Stack Overflow 用户不信任人工智能，但他们仍在使用它

有趣的是，我们有一个叫做‘一九九规则’的规律。

What's interesting is that there is a, we have something called a one-nine-ninety rule.

1% 是那些投入大量时间、心血和汗水来整理知识、在网站上花费时间的硬核用户，9% 是以中等程度参与的用户，而90% 的用户则主要是在浏览和充当旁观者。

1% are the hardcore users who have been spending a lot of time with their blood, sweat and tears, curating knowledge, spending their time on the site, etcetera, contributing 9% doing it in a medium way and then 90% sort of consuming and mostly being a lurker on the site.

关于网站的分布情况，当我们询问人们是否在使用AI时，我们自己的调查数据显示：如果查看Stack Overflow 2025年的调查，超过80%的社区成员正在使用AI或打算使用AI，80%，对吧？

And the distribution of the site, and when we ask people on whether or not they're using AI, our own surveys basically say, if we took a look at the Stack Overflow 2025 survey, over 80% of our community members are using AI or intend to use AI, 80%, right?

但当他们使用AI时，对这一答案的信任度仅有约29%。

But the trust level on that answer when they're using AI is only about 29%.

只有29%的用户真正信任AI输出的内容，这实际上非常合理，因为我们正处于这一新技术的初期阶段，理应对它保持怀疑态度。

Only 29% of our user base actually trusts what's coming out of AI, which is actually quite appropriate considering where we are because there should be skepticism of this new technology.

因此，人们乐于尝试，但并不完全信任。

So there's enthusiasm to try it, but not to fully trust it.

结合这一‘一九九规则’，我认为我们拥有一群核心用户，他们始终是公司原始使命的守护者——即创建一个完全准确的知识库，仅此而已。

And with this one-nine 90 rule, I think what we have is we have a core group of users that are always going to be the protectors of the original mission of the company, which was to create this kind of knowledge base that was completely accurate and do nothing more than just that.

而另一方面，我们也有大量用户，比如年轻开发者或下一代开发者，他们希望利用最新、最前沿的工具。

And then you're going to have, we have a very large number of people who are, let's say younger developers, people, the next generation of developers who are looking to leverage the latest and greatest of tools.

根据我们进行的调查和额外研究，我们很清楚，他们希望使用自然语言作为接口来实现这一点。

And it's very clear to us based on surveys that we've done and additional research that they want to use natural language as the interface to be able to do this.

这是计算机科学发展史上最具意义的范式转变。

It is the most meaningful seed shift sort of change in terms of computer science development.

如果你回溯到几十年前的面向对象编程，那其实并没有带来如此巨大的飞跃。

If you look back to all the way to even object oriented programming many, many decades ago, that wasn't such a huge boom actually.

它并没有真正促成质的飞跃。

Like it didn't actually create the step change.

但现在，我们正处在一个一切都被解锁的时刻。

But now we're in this moment where everything's been unlocked.

因此，我认为这是一次巨大的变革，我们必须决定：既要尊重原始使命，又要将准确性置于核心位置。

So I think it's a huge change effort and we've had to sort of decide that, hey, we've got to be able to respect the original mission, keep accuracy at the heart of it.

因此，我们对使用AI生成答案感到不安，因为它会产生错误内容，会幻觉，这就是信任评分低的原因，但为什么我们不采用自然语言接口呢？

And so we're not comfortable with using, for example, AI for answers, because it will generate slop, because it hallucinates and that's hence the trust score is low, but why don't we incorporate natural language interfaces?

因此，这是更理想的互动方式。

So that's the preferred way to sort of engage.

因此，我们在公共领域和企业领域都采用了这种方式。

So we ended up doing that both on the public side, as well as on the enterprise side of the house.

这得到了绝大多数用户的热烈欢迎，但总会有一部分声音强烈的少数群体反对这种对话方式的融入。

And that's been really well received by the vast majority of users, but there will be always a vocal minority that will push back against incorporation discourse.

有很多人，不仅仅是对网站本身，而是对这一切对工作产生的影响感到更广泛的担忧。

There's a lot of, you know, not beyond just the site, there's just a level of, I think, broader concern about what all this does to jobs.

如果我们把这只猫放出来，接下来会发生什么？

And if we let the, you know, cat out of the bag, then what's going to happen?

你知道的？

You know?

所以我认为，这种担忧显然是可以理解的。

So I think there's that obviously concern also, which is understandable.

让我把这个观点说得更明确一些。

Let me put a pretty fine point on that.

我认为我以一种更清晰的方式理解了这一点。

I I think I understand that in a maybe a sharper way.

如果我是你那1%中的一员，花了大量时间在Stack Overflow上帮助他人，我之所以免费回答问题，是因为我能直接看到我的努力帮助了他人成长，我在帮助别人解决问题。

If I am somebody who in in in your 1%, who spent a lot of time on Stack Overflow, helping other people, and the reason I answer questions for free on your platform, which you monetize in lots of ways, is because I can directly see that my effort helps other people grow, and I'm helping other people solve problem.

这是一个非常自洽的动态。

That is one very self contained dynamic.

是的。

Yeah.

你上次做客节目时，我们整个对话都是关于这个动态的。

The last time you were on the show, our entire conversation was about that dynamic.

是的。

Yes.

对，以及你是如何让人们参与这个动态，以及这个动态的价值。

Right, and how you got people to participate in that dynamic and the value of that dynamic.

但突然间，拥有这个数据库的公司获得了明确的经济利益，因为他们把我的贡献卖给了OpenAI，对吧？这正在整个行业普遍发生。

And then suddenly there's a very clear economic benefit to the company that owns the database because they're selling my effort to OpenAI, right, which is a thing that is happening across the board, right.

我们即将与所有这些AI提供商签订数据授权协议，他们会用我 painstakingly 输入到这个数据库中的答案进行训练，而下一代软件工程师将获得基于我工作的自动补全功能，而我却一无所获。

We're gonna we're gonna do all these data licensing deals with all these AI providers, they're gonna train on the answers that I have painstakingly entered into this database to help other people and now the next generation of software engineers is gonna get auto complete that's based on my work and I've gotten nothing.

我的意思是，我已经从很多人那里听到过这种说法。

I mean I've heard that from lots and lots of people.

对吧？

Right?

在我的社区里，我也听到过这种说法；我认为，当各种媒体公司达成这些协议时，我也感受到了这一点。

I've heard that in our own community, I think I have felt that as various media companies have made these deals.

你如何回应这一点？

How do you respond to that?

因为这感觉像是你提供的一个数据库，你必须以某种方式将其商业化，但人们之间的互动才是真正的价值。

Because that feels like what you were providing was a database that you had to monetize in some ways, but the people, the interaction that the people had was the value.

而现在，另一种经济价值可能正在掩盖、重塑或重新定义人们之间的互动。

And now there's another kind of economic value that is maybe overshadowing it or recasting or recharacterizing the interaction that people have.

这里有几点需要说明。

There are couple of points there.

一方面，如果你思考一下这家公司的原始基因以及人们为何聚在一起做这件事，你知道，当我加入公司时，我问过一个问题：人们为什么愿意花时间做这些？

One is, I think if you think about the original DNA of this company and why people came together to do this thing, you know, when I joined the company, I asked a question like, what's people's incentive to spend time doing this?

我向创始人之一乔尔·斯波尔斯基问过这个问题。

And I asked the founders, specifically Joel Spolsky about this.

他的观点是，软件开发社区非常高效。

And his point is that the software development community is effective.

它非常利他。

It's a very altruistic.

大家只是想互相帮助，因为人们都明白这有多令人沮丧。

Just wanna help each other out, you know, because people understand how frustrating.

我很多年前也写过代码。

I used to write code many years ago.

最近我用一些代码生成工具重新拾起了编程，这很有趣，可以对比一下。

I recently picked it back up with some of the CodeGen tools, which is interesting, compare and contrast.

但我还记得，当你卡在某个问题上时，有多让人沮丧。

But I just remember how frustrating it was if you get stuck on something.

因此，当Stack Overflow诞生时，它无疑极大地促进了这一点。

And so Stack was obviously a huge boon when it was created to unlock this.

因此，这正是其真正的由来。

And so it was truly out of that, that was the reason.

我们还提出了一个问题，即使在ChatGPT出现之前，我们是否应该通过付费来激励用户？

We also asked the question like, should we, even before ChatGPT, should we incentivize users by paying them for that?

我们应该给予他们经济回报吗？

Should we give them monetary benefit?

但这并不是用户群体的高要求。

And that wasn't like a high ask by a user base.

我们去研究了人们。

We went and researched people.

人们并不是为了钱而来，而且这会让事情变得复杂，因为你怎么判断一个JavaScript问题的报酬相对于一个Python问题？

So people were not in for the money, plus it complicates things because how do you judge the payment for a particular JavaScript question relative to a particular Python question?

你知道，这会陷入一个非常、非常深的困境，我认为这是一个不可持续的困境。

You know, it's just a very, very, it goes down a rabbit hole, which is I think a sort of a untenable sort of rabbit hole.

因此，这个商业层面，这就是其中一个原因。

And so this commercial aspect, so that's one.

那么，人们最初聚在一起的真正原因是什么？是因为这个使命吗？

So what is the original reason why people got together and it was about the mission?

其次，从某种意义上说，我认为我们必须这样做，这是否不公平等等。

Secondly, in terms of like, I think why we have to do this and is it unfair and so on.

我们不得不走上网站数据授权这条路的主要原因，是因为互联网的模式已经彻底颠倒了。

The primary reason why we have to go down the site data licensing route or why we've had to do it is because the model of the internet has literally been turned upside down.

人们曾经依赖——我知道你和阿尼尔谈过这个问题，关于DoorDash的困境。

People relied on, and I know you talk about this, Anil, with the DoorDash problem.

我认为互联网的模式已经彻底改变了：人们现在去搜索引擎，去网站，靠广告盈利——这种模式已经完全变了。我非常理解那些严重依赖广告的内容网站，因为我认为大多数内容网站的流量已经下降了30%、40%，差不多就是这样。

I think the model of the internet has literally where people are, you know, go to search engines and go to, you know, websites and, you know, you monetize off of ads, you know, that is completely, really empathize with, you know, content sites that are heavily dependent on advertising, because I think most content sites, their traffic is down 30%, 40%, something like that.

在这种巨大的范式转变下，支持这些平台的公司最终必须——我们终究是一个商业实体。

So with this huge seed shift and where companies that support these platforms have to ultimately, we're a business ultimately.

那么，我们该怎么做？

So what do we have to do?

我们必须采取必要措施，建立新的商业模式，以生存、发展并实现所有目标。

We have to do what is necessary to adopt a new business model to survive and thrive and do all the things.

幸运的是，我们拥有一个独立于这一切的企业业务。

And so thankfully for us, we've had an enterprise business, which is independent of all of this.

幸运的是，我们仍然保留了面向大型广告主的广告业务，他们依然关心我们的社区。

Thankfully for us, we were now able to we still had the advertising business for large advertisers still cared about this, our community.

因此，数据授权在确保我们能够及时变现的同时，也让我们有能力将收益回馈给社区，让那些真正出于正确动机参与的人受益。

And so data licensing only felt right in terms of making sure that we can effectively capitalize in the moment, plus also be able to invest back into our community so that people who are there for the right reasons saw the benefits of that.

因此，我们投入了所有这些新功能，比如我刚刚提到的各种新内容类型、挑战、聊天功能、AI助手等，这些都需要资源来开发。

So we've invested with all these new features I just mentioned, whether that is all these new content types or challenges or chat or AI assist or any of these things all takes resources to go and build.

所以我们不得不利用这些获得的资金来实现这些目标。

And so we've had to go and leverage these funds that we received to be able to go do that.

未来，我们可能会考虑其他方式，比如，我们是否应该向用户付费，让他们分享一部分数据授权的收入？

Now, the future, we may consider other ways to make it, know, for example, should we pay our users, give them a piece of the data licensing revenues?

也许，我们会一直提出这个问题。

Perhaps, you know, we'll ask that question always.

我们总能找到继续前进的方式，但目前这就是我们现有的模式。

There's always ways for us to continue, but right now this is the current setup that we have.

你提到为了达成数据许可协议，你们不得不部署大量反爬虫工具，还得深入到技术栈的二级和三级层面，才能与Databricks及其他供应商达成合作。

You mentioned to to get to the data licensing deals, you had to put up a bunch of anti scraper tools, you had to go, you know, into secondary and tertiary layers of the stack to get deals from Databricks and other kinds of providers.

这些AI公司之前一直在抓取你们的网站。

The AI companies, they were just scraping your site before.

对吧？

Right?

他们很过分，而且很可能现在还在这么做。

They were mean, they they probably still are.

不管他们有没有付钱给你，他们大概还是直接从入口爬取数据，因为确实如此。

Like, whether or not they're paying you, they're they're probably still just going through the front door because that Yeah.

他们似乎都在这么做。

All of them appear to be doing that.

你们是不是得先说‘我们停止你们的行为’，然后再去谈合作？

Did you have to say we're stopping you and then go get the deal?

还是你们说：‘我们知道你们在这么做，但你们必须付钱，否则我们就起诉你们？'

Or did you say, hey, we know you're doing this, but you have to pay us or we're going to start litigating?

我觉得介于两者之间。

Somewhere in between, I would say.

我认为我们很快就开始部署反爬虫工具了。

I think we put up the anti scrapers very quickly.

我们甚至改变了人们获取我们数据转储的方式，因为我们需要在两者之间取得平衡——我们从不希望阻止社区用户为合法需求获取我们的数据，比如做学校项目、博士论文之类的。

We even sort of changed the way in which people received our data dumps and, you know, because we want against as a balance because we never wanted to prevent our community users from grabbing our data for their legitimate needs, you know, for them to do their school projects or PhD theses or anything like that.

因此，我们对社区成员继续保持数据开放，但他们必须是社区成员，而不能是那些企图商业化利用数据的公司。

So we've continued to be open about our data for our community members, but they have to be community members and there can be companies looking to commercialize off the data.

所以我们对政策条款非常明确，部署了技术手段来阻止人们抓取数据。

So we were very specific about the policy terms, putting up technology that prevented people from grabbing it.

因此，我们确切知道谁在爬取数据，谁没有爬取。

So we know exactly who is scraping, who is not scraping.

其中一部分工作是联系那些人，告诉他们：请停止，因为你们的行为明显给服务器带来了巨大压力。

Some parts of those were outreach to those folks who say, look, you know, stand down because you're just, you're obviously putting a lot of pressure on the servers by doing what you're doing.

所以，请放轻松一些，这里有个警告。

So, you know, take it easy and warrant here.

但我认为对这些公司的描述是：它们不在乎，或者有些公司在乎并希望成为好公民，而有些则完全不在乎，宁愿继续偷偷抓取数据。

But I think my characterization of those companies is they don't care Or some of them care and they they wanna be good citizens, and some of them absolutely do not care and they would prefer the smoke.

你可以直接将它们分类。

You can just categorize them.

亚马逊起诉Perplexity是有原因的。

There's a reason Amazon is suing Perplexity.

他们曾告诉Perplexity

Like, they told Perplexity

是的。

Yeah.

停止这种行为，但Perplexity拒绝了。

To stop it and Perplexity won't.

就在今天早上，我们谈话的时候，《纽约时报》也在起诉Perplexity。

As we're speaking today, this morning, The New York Times is suing Perplexity.

还有其他一些参与者采取了不同的做法，并达成了各种不同的协议。

Then there are other players who are acting in different ways and they're striking different kinds of deals.

给我讲讲其中一笔交易的具体情况。

Walk me through one of those deals.

当你和OpenAI达成协议时，是你们说‘我们要阻止你，如果你想开门，就必须付钱给我们’，还是你们说‘这做法不对，我们可以采取所有技术和法律手段，但其实我们不如直接好好谈个协议’？

When you went and struck your deal with OpenAI, was it we're gonna stop you and if you want the door to be opening and you have to pay us or was it you know this is wrong, we can take all the technical and legal measures, but we should actually just get to the deal correctly?

跟我详细说说那次对话的过程。

Like, walk me through that conversation.

对于像他们这样的公司，我们其实已经把OpenAI的技术整合到我们的产品中了。

With some folks like them, we were already, you know, we were incorporating something like OpenAI into our product.

还记得Code Red事件吗？当时我们正要发布我们的AI回复功能。

Remember the Code Red situation where we were about to announce our AI response?

所以我们实际上在使用这项技术，来实现将AI整合到公共平台以及我们的企业产品中的目标。

So, we were actually using that technology to do what we had to do to incorporate AI into the public platform, as well as our enterprise products.

因此，我们和他们早就有合作关系。

So, we had a relationship with them.

我们也明确表示：这种方式行不通，不可持续，这才是未来工作的正确方式。

And we also said, look, this is not gonna work, it's not tenable and this is the new way of working.

所以我们需要一个新的安排，一个关于你使用数据的商业合作方式。

And so meaning we need a new arrangement, a business arrangement for you to use the data.

因此，让我们真正进行一次对话；值得称赞的是，他们在这一点上非常以合作伙伴为中心。

And so let's actually have a conversation and to credit to them, they were very partner centric around that.

我对OpenAI以及像谷歌这样的公司印象深刻，他们都非常愿意就这一话题展开合作，并希望成为负责任的AI合作伙伴。

I was very impressed by both OpenAI and companies like Google who are all very open to engaging on this topic and wanted to be responsible AI partners.

他们立刻就明白了，甚至在我们提出之前就已经理解了，这并不是那种需要从头开始、大费周章地解释为什么必须这么做的对话。

They got it immediately, even before we asked them, it wasn't like this kind of big, let's go have this conversation kind of from the ground up and justify why it had to be done.

我们只是说：看，这种情况必须发生，因为这是一个新的商业模式。

We just said, look, this is what needs to happen because this is a new business model.

我们很快就进入了对话，讨论如何以建设性的方式明确你们究竟想要什么？

And we got into the conversation pretty quickly and that, okay, let's actually have a constructive way to what exactly are you looking for?

你们希望以什么格式抓取内容数据？

Which format of data do you want to scrape the content?

你们希望进行批量上传吗？

Do you want bulk uploads?

你们想要API调用吗？

Do you want API calls?

你们想要什么？

What do you want?

所以我们进入了整个混合状态。

So we got into that whole mix.

当然，还有一场关于这些交易的对话，提醒你一下，Nile，这些都是经常性的收入型交易。

And then of course there is the conversation around, these are recurring, just mind you, Nile, that these are recurring revenue type of deals.

这些不是一次性付款。

These are not one time payments.

所以是的，他们是非常有合作精神的伙伴。

So yes, they were very collaborative partners.

但你说得对，确实有一些参与者是矛盾的。

But you're right, there are players who are contradictory.

他们嘴上这么说，但他们的行为却证明了其他事情，你知道的，就他们参与的方式而言。

They say something and their actions, I think prove others other things, you know, in terms of how they've engaged.

当然有一些人是顽固派，他们言行并不一致。

And so there are people that are holdouts for sure, people that are not exactly consistent with their word.

这很遗憾，我认为像我们这样的公司都必须决定如何应对这种情况。

And that's unfortunate, you know, it's a and I think every company like us has to decide what we do about that.

我们正与不同的人处于这些对话的不同阶段，以确保我们能让他们真正做正确的事。

And, you know, we're in various stages of these conversations with various people on, you know, how to make sure that we extensively get them to sort of do the right thing.

现在你得说出一家这样的公司。

Now you have to name one of those companies.

你觉得哪家公司公开表态和实际行动不一致？

Who who do you think is is holding out differently than their public posture?

我宁愿不点名。

I'd rather not be.

但我认为你实际上已经概括了所有那些常见的对象，你所报道的那些公司，正是我们所遇到的典型情况，我想这么说。

It's but I think you've actually, like, characterized all the usual suspects that you are you that you are covering are the usual suspects that, you know, we are encountering is how I would put that.

是的。

Yes.

让我问你关于持续收入的问题，然后我想转向解码器相关的问题，因为我觉得在讨论完这个之后，这些问题会更有启发性。

Let me ask you about the recurring revenue piece and then I wanna get into the decoder questions because I I I think they'll be illuminating after this conversation.

有一种感觉，对吧，我们已经完成了所有将要进行的预训练。

There's a sense, right, that we've done all the pre training that we're gonna do.

对吗？

Right?

从互联网上抓取数据并不是这些模型的未来。

That scraping the Internet is not the future of these models.

我们需要某种其他的突破。

That there needs to be some other leap.

Stack Overflow 现有的信息库才是有价值的，那里有大量的信息，数据库里有二十年的内容。

Stack Overflow's existing corpus of information is the valuable thing, it's the there's a lot of information, there's twenty years of stuff in that database.

你再次向我们付费以训练 Gemini 或 GPT 的下一个版本，与现有数据库中不断新增的增量信息相比，其价值何在？

What's the value of you have to pay us again to train the next version of Gemini or GPT and the value of there's incremental information being added to the existing database?

因为这在我看来是一个明显的区别。

Because that seems like a clear split to me.

我们对这个问题的看法是，每个正在训练的模型都是在某个信息语料库上进行训练的。

The way that we have thought about this is that, you know, every model that's being trained, you're training it on some corpus of information.

你知道，你从GPD X训练到Y。

You know, you're going from GPD X to Y.

因此，在你训练的新模型中，如果你使用了我们原始的数据或其从前序模型衍生出的版本，你就必须为此向我们付费。

And so in the new model that you're training, if you're leveraging our original data or some derivative of that from a prior model, then you have to pay us for it.

这实际上是能够这样做的法律要求。

That's effectively the legal requirement to be able to do that.

所以这是一个累积的过程，对吧？

And so it's a cumulative aspect, right?

所以别忘了这一点。

So let's not forget that.

因此，人们必须为累积的数据付费。

And so like people have to pay for the cumulative data.

这不仅仅是它曾经在早期被使用过的问题。

It's not just the fact that it was used in back in the day.

是的，显然，相对于二十年来说，一年的信息量会更少，这就是为什么你会得到二十加一，这就是这个思路。

And yes, obviously, relative to twenty years, you know, one year's worth of information is going to be less, that's why you're getting 20 plus one, you know, that's the idea.

所以，这仅仅是法律协议的设定方式。

And so that's just the way that the legal agreement has been set up.

是按年计算的吗？

Is it per year?

是每年的数据都算作一笔费用，还是说这是怎么运作的？

Is it every year's worth of data is a chunk of money, or how does that work?

不是。

No.

它是一个累积的，包括整个历史数据集，以及未来一年内所有新增的数据。

It's just it's a cumulative it's like the whole corpus past, you know, historical data as well as, you know, the the anything that's that's going forward for the following year.

所有这些数据被视作一个累积的数据集，实际上就按一个整体收费。

All that is sort of one accumulated sort of dataset and that's, you know, charged as one effectively.

但今年的数据并不会被纳入刚刚发布的Gemini三号的训练集。

But that doesn't so like this year's data doesn't get pulled into the training set for Gemini three, which just came out.

对吧？

Right?

自Gemini三发布以来，Stack Overflow上每一个新的问答都没有被纳入Gemini三的训练数据中。

Every new question answer in Stack Overflow since Gemini three has come out is not incorporated in Gemini three's training.

所以你实际上是赌他们会不断训练更大规模的模型。

So you're kind of betting that they're just going to train ever bigger models.

没错。

That's right.

在你看来，情况就是这样安排的吗？

That is that how it's structured in your mind?

是的。

Yeah.

而且有些公司提出了非常广泛的应用需求。

And it it and some companies have asked, you know, they've asked for very, very wide use cases.

你知道，有预训练的应用场景，甚至更进一步，你可以以多种方式利用这些数据，用于AI和非AI场景、搜索场景等等。

You know, there's pre training use cases, there are, you know, even beyond that, you you can leverage the data in many, many different ways for AI and non AI use cases, search use cases and so on.

但说得对，我认为在某些情况下，可能会构建更大的模型，而我们的数据将对这些场景有所帮助，但也会有RAG索引、后训练需求，以及各种各样的场景。

But correct, I think that we, in some, there may be scenarios where more larger models are built and our data is going be useful for those scenarios, but there's going be rag indexing, there's going to be post training needs, there's going to be all sorts of scenarios.

看到一些前沿实验室要求非常具体的数据切片，觉得它们非常有用，这非常有趣。

And it's very quite interesting to see some of the frontier labs ask for very specific slices of data that they find it to be, find very useful.

我们能够记住，这不仅仅是问答，还包括评论历史、元数据历史、投票历史，以及用户A走过的路径历史。

And we're able to remember, we've got a lot of, there's not only just a question and answer, but we've got the comment history, we've got the metadata history, we've got the voting history, we've got the user A has gone down this path history.

因此，这对于推理等任务来说提供了大量优秀的上下文，能够真正帮助我们模拟人脑。

So it's a lot of excellent context for things like reasoning and to really sort of be able to be useful to really mimic the human brain.

这几乎相当于记录了一个完整的人类大脑。

It's effectively one human brain that's been documented almost.

我们需要在这里暂停一下，进行短暂的休息。

We have to pause here for another quick break.

我们马上回来。

We'll be back in just a minute.

Decoder的支持来自Superhuman。

Support for Decoder comes from Superhuman.

人工智能承诺了很多，但在实践中，它往往只是你需要管理的又一个标签页，拖慢了你的速度。

AI promises a lot, but in practice, it often ends up being yet another tab you have to keep track of, slowing you down.

我不知道你怎么样，但我生活中需要的是更少的标签页和应用，而不是更多。

I don't know about you, but I need fewer tabs and apps in my life, not more.

欢迎使用Superhuman，这款AI生产力工具能为你在所有工作场景中赋予超能力。

Say hello to Superhuman, the AI productivity sweep that gives you superpowers everywhere you work.

通过Grammarly、Mail和Coda协同工作，你能在整个工作流程中获得主动帮助，从写作到准备会议、演示文稿等等。

With Grammarly, Mail, and Coda working together, you get proactive help across your workflow, from writing to preparing for meetings, presentations, and so much more.

与那些存在于独立窗口中的聊天机器人不同，Superhuman的AI直接融入你已有的应用和标签页中，比如你的邮件、文档以及所有工作场所。

Unlike chatbots that live in separate windows, Superhuman's AI is in the apps and tabs where you already are, like your email, docs, and everywhere you work.

把Superhuman想象成你的AI梦之队，它会主动帮你更快地从待办事项走向完成。

Think of Superhuman as your AI dream team, proactively helping you go from to do to done faster.

Superhuman知道你可能需要什么，并提供建议。

Superhuman knows what you might need and offers suggestions.

它会在关键时刻引导你，让你表现得更出色，并专注于真正重要的事情。

It guides you in the moment so you sound like your best self and stay focused on what matters.

它并没有让你变得超人。

It doesn't make you superhuman.

它为你提供了证明你一直以来就是超人的工具。

It gives you the tools to prove you always were.

释放你的超人潜能，让AI在你工作的地方与你相遇。

Unleash your superhuman potential with AI that meets you where you work.

了解更多，请访问 superhuman.com/podcast。

Learn more at superhuman.com/podcast.

那就是 superhuman.com/podcast。

That's superhuman.com/podcast.

本节目由领英赞助。

Support for this show comes from LinkedIn.

想象一下，如果所有包含‘我需要一个合适的人来做这份工作’这句台词的电影，都妥协为‘我随便找个人就行’。

Imagine if any of the movies that included the line, I need the right person for the job, settled for, I'll just take about anyone.

有多少次劫案会失败？

How many heists would have failed?

有多少笔交易会因此泡汤？

How many deals would have fallen through?

有多少秘密特工任务会以灾难告终？

How many secret spy missions would have ended in disaster?

那么，为什么在为你的企业招聘时，你要接受任何一个人呢？

So why would you accept just anyone when hiring for your business?

当你需要找对的人来做这份工作时，你可以使用LinkedIn职位。

When you need the right person for the job, you can turn to LinkedIn jobs.

现在，LinkedIn职位推出了全新的AI助手，让你有信心找到那些在其他地方都无法找到的顶尖人才。

And now LinkedIn Jobs is stepping things up with their new AI assistant, so you can feel confident you're finding top talent that you can't find anywhere else.

通过LinkedIn职位AI助手，你可以跳过那些令人困惑的步骤和招聘术语。

With LinkedIn Jobs AI assistant, you can skip the confusing steps and recruiting jargon.

它会根据你为职位设定的标准筛选求职者，只呈现最匹配的人选，让你不必被困在堆积如山的简历中。

It filters through applicants based on criteria you've set for your role and surfaces only the best matches so you're not stuck sorting through a mountain of resumes.

一次就招对人。

Hire right the first time.

在 linkedin.com/partner 免费发布职位，然后推广该职位以使用 LinkedIn Jobs 的新 AI 助手，更轻松快捷地找到顶尖候选人。

Post your job for free at linkedin.com/partner, then promote it to use LinkedIn jobs new AI assistant, making it easier and faster to find top candidates.

免费发布职位，请访问 linkedin.com/partner。

That's linkedin.com/partner to post your job for free.

条款和条件适用。

Terms and conditions apply.

欢迎回来。

Welcome back.

我正在与 Stack Overflow 的首席执行官普拉桑特·钱德拉塞卡尔交谈。

I'm talking with Prashanth Chondrasekar, CEO of Stack Overflow.

我们已经讨论了很多关于 AI、模型和颠覆的话题。

We've talked a lot about AI and models and disruption.

这些都是重大问题。

These are all big issues.

那么，这些如何融入 CEO 的基本工作——做出决策并维持公司运转？

So how do they fit into the fundamental work of being a CEO, which is to make decisions and keep the company running?

这让我想到了 Decoder 的问题。

This I think brings me to the Decoder questions.

你重新调整了公司结构，经历了几轮裁员，显然你已真正将重心重新放在 SaaS 业务上，我觉得我们应该谈谈这一点。

You've restructured the company, there's been some rounds of layoffs, you've obviously refocused on the SaaS business in a real way, I think we should talk about that.

但问题是，我们是应该继续训练更大规模的模型，让这成为业务增长的核心，还是仅仅需要一些细分领域，或者实际上，检索增强生成（RAG）才是未来大多数企业的方向。

But the idea that we're going to train ever bigger models and that will be the growing part of the business versus we might just want some slices versus it's actually retrieval augmented generation or RAG that's going to be the future for a lot of these other businesses.

根据哪一个方向会增长得更快，你会做出不同的决策。

You would make different decisions based on which one of those is going grow faster.

我认为没人知道。

I don't think anybody knows.

没错。

Correct.

也许你知道，或者你认识知道的人，但我们现在正处于这一领域发展的非常早期阶段。

Maybe you know, you can tell me if if you you know or you know someone who knows, but we're in a very nascent period for all of this kind of development.

你是如何架构公司的，以平衡所有这些风险，并为未来人们真正需要数据的方式做好准备？

How have you structured the company to even out all of that risk and and be prepared for how do people actually need the data in the future?

因为我不确定，也许你知道，但我觉得没人知道，我觉得我也不知道。

Because I don't maybe you know, but I don't think anybody, I don't think I know.

是的，是的，很难明确预测。

Yeah, yeah, it's hard to predict clearly.

而且，你知道，谁知道呢？我们有一些像谷歌的德米斯·哈萨比斯这样的杰出人才，他们正在开发下一代类似Transformer技术的突破，以持续推进他们共同追求的终极目标——通用人工智能。

And also, you know, who knows with, you know, we've got some brilliant minds like the Demis Hospice at Google and others who are coming up with the next generation leap of whatever the equivalent of transformer technology is to keep going towards this ultimate goal that they're all pursuing, which is AGI.

所以，你说得对。

So yes, you're right.

很难确切知道什么会出现以及何时出现。

It's hard to know exactly what shows up and when.

然而，我们公司的结构实际上是两部分，阿尼拉。

However, the way we're structured as a company is effectively two parts, Anilay.

一部分是企业业务。

The one part is the enterprise business.

我们有一个产品团队和工程团队。

So we have a product team, engineering team.

我们显然有一个专注于该领域的市场推广团队。

We have obviously go to market, a go to market team focused on that.

企业产品业务非常明确。

And that's very, very clear, the enterprise products business.

而公司的另一部分是我们所说的社区产品板块。

And then the other side of the house is what we call the community products side of the house.

社区产品团队专注于公共平台，包括我们迄今为止讨论的所有功能，如AI助手、各种新的主观问题、聊天等。

And the community products team focuses on the public platform, all the features that we talked about so far, AI assist and all the subjective new questions and chat and all these things.

社区网站和数据授权业务也归属于这个团队。

And it's the community site and the data licensing business sits in that group.

因此，它们显然与网站上的用户参与度密切相关。

And so that they are obviously tied to the kind of the engagement on the site.

因此，这里存在一种良性循环。

And so there's a kind of a virtuous cycle there.

这就是我们的分工方式。

So that's how we're split.

这还包括产品资源、一个小的市场推广团队，以及确保相关人员等。

And that includes again, product resources, a small go to market team, ensuring folks, etcetera.

因此，我们的社区管理团队也花大量时间与版主等互动，以促进参与。

So that's, and then also our community management team, which spends a lot of time with the moderators, etcetera, to engage there.

所以这是对半分的，当然我们的其他职能部门也支持这两边。

So that's so it's split down the middle and of course our our other functions support both.

Stack Overflow今天有多大？

How big is Stack Overflow today?

我知道你们在2023年因流量下降裁掉了近四分之一的员工。

I know you you laid off almost a quarter of the company in 2023 because of traffic declines.

你们还建立了其他业务。

You've built other businesses.

你们现在有多大？

How big are you today?

我们大约有300人左右。

We are about 300 people or so.

你认为数据授权或你的SaaS业务带来的收入会允许你们再次增长吗？

And do you think the revenue you're gonna see from data licensing or your SaaS business is gonna allow you to grow again?

我们相信会的。

We believe so.

是的。

Yeah.

我们是一家正在成长的公司，并且实现了盈利。

We're a growing company, we're profitable.

所以，你知道，在财务上，我们幸运地处于非常良好的状态。

So, you know, financially, in, thankfully, in a very good spot.

现在，一切都在于押注最具增长潜力的机会。

And now it's all about placing bets on the highest growth opportunities.

我们相信，通过我们的内部产品在企业内部构建这一知识智能层，是一个绝佳的增长机会，因为客户正主动向这个方向推动我们，这真是太好了。

And we believe creating this knowledge intelligence layer inside the enterprise through our stack internal product is a phenomenal growth option because customers are pulling us in that direction, which is fantastic to see.

所以

So

这就是你们正在走向的方向。

That's where you're headed.

这些是我的一些公告。

Those are some my announcements.

我想在结束前谈谈这些内容。

I do wanna wrap up by talking about those.

我只关注未来：企业的SaaS业务，未来的方向是数据授权。

I'm just focused on the future is the SaaS business for enterprise, the future is data licensing.

你们还在看到公共网站流量的下降吗？

Are you still seeing declines on the public website?

可以说过去几个月已经稳定了。

Would say that it's been stabilized for a few months.

所以我认为网站上的参与度和活跃度实际上相当稳定。

So I think the engagement and the activity on the site are actually pretty stable.

我之前提到的提问数量下降，都是那些简单的问题，现在似乎已经趋于平稳，人们开始提出更复杂的问题了。

The drop in questions that I've mentioned previously were all the simple questions and it seems to be now has sort of come to a place where it is, the complex questions are being asked.

我们网站每天都有稳定数量的用户。

We have a consistent number of people on the site every day.

我们有一个叫做‘心跳’的指标。

We have something called a heartbeat.

事实上，任何人都可以去查看。

In fact, anybody can go check it out.

如果你访问 stackofill.com，就会在页面底部看到当前在线用户数量之类的信息。

There's always, if you go to stackofill.com, you'll see it at the bottom, how many users are online at the moment sort of thing.

所以你总会看到那里有一个非常稳定的数据。

And so you'll always see, you know, very consistent number there.

因此，我认为要预测未来很难，但毫无疑问，最糟糕的时期肯定是在2324年。

So that's so it's I think it's it's hard to know how to predict the future, but certainly, I think the the the worst of it was back in, you know, 2324 for sure.

我在《Decoder》节目中问每个人的问题，正如你所知，是：你是如何做决策的？

The question I ask everyone in Decoder, as you well know, is is how do you make decisions?

上次你做客节目时，你说你希望尽可能身处前线，并且希望从一线人员那里获取信息。

Last time you were on the show, you said that you wanted to be on the front lines as much as possible, and you wanted to be informed by people who are on the ground.

过去三年里，这种情况有改变吗？

Has that changed in the past three years?

你的决策过程是怎样的？

What's your decision making process?

嗯。

Yeah.

其实没有。

Not not really.

我认为，对于领导者和像CEO这样的人而言，掌握完整的信息至关重要，因为这些决策不能拖延，不能只依赖过滤后的信息。

I think it's very important for leaders and people like CEOs, etcetera, to have the full context because these decisions cannot be delayed, kind of, you can't have filtered information.

所以我花了很多时间与用户和客户交流，真正理解他们关心什么。

So, spent a lot of time with users, a lot of time with customers, really understanding what they care about.

正是基于这一点，我们才决定推出像AI辅助功能这样颇具争议的对话功能。

And that's how we even decided even something like as controversial as the AI assist feature to our conversation.

如果你不倾听一线的声音，根本不会想到要去开发这个功能。

It wasn't obvious if you were not listening to ground to say, let's go and, you know, build that.

因为如果你只听 headline 的说法，似乎人们并不希望将 AI 集成到 Stack Overflow 中。

Because if you just listen to the headline statements, it seems like, hey, people don't want that, want to integrate AI into Stack Overflow.

但事实是，许多用户——我之前提到的 90%——更喜欢自然语言界面。

But the reality is that many, many users, the 90% I was mentioning, you know, a natural language interface.

这正是他们如今习惯使用的工具，也是他们真正想要的。

That's what they are comfortable using these days and that's what they wanted.

所以我们决定这么做。

So that's why we decided to do that.

我到处都看到这种分裂现象。

One of the things that I I see everywhere is that split.

你之前提到过，九比一，有一个非常吵闹的少数群体。

You you mentioned it before, one nine ninety, right, there's very vocal minority.

我们在自己的网站流量中也看到了这种情况。

We see it in our own traffic on the verge.

我们深入报道 AI，却总有人说人人都讨厌它，我理解为什么，我明白这些评论，这就是我想说的，我懂。

We we cover AI deeply, we are told that everyone hates it, I understand why, I understand the I understand the comments, that's what I'll say, I get it.

然后我看到使用数据，看到我们对AI工具的覆盖流量，看到像你们这样的公司说每个人都在使用它，这其中存在巨大的分裂，这在我看来是过去十五年报道科技时从未遇到过的：人人都说不喜欢，却用得停不下来。

And then I see the usage numbers, I see the traffic on our coverage of AI tools, I I see companies like yours saying everyone's using it and there's some gigantic split there that is unlike any other split I I think I've ever encountered in covering technology for the last fifteen years, where everyone says they don't like it and then they're using the hell out of it.

是的。

Yeah.

我能想到的唯一一个稍微有点类似的例子是。

The only one other one I can think of that is lightly comparable Yeah.

就是人们对Adobe的看法。

Is how people feel about Adobe.

每个人都用这些工具，但每个人都对Creative Cloud的订阅费感到愤怒。

Everyone uses the tools and everyone's mad at the Creative Cloud subscription fee.

这几乎是唯一的类比。

It's basically the only comparison I have.

是的。

Yeah.

这并不是一个完美的类比，无法一一对应，但这是我能找到的最接近这种分裂的例子。

It's not a good one, it doesn't map one to one, it's as close as I've come to that split.

在你看来，是什么原因导致了人们对AI的这种矛盾现象？人们嘴上说不喜欢，非常强烈地表达不满，但数据却显示，人人都在用？

What in your mind accounts for that split with AI, where people don't like it, they're very vocal about not liking it, and then we see the numbers and everyone's kinda using it anyway?

我认为这跟之前我提到的一个数据点有关：我们超过80%的用户希望使用AI，或者已经在使用AI来处理与代码相关的问题，但其中只有29%的人信任AI。

I think it comes down to that data point I shared earlier, which is that 80 plus percent of our user base wants to use AI or are already using AI for code related topics, but only 29% of that population trusts AI.

‘信任’这个词含义很深，你想想，为什么你不信任某样东西？

And trust is a very deep word, you know, like kind of like, why don't you trust something?

你不信任它，是因为你觉得它提供的答案缺乏可靠性、不准确。

You don't trust something because you don't think it's producing high integrity answers, accurate answers.

你可能也不信任它，是因为你担心它有一天会取代你，而你并不喜欢这种可能性。

You may not trust it because it may replace you one day and you don't like that either.

但与此同时，你显然也会好奇：这个即将成为标志性力量的东西究竟是什么？

So, but at the same time, you're obviously going to be curious on what is this force that's going to be such an iconic force.

所以你会不断尝试、使用它，努力提升自己，最终希望利用它来为自己谋利，让自己在未来作为一名开发者依然保持竞争力，能够更快地完成工作等等。

And so you want to keep trying it and using it and perhaps getting better and ultimately, hopefully leveraging it to your benefit in a way that you can be relevant as an individual developer in the future, being able to go a lot faster and so on.

所以我认为这可能就是原因，尤其是开发者群体，他们非常挑剔，可以说是高度理性的群体，如果某些东西不够确定、不够可靠，他们就会很反感，对吧？

So I think that's probably the reason because I think especially the developer audience, I think they're very discerning, a kind of a, let's say analytical audience and they can be prickly if things are not, let's say deterministic, right?

长期以来一直都是这样。

The way it has been for a long time.

所以这是一种非常概率性的技术。

So this is a very probabilistic sort of technology.

所以，这就像是去赌场玩轮盘赌一样，每次都会得到不同的答案。

So the fact that you, it's almost like, you know, going to a casino and using a roulette wheel, you're gonna get a different answer every time.

对于编写非常具体代码、寻求特定结果的人来说，这并不一定令人安心。

It's not necessarily comforting for somebody who's writing very, very specific code, looking for very specific outcomes.

我认为人们会随着时间的推移逐渐适应这一点。

And I think that people will get used to that over time.

这对编写软件的人来说是一种思维上的转变。

It is a mind shift change for people writing software.

这可能就是人们对此感到好奇的原因，因为这项技术如此强大。

And that may be the reason that why people are intrigued because it is so powerful as a technology.

别误会我的意思。

Don't get me wrong.

我们在 Stack 的各个地方都在使用氛围编码，对吧？

We use vibe coding all over the place at Stack, right?

我跟你提到的所有功能，我们设计师和产品经理都是先用氛围编码来展示，获取用户反馈，然后再去开发。

All the features I mentioned to you, we are designers and our product managers vibe coded it first to show it and get user feedback before we went and built it.

因此，我们内部已经接受了这些工具带来的好处。

So we've embraced these tools internally for their benefit.

所以你会有一些方式让自己用起来更安心，但我认为这仍然是核心原因。

So there will be ways in which you can feel comfortable using it, but still, I think that's that's the core reason.

我其实想谈谈这种依赖性。

I actually wanna talk about that dependency.

对吧？

Right?

你知道它不可靠，但你仍然在用它来构建产品。

You know that it's not trustworthy, but you are building products with it.

你正在构建产品来推动它的应用。

You are building products to enable it.

你已经多次提到过它。

You've mentioned it several times.

本周的重大发布是 Stack Overflow AI Assist，你在整个对话中多次提到过它。

The big announcement this week is Stack Overflow AI Assist, you've talked about it several times throughout this conversation.

你相信这是人们真正需要的，对吧？

You're betting that this is what people want, right?

你相信 Stack Overflow 上的 AI 工具能帮助更多人，但也许这个工具会疯狂地幻觉，给人们错误的答案。

You're betting that an AI powered tool on Stack Overflow will help more people and then maybe that thing is gonna hallucinate like crazy and give people the wrong answer.

当你知道用户并不信任它，但你又必须推出这些工具，因为行业趋势就是如此，你如何做出这样的赌注？

How do you make that bet when you know that the users don't trust it, but you still have to roll out the tools because that's where the industry is going?

我们相信，实际上已经解决了信任问题中的一个关键方面，那就是我们的 AI Assist 功能通过 RAG 加 LLM 方案来应对。

We believe we've actually unlocked a very important aspect of, you know, that trust issue, which is and responding to it by our AI assist feature provides, it's a Rag plus LLM solution.

因此，它会首先在我们数千万个问答的语料库中查找信息来生成答案。

So effectively it provides an answer that first goes out into our corpus of information of tens of millions of questions and answers.

我们拥有八千到九千万个。

We have 80 to 90,000,000.

这些信息首先用于生成回答。

And those are first used to produce a response.

如果无法找到答案，则会启用备用方案，利用我们的合作伙伴OpenAI等，从网络其他部分获取可信的知识。

And then if they don't, then there's a fallback option where it goes and leverages OpenAI, for example, who's our partner, to be able to go and produce trustworthy knowledge from other parts of the web.

因此，它首先在我们可信的、有出处的知识库中进行搜索，并生成链接，以便用户可以进一步了解相关信息。

And so it's first searching through our trusted attributed knowledge base, it produces the links so people can go down that path and learn more about it.

因此，这对我们的团队来说非常重要，对吧？

And so that was just very important to us, right?

出处以及其他方面。

Attribution and so on.

这就是我们应对幻觉问题的方式，我们一直在持续测试，但它并不完美，对吧？

And so that's how we are navigating this element of hallucinations and we're constantly testing it and it's not perfect, right?

总会不断改进，但我们也在关注世界的发展方向。

There will always be improvements, but we're also looking at where the world's headed.

如果这些模型持续进步，那么我们也应该能从中受益。

And if these models continue to get better, then this should, we should benefit from those improvements.

最终，我们应该拥有最佳的解决方案，因为你既有扎实的人类背景，又具备这些优势。

And ultimately we should have the best solution because you've got grounded human context plus the strengths as well.

我认为我最感兴趣的是，模型会持续变得更好的这种信念。

I think the thing that I'm most interested in is the faith that the models will continue to get better.

我不完全确定这是真的，我不确定大语言模型作为一种核心技术真的能具备智能。

I'm not a 100% sure that's true, I'm not sure that LLM technology as a core technology can actually be intelligent.

正如你所说，人们非常被这些模型的自然语言组件以及正在发生的界面转变所吸引——我们所有人将来都会用自然语言开发软件，或者让大语言模型进行推理，基本上自我提示以得出答案，对吧？

As you're saying, people are very attracted to natural language component of these models and the interface shift that's happening, the platform shift that we're all gonna develop software with natural language or let the LLMs reason, basically self prompt themselves into an answer, right?

这其中似乎存在某种风险。

There's something there that seems risky.

你今天是否感知到这种风险？

Are you perceiving that risk today?

你是否在考虑这一点，还是说这是我们现在所处的状态，必须继续推进，直到出现变化？

Are you factoring that in or are you saying this is where we're at now and we have to continue until something changes?

关于大语言模型的改进，我想首先谈一点。

I think the first thing around the improvement on the LLM, let's call it plain.

我同意你的观点。

I'm with you.

对吧？

Right?

当你觉得过去六个月已经停滞不前时，很难预测未来会如何进步。

Like, it's hard to know how things are gonna improve when you when you just think that, okay, it's plateaued when you thought about the past six months.

砰。

Boom.

Gemini 3来了。

Here comes Gemini three.

没错。

Yeah.

我们再次为成为谷歌的合作伙伴感到自豪。

And again, we're proud partners of Google.

我们又开始了。

And here we go.

你知道，这是一次根本性的转变。

You know, it's it is a seed shift.

它把其他所有模型都远远甩在了后面。

It blows every other model out of the water.

现在，其他公司、其他大语言模型都进入了红色警报状态。

And now we've got a Code Red situation in other companies, other LM.

这简直是每个萨姆都得面对的情况。

You it's every Sam

艾尔曼确实说了‘红色警报’这个词，就是这种感觉。

Allman did type the words code red, the way, wanna be that.

所以这非常好。

So that's very good.

也许在过去，这正是我该走的路，但我想说的是，令人惊讶的是，当表面上一切似乎已陷入停滞时，你们竟能实现如此巨大的飞跃。

Perhaps that was the way for me to go back in the day, but I would say, but the point being that that is, it is surprising that you're able to produce that sort of a leap when seemingly things have plateaued.

所以我不知道，目前我无法预测，内莉，因为这些家伙——丹尼斯和其他人——对这个领域了解得非常深入。

So I don't know, I can't predict that right now, Nelly, because these folks are deep, deep in the subject, Dennis and others.

所以这是真的，但我也相信，还会有其他我们尚未知晓的创新出现。

And so that's true, but there's also going to be other, I'm sure innovations that we are not even privy to.

就像我之前解释的，比如Transformer显然是这个领域的一大突破，也许这些AI研究实验室会提出一些我们完全不了解的新技术，最终推动整个领域向前发展。

Like I was explaining previously, like if, you know, transformers were obviously a huge development in this space, there may be something that these AI research labs come up with that we're not even aware of that's going to be ultimately pushing things out.

我们知道，复利效应是真实存在的。

And we know that the compounding effects are very real.

你知道，我们现在拥有无限的计算能力，还有极其强大的芯片和GPU，而且它们的成本还在不断降低。

You know, we've got unlimited compute, you've got, you know, extremely powerful chips and GPUs that are now even lowering their costs.

这周我参加了AWS re:Invent大会，他们提到了正在研发的Tranium芯片，包括Tranium 3和Tranium 4。

And I was at, you know, AWS re:Invent this week where they talked about the Tranium chips, Tranium three and Tranium four being built out.

因此，这些计算组件的普及只会越来越多。

So, there's going to be just the proliferation of these compounds.

再加上我们已经讨论过的数据获取优势。

And then you got access to data that we've already talked about.

当这些因素相互结合并产生复利效应时，将会带来非常神奇的结果。

And so these things, when they combine and compound together, it's going to produce very magical outcomes.

所以我认为这就是信念的来源，也是我之所以坚信它会进步的原因。

So I think that's the belief and why it's rooted, know, why my why why my, you know, my own assumptions are rooted in the fact that it's gonna improve.

我之所以在你所构建的工具以及只有29%的人信任这些工具的背景下提出这个问题，是因为你必须大幅提升这个比例，才能达到每家投资AI的公司——包括你们公司——都试图实现的回报。

The split, and the reason I'm asking about this in the context of the tools you're building and everyone using it and only 29% of people trusting it, is you've got to bring that number way up to reach the returns that every company investing in AI is is trying to reach, including yours.

我不知道核心技术是否能做到这一点，也不知道是否能通过堆叠多种技术来实现，但我确实知道，软件开发的某种未来图景是：一切永远都是‘氛围编码’。

I don't know if the core technology can do that, I don't know if you can stack a bunch of technologies to do that, but I do know that one version of the future of software development looks like everything is vibe coding all the time.

对。

Right.

另一种软件开发的未来图景是：编写极其冗长的提示词。

Another version of the future of software development looks like writing intensely long prompts Right.

针对那些本身就有好几页内容的模型，这在我看来很荒谬，但也许这正是未来。

For models that are pages and pages themselves, which seems ridiculous to me, but you know, maybe that is the future.

还有一种未来图景是：人类重新回到软件开发的前沿，与AI模型协同开发，然后可能向堆栈提问。

And another version looks like, oh, we returned to humans at the cutting edge of software development and they are co developing with an AI model and then maybe asking stack.

这种未来感觉更丰富、更有趣，但我不清楚我们目前处于这个谱系的哪个位置，也不清楚它究竟会如何展开。

And like that feels like a richer, more interesting future, but it's unclear to me where we are on that spectrum or how that even plays out.

我们显然不仅从公共社区的宏观视角看到了这个29%的数据点，而且作为企业客户，我们也清楚地了解客户的真实情况，对吧？

We obviously have not only a bird's eye view in the context of our public community with this 29% data point as an example, but we are enterprise customers, give us a clear view on where they are, right?

因为公司内部正在强烈地提出投资回报率的问题。

Because the ROI questions being asked very heavily inside companies.

显然，我认为2026年将是理性调整之年。

And clearly, you know, think you hit the 2026 is going be the year of rationalization.

如果2025年是代理工具之年——至少在这些公司里，每个工具都在被尝试，CTO们有着非常开放的环境去采购和测试各种工具。

If 2025 was the year of agents, at least where every tool is being tried out inside these companies, there's a very open landscape for CTOs to go and buy various tools, test various tools.

因此，我认为这对一些正在构建这些工具的公司来说是绝佳的时期。

And so I think it's been a tremendous time for some of the companies building these tools.

但到了2026年，CFO们会施加压力，说：‘好吧，生产力的提升必须来自这些工具。’

But '26, the CFO pressure with, Hey, look, okay, there's productivity improvements have to come from these.

在年度规划期间，我们将减少招聘人数。

We're going to hire less people as annual planning happens.

系统将面临巨大压力，必须证明这些工具的真实价值。

There's going to be tremendous pressure in the system to prove out what the real value is.

我接触过的每一位高层人士都承认，这是一次巨大的转变，而且他们都在积极推动这一变革。

Everybody at a senior level that I've talked to acknowledges that this is a big shift and they all are leading it pretty hard.

他们都在等待这些改进的成果。

They're all waiting for the improvements.

我认为，大多数人和公司都会说，他们在小范围测试这些工具时看到了一些改进，但这是一个自我选择的群体，因为他们本身就是爱好者等等。

I think most people, most companies will say that they have seen improvements in the small groups where they've tested these tools, but that's a self selecting group because they're enthusiasts and so on.

他们会看到显著的生产力提升，这可能是真的，但当你考虑在整个企业范围内推广时，生产力会出现断崖式下降。

And they will see great productivity gains, which is probably true, but there's an absolute drop off in productivity as you think about the adoption across the enterprise.

而且可能原因相同，你知道，你让员工使用这些工具，而这些工具可能会让他们丢掉工作。

And probably for the same reasons, by the way, know, you're telling employees to use tools, which may put them out of a job.

那他们为什么要这么做呢？

So why would they want to do it?

或者更根本地说，如果这些工具并不完美，会幻觉，而且使用者还要为此负责，那情况也不妙。

Or more fundamentally, if these tools are not perfect and they're hallucinating and, you know, they're going be held accountable, then that's not good either.

因此，随着流程的改变、思维的转变，以及必须彻底重构工作流程，所有这些企业变革管理的工作，正是为什么我们的解决方案Stack Internal正在构建这个人为主导的审核层和知识智能层，以便你可以在其上部署MCP服务器，中间放知识库，再通过我们从公司其他部分摄入知识的能力，创建这些高度实用的原子化问答，从而锚定你的企业知识，并通过这些AI代理发挥作用。

So, and then of course, as the process change, the mindset change, the ability to have to completely change the workflows of how you work, all the enterprise change management work, which is why our solution, is Stack Internal is building this human curation layer, knowledge intelligence layer so that you can put the MCP server on top of it, knowledge base in the middle, and then our ability to ingest knowledge from other parts of the company to create these atomic Q and A that are extremely helpful to root your enterprise knowledge and through these AI agents.

这就是我们的解决方案，因此我们全力以赴地开发它。

That is that solution that we that's why we've gone and really gone hard at producing that.

我们从客户那里得到了非常强烈的反馈。

And we've seen really, really strong response from our customers.

我们有一些全球最大的公司，比如惠普、礼来和施乐等，正在与我们合作使用、测试和构建这一系统。

We have some of the world's largest companies you know, leveraging and testing and building this with us, you know, HP and Eli Lilly and all these companies Xerox.

看到他们主动选择我们，因为我们满足了你所提到的ROI目标，这真是太棒了。

And so it's been amazing to see them gravitate to us because they want to fulfill that ROI point that you're making.

那么，还有哪些差距呢？

And what are the gaps?

依然是信任问题。

It's trust again.

因此，他们希望借助像Stack这样的公司建立一个信任层，将其插入到他们的数据和AI工具之间。

And so they want this trust layer through a company like Stack that they can actually insert in between their data data as well as their AI tools.

当你提到‘理性化时代’时，我理解为你认为2026年泡沫将破裂。

When you say the age of rationalization, what I hear is you think the bubble is gonna pop in 2026.

是的。

Yeah.

我认为，那种狂热情绪以及随意尝试各种工具、无限预算投入AI的做法，最终一定会付出代价。

I think certainly the exuberance and just trying out various tools and unlimited budgets on this AI budget, think will ultimately, it'll come to roost.

我不确定泡沫是否会破裂。

I'm not sure about the bubble bursting.

但毫无疑问，过程中一定会出现调整。

Like, think there's definitely going to be corrections along the way.

这一点毫无疑问。

There's no question about it.

如果你看看历史，就会发现，向这些公司销售产品的供应商数量众多，但让我非常惊讶的是，它们的功能竟然如此相似。

That is like, if you look at history, but the number of vendors that are selling into these companies, what I'm very surprised by is that there's similar functionality.

这些公司内部正在测试其中四家供应商的产品。

There's four of them being tested within these companies.

当然，这些公司可能都有上亿美元的年经常性收入，但终有一天，当CTO决定：算了，我只用一个，再留一个作为备份就好——就像云计算一样，届时必然会出现客户流失。

Certainly all these companies may have received what if they've got a 100,000,000 in ARR, but at some point there's gonna be churn when the CTO decides, you know what, I'm only gonna use maybe one and maybe a second one as a backup, no different from the cloud.

你知道，回想起多云时代，人们一开始并没有同时使用三个云。

You know, if you think with the multi cloud world back in the day, you know, people didn't have three clouds out of the gates.

我的意思是，现在你可能有一个主云和一个备用云。

I mean, now you have maybe one primary cloud and one secondary cloud.

这不一样，但同时，我认为你不会拥有四个不同的AI编码工具，依我之见。

This is different, but at the same time, I don't think you're gonna have like four different vibe coding tools in my opinion.

当你想到Stack Overflow作为信任层时，对吧？这就是它的价值所在，也许长期来看，你可以因此收取溢价。

When When you think about Stack Overflow as being that trust layer, right, I mean, that's the value add, that's what maybe you can charge a premium for it over time.

你仍然依赖于只有29%的用户信任的工具。

You're still dependent on tool only 29% of your users trust.

你和我都知道，你谈的是RAG和其他用于实现这一点的系统。

How do you and I know you're talking about Rag and your other systems for doing that.

你打算如何通过Stack Overflow提升这个比例？

How do you think you bring that number up with Stack Overflow?

这有可能由你来实现吗？还是必须由生态系统来完成？

Is is that possible for you to do or is that the ecosystem has to do it for you?

我认为这更普遍来说是一个生态系统的问题，因为这更多反映了人们在Stack之外能接触到的东西，对吧？

I think it's an ecosystem point more generally speaking because, know, that is more a reflection of people's, you know, what they have access to beyond stack, right?

他们可以接触到所有这些其他的选项。

They've got access to all these other options.

因此，我们能关注的是成为技术人士最重要的信息来源。

And so what we can focus on is being the most vital source for technologists.

对我们来说，关键是确保这些内容卓越、高质量，同时也是一个让人们再次培养社区、相互连接、学习和职业成长的好地方。

And so for us, it's about making sure this content is excellent, it's high quality, but also it's a great place for people to again, cultivate community, connect with each other, learn and grow in their careers.

但我认为我们能做到这一点的方式，是通过与所有这些大型AI实验室合作。

But I think the way in which we can do it is through the fact that we are working with all these big AI labs.

而我们经过人工精心策划的可信知识，将流入这些大语言模型，最终产生可信的答案。

And the fact that our trustworthy knowledge that's been human curated painstakingly is going to flow into these LLMs, which ultimately produce trustworthy answers.

所以我们某种程度上是在背后一层，但我想这就是我们的定位。

So we're sort of one layer behind, but that's I think where we operate.

在大语言模型的语境中，我们运作于信任层或数据层。

We operate in that trust layer or the data layer, if you will, in the context of LLMs.

因此，这是我们对该29%的间接贡献。

So that's our indirect contribution to that 29%.

普拉尚特，这次对话非常精彩。

Prashant, this has been a great conversation.

下次你得比三年后更早回来。

You're going to have to come back sooner than three years next time.

Stack Overflow接下来有什么计划？

What's next for Stack Overflow?

人们应该关注什么？

What should people be looking for?

我们最大的重点将是为企业构建一个企业级知识智能层，使他们能够以可信赖的方式使用AI代理。

Our biggest focus is going to be making sure that we build this enterprise knowledge intelligence layer for companies to truly use AI agents in a trustworthy way.

事实上，我们几周前在微软Ignite上推出的内部产品，让我们在企业端感到非常兴奋，同时正如我前面提到的，在公共平台上，我们也将帮助社区用户相互连接、真正学习并促进职业发展。

So our Internal product that we launched a couple of weeks ago at Microsoft Ignite, in fact, very, very excited about that on the enterprise side, as well as of course on the public platform, as I've mentioned throughout to help our community users connect with each other, really learn as well as grow their careers.

未来将有众多新途径和入口，比如我们的AI辅助、主观内容、聊天功能以及其他一些东西，人们在周围环境快速变化时， hopefully 会发现它们非常有用。

And there are going to be so many avenues and new entry points like our AI assist and our subjective content and chat and other things that people hopefully find very useful as things change around them very rapidly.

而且，你知道，他们可以成为这个精彩社区的一部分，互相帮助。

And, you know, they can be part of this amazing community and and help each other out.

所以我们的两个重点是企业市场和我们的公共社区。

So those are the two focuses, enterprise as well as our public community.

好的。

Alright.

好吧，明年泡沫破裂时，我们还要请你回来。

Well, when the bubble pops next year, we're gonna have we're gonna have you come back.

太棒了。

That's awesome.

你怎么预测的？

How you predict it.

谢谢你，尼莱。

Thank you, Nilay.

非常感谢。

Appreciate it.

我要感谢普拉桑特做客《解码器》节目，也感谢大家的收听。

I'd to like to thank Prasanth for joining me on Decoder, and thank you for listening.

希望你们喜欢这期节目。

I hope you enjoyed it.

如果你对本集或其他任何内容有任何想法，请告诉我们。

If you'd to let us know what you thought about this episode or really anything else at all, drop us a line.

你可以发送邮件至 Decoder@TheVerge.com。

You can email us at Decoder@TheVerge.com.

我们真的会阅读每一封邮件。

We really do read all the emails.

你也可以在 Threads 或 Blue Sky 上联系我。

You can also hit me up on threads or blue sky.

你还可以在我们全新的 YouTube 频道留下评论。

You can also leave a comment on our fancy new YouTube.

你可以观看完整的节目片段，我们也有 TikTok 和 Instagram 账号。

You can watch full episodes Dakota pod, and we have a TikTok and an Instagram.

它们也在Dakota播客上。

They're also at Dakota pod.

它们非常有趣。

They're lot of fun.

如果你喜欢Dakota，请分享给你的朋友，并在你收听播客的平台订阅我们。

If you like Dakota, please share with your friends and subscribe wherever you get your podcasts.

Dakota是The Verge的制作作品，属于Box Media播客网络的一部分。

Dakota is production of verge and part of the Box Media Podcast Network.

本节目由凯特·考克斯和尼克·斯塔特制作。

The show is produced by Kate Cox and Nick Statt.

本节目由乌尔萨·赖特剪辑。

It's edited by Ursa Wright.

我们的编辑总监是凯文·麦克肖恩。

Our editorial director is Kevin McShane.

《Decoder》的片头音乐由Breakmaster Cylinder创作。

The Decoder music is by Breakmaster Cylinder.

我们下次见。

We'll see you next time.