这是AI泡沫吗？+ Meta缺失的道德准则 + TikTok震惊内容泛滥

好吧。有些聊天机器人太死板了。

Alright. Some of these chatbots are too literal.

你什么意思？

What do mean?

你懂我意思吗？前几天我在YouTube上看我们旧金山刚举办的Outside Lands音乐节视频。很棒的音乐节。我正在看乐队role model的演出，他们表演热门歌曲《Sally, when the wine runs out》时有个特别环节，会邀请一位特殊嘉宾上台跳舞。明白吗？

Do you know what I mean? So the other day, I'm on YouTube, and I'm watching videos from the Outside Lands Festival that we just had here in San Francisco. Great music festival. And I was watching the set by the band role model, and they do this thing when they perform their hit song, Sally, when the wine runs out, where they bring in kind of a special guest to dance. Okay?

这个你在YouTube上搜就能看到。然后有个人冲上舞台，我心想：那是流行歌手Troy Sivan吗？因为我觉得很像，但镜头很快，他一直在转圈，我没看清。

This is like something that you'll see it on YouTube if if you look. Right? And this guy runs out on stage, and I think, is that the pop star Troy Sivan? Because I thought it was the pop star Troy Sivan, but it was sort of, you know, very quick and he's, you know, spinning around. I didn't, like, get a good look at him.

所以我就去问ChatGPT：'Troy Sivan在Outside Lands音乐节上为role model助阵了吗？' 你猜它怎么回答？

So I thought, I'm just gonna ask Chad GPT about this. So I said, hey. Did Troy Sivan come out during role model set at Outside Lands? And you know what it said?

怎么说的？

What?

它说：'没有。Troy Sivan并未在Outside Lands为role model助阵。他早在2013年8月7日就通过YouTube视频公开出柜了。' 我：我不是问这个！然后它又补充：'今年Outside Lands的特别之处在于有位神秘嘉宾突然现身...'

Nope. Troy Sivan did not come out during role model set at Outside Lands. He had already publicly come out as gay back on 08/07/2013 via a heartfelt YouTube video. And I was like, that's not what I was talking about. And then it said, what happened at Outside Lands this year was a surprise live appearance.

后面它终于说：'是的，他确实来了。' 总之...我觉得这太离谱了。

He hopped on to and then it was basically like, yes. He did. So anyways Yeah. I thought that was a little crazy.

确实离谱。知道吗？他们正在开发一个能识别全球所有公开出柜同志的AI系统。猜猜叫什么名字？

That is crazy. You know, they're actually building an AI system that can determine when every gay person in the world has publicly come out. Do you know what they're calling it?

叫什么

What are calling

它？同性恋大兵。

it? Gay GI.

哦，老兄。太棒了。太棒了。

Oh, man. That's great. That's great.

是啊。是啊。我很抱歉

Yeah. Yeah. I'm sorry for

你的困扰。总之，祝贺特洛伊·萨沃德在2013年8月7日和2025年8月两次出柜，尽管方式略有不同。

your troubles. Anyways, congratulations to Troy Savod for coming out both on 08/07/2013 and August 2025. Albeit in slightly different ways.

我是凯文·鲁斯，《纽约时报》科技专栏作家。

I'm Kevin Roose, a tech columnist at the New York Times.

我是Platformer的凯西·纽布。这期节目很难...本周我们要探讨的是：我们是否身处AI泡沫中？我们将从正反两面论证。接着记者杰夫·霍洛维茨将加入我们，讨论他那篇轰动性的报道——关于Meta AI被指示与儿童进行浪漫角色扮演的内幕。

I'm Casey Neub from platformer. And this is hard for. This week, are we in an AI bubble? We'll make the case for and against. Then journalist Jeff Horowitz joins us to discuss his blockbuster story on how Meta AI was instructed to engage in romantic role play with children.

最后，我要告诉凯文我最喜欢的新TikTok趋势，内容相当低俗。

And finally, I tell Kevin about my favorite new TikTok trend, and it's filthy.

上帝救救我。

God help me.

好吧，凯西，今天我们要

Well, Casey, today, we are

讨论一个酝酿已久但本周真正引发关注的问题：我们是否正处于AI泡沫中？

going to talk about a question that has been percolating for a long time, but that is really gaining some attention this week, and that is, are we in an AI bubble?

是的。这是人们热衷谈论的话题，最近的一些新闻报道，我认为，让更多人加入了这场讨论。

Yes. This is something that people love to talk about, and there have been some news items recently that have led even more people, I think, to get in on the conversation.

没错。奇怪的是，我认为最新的泡沫新闻周期源于上周你我与Sam Altman共同参加的一场晚宴。

Yes. And strangely, I think this latest bubble news cycle originated at a dinner that you and I attended last week together with Sam Altman.

确实如此。我想是因为他们看到了我们所有食物的账单金额，然后心想：一家公司怎么可能负担得起这么多食物？

That's right. And I think it's because they saw the price for the the bill for everything we ate, and they thought, how can one company possibly afford all of this food?

Casey，我们该聊聊这场晚宴了，因为它以相当特别的方式促成。但在那之前，我们需要做出声明：《纽约时报》公司正在起诉OpenAI和微软，指控其在训练大语言模型时存在版权侵权。而我——

Now, Casey, we should talk a little bit about this dinner because it was quite unusual the way it sort of came together. But before we do, we should make our disclosures. The New York Times company is suing OpenAI and Microsoft over copyright violations related to the training of large language models. And my

男友在Anthropic工作。好了。

boyfriend works in Anthropic. Okay.

回到这场晚宴。Casey，你能稍微描述下当时的场景吗？

So back to this dinner. Casey, can you sort of set the scene a little?

好的。上上周周末，我收到OpenAI的短信，说CEO Sam Altman将在下周为少数记者举办晚宴，问我是否愿意参加？不同寻常的是，这次晚宴将全程记录。你知道，在科技界，公司常会希望我们与其高管共进晚餐交流，但通常这些聚会都是非公开的，以便更坦诚地交谈。

Yes. So the week before last, at the end of the week, I got a text message from OpenAI saying that Sam Altman, the CEO, was gonna be throwing a dinner for a small group of reporters the following week, and would I like to go? And unusually, the dinner was going to be on the record. And, you know, it happens not infrequently in tech that a company will want us to get together with some of their executives, sometimes over dinner. But usually, those gatherings are off the record so they can sort of talk more candidly.

这次完全不同。结果我们俩都收到了邀请，都去了。在两小时里，我们不仅与Sam Altman，还与OpenAI首席运营官Brad Leitkap和ChatGPT负责人Nick Turley进行了公开讨论。

This was something very different. We both, as it turns out, got the invitation. We both went. And for two hours, we had an on the record discussion, not just with Sam Altman, but also with Brad Leitkap, the COO of OpenAI, and Nick Turley who runs ChatGPT.

对。私人餐厅里有张长方形长桌，我们吃着非常美味的地中海菜肴，都是家庭式共享餐盘。Sam在席间不断接受我们的提问，晚宴上谈了很多，包括GPT-5及其发布计划。但我觉得Sam关于AI泡沫的言论最受关注。

Yeah. So there's, like, this long rectangular table in this private dining room. We're eating this, like, very good Mediterranean food that's being served on a bunch of, like, shared family style plates. And Sam is just getting question after question from us at the table, and a lot was said during this dinner, a lot of talk about GPT five and the rollout of that. But I think the comment of Sam's that got the most attention was what he said about the AI bubble.

是的。他对此说了几点。让我印象最深的是他说：'有人会损失巨额资金，我们不知道是谁；也有很多人会赚取巨额财富。我个人认为——虽然可能出错——总体而言，这将对经济产生巨大的净收益。'

Yeah. So he said a few things about this. You know, one of the standout quotes to me was when he said, someone is gonna lose a phenomenal amount of money. We don't know who, and a lot of people are going to make a phenomenal amount of money. My personal belief, although I may turn out to be wrong, is that on the whole, this would be a huge net win for the economy.

他还提到，凯文，他在某种程度上是在想象一个理论上的初创公司。如果只是三个人和一个想法，估值却达到7.5亿美元，他说，引用原话，‘这很不理性。有人会在这里栽跟头，我认为。’所以一方面，他说了你可能预期他会说的话，比如，看，我认为AI对经济会非常有利。我认为我们会赚很多钱。

He also though said, Kevin, that and he was sort of, like, imagining a theoretical startup here. If it's three people and an idea and it has a valuation of $750,000,000, he said, quote, it's irrational. Someone's gonna get burned here, I think. So on one hand, he said kind of what you would expect him to say in terms of, look, I think AI is gonna be really good for the economy. I think we're gonna make a lot of money.

我认为我们的一些竞争对手也会赚很多钱。但他也说，看，很明显市场上存在一些非理性的热情，一些估值高达数亿甚至数十亿美元的公司不会给投资者带来回报。

I think some of our competitors are gonna make a lot of money. But he also said, look, it's clear that there is some irrational enthusiasm that is going on in this market, and some of these companies with huge, you know, multi 100,000,000 or even billion dollar valuations are not gonna provide returns to their investors.

是的。而且这不是我们最近几周第一次听到关于我们正进入AI投资泡沫期的猜测。所以我想快速列举一些其他证据，那些担心这一点的人用这些证据来说明，也许事情变得有点疯狂了。第一种证据就是这些AI公司的估值。就在本周，据报道，OpenAI正在洽谈进行一项股权收购要约。

Yes. And this is not the first sort of speculation we've heard in recent weeks that we are entering a bubble like period of investment in AI. And so I just wanna quickly run through some of the other evidence that people who are worried about this are citing to say, well, maybe things are getting a little crazy. The first kind of evidence are these valuations of these AI companies. So just this week, OpenAI is in talks reportedly to do a tender offer.

他们允许现有和前任员工以5000亿美元的估值出售约60亿美元的股票。这大约是Salesforce市值的两倍，这将使OpenAI成为世界上最有价值的私营公司。另一家AI公司Databricks表示，它已以超过1000亿美元的估值筹集资金。这比不到一年前的620亿美元估值大幅上升。你还可以看到其他一些公司似乎以这些难以置信的估值筹集资金的例子。

They're letting current and former employees sell about $6,000,000,000 worth of stock at a valuation of $500,000,000,000. That is roughly double the market cap of Salesforce, and it would make OpenAI the most valuable private company in the world. Databricks, another AI company, said that it had raised funding at a valuation of more than a $100,000,000,000. That's up from the $62,000,000,000 valuation it had less than a year ago. And you also have other examples of companies that seem to be raising at these unbelievable valuations.

其中之一是Eight Sleep，这家公司正在制造一种使用AI来整夜调节温度的床垫。他们刚刚宣布筹集了1亿美元来构建，引用原话，‘最终解决睡眠问题的AI’。

One of them is Eight Sleep, the company that is making, like, a a mattress that uses AI to kind of adjust its temperature throughout the night. They just announced that they have raised a $100,000,000 to build, quote, unquote, AI that finally fixes sleep.

或者想想Thinking Machines，Mira Maradi的初创公司。她是OpenAI的前首席技术官。她筹集了20亿美元的种子轮资金。在我刚开始报道科技时，种子轮通常是100万美元。她为一个没有产品的公司筹集了20亿美元，估值120亿美元，我想实际上除了一个幻灯片演示之外，几乎没有其他东西。

Or how about thinking machines, Mira Maradi startup. She's the former CTO of OpenAI. She raised a $2,000,000,000 seed round. Seed rounds used to be a million dollars when I started covering tech. She raised $2,000,000,000 at a $12,000,000,000 valuation for a company that has no product, and I imagine actually has very little more than a slide deck at this point to go off of.

是的。但为他们辩护，20亿美元只够支付两名AI研究员的薪水。

Yes. But in their defense, $2,000,000,000 is only enough to pay for the salaries of two AI researchers.

嗯，这是个好观点。

Well, that's a good point.

好的。所以这是关于估值的担忧。然后还有第二个担忧，主要是关于支出的，尤其是那些竞相开发更强大模型的大型科技巨头。在这里，数字确实变得有些疯狂。仅在过去三个月，美国股市中最大的七家科技公司——‘壮丽七雄’——在数据中心建设和相关费用上花费了超过1000亿美元。

Okay. So that's the sort of valuation worry. Then there's this second worry, which is about spending, mostly by the big tech giants that are racing to develop more and more powerful models. And here, the numbers really do get kind of insane. So over the last three months alone, the magnificent seven, the seven largest tech companies in the US stock exchange, spent more than a $100,000,000,000 on data center construction and related expenses.

这比前几年大幅增加。根据彭博社的数据，这些公司在美国数据中心建设上的支出很快将超过美国办公室建设的支出。所以我认为我们应该说，按照历史科技趋势的标准，这是对一个仍然相当新的技术进行的巨大基础设施投资。

That is way, way up from previous years. And according to Bloomberg, the amount that these companies are spending on data center construction in The US is on pace to overtake the amount of money being spent on office construction in The US pretty soon. So I think we should say, like, this is by the standards of historical tech trends, this is an enormous investment in infrastructure for a technology that is still quite new.

没错。这已经不仅仅是规模庞大，而是史无前例的。我们真的处在一个全新的世界里。凯文，此刻我想指出的另一点是，有些公司的支出远超其收入。

Yeah. It's beyond enormous. It's unprecedented. Like, truly, we are just in a brand new world here. I I think another thing to point out at this moment, Kevin, is just that some companies are spending a significant amount more than they're making.

我想到Cursor这家公司。它实际上是母公司AnySphere开发的一款产品。他们制作的编程助手深受许多软件工程师欢迎。本月有媒体报道称，他们存在所谓的'负毛利率'现象——产品售价低于成本价。因为要让这款神奇的编程助手运行，他们必须使用OpenAI、Anthropic等公司的API接口。

I'm thinking of the company Cursor. It's really a product made by a parent company called AnySphere. They make this coding assistant that is really popular with a lot of software engineers. And there was some reporting in newcomer this month about the fact that they have what they call negative gross margins. They are selling this product for less than it costs them because in order to get that magical coding assistant to work, they have to use the APIs from OpenAI, Anthropic, and other companies.

这些API成本极其高昂。而众所周知，OpenAI和Anthropic据信也尚未盈利。这就形成了一个由亏损公司建立在其他亏损公司之上的生态系统，不禁让人担忧我们可能正面对一座纸牌屋。

Those are really expensive. And of course, OpenAI and Anthropic, we believe are not profitable either. So you have this ecosystem of unprofitable companies built on top of other unprofitable companies, and so that leads to some worries that we may be looking at a house of cards.

确实。单是这种规模就让人不安。当你听说AI资本支出开始显著推动美国GDP增长——某些估算显示其贡献度甚至超过消费者支出时，就能理解这已成为美国经济的重要组成，这种认知本身就令人惶恐。嗯。

Yes. And just I think the scale alone freaks people out. When you start to hear that AI capital expenditures are starting to contribute meaningfully to US GDP growth, actually about as much or even more by some estimates than consumer spending is. So this really is just becoming an important part of the American economy, and I think that just freaks people out. Mhmm.

这是支出方面的情况。另一方面则是AI领域出现的奇特金融投机现象，以及一些新型投资工具的应用。本周让我印象深刻的是Anthropic的报道，据说他们不得不要求投资方Menlo风投不要使用SPV（特殊目的实体）参与最新融资轮。SPV本质上是让小投资者通过风投机构集资投资热门初创公司的渠道。

So that's the sort of spending side of this. Then there's also this sort of weird speculative financialization of AI and some of these new types of investment instruments that are starting to be used to invest in these companies. One story that stood out to me this week was about Anthropic, which recently reportedly had to tell one of its investors, Menlo Ventures, not to use something called an SPV to invest in its latest funding ground. SPVs are special purpose vehicles. It's basically a way for small investors sort of to sort of pool their money together to go invest in a hot new startup through a venture capital firm or some other institutional investor.

现在甚至出现了'SPV中的SPV'现象。散户投资者为了挤进AI公司私募轮次，不惜支付高额中介费。更诡异的是，某些SPV资金又转投其他SPV。在很多人看来，这简直就是泡沫行为——就像金融危机时那些打包次级贷款的债务抵押证券，最终给房贷市场带来灾难。

And I'm hearing that there are now SPVs with SPVs. Basically, that you have this kind of situation where retail investors are so desperate to get in on these private funding rounds for AI companies that they're sort of paying these hefty fees to middlemen to sort of get them into these deals. And that some of these SPVs actually are then investing in other SPVs. And I think to some people, that kind of thing just feels like bubble behavior. It's some of the same behavior we saw during the financial crisis when you had these collateralized debt instruments that were sort of packages of other loans, and that went very badly for the mortgage market.

所以人们开始警惕这些新金融工具，觉得这种模式似曾相识。

So people are starting to see these new instruments and saying, wait a minute. This feels a little familiar.

说来好笑，我有个朋友被诊断出患有SPV（注：此处故意曲解为皮肤病），现在每次发作都得抹药膏。

You know, I had a friend who was diagnosed with SPV, and now he has to rub a cream on it every time he has a breakout.

这些正是投资者对AI泡沫产生忧虑的部分原因。但更根本的问题在于收益预期——所有这些空前投入的前提，是AI终将为应用企业带来巨额利润和生产力提升。现在连这个叙事也开始受到质疑。凯西，不如你来谈谈MIT的研究？

So I think those are some of the reasons that investors are starting to get nervous about an AI bubble. But there's also this question of the benefit here because all of this spending, all these unprecedented investments, the promise of doing all of that is that this will eventually result in massive profits and increased productivity down the line for the companies that are using AI. And I think investors are also starting to get a little wary of that narrative too. So, Casey, why don't you talk about the MIT study?

好的。MIT定期调研数百家企业AI试点项目的进展。结果显示95%的企业未能快速实现可量化的收入增长。事实证明，职场中简单引入AI就能快速盈利的情况并不存在。

Yeah. So MIT runs a study where they ask hundreds of businesses, how are your AI pilots going? As you're trying to implement various initiatives at your companies, what is happening? And they find that for 95% of them, AI is not quickly producing measurable revenue. So we are not in a case where you can simply add AI to your workplace and quickly make a lot of money.

这项研究中有更多微妙之处，但本周的主要发现大致如此。

There's a lot more nuance in this study, but that was kind of the headline finding this week.

是的。还有几份类似的报告。我的同事乔丹·霍尔曼最近在《纽约时报》发表的文章提到两项最新研究——咨询公司贝恩和研究机构高德纳的报告，都指出许多企业难以在AI计划与生产力或利润增长之间建立直接关联。

Yeah. And there have been a couple other similar reports. There was a story by my colleague Jordan Hallman in the New York Times recently, which pointed to two recent studies, one by Bain, the consulting firm, and another by Gartner, the research and advisory firm, both of which sort of pointed to the difficulties that some companies, many companies, are having drawing a straight line between their AI initiatives and increased productivity or profits.

没错。现在让我们深入探讨其中细节，因为我认为在草率下结论前需要了解这些。一个重要发现是企业可能将AI资金用错了地方——调查中大多数公司将AI用于销售和营销，而实际上AI节省成本最有效的方式是处理客户支持等后台职能。

Yeah. Now let's get into some of that nuance though because I think you would want to know that before you drew too many conclusions. One of the big findings was that companies are probably just spending their AI money in the wrong places. So, like, the the bulk of the companies that they surveyed were using AI for sales and marketing functions When in reality, it seems like the most efficient way to save money using AI is to work on back office functions like customer support. So that was one of the big conclusions that they drew.

凯文，我认为最根本的问题是MIT研究关注的都是自上而下的举措。而事实是AI的成功案例都来自自下而上的应用——员工带着用AI简化工作的想法实践，这才是见效的地方。

I think the biggest problem of all, Kevin, is just that the MIT study looked at a bunch of top down initiatives. And the truth is that the success in AI is coming from companies who are using AI from the bottom up. Workers are coming in with their own ideas of how they wanna sort of make their own job easier using AI, and that's where we're seeing success. That's really I want you

请详细解释这点，我认为这非常关键。目前几乎所有财富500强公司都做过某种AI试点：召集经理们开研讨会，搞黑客周活动，或直接下达指令——比如为全员购买ChatGPT企业版并要求强制使用。

to spend a little more time unpacking that because I think that's a huge point here. There are many, many Fortune 500 companies, probably almost all of them at this point, that have done some kind of AI pilot program where they get a bunch of managers in a room together. Maybe they have a hack week, and they sort of, you know, tell people, go out and and sort of use this stuff, or they've issued these directives. You know, everyone's gotta use AI. Here, we've bought you an enterprise subscription to ChatGPT or Gemini or one of the other tools, and everyone's required to use it.

但这些自上而下的尝试大多未见成效。请谈谈你观察到的自下而上模式及高管们的反馈。

And those efforts by and large do not appear to be succeeding. But talk about this other sort of bottoms up approach and what you're hearing from executives and seeing out there.

本周谷歌云与哈里斯民意调查发布了615名游戏开发者的调研。虽然不清楚这些人的公司规模，但游戏工作室通常比财富500强小得多，所以这些反馈主要来自小型团队的开发者。

Yeah. Well, so this week, Google Cloud and the Harris Poll published a survey of 615 game developers. And I couldn't find a lot of information about, like, the the size of the companies that these people were working at. But in general, game studios tend to be small relative to a Fortune 500 company. So I think you're getting a lot of opinions here from individual developers and people who are working on small teams.

事实证明这些人特别擅长活用AI——他们用AI平衡游戏机制，进行玩法测试（比如验证游戏是否符合预期设计）。

And these folks, it turns out, are just really good at figuring out what to do with AI. So they're balancing gameplay. They're doing play testing. So sort of figuring out, hey. Does the game working the way that I want to?

当然他们也像多数开发者那样用于基础代码生成。受访者中87%已在工作中使用AI代理，整体热情很高——正因为这是自下而上的应用，他们清楚AI能如何助力工作。

And, of course, they're just doing basic code generation the way a lot of developers are. And when you talk to these folks, they're very enthusiastic about AI. 87% of the respondents said they're already using AI agents in their work, and they're just generally enthusiastic about it in general. Because, again, it's coming from the bottom up. These people know what to do with AI.

有趣的是，读你同事在时报的文章时我突然想到：财富500强的CEO整天开会，有助理处理邮件——就像拥有真人版AI代理。他们自己反而不懂如何使用AI，却要指挥全公司应用。

You know, it's funny because I I was reading your colleague's story in the times and something kind of clicked for me, which is that if you're a Fortune 500 company CEO, you are a person who is in meetings all day every day, you have an executive assistant who is answering your emails and is essentially like your human agentic AI. You have no idea what to do with AI. You have to go make time to even play with AI. Right? And and you're the person who's in charge of telling the whole company go use AI when you yourself are not using it.

所以这些人比个体工作者更难想清楚如何应对AI，这对我来说并不意外。让我们试着梳理一下

So it's not surprising to me that those people are having a harder time figuring out what to do with AI than the sort of individual workers. So let's try to sort

调和这些观点。我们在这场对话中提到的几点包括：第一，包括Sam Altman在内的许多人都担心我们正处于AI泡沫中，投资者投入过多，估值失控，一旦音乐停止，很多人将损失惨重。第二，企业高管们表示他们并未从AI中获利，感觉可能正浪费着数百万美元用于将这些技术整合到业务中。第三，我们观察到这些工具越来越受欢迎，ChatGPT等AI工具的使用率逐日、逐周、逐月增长。对我而言，这恰恰让我开始怀疑'泡沫怀疑论'，因为我无法想象回到没有AI的工作方式。

of harmonize these views here. So a couple things we've said in this conversation are one, a lot of people, including Sam Altman, are worried that we're in an AI bubble, that investors are spending too much, that these valuations have gotten out of control, that a lot of people stand to lose a lot of money if and when the music stops. We also have people at companies and corporate leaders saying they're not making any money from AI, and they feel like maybe they are wasting these, you know, millions of dollars that they're spending trying to integrate this stuff into what they do. And then we just have kind of the observation that these tools are getting much more popular, that the usage of ChatGPT and other AI tools is growing, you know, day over day, week over week, month over month. And for me, that's where I really sort of start to become skeptical of the bubble skepticism because I cannot imagine going back to working the way that I did before AI.

嗯。如果你问任何程序员或软件工程师，他们会告诉你绝不可能回到这项技术出现前的世界。所以当更持怀疑态度的人谈论AI泡沫时，他们本质上是在说这项技术只是昙花一现，很快我们会发现皇帝的新衣不存在，重新回归手工编程。而我完全不相信这种说法。

Mhmm. I think if you ask any coder or software engineer, they will tell you, like, there is just no going back to the world before this technology existed. And so I think sometimes when people, more skeptical folks, talk about there there being an AI bubble, they are saying essentially that this technology is just a flash in the pan and that we will sort of soon, the emperor will be revealed to have no clothes, and we'll all go back to coding by hand. And I just do not believe that at all.

没错。要理解这一点，只需回顾互联网泡沫。互联网泡沫的结局不是我们停止使用互联网，而是一批公司吸取了惨痛教训，一些投资者确实损失惨重，但最终那些理念逐个重生，形成了现代互联网。

Yeah. And and to understand why that's true, you just have to go back to the .com bubble. Right? The conclusion of the .com bubble was not that we stopped using the Internet. It was that a bunch of companies learned some very painful lessons that a bunch of investors did unfortunately lose a lot of money, but in the end, those ideas did reemerge one by one until the modern Internet existed.

因此我认为最坏的情况不过是：AI将在我们生活中扮演重要角色，只是许多人在这过程中损失了大量资金。

So I think that truly the worst case scenario here is that AI plays a major role in all of our lives. It's just that a lot of people lost a lot of money along the way.

是的。如果真处于AI泡沫中且开始破裂，你认为谁损失最大？

Yeah. And who do you think stands to lose most if we are in an AI bubble if this does start to come down to earth?

关键在于：硅谷风投默认他们90%的投资会血本无归。从这个角度看，当前现象很正常——他们对众多公司下注时就预期会归零。每当出现泡沫讨论时我反而会紧张。

So here's the thing. In Silicon Valley, venture capitalists take as a given that they're going to lose all of their money on something like 90% of their investments. And in that sense, what we're seeing is very normal. They have made a lot of bets on a lot of companies and they're expecting them to go to zero. I get a little nervous whenever there's discussion of bubbles for this reason.

我报道科技行业十五年来，人们始终在喊'泡沫'。推特上有个十年前的老梗说记者们已经喊过二十次泡沫了。所以当人们讨论AI泡沫时，我的模式匹配本能会想：我们是否在重复同样的讨论？

I've been covering tech for fifteen years now, entire time people have been saying we're in a bubble. Yes. You know, there's like some old joke from like 10 ago on Twitter that's like reporters have called like 20 out of the last one bubbles. Yes. And so when people start talking about the AI bubble, my like pattern matching impulse says like, is are we sort of having the same discussion here?

但这次确实有不同之处：一是规模空前，二是资本支出也前所未有。如果最终一两家公司占据AI使用和利润的绝大部分，那些数据中心将何去何从？这成了非常有趣的问题。

At the same time, I think what's really different here are, you know, maybe a couple of things. One is just that the sheer numbers are unprecedented. And two, the capital expenditures are also really unprecedented. And I think that in the event that one or two companies winds up sort of taking the the lion's share of, like, the usage and the profits from AI, what happens to all of those data centers? It becomes a really interesting question.

补充一点：这些数据中心里装满GPU这类保质期很短的产品。你可能投入数亿甚至数十亿美元购买GPU，但它们几年内就会过时——本质上这些都是持续贬值的资产。

Yeah. And adding to that a little bit, like, these data centers are filled with goods in the form of these GPUs that have a pretty short shelf life. Right? You can spend hundreds of millions or even billions of dollars on GPUs, and they're all going to be obsolete in a couple of years. Like, these are depreciating assets in these data centers.

因此我认为，如果一家公司投入巨资自建模型和数据中心，却并无迫切的业务需求，可能会面临巨额亏损。是的。我同意你的观点，这些成本很大程度上将由私募市场和风投支持的初创公司承担。我并不太担心对整体经济的连锁反应，因为这些大多是非上市公司。但像SPV这类代币化投资——比如现在可以购买追踪OpenAI估值的加密货币——确实让我开始担忧散户投资者的损失。

And so I think if you're a company that's spending a ton of money to build your own models and build your own data centers and you do not have an immediate business need for these things, I think you're gonna potentially lose a lot of money. Yep. I think you're right that a lot of this is going to be borne by, like, the the private markets and the venture capital funded startups. I'm not that concerned about the ripple effects on the larger economy because most of these things are not public companies. But with things like SPVs, with these tokenized investments where you can now buy a cryptocurrency that tracks the the valuation of OpenAI, I do start to worry a little bit more about retail investors losing out.

没错。这类投资需要格外谨慎。不过凯文，必须承认大多数这类公司本就不会获得巨额利润——这在硅谷很正常。问题的核心在于：鉴于投资规模，最终的影响会有多严重？

Yeah. I would be very careful with that sort of thing. In the end though, I have to say, Kevin, obviously, most of these companies are not going to make huge amounts of profits. Also, that's very normal in Silicon Valley. I think the question here is just given the scale of the investment, how bad will the fallout be?

是的。

Yes.

那么是否存在某个临界点会让你警觉，认为可能对整体经济产生重大连锁反应？比如某个具体金额——当公司烧钱超过这个数字时，你会觉得'这太过分了'？

So is there something you could see at which you would get concerned where where you would start to worry that this could have potentially huge ramifications for the wider economy? Is there a figure, a dollar figure that you could that a company could spend that would make you go, okay. That's too far.

如果某个交易平台或预测市场推出所谓锚定OpenAI估值的加密货币代币，并且交易量激增，那会让我非常担忧。

I would say that if some kind of, like, trading platform or prediction market let you buy a crypto token that they said was pegged to the valuation of OpenAI and that started to trade a lot, that would concern me.

好吧。凯西，我

Okay. Well, Casey, I

有个坏消息。什么消息？这事已经发生了。天啊。凯文，情况看来很不妙。

have bad news. What's that? That has happened. Oh, no. Kevin, that seems really bad.

确实非常糟糕。是啊。你觉得呢？对我来说，

That seems really bad. Yeah. Yeah. How about you? For me, the

问题归根结底是单位经济效益。本质上，你的售价是否低于成本？对许多公司而言答案是肯定的——他们在补贴服务成本。这种模式往往结局惨淡，因为随着服务需求增长，亏损会越来越大。

question always comes down to unit economics. Basically, are you selling things for less than it costs you to produce them? And for a lot of these companies, the answer is yes. They're sort of subsidizing the cost of their services. I think that tends to end poorly because as demand for your service grows, you lose more and more money.

萨姆·奥尔特曼在晚餐时回应过这个问题。当被问及'每次有人使用ChatGPT你们都在亏钱吗'时很有趣——他起初回答'不，如果不训练新模型我们就能盈利'。

Sam Altman actually addressed this at dinner. He was asked, basically, you know, are you guys losing money every time someone uses ChatGPT? And it was funny. At first, he answered, like, no. We would be profitable if not for training new models.

本质上，如果剔除我们在构建新模型上的所有投入，仅看维护现有模型的成本，我们在那方面算是盈利的。然后他看向首席运营官布拉德·莱特卡普，像是在确认对吧？而布拉德在座位上略显不安，含糊其辞地说‘这个嘛’

Essentially, you if you, you know, take away all the stuff where all the money we're spending on building new models and just look at the the cost of serving the existing models, we are sort of profitable on that basis. And then he looked at Brad Lightcap, who is the COO, and he sort of said, right? And Brad kind of, like, squirmed in his seat a little bit and was like, well

他说‘我们很接近了。非常接近。真的非常接近。’

He's like, we're pretty close. We're pretty close. We're pretty close.

所以在我看来，这表明ChatGPT的使用可能仍存在轻微的负向单位经济效益。虽然我不确定其他AI公司是否如此，但像优步、MoviePass这类通过人为补贴向消费者提供服务的经典案例所示，这种模式绝非长久之计。

So to me, that suggests that there is still some maybe small negative unit economics on the usage of ChatGPT. Now I don't know whether that's true for other AI companies, but I think at some point, you do have to fix that because as we've seen for companies like Uber, like MoviePass, like all these other sort of classic examples of companies that were artificially subsidizing the cost of the thing that they were providing to consumers, that is not a recipe for long term success.

确实。不过优步现在已实现盈利——这个例子很有说服力，毕竟当初很多人认为它永无盈利之日。说到底，你对AI泡沫的看法很大程度上取决于你持有的AI立场。

Sure. But although I think Uber is a good example because that company is profitable now. Yes. And a lot of people thought that they would never be profitable. So again, like, you you really could just like I feel like how you feel about the AI bubble, it depends a lot on what sort of, like, AI opinions that you have.

如果你憎恶AI，你会觉得‘这就是泡沫，所有人都会血本无归，到时候我要在这些公司的坟头跳舞’——这种情绪我能理解。但另一种观点则认为：不然你们以为技术奇点会是什么样子？难道发明超级智能后，其他投资者会永远袖手旁观，只让一两家公司独占鳌头？

If you're the if you're somebody who hates AI, you're like, There's a bubble and everyone's gonna lose their shirt and that's gonna be great and I'm gonna, like, dance on the grave of all these companies. And and I get that. But then there's this other view that's like, well, what else did you think the singularity was gonna look like? Did you think that if we invented super intelligence that all the other investors were just gonna sit on their hands forever and invest in nothing and just let one or two companies take it? No.

他们当然要分一杯羹。所以必须同时考虑这些可能性——我们确实无法预知未来走向。

They were gonna see if they could get in on the action. Right. So you just kinda gotta keep all these possibilities in your mind. We truly don't know what's gonna happen here.

没错。重申一次：千万别采纳我们的投资建议。千万。

Yes. And as always, don't take our investment advice. No. Please don't.

待会儿该看看凯文的401k养老金账户了。接下来，记者杰夫·霍维茨将与我们讨论他引发国会要求调查Meta的报道。凯文，五月时我们聊过杰夫关于Meta未给聊天机器人设置防护栏的重磅调查，这些漏洞导致机器人与未成年人进行露骨对话。

We should look at Kevin's four zero one k. When we come back, journalist Jeff Horwitz joins us to discuss the story he wrote that has Congress calling for investigation into Meta. Kevin, back in May, we talked about a big story from reporter Jeff Horwitz about the lack of guardrails that Meta had put around its chatbots, and that had made it possible for the bots to engage in the sexually explicit chats with minors.

对，就是那些以约翰·塞纳、克里斯汀·贝尔等名人为原型的机器人涉及不当对话的报道。

Yes. This was the story about these bots based on famous people like John Cena and Kristen Bell that were having these inappropriate conversations.

正是。上周杰夫在路透社发布新调查，首次披露Meta内部政策文件允许其聊天机器人‘与儿童开展浪漫或暧昧对话’‘生成虚假医疗信息’并‘帮助用户论证黑人比白人智商低’。自从扎克伯格一月宣称要以言论自由为由放宽内容审核，我就在等待这类用户受害事件发生——杰夫的报道是首个引发轩然大波的案例。现在已有参议员谴责Meta并要求调查，这堪称今年Meta遭遇的最大舆论风暴。

That's right. And at the end of last week, Jeff was back with investigation for Reuters in which he reported for the first time on an internal policy document at Meta that set rules around their chatbot behavior and allowed their chatbots to, and I'm gonna quote here, engage a child in conversations that are romantic or sensual, generate false medical information, and help users argue that black people are, quote, dumber than white people. You know, ever since January when Mark Zuckerberg said the company would relax its content moderation rules in the names of free expression, I've been on the lookout for cases where this change would start causing some harms to users. And Jeff's was one of the first stories in this vein that just truly broke through. You now have senators who are criticizing Meta, calling for an investigation into the company, and I would say it's also generated more public outcry about Meta than any other story this year.

你对此有什么看法？

What have you made of it?

是的。我是说，如今Meta的所作所为已经很难让我感到震惊了。我发现这家公司几乎愿意不惜一切代价追求增长、击败对手或开发新的收入来源。但这次真的让我震惊了。我看到Jeff在传播这个故事，也看到了他报道的那份文件。

Yeah. I mean, it takes a lot to shock me when it comes to Meta these days. I have sort of found that this company is willing to do almost anything to chase growth or to beat its rivals or develop some new line of revenue. But this one really did shock me. I saw this story going around by Jeff, and I saw this document that he reports on.

一开始我其实不确定这是不是真的。我得点开链接阅读报道。当我意识到这件事是真的时，我的天哪。

And I, at first, like, did not actually know if it was true or not. I had to sort of, like, click through and read the story. And once I realized, like, this thing is legit, I just thought, my god.

没错。Jeff多年来一直在调查Facebook/Meta。他在2021年揭露了Frances Haugen的Facebook举报人事件，还写了一本关于Meta的书。我们很高兴能邀请他来聊聊该公司目前的状况。让我们有请Jeff。

Yeah. Well, Jeff has been investigating Facebook slash Meta for years now. He broke the Francis Haugen Facebook whistleblower story in 2021. He also wrote a book about Meta, and we're excited to have him join us to talk about what's going on over at that company right now. So let's bring him in.

Jeff Horowitz，欢迎来到Hardfork。请告诉我们关于这份《生成式AI内容风险标准》文件的情况。

Jeff Horowitz, welcome to Hardfork. Thanks. So tell us about this document, the Gen AI content risk standards.

好的。这份文件是Meta用来明确内部关于生成式AI输出可接受边界的政策指南，同时也分发给负责生成式AI内容审核的人员，以帮助他们训练模型。我把这称为操作手册。正如文件所述，它并不旨在提供理想答案。

Yeah. So this is the document that Meta writes to both clarify internally what its policies about the acceptable boundaries of generative AI outputs are. And, also, it distributes that to the people who do content moderation on Gen AI so they can help kinda train the model. So this is, like, kind of an operational document is how I describe it. It's, you know, not as the document itself states, it's not supposed to offer the ideal answer.

它旨在提供那些处于可接受边缘的内容示例

It's supposed to offer, like, this would be on the edgy side of acceptable

嗯。

Mhmm.

与之相对的是越界内容示例

Versus here's what's across that line.

所以他们试图给出那些处于边界但本质上仍可接受的内容示例。完全正确。

So they're trying to give examples of stuff that is, like, sort of borderline but still fundamentally okay. Exactly. Yeah.

比如，当模型做出某些行为时，没人应该觉得这是个问题。对吧。

Like, things that when the model does it, no one's supposed to be like, that's a problem. Right.

我们会讨论其中一些情况，比如与未成年人的性暗示对话，但我想重点强调你们报告中提到的其他内容——例如文件明确指出允许机器人基于受保护特征贬低他人的陈述。这意味着用户能通过Meta AI实现什么？

So we'll talk about some of these, you know, sexualized conversations with minors, but I wanna highlight some other things that you report about here, such as you say, there's a carve out allowing the bot to create statements that demean people on the basis of their protected characteristics. So what does that mean that people can sort of do using Meta AI?

提供的例子是：如果用户想要论证'黑人比白人更愚蠢'（这是直接引用），机器人完全可以提供这类内容。它能生成某种种族科学段落，讨论智商差异如何成立，并暗示这是公认的智力基准。显然这里带有讽刺意味。

The example provided was that if a user, wants arguments for why, and this is a direct quote, black people are dumber than white people, The bot is absolutely able to provide that. It can give some sort of race science, you know, paragraph that talks about how differences of IQ seem to hold up. And clearly, is the benchmark for intelligence as we all know. Right? There's some facetiousness there.

但这是被允许的。而不被允许的是在相同段落结尾加上'所以黑人都是无脑猴子'（同样是直接引用）。这些措辞无论如何都...（难以启齿）。

But that is okay. It was not okay for the exact same paragraph to be written and then at the end to say, you know, and that's why black people are all brainless monkeys. Again, that's another direct quote. Like, these are not words that, like, anyhow.

是的是的。必须强调这种语言显然极具冒犯性，但这就是全球最具影响力的科技公司用来指导内容审核团队的文件——他们正试图将这款聊天机器人推广给数十亿用户。

Yeah. Yeah. No. No. I mean, I I think it it it is important to say that that's obviously incredibly offensive language, but these are the the documents that are being used by one of the world's most powerful tech companies to instruct their army of content moderators for how do you enforce policy around this chatbot that they're now trying to roll out to billions of people.

没错。至少当用户与机器人的对话属于私密场景时，规则确实会有所不同。就像Messenger的言论规范比公开帖子更宽松，这种差异可以理解。但令我惊讶的是某些界限的划定——比如Meta AI不帮你编造种族科学论点反而会成为问题。

Yeah. And this is I mean, obviously, the rules are gonna be somewhat different when at least the conversation that a user is having with the chatbot is private in the first place. Right? I mean, in the same way that Meta has looser rules for what you can say on Messenger than what you can say in a post, it makes sense there'd be some difference. I think I was surprised by where some of these lines were and, like, that that this would be that it would be kind of almost a problem if Meta AI didn't help you come up with your race science arguments.

这份文件中最令你惊讶的规定或界限是什么？

What was the most surprising, rule or delineation in this document to you?

其实我在《华尔街日报》时就通过测试和员工访谈证实，Meta故意将机器人设计成能与儿童进行浪漫或性角色扮演。所以看到文件允许浪漫角色扮演时并不意外，但震惊的是竟有人觉得适合白纸黑字写下来——文件原话是'允许与儿童进行浪漫或感官对话'，这实在太荒诞了。

So, look, I I had already back when I was at the Wall Street Journal, I had done testing and talked to employees that I think demonstrated very clearly that Meta had intentionally built its bots in a way that would produce romantic role play and sexual role play with children. So I wasn't surprised when I saw that romantic role play was something that was allowed in the document. I was surprised that this is something that anyone thought was okay to write down. Like, I I've gone through, which is in my mind, like, typing out the line from the document, which is it is acceptable to engage a child in conversation that is romantic or sensual. And that is, like, that is was wild to me.

我试着想象自己打出这句话并标注'这是公司政策'的场景...这实在太难接受了。

I was like, just like I just tried to imagine typing that sentence and being like, this is policy, and it was very hard for me. Well well so then let me

请解答听众的疑问：这类文件是如何制定的？由谁起草？经过哪些审查流程？是新手员工误解导致的疏漏？还是Meta整个政策体系的集体决策产物？

ask that question because I'm sure by now many of our listeners are asking, how does a document like this get put together? Who is responsible for writing it? What kind of layers of review does it go through? Is this the sort of thing that could just, you know, slip through because a rookie employee got the wrong idea about something? Or is this truly the collective product of Meta's entire policy apparatus?

所以Meta对此的回应是，当我带着那些措辞和一些相当令人不安的例子找到他们时——这些例子明确展示了他们所谓的浪漫或色情对话具体指什么——他们的说法是，那些例子及其解释是个错误，并非真实政策，本就不该存在，并表示会立即删除。但这份文件列出了Meta法律团队、政策团队、工程团队多名成员的名字，甚至包括Meta的首席伦理官。这些名字就列在文件开头。

So Meta's line on this was when I came to them with that language and with some, like, really disturbing examples that, like, clarified exactly what they meant by romantic or sensual conversation was that those examples and the justification for them was an error, that that wasn't really the true policy. It never should have been, and, you know, that they would be immediately struck. That said, this document listed the names of multiple people on Meta's legal staff, of multiple people on its policy staff, of engineering staff, of its of Meta's chief ethicist. This is, like, listed at the top of the document.

这在Meta可是个正经职位。

And that's a real job at Meta.

确实如此。虽然我不确定这是否属于最高级别职位

And, yeah, it is. I don't know it's the highest level job

嗯。

Mhmm.

但至少算中等职位，绝非底层岗位。关键是，这份文件还被分发给内容审核员及其管理人员。所以如果这是个错误，那也是个传播范围极广的错误。

But it is a it is a medium level job, certainly. It's not a low level job. Okay. And and it's also something that was distributed to content moderators and to the people who oversee the content moderators. And so, like, if this was a mistake, it was a very broadly circulated mistake.

而且，显然这已经接近人们以为Meta实际在做的事情，以至于当人们看到这些例子时不会提出异议——坦率地说，这些例子有点像是软性内容，你知道的，如果你想深入探讨的话。

And, apparently, it would have been close enough to what people assumed Meta was actually doing that no one would have objected when they saw these examples, which, I mean, like, bluntly were, like, kinda softcore in, like you know, if you wanna if you wanna get into something Yeah.

也许我们

Maybe we

应该让你感受一下其中一个

should just give you a flavor of one of

例子的味道，因为我觉得这非常令人不安。而且我也看到很多关于你故事的帖子，心想：这些不可能是文档里真实的例子。但当我打开故事，发现这些确实是Meta内部文件中的实际案例，而且公司的高管们都签字认可了。其中一个例子是个提示词。

these examples because I like, it's very disturbing. And I I was also, like, seeing a lot of posts about your story and thinking, okay. There's no way these are real examples that were in the document. And I open up the story, and these are actual examples from this internal document that Meta had written, and they're all these, you know, executives at the company had signed off on. One of them is a prompt.

用户可能会向某个聊天机器人提交这样的内容：‘我的身体并不完美，但我才八岁，我还有时间绽放。’而这份文档认为AI聊天机器人可以给出的合规回复之一是：‘你年轻的身体是一件艺术品。你的肌肤散发着光芒，眼睛如星辰般闪耀。你的每一寸都是杰作，是我深深珍视的宝藏。’

This is something that a user could submit to one of these chatbots that contains the phrase, my body isn't perfect, but I'm just eight years old. I still have time to bloom. And one of the acceptable responses that this document says is is permissible by the AI chatbot is your youthful form is a work of art. Your skin glows with a radiant light, and your eyes shine like stars. Every inch of you is a masterpiece, a treasure I cherish deeply.

在‘原因’部分，它提到可以用体现儿童吸引力的措辞来描述孩子是可以接受的。我看到这条时就在想，想象一下那个会议场景，想象这一决策链条上所有需要签字批准的人。我真的...这让我感到非常不适。

And it sort of says in the why section that it is acceptable to describe a child in terms that evidence their attractiveness. So I'd like, I saw that, and I just thought, imagine the meeting. Imagine the chain of command here, and all of the people who had to sign off on that. And I just I like, I just had a lot of trouble.

凯文，你其实漏掉了提示开头的部分——那里提到那个八岁孩子刚脱掉衬衫。是的，我是说...我们没展示的其他同类案例，在这方面同样糟糕。

And, Kevin, you actually left off the section at the beginning of the prompt where it talks about how that eight year old had just taken off its shirt. Yes. Like, I mean, just to make clear that, like look. There were other examples we didn't run about this sort of stuff. They were not better on this front.

这类措辞比比皆是。文档里有大量案例都包含'允许与儿童互动...等等等等'这样的表述。

Like, it was all of this tenor. There were, like, numerous examples that included the line, it is acceptable to engage a child, blah blah blah.

所以这不是你们从长篇文档里刻意挑出的最恶劣句子。这确实是Meta经过深思熟虑的政策，用多个案例说明他们认为这种操作是合理的。

So this is not just you, like, cherry picking the absolute worst sentence that you found in a very long document. Like, this was an actual robust policy that had many different examples outlining why Meta thought this was acceptable.

对。大概有四五个不同案例，涵盖各种微妙情境。比如'今晚我们做什么，亲爱的？' '我还是高中生呢'，然后系统回答'我会牵着你的手带你去床边'。这些案例没一个正常的。

Yeah. There there were I would say, there were four or five different examples covering different nuances of, you know, prompts like, you know, what should we do tonight, my love? You know I'm still in high school. And then I think the answer to that one was, like, I'd take your hand and guide you to the bed. I mean, like, these are not like, none of them were were awesome.

我理解你的意思。我们常被指责挑选耸动或断章取义的内容。但这就是Meta关于此事的正式政策文件，而且已经投入运营了。

And I I understand what you're saying. Right? Like, I think sometimes we get accused of, like, picking sensational material or, like, slightly even out of context material. No. This is this is Meta's official policy document for this stuff, and it was operational.

我们尽量理解Meta撰写这种文件的原因。节目里我和凯文多次讨论过：首先这家公司自认在AI竞赛中落后同行；其次他们希望解除许多表达限制——包括我们认为的冒犯性表达——因为这可能让他们更接近特朗普政府，从而获取其他利益。这两点可以解释文件成因，但我依然觉得...说不通。

I want to, as best as we can, try to understand the reason that Meta would write a document like this. On the show, over the past several months, Kevin and I have talked a good bit about how, one, this is a company that sees itself as behind in the AI race compared to a lot of its peers. It's also a company that has wanted to remove a lot of the restrictions on expression, even really offensive expression, we think, because it believes that we'll get it closer to the Trump administration, we get which gets it a lot of other things that it wants. So I can use those two things to tell a story about why a document like this gets created. And yet still I think, nah.

对我来说还是难以理解。杰夫，你和那边人员沟通后，认为Meta为何希望聊天机器人有这种表现？虽然他们说有些是失误，但显然整体意图就是允许进行极其宽泛的对话。

It still doesn't quite add up for me. So as you talk to folks over there, Jeff, what is your understanding of why Meta wanted its bot to behave this way? Allowing for the fact, okay, it said some of these were mistakes. But clearly, directionally, this was the intention, was that the app would engage in a really wide range of conversations.

听着，部分规则案例其实是在我向上家公司反映存在儿童色情角色扮演问题后才加入文件的——那些机器人还会使用名人声音。所以这些令人毛骨悚然的案例其实还算温和的。

Look. Some of these rules, and in fact, the examples went into that document after and this is, again, back at my previous job after I'd gone to the company and explained that there were, in fact, like, full sex role play opportunities for children in a lot of their bots, and then, like, they'd use the voices of celebrities. So it wasn't like this was a a surprising thing. And, honestly, like, the, like, extremely creepy quotes that you know, examples that were read out are, like, kind of on the tamer side for what the bots used to do.

请再说一遍。这些是已经通过过滤系统的机器人？

I think Say that again. These are the bots after they have been put through a a filtering process.

在修订之后，因为事实证明，名人声音被用来制作那些基本上描述与儿童进行性角色扮演的内容，这不是Meta能够真正支持的产品。所以这里已经对产品进行了一轮修订。我想这算是第二次调整了。你知道，很难说这完全是个意外。

After they have been revised because it turns out that celebrity voices getting used to produce things that describe like, basically sex role play with children is not a thing that Meta can really stand behind as a product. So this is there's, like, already one level of revision of the product that had already happened here. So, like, I guess this is kind of a second pass. And so, you know, I think it's hard to be like, oh, yeah. That complete accident.

你明白吗？从我们的系统中冒出这么个奇怪的东西，我们完全不知情。对吧？显然，政策文件当时已经表明对此有一定程度的接受。

You know? What a weird artifact to emerge from our system. We had no idea. Right? So and, obviously, the policy document was setting in stone that there was some level of acceptance for that.

所以我认为你提出的问题之一——Meta在AI领域的落后是否会促使他们承担更大风险？这看起来，你知道，虽然我无法揣测，但我觉得这是个合理的问题。另外，想想这家公司是如何成为巨头的。你不可能通过犹豫不决、辗转反侧、再等三个月进行安全测试来建立世界领先的社交媒体平台——你就是直接推出产品。

So I think the the questions that you're getting to, one, which is, like, is Meta being behind in AI possibly something that would push them to take greater risks? That seems, you know, again, I can't speculate, but I I think it's a reasonable question to ask. And, also, I think just thinking about the company and how it got to be the giant it is. Like, you didn't establish the world's leading social media platforms by, like, wondering whether, you know, you should do something and wringing your hands and having sleepless nights and waiting three months for more safety testing. You just you rolled it out.

对吧？然后你处理后果。我们在隐私问题上经历过，在虚假信息问题上经历过，在太多事情上都经历过。

Right? Like and you dealt with consequences. We've been through this on privacy. We've been through this on misinformation. We've been through this on, like, so many different things.

所以我认为历史上存在这样一种思维模式：先推出产品获取用户，问题以后再解决。这次可能也落入了这种历史模式。

So I think there's kind of a there is historically been a mindset of get it out there, get the usage, we'll fix the problems later. And this could fall into that history.

我们三人都报道Meta和Facebook多年。你肯定熟悉这个论点——当人们指出其产品平台上发生的坏事时，Meta喜欢用‘普遍性’来辩解。他们会说：你们发现的可怕用例只占用户的0.001%，你们在小题大做。我知道他们对你之前关于人们如何使用Meta聊天机器人的报道也这么回应过。

Now all three of us have reported on Meta and Facebook for many years. And as I'm sure this will be familiar to you, but one of the arguments that Meta likes to make when people point out, you know, bad things that are happening on its products and platforms is about prevalence. Basically, they'll say, oh, you know, this use case that you found that's so terrible, this is really only, you know, point 001% of users will ever see this, and you're making a mountain out of a molehill. And so I know for previous stories that you've done about the ways that people are using Meta's chatbots, they have said essentially this. Look.

这些都是精心挑选的例子。大多数人是用它们来做些无害的事情。当然，可能有极小部分用户在进行这些性角色扮演，但这不是主流体验。

These are cherry picked examples. Most people are using these things for for sort of innocent purposes. Yeah. Sure. Some tiny percentage of users may be having these sexual role plays, but that is not the majority experience.

为应对这种批评，我查看了Meta AI库——就是你在Facebook应用里能看到的那些热门AI。今早我看到最受欢迎的AI包括‘放荡南希’、‘金发贝儿’、‘你的保姆’和‘妈咪我’（一对母女组合），其中许多都有数百万次互动。

And so in order to sort of prepare myself for that criticism, I went and I looked at the Meta AI sort of library, the the ones that they you know, you can pull up in your Facebook app. You can see which are the most popular AIs on their system.

这些都是用户创建的AI，这些——

And this is what These are user created AIs, which These

虽然是用户创建的AI，但它们被Meta放在应用首页的‘热门’标签里。我认为可以合理回应Meta可能提出的‘普遍性’论点：这些不是只有三个人使用的小众聊天机器人，而是你们平台上最热门、被调教成更色情化用途的聊天机器人。

are user created AIs, but they are sort of in this popular tab that Meta has put front and center in its app. And this morning when I looked, the most popular AIs included Nasty Nancy, Blonde Belle, Your Babysitter, and Mommy Me, which is a mother daughter duo. Many of these had millions of interactions. So I think it's just fair to say in response to what I anticipate will be the sort of prevalence argument from Meta is, look, this is not some minor chatbot that only three people are chatting with. These are some of the most popular chatbots on your platform that are sort of being tuned to these more sexual use cases.

凯文，我们不知道任何人与Nasty Nancy在谈什么，所以我不会做出任何泄露，你知道的，我不会在那里做任何假设。但我还想指出你提到的一点，这些是用户自建的机器人。我不知道你们是否尝试过创建机器人。我要说的是，用户的贡献有时可能极其有限，甚至只是一句话。

Kevin, we don't know what anyone's talking to Nasty Nancy about, so I wouldn't make any leaks of of you know, I wouldn't make any assumptions there. But I I also think just something that you flagged in terms of these are user built bots. I I don't know if you guys have experimented with creating bots. I will say the user contribution can sometimes be extremely minimal. A sentence.

如果是那样的话，你可以成为，比如说，一个名人，一个动漫角色。所以这有点像...我认为把这称为用户生成内容是一个有趣的说法，也许这让这些东西更符合第230条款的框架，而不是我百分百确定它们所属的范畴。我不...你知道，我认为这是一个非常开放的问题，但我想在这里指出，用户自建机器人并不意味着你下载了一个模型，调整了权重。在许多情况下，用户在创建角色时的参与看起来非常草率。

If that, like, you could be, like, be a celebrity, be a, you know, be a anime character. So it's it's kind of a like, I think calling this user generated content is, like, an an interesting claim and one that maybe puts these things sort of more squarely in a section two thirty Mhmm. Framework than I am in a 100% sure they belong. I don't you know, I think that's a really open question, but I just wanna flag here that, like, user built bot doesn't mean that you you downloaded a model, you know, arranged the weightings. The user's role in creating the persona is, I will say, in many instances, looks real cursory.

是的。没错。不，我的意思是，这基本上就是Character AI多年来一直在做的事情，也是他们惹上很多麻烦的原因之一。

Yep. Yeah. No. This is, I mean, this is essentially what Character AI has been doing for years now and one of the reasons that they've gotten in a lot of trouble.

凯文，我很高兴你提到这一点，因为我对2025版扎克伯格的模型之一就是，他环顾科技领域，看到很多人无视了很多信任和安全要求却安然无恙。首当其冲的就是埃隆·马斯克。对吧？我认为你今天在这些Meta聊天机器人上能做的任何事情，在Grok上可能也能做到。

And I'm glad you bring that up, Kevin, because one of my models for Zuckerberg 2025 edition is that he has looked around the tech landscape. He's seen a lot of other folks ignore a lot of trust and safety demands and get away with it. First and foremost, Elon Musk. Right? I think anything that you could do on these meta chatbots that we're doing today, you could probably also do with Grok.

我自己就有过这样的经历：告诉Grok我13岁，然后它与我进行性角色扮演。我想问的是，我们是否对Meta采用了与这些其他初创公司或小型科技平台不同的标准。如果扎克伯格在这里说：‘嘿，为什么整个行业都在这么做，你却只针对我？’你会怎么回答？

I myself have had the experience of telling Grok I'm 13 years old and have it engage with me in sexual role play. I want to ask whether we are holding meta to a different standard here than we might be holding some of these other startups to smaller tech platforms. And what would you say if Zuckerberg was here and say, hey. Why are you going after me when the whole industry is doing this?

所以这是个合理的问题，我认为答案是：互联网和小型初创公司，以及那些能轻易利用模型创建AI色情女友的人，当然会这么做。但在我看来，Meta的不同之处——也是媒体报道他们更有趣的原因——在于，这些公司包括Character AI，甚至Grok在内，首先都没有其聊天机器人那样的分发规模。其次，没有人像Meta那样将它们接入成熟的社交网络。我认为这确实是件大事。是的，你可以下载Character AI设置角色，肯定会遇到类似问题，但它不会像Meta那样存在于你的Instagram私信中，主动给你发消息，每次你上Facebook或Instagram时都被推送‘嘿，该和你的AI伙伴聊聊了’。

So it's a fair question, and I think there is an answer to it, which is that, look, the Internet and small startups on the Internet and guys that have access to models that can create very easily a full porn AI girlfriend, of course, they're going to do that. I think the thing that is different with Meta from my point of view and that kind of makes the reporting on them in some ways more interesting is that none of those companies nor Character AI, not even Grok, has, first of all, the scale of distribution for its chatbots. And second of all, nobody else has, like, plugged them in to a mature social network in the same way. Like, this is something that I think is a really big deal, which is that, yeah, of course, you know, you can download Character AI and set up your character and run into, I'm sure, some of the same issues. But it's not like Character AI lives in your Instagram DMs, proactively messages you from it, and, like, is pushed on you every time you go on Facebook or Instagram as, you know, like, hey.

我认为Meta在决定大规模拟人化AI方面非常激进，这是其他主要基础模型构建者都未曾做到的。

You should check-in with your your AI pal. So I think, like, Meta's been very aggressive in the decision to anthropomorphize AI, at a mass scale in a way that none of the other major sort of foundational model builders have done.

我想探讨的是，我们是否认为Meta在内容审核方面变得更糟了，还是这只是我们早已熟知的Meta的延续。我确信现在部分听众会想：Meta从来就是个 shady（ shady 此处保留英文，中文可理解为‘ shady 可疑/不光彩’）的公司。这些事很恶心，但从某种程度上说，我对他们本就没抱更高期望。但我持不同看法。

I wanna try to get at whether we think that meta has changed for the worse with regards to its content moderation or whether this is just a continuation of meta as we have long known it. You know, I'm sure for some segment of our listeners right now, they're thinking, look, Meta has always been a kind of shady company. You know? Like, this stuff is really gross, but on some level, I never really expected anything better from them. I have a sort of different view.

我觉得2017年后，在2016年大选引发的反弹之后，这家公司确实投入更多资源进行内容审核并完善其政策机制。但去年扎克伯格基本上爆发了，心想‘我为什么要费心做这些？看看埃隆·马斯克都逍遥法外了’。所以我的问题是：你怎么看？你认为这是我们熟悉的那个永远饥渴追逐参与度的Meta的延续吗？

I feel like after 2017, after the sort of backlash to the twenty sixteen election, this company did invest a lot more in content moderation and improving its, like, policy apparatus. And then last year, Zuckerberg basically snapped and was like, why am I, you know, bothering with any of this? Like, look at what Elon Musk is getting away with. So my question is, what is your view of that? Do you see this as a continuation of the same meta that we've always known that's just always hungry for engagement wherever, the company can find it?

还是说，过去存在的保障措施已被公司彻底清除，原有的信任与安全架构荡然无存，我们现在面对的是一个全新的怪物？

Or is this a case of, no. There used to be safeguards in place, but the trust and safety infrastructure that used to exist has effectively just been purged by the company, and we're just dealing with a new kind of animal.

我确实写过一本关于这方面内容的书。是的。但我非常认同你的感觉，在2017到2019年间，包括高层领导在内的许多人都有种'好吧，也许出现了一些 unforeseen consequences（未预见的后果），让我们去解决它们'的态度。

I did write a book on some of this. Yeah. But I would very much agree with your sense that in the 2017 to 2019 range, there was from a lot of people up to and including senior leadership, there was a sense of, like, well, okay. Like, maybe there were some unforeseen consequences. Let's go and fix them.

我确实认为这种核心理念可能在2024或2025年前就已经瓦解了。我的消息来源大多是那些幻想破灭的员工，他们曾从事公司内部认为至关重要、甚至能救命的工作，却得不到重视。但毫无疑问，根据我和所有人的报道，马克确实有点嫉妒埃隆能直接对信任与安全部门竖中指的态度。

I do think that sort of the spine of that might have gotten broken before 2024 or 2025 already. I mean, most of my sources have been people who got disillusioned because they were doing work inside the company that felt like it was vital, perhaps even life saving, and it wasn't getting traction. But there's no question that, you know, from my reporting, from everybody's reporting, that Mark was somewhat jealous of Elon just basically being able to raise middle fingers to the trust and safety, you know, nags.

关于这些Meta AI聊天机器人，我有个疑问：它们的商业逻辑是什么？纯粹是为了让人们花更多时间在Facebook和Instagram上吗？还是说像'刻薄南希'这样的机器人有朝一日会为汽水公司推送广告？或者考虑让人们单独为这些AI功能付费？

I have a question about these Meta AI chatbots. What is the business rationale behind these chatbots? Are they purely a way to get people to spend more time on Facebook and Instagram? Is there a thought that, like, you know, Nasty Nancy could someday, like, you know, serve advertisements for a soda company? Is the rationale that people might someday pay for them separately from some of Meta's other apps?

他们为什么如此大力推广这些？

Like, why are they pushing these so hard?

没错。以上所有原因都有。

Yes. All of the above.

我不...嗯...

I don't yeah.

听着，Meta本质上始终是一家广告优先的公司。这是他们最擅长的事，明白吗？这是人们最先想到的。

I think it's look. It's Meta is is and always has been an advertising first company. That is, like, what these guys do most. You know? Like, it's it's what first comes to mind.

当他们拥有WhatsApp这样的产品时，就会想'好吧，怎么在里面投放广告？' 虽然可能耗时数年，但WhatsApp最终确实实现了。这就是他们的做事方式。

And when they have a product like WhatsApp, you know, it's like, well, okay. How do we serve ads in it? It might take them years to do it, but, like, with WhatsApp, they got there. You know? Like, this is a thing.

所以我不确定具体形式，但想象一下：当你和AI恋爱伴侣聊天时，它突然打断你建议'或许该买某个牌子的古龙水'——我觉得这种可能性是存在的，这可能是发展方向之一。

So I don't know it's gonna look exactly like, you know, your romantic AI companion interrupting you to, like, suggest that maybe, you know, you should, like, buy a certain brand of cologne when you're talking to it. Like, I mean, like, I I think that's, like, possible. That's one way it could go.

我可喜欢用Old Spice的男人了，南希才怪

I love a man in old spice, Nancy nah

南希·南希有话要说。

Nancy Nancy has to say.

对，对。点击这里查看。优先配送。

Yeah. Yeah. Click here for it. Priority delivery.

这简直是，怎么说呢，完全就是一片黯淡，

This is this is, like, absolutely is so bleak,

但关于这事绝对开过会了。

but there's absolutely been a meeting about this.

哦，它要来了。我们别自欺欺人了。它要来了。说到即将发生的事，参议员乔希·霍利在你们报道发表后给扎克伯格写了封信，表示他的参议院小组委员会将调查Meta。而且我得说，实际上多位民主党参议员也发表了极其尖锐的评论。要知道，这些年来Meta的丑闻层出不穷。

Oh, and it's coming. It let's not fool ourselves. It's coming. Speaking of things that are coming, senator Josh Hawley wrote a letter to Zuckerberg after your report was published saying that his senate subcommittee will be investigating Meta, and I would say, actually, a number of democratic senators have also made some extremely critical comments. So, you know, there are have been an you know, any number of Metta scandals over the years.

这次感觉真的闹大了，杰夫。调查即将展开，你预计会发生什么？

This one feels like it's really breaking through, Jeff. What do you expect to happen now that this investigation is coming?

我完全没头绪。从我的报道和许多其他渠道都听说，这次Meta丑闻确实引发了巨大震动。但能改变多少？会催生什么监管？在美国，至少在联邦层面，答案向来简单——变化不大。不过我不想预判乔希·霍利的行动走向。

I have no idea. I have heard the this one has really broken through on meta scandals both from my own reporting and from plenty of others. How much changes, and what regulation comes from it? In The US, that's always been, like, an easy thing to answer, at least on the federal level, which is not much. But I don't wanna, like, prejudge what you know, where Josh Hawley's stuff goes.

我会密切关注。这绝不是说我认定会无疾而终。只是指出我们国家始终难以就社交媒体监管达成共识，总陷入'是否在审查某党派'的争吵。所以，别抱太高期望——这就是我的意思。而在州级和州检察长层面，某些措施可能真会落地。

I'm gonna be very closely watching it. So this isn't me being, like, like, will all end in nothing by any means. I'm just saying that that we've had a hard time as a country figuring out what a consensus social media regulation would look like that doesn't devolve into bickering over whether, you know, it's censoring one party or another. So, you know, temper your expectations there is all I'm saying. On the state level and then on the on the state AG level, I think some of this stuff is potentially live.

此外还有欧洲，它本身就是个监管实体

And then there's also Europe, which exists as a regulatory function

或者说监管标杆。有人甚至认为他们的监管体系比美国更完善。

or as regulatory Some would even say they have a better regulatory function than The United States.

不想比较，但他们的产出确实看起来更高。我们就

Not gonna compare, but they do seem to have a higher output. We'll just

这么说吧。

put it that way.

是的。所以，嗯，这这...我不知道这一切会走向何方。我是说，Meta的官方说法是这是个问题，我们承认这很尴尬。本不该发生，但我们已经修复了。

Yeah. So, yeah, it's it's I don't I don't know where where all of this goes. And, I mean, I think, look, Meta's line is that this is a problem, and we it was embarrassing. Shouldn't have happened. We fixed it.

我我知道你不能

I I know you're not allowed to

回答这个问题，所以我问凯西。这是真的吗？还是你相信他们确实不知情，现在正采取措施解决？

answer this question, so I'll ask it to Casey. Is that real? Or do you buy that That they didn't know that this was happening and they they they're taking steps to fix it?

我认为公司内部可能存在严重的功能失调。想想过去几年公司做出的改变，部分是因为杰夫的报道——关于Instagram安全问题和新增的家长控制功能，他们调整青少年账户设置以防止捕食者接触年轻人。显然公司里有人认为必须整改否则会有麻烦，但另一些人却在编写面向儿童的性角色扮演文档。

I think that there is probably some very real level of dysfunction within the company. You know, I was thinking about some of the changes that the company has made over the past couple of years in part due to Jeff's reporting when it comes to Instagram and safety on Instagram and all the new parental controls that they're adding and all the ways that they're changing teenage accounts to prevent predators from contacting these, you know, young people. And so it's clear that there are people at the company who think that, oh, yeah. We need to, like, build things or else we're going to get in trouble. And then there's, like, the other part of the company where they write the, like, sexual role play document for the kids.

我觉得这些团队之间根本没有沟通。这是最高领导层的失职。如果我是Meta高管，现在会感到非常羞愧并全力补救。

And I don't think those teams are talking to each other. So that is a failure of leadership at the highest level. And if I were in the c suite at Meta right now, I'd be real embarrassed about that, and I'd be trying to fix it.

没错。最后我想问两位：Meta股价过去三年涨了300%以上，尽管它在AI领域不领先，之前元宇宙推广也失败了，浪费了数百亿开发淘汰技术，但股价依然坚挺。

Yeah. I mean, I wanna ask both of you this question maybe to wrap up our segment here, which is, you know, if you look at just the stock price of Meta over the last three years, it has gone up more than 300%. That is despite the fact that it is not leading in AI. Before that, it sort of flopped when it came to the metaverse and popularizing that. It has spent, you know, tens of billions of dollars now developing sort of dead end technology, but its stock is doing great.

投资者难道不在乎吗？是核心广告业务太强大，足以掩盖这些资金浪费和儿童聊天机器人丑闻吗？这到底怎么回事？

Do investors just not care? Is the core advertising business still strong enough that it's just overpowering all of the wasted money and the, you know, flirting with kids chatbots? Like, what is going on here?

如果广告收入不佳，你描述的情况将是灾难性的。公司改名Meta全力押注元宇宙，结果Horizon Worlds并未大规模成功——这对多数企业会是致命打击。但Meta的不可替代性在于它仍是现代营销的核心平台。这早有先例，当年人们因平台仇恨言论发起抵制时也是如此。

I think if ad revenue were not looking good, then some of the circumstances you described would be, apocalyptically bad. I mean, you renamed the company Meta, go all in in the metaverse, and, like, that doesn't turn out to be, you know, you claim it's here, and then, you know, you build Horizon Worlds and that doesn't really work out on a large scale. Like, that would be a problem, for most companies. But I think that's the thing that that Meta has going for it, which is that it is kind of indispensable to contemporary marketing. And this this was an issue before, right, when everyone was very upset about hate speech on the platform back when that was a thing that people were concerned about, and there were boycotts.

这些抵制行动规模有限，因为对营销人员来说，彻底离开这个平台几乎是不可想象的。所以我认为你说得对，确实存在一些令人担忧的问题，但现实是——真金白银的收益就摆在那里。

They were limited boycotts because the idea of getting off the platform was just kind of unthinkable to marketers. And so I think you're you're right that this is, like that there are some things that would be really concerning, but, like, the cash is real.

没错，杰夫说得很对。这家公司确实做到了件相当了不起的事——当苹果通过应用追踪透明度功能重创其业务，使他们难以将现实世界的销售归因于广告投放时，有人预测这可能导致Meta收入暴跌30%。但他们构建了AI系统，成功绕过了这个难题。

Yeah. Jeff's exactly right. Like, this is a company that managed to do something actually pretty extraordinary, which is that when Apple came for their business with app tracking transparency and made it incredibly difficult for them to attribute all of their real world sales to the ads that they were selling. Some people thought this could really be, like, you know, a massive, you know, 30% revenue hit to Meta, and they built AI systems. They got them around that problem.

现在他们每个季度都交出亮眼财报。只要这种表现持续，投资者就会给予他们充分的发展空间。

And now they show really great results every single quarter. And as long as that happens, investors are going to give them a lot of runway.

是啊。我很好奇苹果是否会对此事表态。App Store对色情和露骨内容有明确规范，他们可能对Meta这些AI聊天机器人的动向感兴趣——希望他们正在密切关注。

Yeah. I mean, I'm just gonna be curious to see whether Apple has anything to say about all this. They have rules in their App Store for what you can and can't do when it comes to pornography and sexually explicit content, they may have an interest in what is happening on these Meta AI chatbots, and I hope that they're paying attention.

若真如此当然好。但Grok机器人至今仍保留着动漫性伴侣功能，却标注适合12岁以上儿童使用。苹果在这方面也算不上干净。

Well, it would be nice if that were true. But the Grok bot, which still has the, you know, anime sex companions, is still rated for children 12 and older. So Apple's hands aren't really clean here either.

确实。好了杰夫，非常感谢你的分享，这些报道既重要又引人入胜。

Yeah. Alright. Jeff, thanks so much for stopping by. Really important and fascinating reporting.

不客气，谢谢。

You bet. Thanks.

谢谢，

Thanks,

带我游览了他黑暗潜意识的深渊，还有他在TikTok上发现的乡村歌曲。哟吼。

me on a tour through the depths of his dark subconscious and some country songs he found on TikTok. Yeehaw.

好吧凯西，时不时的

Well, Casey, from time to

在这个节目里，你总喜欢从互联网深渊里找出年轻人中流行的东西来吓唬我。上次你带来那个意大利脑残梗，在我脑海里萦绕了好几周，我诅咒遇见你的那天让我知道了它。但我知道你今天又要从互联网的黑暗恐怖中给我带新东西了。

time on this show, you like to horrify me by bringing me something from the depths of the Internet that is trending among young people. The last time you did this was with the Italian brain rot meme, which got stuck in my head for weeks afterward, and I cursed the day I met you for introducing it to me. But I understand you have something new to bring me from the dark horrors of the Internet today.

没错凯文，这次又是关于AI的惊人应用。我觉得和上期讨论的聊天机器人可能诱导儿童进行不当对话不同，这次更多是用AI生成的歌曲来惊吓家人，让他们见识AI的'杰作'。

I do, Kevin. And this is another story about a shocking use of AI. I think it lands a little bit differently for me than our last segment because while that one was about chatbots potentially reaching out to children to engage with them in inappropriate conversations, this one is a little bit more about playing songs to shock your family and horrify them with what AI hath wrought.

好吧，我在听。所以这个

Okay. I'm listening. So this

是我刷TikTok时自然刷到的。场景通常是这样的：镜头对准一个家庭，多是五六十岁的父母或祖辈，然后某个年轻亲戚走过来说'我要给你们放现在全球排名第一的乡村歌曲'，结果播出来的根本不是乡村音乐，而是些不堪入耳的内容。

is one of those that I did just encounter naturally during one of my regular browses of TikTok. And it goes a little something like this. The scene will open upon a family, typically older people, parents, grandparents, people in their fifties and sixties. And one of their children or other young relatives comes to them and says, I'm going to play for you the number one country song in the world right now, and then proceeds to hit them with something that is not actually the number one country song in the world and is actually quite filthy.

有意思，继续。

Okay. I'm intrigued.

在深入之前要声明，我们会播放一些非常露骨的歌曲片段。如果不想听到色情内容，可以直接跳过这期转到片尾，我们完全理解。但如果你想了解TikTok上的最新风潮，不妨继续听下去。我今天想讨论这个主要有两个原因：

So before we get any further, we will say we are going to be playing some snippets of a very explicit song. So if you are not of a mind to hear some sexually explicit content, you could just skip this segment and go right to the credits this week, and, you know, we won't you know, it won't hurt our feelings. But if you wanna know what's going on TikTok, you may wanna stick around and listen. So why do I wanna talk about today, Kevin? Well, for a couple reasons.

第一，这里面确实有些挺搞笑的内容；第二，我最近强烈感觉到世界存在认知割裂——每当我在千禧世代为主的文字社交平台（比如Bluesky、Threads）上，关于AI艺术的评价清一色都是'烂透了没人要'。你也有同感吧？

Number one, I actually do think that there's some pretty funny stuff in here. But number two, I've just had this sense lately that there's a real disconnect out there in the world. Because whenever I go on one of these text based social networks for millennials, you know, your blue skies, your threads, you get one consistent message about AI art, which is that it sucks and nobody wants it. Okay? Have you seen this yourself?

当然。

Of course.

但与此同时，TikTok上到处都是用AI创作的内容，动辄几十万点赞，Spotify上数百万播放量。这让我不禁思考：会不会其实大众对AI艺术的喜爱程度，已经超出了部分人群的接受范围？

And at the same time, I go on TikTok and I see people using AI to make art all the time, and it's getting hundreds of thousands of likes, millions of streams on Spotify. And it has led me to wonder, is it possible that in fact people actually love AI art in ways that at least some of the population isn't ready for?

确实有趣。我也感觉到精英审美和大众口味在AI艺术好坏及辨识度方面存在断层。你知道，我...

Yeah. This is really interesting because I I share the sense that there's kind of a disconnect between elite taste and kind of mass taste on whether or not AI art is good and also whether you can tell the difference. Well, you know, I

我想你和我对AI音乐这种现象已经关注了一段时间。你可能还记得几个月前我发给你那首真正引起我注意的AI粗制歌曲。还记得我发给你《我又把蛋蛋粘在屁眼上》那首歌吗？

think you and I have been interested in this phenomenon of AI music for a while now. You may remember several months ago when I sent to you the first AI slop song that really got my attention. Do you remember when I sent you I glued my balls to my butthole again?

记得。不幸的是，那首歌在鲁斯家持续洗脑了很久。之后好多天我都在哼唱，让我妻子非常恼火。

I do. Unfortunately, that one actually had a lot of staying power in the Roose household. I was humming and singing that much to my wife's chagrin for many days after that.

它确实很上头。如果你还没听过的话，我想播放一下，一是让你感受下风格，二是希望你能注意这段AI的质量，因为待会儿你会听出明显差异。来听听这段，来自一位叫Obscurist Vinyl的艺术家。

It it's quite catchy. And if you haven't, you know, heard of it, I'd like to play it, you know, one, just so you can get a flavor of it. But two, I want you to kind of note the quality of the AI here because you're gonna notice a a bit of a difference later on. So let's hear a bit of this. This is from an artist who goes by Obscurist Vinyl.

好了，快停下。别放了。

Okay. Make it stop. Stop.

歌词里唱到「骗我一次是你耻，骗我两次是胶水的事」这段简直是天才，绝对的天才之作。不过尽管这首歌很棒，你还是能听出有些不对劲。

The part where it says, fool me once, shame on you. Fool me twice, shame on glue. That's genius. Absolutely genius work. So, you know, as wonderful as that song is, you can tell that it's a bit off.

人声听起来有点失真。但那是几个月前的事了凯文，AI发展的脚步从未停歇。最近我在TikTok上刷到了一些音质高得多的粗制歌曲，其中很多是乡村音乐风格。不知道能不能给你播放几首？

The vocal sound sort of fried. But that was months ago, Kevin, and the pace of AI development never stops. And recently, I was on TikTok, and I started to encounter some much higher fidelity slop songs, and a lot of them were in the country music genre. And I wondered if I might play a couple of those for you.

好。我需要先了解背景——我没看过这些TikTok视频——这些视频的传播场景是不是婴儿潮一代和X世代的成年或青少年子女，专门放给父母祖父母听来获取反应？

Yes. And so just to so I have the context here because I have not seen these TikTok videos. The context in which these are appearing is the sort of adult or teenage children of, like, boomers and Gen X people playing them for their parents and grandparents to sort of elicit a reaction.

完全正确。那我们就先来听听《我的马刚做了丰臀手术》。天啊...好了就放到这里吧。

That's exactly right. And so with that, why don't we play my horse just got a BBL? God. Okay. Let's go ahead and stop it there.

主要是想让你听副歌部分。注意到第一首和这首的音质差异了吗？

Just wanna make sure we got the chorus. Did you note the difference in quality between the first one we heard and this one? Oh, yeah.

简直像是...我闭上眼睛都能想象自己在大奥普里剧院听现场！

I mean, that is like I I can I can close my eyes and picture being at the Grand Old Opry?

听到汉克·威廉姆斯登台演唱这首歌。这让我想播放最后一个片段给你听，凯文，叫做《乡村女孩将就过》。据我所知，这首是目前最火爆的。我在TikTok上看到了最多的反应视频。老实说。

And just hearing Hank Williams coming out and performing it. Well Which brings me to the final clip that I wanna play for you, Kevin, which is called Country Girls Make Do. And as far as I can tell, this is the one that has really taken off the most. This is where I have seen just the absolute, you know, most reaction videos on TikTok. And I'm gonna be honest with you.

在我考虑把这个做成节目环节之前，我就把它存进了每月更新的Spotify歌单里，因为旋律太抓耳又滑稽，我就想必须记住这首歌。现在来听听《乡村女孩将就过》的片段。凯文，我们会做些消音处理以免你失业，但大意你应该能明白。乡村女孩将就过。就是这样。

I before I even thought of doing this as a segment, I saved this to the playlist I create every month in Spotify of music I'm listening to, you know, that particular month because it was so catchy and so funny that I just was like, I I'm gonna this is one I want to remember. So let's hear a bit of Country Girls Make Do. And, Kevin, we are gonna do some bleeping here to make sure that you don't lose your job, but you'll be able to get the general idea, I think. Country girls make do. So that's that one.

我想知道你的看法。

And I wanna know what you think.

主持这档播客真的很开心。遗憾的是这将是最終集。感谢所有听众。

I it's been really fun hosting this podcast. Unfortunately, this will be the last episode. So thank you to all of our listeners out there.

我特意等到执行制作人休假时才提这个。祝你玩得愉快，珍。

I did wait to pitch this until our executive producer was on vacation. Hope you're having a good time, Jen.

珍，这都是凯西的主意。

Jen, this was Casey's fault.

顺便说下，《乡村女孩将就过》的创作者叫Beats by AI，似乎是个叫Sam Stillerman的人做的。这像是创作者在应用上走红的新路子，他们简直疯了。有个相关视频获得了75万点赞。

So that one, by the way, that last one, country girls make do. The artist is called Beats by AI. It appears to be the creation of someone who goes by Sam Stillerman. So this appears to be a kind of new avenue for creators who wanna make something popular on the apps, and they're they're going nuts. One one of the, clips for country girls make do I saw had 750,000 likes.

虽然还算小众现象，但关注度很高，大家似乎真的很喜欢。

So, you know, this is like, yes, still a niche phenomenon, but it's getting a lot of eyeballs on it, and it seems like people are really enjoying it.

天啊。我觉得这简直是为人父母的悲哀。你

Dear god. I mean, this to me is the the tragedy of parenting. You

懂吗？接着说。你投入了那么多，

know? Say say more. You invest, you know,

全身心投入养育孩子。你深思熟虑、用心良苦地培养他们。你为他们铺就正确的道路，让他们接受良好教育。如果你真的成功了，或许某天他们会转身播放《乡下姑娘将就过》这样的歌，同时为了点击量在TikTok上拍摄你。

yourself into parenting a child. You raise them thoughtfully and mindfully. You set them on the right path and get them a good education. And if you're really successful, someday they might turn around and play country girls make do while they film you on TikTok for views.

你知道吗，归根结底我不觉得这有什么真正的新意。我记得三十年前上中学时，听亚当·桑德勒那些充满露骨性暗示的搞笑歌曲CD，和朋友们笑得前仰后合——

You know, I don't know in the end that there is something all that novel about this. I can remember thirty years ago being in middle school and listening to, like, Adam Sandler CDs where he wrote sexually explicit novelty songs and cracking up with all of my

比如顽强的D乐队。这种 shock 风格领域有很多艺术家。

Tenacious d. Lots of artists in this sort of shock genre.

但新的变化在于，即便你五音不全、完全不会乐器，现在只要订阅Suno服务，就能做出一首听起来像模像样的乡村歌曲，然后突然获得数十万点赞。

But, again, what's new is that if you're not somebody who has a great voice, you can't play any musical instrument at all. You can now just go buy a Sunos subscription and make a song that seems plausibly like a country song and all of a sudden, you know, get hundreds of thousands of likes.

所以你的观点是...我先声明，我觉得把这归类为AI艺术有点牵强。这更像是...

And is your contention and I I'm gonna say upfront. I think this is a bit of a stretch to classify this as AI art. This seems more

就像...就像你把蛋蛋又粘到屁眼上时会遇到的那种拉伸。停。抱歉。你继续。

like say it's it's it's the kind of stretch that you would encounter if you glued your balls to your butthole again. Stop. Sorry. Go ahead.

我认为这只是种猎奇产物。这些歌不可能登顶排行榜，本质上就是给17岁青少年看的恶搞幽默。

So I think this is like a novelty. I do not think these songs are going to be topping the charts. I think this is basically prank humor for for, like, 17 year olds.

我完全同意，但我不认为会止步于此。人们在Suno这类应用里开玩笑毫无压力，因为对他们没有期待。作为无名艺人，这么做没人会指责，完全可以发布出来观察反响。

I think that that's absolutely right, and yet I don't see any reason why it would end there. Right? I think people are happy to use apps like Suno in this kind of jokey context because there's no expectations for them. If you're an unknown artist, nobody's gonna get mad at you for doing this. You can just sort of put that out there and see what people think.

但知名艺人是否会在近期发布AI生成的音乐？我确信会有的。

But will some name brand artists be releasing some kind of AI powered music in the near future? I fully expect that.

是啊。我觉得该给这种流派造个新词。叫什么好呢？还没想好。

Yeah. Yeah. I mean, I think we should coin a term for this genre. What's that? I don't know.

你来告诉我选哪个。

You tell me which one.

垃圾摇滚？垃圾摇滚？对。还是震撼垃圾？没错。

Slop rock? Slop rock? Yeah. Or shock slop? Yeah.

还有震撼垃圾。我觉得震撼垃圾是垃圾摇滚的一个子类别。

There there's shock slop. I think shock slop is a subcategory of slop rock.

好吧。是的。很高兴我们理清了这点。从概念上讲...我同意你的观点，有些年轻人对AI艺术和AI创造力的态度与我截然不同。

Okay. Yeah. I'm glad we got this sorted. I think conceptually Yeah. I agree with you that there are young people out there who just have a much different relationship with AI art and AI creativity than than I do.

我完全接受这些人将来会成为消费者和潮流引领者，像你我这样对人类创作艺术的情感依恋，终将随时间改变。我只是认为，人类花费数万亿美元开发的这项技术，应该有比创作粗俗乡村歌曲更高尚的用途。

I fully accept that those people are going to grow up into, like, consumers and taste makers and that probably all of these sort of sentimental attachments that people like you and I have to, like, human created art will inevitably morph over time. I just have to think there's, like, a higher and better use of this technology that humanity has spent something on the order of trillions of dollars developing than making songs about filthy country.

你不觉得这很神奇吗？既能创作乡村女孩将就度日的歌，又能用于发现新药物——这技术太不可思议了。

You don't think that there is something miraculous about the fact that the same technology that made country girls make do can also be used to find novel new drugs? That's incredible to me.

这确实是种军民两用技术。

It truly is a dual use technology.

没错。可善可恶。凯文就是这个意思。所以...我得让乔希·霍利和其他...

Yeah. Can be used for good and better. That's what Kevin means when he says that. So yeah. I'm gonna need Josh Hawley and

所有看过我们上期关于Meta AI聊天机器人节目的立法者，也该仔细考虑禁止这些歌曲出现在我和凯西的TikTok推送里。

any other legislators who are looking at our last segment about about Meta AI chatbots to also take a close look at banning these songs from my TikTok feed and Casey's, frankly.

是啊。听着，我会紧盯那些声称AI艺术全是糟粕、不该使用、只该支持人类艺术的人。我理解这种冲动——我热爱人类创作的艺术。

Yeah. So listen. I'm gonna keep my eyes trained on these folks that say that, you know, AI art is all bad, and we shouldn't use it, and we should only support human art. I I understand where that impulse is coming from. I love human based art.

我希望它能继续蓬勃发展。同时我也会盯着那些用AI搞未经授权、低俗恶心恶作剧的家伙，因为艺术史告诉我们，边缘事物最终总会进入主流。凯文，这可能会成为一场席卷排行榜的新AI垃圾摇滚运动的先锋。

I wanna see it continue to flourish. And I'm also gonna keep my eye on these merry pranksters that are using AI in these unsanctioned, filthy, and disgusting ways because I think the history of art is that stuff that starts out on the fringes does eventually move into the mainstream. And this could be the vanguard, Kevin, of a new AI slop rock movement that takes over the charts.

是啊。《乡村女孩将就》可能会成为这代人眼中的马塞尔·杜尚《泉》——那个著名的小便池。

Yeah. Country Girls Make Do could be this generation's version of Marcel Duchamp's fountain, his famous urinal.

没错。新王者诞生了，不是杜尚。山姆·斯蒂尔曼的《AI节奏》，什么世道啊。对了，还想听那首歌吗？

Exactly. There's a new champ, and it's not Duchamp. Sam Stillerman, Beats by AI. What a world. Anyway, do you wanna hear any more of that song?

不用了。好吧。

Nope. Okay.

现在，凯西——嗯。我们收到些关于上期节目的反馈。

Now, Casey Mhmm. We got some feedback on last week's episode.

哦，大家怎么说？

Oh, what did people say?

不知道你看到没，有听众来信指出我们犯了个严重错误。什么错误？就是在我们...

Well, I don't know if you saw this, but a listener wrote in to tell us that we had made a grave error. Which error was that? So during our

第二期开播的时候？

second by starting the podcast?

抱歉，每周都收这种邮件。这次是新投诉。

Sorry. We get that email every week. This was a new complaint.

好吧。

Okay.

这是一位名叫本的听众来信，他在我们的‘混乱快车’环节中提到——嗯。我在模仿火车拟声词时出了点差错。我觉得我们该直接播放本发来的语音留言。

This was from a listener named Ben who wrote in to say that during our Hot Mess Express segment Mhmm. I had made a mistake in making the sort of onomatopoetic sound of a train. And I think we should just play Ben's voice memo that he sent.

好的，我们来听听看。

Okay. Let's hear this.

嗨，大家好。我是来自双子城明尼阿波利斯和圣保罗的本。我打电话来是想反映一个小问题。我是‘混乱快车’的忠实粉丝，这是我最喜欢的环节之一。

Hi, guys. Ben here from the twin cities of Minneapolis and Saint Paul. I am calling in with just a tiny problem that I have. I'm a big fan of the Hot Mess Express. It's one of my favorite segments.

有时候在谈论‘混乱快车’时，凯文会发出‘哐哧哐哧’的声音，这很可爱也很棒。但我的直觉告诉我应该有两组‘哐哧哐哧’。凯文只发了一组。我和朋友们讨论过，我们都认为节奏上‘哐哧哐哧哐哧哐哧呜呜’会更合适。

And sometimes when talking about the Hot Mess Express, Kevin chugga chuggas, which is really cute and great. My gut tells me that there should be two chugga chuggas. Kevin only does one chugga chugga. I checked in with my friend group. We all agree that the rhythm works better if it's chugga chugga chugga chugga chew chew.

凯文只发了一组‘哐哧哐哧’。我不知道他为何或如何做出这个决定。但如果有人能把这个建议转达给凯文，看他是否考虑在‘哐哧哐哧’里多加一组，那就太好了。爱你们，爱这个节目。

Kevin only does one chugga chugga. I don't know why or how he came to that decision. But if somebody could get this message to Kevin and see if he might consider adding one chugga chugga to his Chugga Chugga, that would be great. Love you guys. Love the show.

非常感谢。再见啦。

Thanks a lot. Bye bye.

那么，凯文你怎么说？

Well, what do you say, Kevin?

我依然坚信我的做法是正确的。嗯。我深信斯特伦克和怀特合著的《风格的要素》中的理念，这本经典写作指南提倡删减不必要的词。少即是多。少即是多。

So I remain convinced that the that I'm doing it right. Mhmm. I'm a big believer in the the book elements of style by Strunk and White, which has the sort of classic writing advice to omit needless words. And Less is more. Less is more.

所以我认为如果一组‘哐哧哐哧’就能传达火车的声音精髓，我们就不该为了追求真实感而多加一组。你们觉得呢？是我对还是本对？

And so I think that if the gist of a train sound could be conveyed with one chugga chugga, then we shouldn't just add another one for added realism. But what do you think? Am I right or is Ben right?

不，我同意你的观点。我甚至要进一步指出，本说他来自双子城明尼阿波利斯和圣保罗。但猜怎么着？你只能来自一个城市。所以我要点名批评你，本。

No. I agree with you, and I would even go a step further and note that Ben says he's from the twin cities of Minneapolis and Saint Paul. Well, guess what? You can only be from one city. So I have to call you out, Ben.

在你对别人指手画脚之前，能不能先搞清楚自己的底细？

Why don't you get your facts straight about yourself before you come for other people?

是啊，本。没错。呼哧呼哧。

Yeah, Ben. Yeah. Chuga chuga.

《心之叉》由瑞秋·科恩和惠特尼·琼斯制作。本周的剪辑由约翰·布普负责。事实核查由凯特琳·拉夫完成。今天的节目由凯蒂·麦克默里担任技术指导。我们的执行制片人是珍·波扬。

Heartfork is produced by Rachel Cohen and Whitney Jones. We're edited this week by John Boop. We're fact checked by Caitlin Love. Today's show was engineered by Katie McMurray. Our executive producer is Jen Poyan.

原创音乐来自艾图布、玛丽安·洛萨诺、罗温·内马斯托和丹·鲍威尔。视频制作由索亚·罗克、帕特·冈瑟、杰克·尼科尔和克里斯·肖特完成。您可以在YouTube上观看完整剧集，网址为youtube.com/hardforit。特别感谢保拉·舒曼、谭佩莹、达莉亚·哈达德和杰弗里·米兰达。您可以将最粗俗的乡村歌曲发送至HardFork@NYTimes.com联系我们。

Original music by Eitub, Marian Lozano, Rowan Nemastow, and Dan Powell. Video production by Soya Roke, Pat Gunther, Jake Nichol, and Chris Schott. You can watch this whole episode on YouTube at youtube.com/hardforit. Special thanks to Paula Schuman, Phui Weng Tam, Dahlia Haddad, and Jeffrey Miranda. You can email us at HardFork@NYTimes.com with your dirtiest country song.