better funded at, I think, $11,000,000,000 in the latest round.

DATA GON以45亿美元完成了招标。

DATA GON did a tender at four and a half billion.

Intercom融资了2.5亿美元，但价格我不知道，Roy。

Intercom raised 250,000,000 at I don't know the price, Roy.

你可能知道，但他们没有公开宣布。

You might know it, but they didn't announce it.

创始人基金正在接近60亿美元的新基金募集。

Founders Fund are closing in on $6,000,000,000 of new funds.

他们在过去的11个月里投资了33亿美元，投向了像SpaceX、Stripe和Angeril这样的杰出企业。

They invested the last 3.3 in eleven months in amazing businesses like SpaceX and Stripe and Angeril.

如果我是有限合伙人，我会非常开心。

I'd be very happy if I was an LP.

Base四十四的年度经常性收入达到了1亿美元。

Base forty four hit a 100,000,000 in ARR.

不知道你对这个有什么看法，杰森，毕竟你跟Ratplat有关。

Don't know if you have anything to say about that, Jason, given Ratplat.

这能救Wix吗？

Does it save Wix?

我们议程上有一项：CrowdStrike季度表现优异，但股价下跌。

Well, we had on the agenda, CrowdStrike crushes the quarter, trades down.

对吧？

Right?

是的。

Yeah.

我本来打算接着讲我们的预测。

I was gonna move on to that with our predictions.

但这和Wix的事情有关。

But it's related to the Wix thing.

然后我看了Cloudflare，它在我最关注的前四名里。

And then I was looking at Cloudflare, which is on my top top four.

对吧？

Right?

Cloudflare一年前增长27%，上个季度增长了34%。

Cloudflare, 27% growth a year ago, 34% growth last quarter.

我的意思是，这个增速相当不错。

I mean, that's pretty good acceleration.

这增速简直太棒了。

That's pretty damn good acceleration.

净新增客户同比增长40%。

40% year over year increase in net new customers.

公开市场方面，我们知道今年发生了些变化。

The public markets, you know, we were wondering what's happened this year.

你或者你上周一请她来过，当时她提到了这一点。

And you or you had her on from Monday on on a week ago on on twenty BC, and he said it.

我们必须加快步伐，才能重新赢得公开市场的信任。

We've gotta we've gotta accelerate to get back credibility with the public markets.

加快。

Accelerate.

不要只是温和地放缓，像2022年到2025年底那样。

Not manage a genteel deceleration, which was '22 to late twenty twenty five.

不要试图通过提高净利率来实现温和放缓，以前大家都觉得规模化之后就该这么做。

Not not manage a gentle deceleration with higher net margins, which everyone thought was the job at scale.

现在，事实是你们必须拥抱云，这不只是Pounder的事。

Now, the truth is you've gotta be cloud it's not just Pounder.

你们必须成为Cloudflare，或者做得更好。

You gotta be Cloudflare or better.

所以我不愿意去运营Wix，因为从44亿的基础增长到10亿，确实很棒，也证明了交叉销售是可行的，对吧？

And so I wouldn't wanna be running Wix because a 100,000,000 from base 44, it's great and it shows you can cross sell, right?

这还展示了大量新事物，但他们的大客户数量却持平或下降。

And it shows a lot of new things, but their larger customer count is flat to down.

所以他们面临着核心业务下滑的巨大惯性。

So they've got the gravity of the decline of the core business.

他们必须重新实现如此强劲的增长，以至于10亿的收入根本不足以产生影响。

Like, they've gotta get back to so much growth that a 100,000,000 just it's not enough.

罗里可以帮我们算一下这笔账，但我认为，以44亿为基础，再加上一些相互侵蚀，Wix要想在核心增长放缓的情况下产生实质性影响，必须做到5061亿美元。

Rory could help trail out the math, but I think base 44, and there's some cannibalization there, gotta be doing $506,100,000,000 dollars to move the needle for Wix with its core growth decelerating.

这实在太难了。

It's just so hard.

对于初创公司来说，今年你就得开始思考，如果它们没有加速增长，什么时候该放弃投资组合中的项目。

And for startups, you've to start wondering this year when you should just give up on your portfolio if they're not accelerating.

到底在什么时刻，你才该彻底放弃？

At what point do you just give up?

因为如果公开市场无法容忍温和的放缓，私人市场又怎能容忍呢？

Because if the public markets won't tolerate gentle deceleration, how will the private markets tolerate it?

温和放缓的时代已经结束了。

The era of gentle deceleration has ended.

它已经死了。

It's dead.

我们并不认为2022到2025年是个特别好的时期，但至少那时温和放缓还能被接受。

We didn't think it was such a great period, twenty two to twenty five, but it was pretty nice that gentle deceleration was tolerable.

again，有太多需要梳理了。

Again, so much to unpick.

首先，明确一点，温和放缓就像熵。

First of all, to be clear, gentle deceleration is like entropy.

它是宇宙的最终状态，因为一切都会放缓至GDP增速。

It is the end state of the universe because everything decelerates to GDP growth.

嗯，但就像，宇宙在这些事情发生之前就已经死了。

Well, yeah, but but, like, the universe will be dead before some of it happens.

同意。

Agree.

所以我想说得精确一点，因为你了解我。

So I just wanna be precise because you know me.

我就是那种人。

I'm that guy.

对吧？

Right?

你真正想说的是，当年增长率达到10倍时，Anthropic有能力放缓。

What you're really saying is at 10 x growth year on year, Anthropic can afford to decelerate.

当增长率只有40%时，就像Figma几个月前那样，你确实放缓了30个百分点。

At 40% growth, like FIGMA had a couple of quarters ago, yeah, you've de accelerated 30.

你会受到一点影响。

You're gonna take a little hit.

但你说得完全对，杰森。

But you're exactly right, Jason.

你真正想说的是，在10%的增长率下，你们已经接近GDP水平，必须重新加速。

What you're really saying is at 10% growth, you're now so close to GDP that you've got to reaccelerate.

所以这正是我所说的精确性。

So that's really what it Just being precise.

我不

I don't

我不知道，即使公开市场认为Figma放缓是可以接受的。

know if even though the public market think it's okay that FIGMA might decelerate.

对吧？

Right?

但我再说一遍，停一下。

But I again, stop.

我得继续深入这一点。

I gotta push on that.

这会不会也适用于CrowdStrike？

Could this take CrowdStrike?

对吧？

Right?

略微放缓，27%的收入增长，2023年的指引。

Slight deceleration, 27% revenue growth, guidance of '23.

他们可能会超出两个百分点。

They'll probably beat that by two.

当你达到一定规模时，增长25%是可以接受的。

It's okay to grow 25% when you have scale.

我的意思是，你知道，微软和谷歌都在增长13%。

I mean, you know, Microsoft and Google are all going at 13.

对吧？

Right?

所以，是的，你必须保持客观。

So, yeah, it you just have to be objective.

当你有足够的规模时，25%的增长已经足够重要了。

When you have enough scale, 25% is enough to matter.

不是所有事情都会一直加速，杰森。

Not everything's gonna accelerate all the time, Jason.

没错。

No.

但今年市场一直很残酷。

But the markets have been brutal this year.

我认为你分析或实证上并没有错，但市场已经放弃了那些未能加速增长的公司。

And I think it's not that you're wrong analytically or empirically, but the markets have given up on folks not accelerating.

它们已经完全抛弃了这些公司。

They've just entirely abandoned them.

嗯，是的。

Well, yes.

再次同意。

Again, agreed.

我是在精确表达。

I'm being precise.

他们所做的就是，三年来一直认为这是暂时的，但现在我意识到这是永久性的。

What they've done is they've said, for three years, I thought it was temporary, and now I recognize it's permanent.

现在我会根据这个增长率来估值，扣除技术过时带来的影响，再扣除自由现金流的折价，然后给你八到九倍的收入或EBITDA估值。

And now I'm going to value based on this growth rate, take a hit off you for technical obsolescence and the terminal value, take another hit off you for free cash flow, and I'm gonna give you eight or nine times revenue EBITDA revenues.

我刚才那番话挺到位的，对吧？

That was sweet that I said that, wasn't it?

那些是好的，没错。

Those are the good No.

你说得完全正确。

You're exactly right.

所以他们已经修正到了真正价值的水平，没有任何魔法加持，但从长期来看，这确实是个艰难的处境。

So they've corrected to fundamentally what they're worth with no pixie dust, and that's just a tough place to be long term.

你，我同意。

You I agree.

那些公司，负面的说法是他们必须重新加速增长。

Those guys, the negative spin is they have to reaccelerate.

积极的一面是，即使只是轻微的重新加速，也会让你从现在的位置获得一些提升，但你必须做到。

The positive spin is even a small amount of reacceleration, you know, will get you some lift from here, But you gotta do it.

无论如何，Wix 就是一个很好的例子。

Anyway, Wix is a good example of that.

这挺有意思的。

It's it's funny.

我们来拆解一下 Wix 吧。

Let's try and take apart Wix.

我这次没做准备，但是

I hadn't prepared on this one, but

有20亿，增长率是13%。

There are 2,000,000,000 growing 13%.

如果这1亿美元的业务能像Lovable或Repless那样保持同样的增长轨迹，我来算一下，你们都知道这些数字，但我们还是来推演一下。

If that $100,000,000 business grew at the same trajectory as Lovable or Repless I'm gonna do some ahead, and you guys know the numbers, but let's just play it out.

他们一年内从1亿增长到了3亿左右。

They went from a 100 to 300 plus or minus in a year.

公平吗？

Fair?

是的。

Yep.

而且再过一年，至少会从300增长到600左右。

And 300 to 600 plus or minus in another year at least.

如果Wix的基数是20亿，今年从1亿增长到3亿，那就说明它现在还很小。

If base 10, which is inside Wix at 2,000,000,000, went from 100 to 300 this year, that would mean it's small today.

从10开始，问题在于：

From 10 the problem is this.

从10到100，会被淹没在噪音中。

From 10 to 100, it gets lost in the noise.

从100到300，为Wix的增长率带来了10%的提升。

From 100 to 300, it's 10% lift to Wix's growth rate.

所以如果原来是10，现在就是20。

So if it's 10 before, now it's 20.

从300到600，那就是15%的增长。

And from 300 to, say, 600, it's 15% lift.

所以你不能仅仅说它不够大，因为小事物的本质就是起初很小，然后呈指数级增长。

So you can't just say it's not big enough because the nature of a small thing is it's small and then it exponentially compounds.

所以我会反驳说，如果我

So I would push back and say if I

我也会说，从比例上看，Lovable现在是3亿，预计今年年底将达到100亿。

would I would also just say, for proportions also just to say, Lovable's at 300, and it projects it will be at 1,000,000,000 by the end of this year.

我同意。

I agree.

它就会朝那个方向发展。

That's where it's gonna go.

从零到一百并不差，杰森。

It's not that zero to 100 is bad, Jason.

但你正确的地方在于这一点。

But where you are correct is this.

如果你是一家传统SaaS公司，内心有一个从零增长到一百的亮眼项目，对吧？你就必须以一种能让它跟上Lovable和Replies的方式去运营它，它必须从一百增长到三百，再到六百，甚至可能到十亿。

If you're an old school SaaS company and you have this bright shiny thing inside you that's growing from zero to 100 in one year, right, you have to run it in such a way that it can keep up with the lovable and the replies, and it has to go 100 to three hundred and three hundred, maybe a billion, maybe 600.

你所说的没错，如果这个项目被整个公司、官僚体系、SaaS流程或其他杂七杂八的东西拖累，那你就无法推动增长。

And what you're saying, and it is correct, is if it ends up hobbled by the the wider company or the bureaucracy or the SaaS or the blah blah, then you're not gonna get the needle.

但挑战在于，他们的核心客户群，也就是付费客户，去年下降了1.2%。

Well, challenge is their core customer base, their core paid customers declined 1.2% last year.

所以他们正处于许多B2B公司都会遇到的终点状态。

So they're at a terminal state where a lot of B2B companies are.

这不仅仅是AI的问题，而是整个生命周期的问题——他们的核心客户群已经到达了瓶颈，对吧？

And it's not just AI, it is life, where they have reached a terminal state for their core customer base, right?

他们比大多数上市公司做得更好的地方是，通过一次出色的收购，额外获得了1亿美元的AI收入。

And what they've done better than most public companies is hell, with a great acquisition, they got a 100,000,000 of AI revenue on top of it.

这已经比大多数公司强了，但还不够。不仅你的数学计算如此，核心业务每年还在下降1%到2%，听起来不多，但天啊，这真的很严峻。

That's like better than most, but it is not enough yet, Not only your math, but the core is declining one to 2% a year, which doesn't sound like a lot, but, man, that's rough.

对吧？

Right?

但是，杰森，我觉得我们实际上说的是同一件事。

But, Jason, just doing I I think we're actually ending up saying the same thing.

所以你每年流失两千万美元。

So you're declining $20,000,000 a year.

这有什么大不了的。

Whoop de do.

问题是，你没有新增客户。

Well, you're not adding new customers is the problem.

对吧？

Right?

是的。

Yeah.

这无关紧要。

It doesn't matter.

别说了。

Stop.

这并不重要。

It doesn't matter.

如果Base24能像Lovable或Repless那样增长，那么两年内，他们就能摆脱困境，成为一家年增长30%的公司。

If Base24 could grow like Lovable or Repless, then within two years, they'd be out of the woods, and they'd be a 30% growth company.

所以你说得完全对。

So you're exactly right.

他们做出了正确的决定，但你实际上是在说，在另一家公司内部，他们可能做不到。

They made the right move, but what you're effectively saying is inside that other company, they might not be able to.

所以我认为，作为投资者，你应该对自己说：如果我们认为Base44有相同的增长轨迹，那太好了。

So I think as an investor, you should be saying to yourself, if we think Base forty four has the same trajectory, then great.

那么你这里就发现了一颗璞玉。

Then you have a diamond in the rough here.

你免费获得了现有业务，同时还获得了新业务。

You get the existing business for free, and you get the new business.

但我认为，你的董事会潜意识里是在说，它不会按这个轨迹增长。

But I think implicitly, your board's saying that it's not going to grow at that trajectory.

它可能会长到100%到200%，所以增长率可能会回升到百分之十几，但这还不足以带来巨大改变。

It's probably gonna grow 100 to 200 to so it's probably gonna get the growth rate back up to the mid teens, but it's not gonna be enough to dramatically change

这份工作。

the job.

我认为时间会证明一切。

I actually think there is a time will tell.

这么说吧，如果我们以Wix作为案例研究，那么每个规模化B2B公司都应该希望Base44能成功，原因是，你看，我非常喜欢Repo和Lovable。

Here's what and if we're using Wix as a case study, here's why, like, every B2B company at scale should hope that Base 44 works, which is, look, I love Repo, Lovable.

它们都很棒。

They're great.

明白吗？

Okay?

但Wix有611万客户。

But Wix has 6,110,000 customers.

我敢打赌，上过哈里节目那位创始人会反驳我，但我知道其中一半成功的原因。

I'm sure the founder who's been on Harry show will challenge me, but I know half the reason it's working.

这是一个好产品。

It's a good product.

对吧？

Right?

他们有611万客户可以销售给。

Is they have 6,110,000 customers to sell it to.

是的。

Yeah.

这没问题。

And that's okay.

不。

No.

这没问题。

It is okay.

所以他们应该能成功实现。

So they should be able to pull it off.

一亿根本不够。

It's not a 100,000,000 ain't enough.

比如，所有关于它有多棒的推文，但规模实在太吓人了。

Like, all the tweets about how great it's just the the scale is so intimidating.

这很难。

It's tough.

你得像马克那样，再买个Informatica来填补差距。

You gotta, like, do what Mark did and buy an Informatica too to bridge the gap.

如果你是一家现有的SaaS公司，杰森实际上是在说，即使你做出了正确的战略决策，也还是不够，因为你由于某些制度性原因无法实现目标，这就意味着对大多数公司来说前景黯淡。

If you're in a game as an existing SaaS company, what Jason's effectively saying is even when you make the right strategic moves, it's not enough because you can't pull it off for whatever institutional reasons, then that points to a grim conclusion for most of those companies.

这个案例研究是Base44，它是个好产品。

The case study will be Base 44 is a good product.

我承认，我不会争论它在 vibe 编码等级中的位置。

I will stip I'm not gonna argue where it stands on the hierarchy of vibe coding.

我用过它。

I have used it.

这是一个好产品。

It is a good product.

团队很精简，但产品很好。

It's a lean team, but it's a good product.

如果Wix都无法向610万用户成功交叉销售，那你还有什么希望？

If Wix can't frack and cross sell that to 6,100,000 people, what hope is there for you?

这对许多创始人来说是个反问句。

This is a rhetorical question to many founders.

这是一个令人沮丧的问题，因为我可以告诉你，任何人都更愿意用振动编码来搭建网站，而不是处理它自带的糟糕模板。

It's a bleak question because I can tell you, anyone would rather vibe code a website than deal with the crappy templates it comes with.

所以这是一个绝佳的使用场景，对吧？

So it's a great use case, right?

611万。

6,110,000.

如果它不工作

It's got if it doesn't work

杰森，看看接下来会发生什么。

Jason, see what's coming.

我觉得你说得完全对。

I think you're exactly right.

我觉得Wix是一个完美的使用案例。

I think Wix I think it's a perfect specimen use case.

如果你无法将新的AI产品交叉销售给那些正试图用旧技术做同样事情的现有客户，那么交叉销售、获取和交叉销售的故事就站不住脚。

If you can't cross sell the new AI product to existing customers who are trying to do the same thing just with an older technology, then the cross sell acquisition and cross sell story doesn't work.

如果你是一家传统SaaS公司，除了管理客户之外，这已经是最好的情况了。

This is as good as it's going to get if you're an old school SaaS company other than managing the client.

所以你是这样的。

So you are.

这完全是一次实验室实验。

It's a total lab experiment.

当你观察Salesforce明智地收购几家下一代AI优先公司时，我认为它们也必须跳同样的舞，只是每个数字都要多一个零。

As you're watching Salesforce, in my view, wisely buying a few next generation AI first companies, they've got to do the same dance, just with an extra zero everywhere.

他们必须把这个400亿美元的庞然大物，找到办法再卖出200亿美元的产品。

They've got to take that $40,000,000,000 behemoth and find a way to up sell another $20,000,000,000 worth of stuff.

至少他们在市场上还有点东西可以玩。

At least they have something in this market to play with.

我看看你们这些Webflow、Squarespace之类的产品，就觉得头疼。

I I look at your web flows of the world in your Squarespaces, and I go, ouch.

是的。

Yeah.

真不明白他们为什么不去尝试。

It's it's a mystery why they're not trying.

我这周做了一件事，第一次真正试用了Figma Make。

One thing that happened to me this week is I tried Figma Make for real for the first time.

我知道Harry在Twitter上看到过。

I know Harry saw it on Twitter.

但对我而言，它的体验太糟糕了。

And for my use case, it was terrible.

这比Base 44差多了，比我用过的任何振动涂层产品都差，Figma Make也不例外。

It was much worse than Base 44, much worse than any vibe coating product I've ever used Figma Make.

太糟糕了。

It was terrible.

它根本没经过精心设计，更糟的是，所有这些产品都已经进步了。

It was undesigned, and even worse, all these products have advanced.

所以我做了这个测试，很多人笑话我。

So I do this test and a lot of people make fun of me.

他们说：‘你真是个傻瓜。’

They're like, Oh, you're an idiot.

你根本不会写提示词。

You don't know how to do a prompt.

我可是写过几个提示词的。

I've done a few prompts.

好吧。

Okay.