AI 安全与加速主义之争：Dario Amodei 谈 Claude 3.7 与 AI 未来

我担心了大概十年，AI 可能成为专制统治的引擎。想想专制政府，他们能有多专制，通常取决于他们能让执法者——人类执法者——做什么。但如果执法者不再是人类，那就开始描绘一些非常黑暗的可能性。越来越多地，我们看到像 Meta 这样的公司展示肌肉，说：‘嘿，你要么接受，要么走人。’而这是一个真正的‘走人’时刻：我们要裁掉你们中的 5%，并给自己发奖金。你正在和造型师合作？是的，没错。这是他们的作品吗？不，我们下周才第一次见面。好吧。你会用 AI 吗？你让我思考，也许我可以带一堆不可能实现的设计来惹恼我的造型师。是的。这个周末我参加了两个 AI 活动。它们有点像 AI 光谱的两极。一个是有效利他主义者的年度大会，然后周五晚上我出去了——你会很为我骄傲——我待到很晚，凌晨 2 点才回家。我去了一场 AI 锐舞派对，它算是非正式地与马克·扎克伯格有关联。它叫‘扎克锐舞’。你说‘非正式关联’，马克·扎克伯格没有参与，我猜他都不知道这事。没错。更准确的说法是‘毫无关联’。他——这算是一场致敬扎克伯格的锐舞，由一群加速主义者举办，他们希望 AI 发展——另一个词是未经许可使用他的肖像。是的，但名人有时会遇到这种事。是的。在扎克锐舞上，我觉得没多少人在跳舞。有个舞池，但人很少。他们有个东西，上面有摄像头对着舞池，如果你站在合适的位置，它会把你的脸变成马克·扎克伯格的脸，显示在大屏幕上——这么说吧，你在嗑药时绝对不想遇到这种事，因为那可能非常让人不安。是的。还有一列室内玩具火车可以坐，速度还挺快。这场锐舞的目的是什么？嗑药。这就是目的。我是凯文·罗斯，《纽约时报》科技专栏作家。我是凯西·牛顿，来自 Platformer。这里是 Hard Fork。本周，Anthropic 的 CEO 达里奥·阿莫代伊回归节目，进行一次超长访谈，讨论新的 Claude、与中国的 AI 竞赛，以及他对 AI 未来的希望和担忧。然后我们以一轮 Hat GPT 大秀结束本周。凯文、凯西，你们注意到 AI 公司现在周末也发布东西了吗？是啊，五天工作制怎么了？是的，他们不尊重记者的工作时间。公司总是在周六周日和不同时区发布消息。这很麻烦，真的。但这个周日，我收到一条激动人心的消息，说 Anthropic 的 CEO 达里奥·阿莫代伊有一些新闻要谈，他想上 Hard Fork 来谈。是啊，差不多同时，我收到 Anthropic 的邮件，说可以预览他们的最新模型，所以我整个周末都在试用。是的，老听众会记得达里奥是节目的常客。2023 年，我们请他聊过他在 Anthropic 的工作、他对 AI 安全的愿景以及这一切的走向。我很高兴再次和他交谈，有几个原因。第一，我觉得他非常有趣且有思想。他思考 AI 的时间比几乎任何人都长。早在 2016 年，他就在写关于 AI 安全中潜在可怕问题的论文。他在 Google 工作过，在 OpenAI 工作过，现在是 Anthropic 的 CEO。所以他是 AI 领域的终极内部人士。凯文，我认为达里奥之所以重要还有另一个原因：在所有领导大型 AI 实验室的人中，他是公开对可能出错的事情最担忧的一个。他一直如此。然而，过去几个月，正如我们在节目中提到的，钟摆似乎已经从关心 AI 安全摆向了那种冲冲冲的加速主义，副总统 JD·万斯前几天在法国的演讲就体现了这一点。因此，我认为让他来谈谈很重要，也许能稍微把钟摆拉回来一点，提醒人们利害攸关的是什么。是啊，或者至少听听他对钟摆摆动的看法，以及他认为为什么未来可能会摆回来。所以今天，我们将和达里奥聊聊 Anthropic 刚发布的新模型 Claude 3.7 Sonnet。但我们还想进行更广泛的对话，因为 AI 领域现在发生了太多事。凯文，还有一件事要提，这次达里奥身上有一件事是上次他来节目时没有的：我的男朋友现在在他公司工作。是啊，凯西的男人在 Anthropic。我的 Anthropic 就是那个 Anthropic。关于这个，我在 Platformer 上有一份很长的披露声明。

I've always worried, you know, maybe for a decade, that AI could be an engine of autocracy. If you think about repressive governments, the limits to how repressive they can be are generally set by what they can get their enforcers, their human enforcers, to do. But if their enforcers are no longer human, that starts painting some very dark possibilities. More and more you see companies like Meta flexing their muscles and saying, 'Hey, you can either like it or you can take a hike.' And this was a true 'take a hike' moment: we're getting rid of 5% of you and we're giving ourselves a bonus for it. You are working with a stylist? I am, yes, that's right. Are you—is this their handiwork? No, we have our first meeting next week. Okay. And are you going to use AI now? You have me thinking, and maybe I could exasperate my stylist by bringing in a bunch of impossible-to-create designs. Yes. So I went to two AI events this weekend. It was sort of polar opposites of the kind of AI spectrum. There was the effective altruists' big annual conference, and then on Friday night I went out—you'd be very proud of me—I stayed out so late, I stayed out till 2 AM. I went to an AI rave that was sort of unofficially affiliated with Mark Zuckerberg. It was called the Zuck Rave. Now when you say 'unofficially affiliated,' Mark Zuckerberg had no involvement in this, and my assumption is he did not know it was happening. Correct. A better word for what his involvement is would be 'no involvement.' He was—it was sort of a tribute rave to Mark Zuckerberg thrown by a bunch of accelerationist people who want AI to go—another word for it would be using his likeness without permission. Yes, but that happens to famous people sometimes. Yes. So at the Zuck Rave, I would say there was not much raving going on. There was a dance floor, but it was very sparsely populated. They did have a thing there that had a camera pointing at the dance floor, and if you sort of stood in the right place, it would turn your face into Mark Zuckerberg's on a big screen, which let's just say is not something you want to happen to you while you're on mushrooms, because that could be a very destabilizing event. Yes. There was a train—an indoor toy train—that you could ride on. It was going actually quite fast. And what was the point of this rave? To do drugs. That was the point of this rave. I'm Kevin Rose, a tech columnist at the New York Times. I'm Casey Newton from Platformer, and this is Hard Fork. This week, Anthropic CEO Dario Amodei returns to the show for a supersized interview about the new Claude, the AI race against China, and his hopes and fears for the future of AI. Then we close it out with a round of Hat GPT Big Show this week. Kevin, Casey, have you noticed that the AI companies do stuff on the weekends now? Yeah, whatever happened to just five days a week? Yes, they are not respectful of reporters and their work hours. Companies are always announcing stuff on Saturdays and Sundays and in different time zones. It's a big pain, it really is. But this weekend, I got an exciting message on Sunday saying that Dario Amodei, the CEO of Anthropic, had some news to talk about and he wanted to come on Hard Fork to do it. Yeah, and around the same time, I got an email from Anthropic telling me I could preview their latest model, and so I spent the weekend actually trying it out. Yeah, so longtime listeners will remember that Dario is a repeat guest on this show. Back in 2023, we had him on to talk about his work at Anthropic and his vision of AI safety and where all of this was headed. And I was really excited to talk to him again for a few reasons. One, I just think he's a very interesting and thoughtful guy. He's been thinking about AI for longer than almost anyone. He was writing papers about potentially scary things in AI safety all the way back in 2016. He's been at Google, he's been at OpenAI, he's now the CEO of Anthropic. So he is really the ultimate insider when it comes to AI. And you know, Kevin, I think Dario is an important figure for another reason, which is that of all the folks leading the big AI labs, he is the one who seems the most publicly worried about the things that could go wrong. That's been the case with him for a long time. And yet over the past several months, as we've noted on the show, it feels like the pendulum has really swung away from caring about AI safety to just this sort of go-go-go accelerationism that was embodied by the speech that Vice President JD Vance gave in France the other day. And for that reason, I think it's important to bring him in here and maybe see if we can shift that pendulum back a little bit and remind folks of what's at stake here. Yeah, or at least get his take on the pendulum swinging and why he thinks it may swing back in the future. So today, we're going to talk to Dario about the new model that Anthropic just released, Claude 3.7 Sonnet. But we also want to have a broader conversation because there's just so much going on in AI right now. And Kevin, something else that we should note, something that is true of Dario this time that was not true the last time that he came on the show, is that my boyfriend now works at his company. Yeah, Casey's man is at Anthropic. My Anthropic is that Anthropic. And I have a whole sort of long disclosure about this that you can read at Platformer.

这周可能值得聊聊新闻/伦理，你知道，我们总是喜欢提醒大家这一点。好了，话不多说，有请 Dario Amodei。Dario Amodei，欢迎回到 Hard Fork。

News/ethics might be worth doing this week, you know, we always like reminding folks of that. All right, with that, let's bring in Dario Amodei. Dario Amodei, welcome back to Hard Fork.

谢谢你再次邀请我。

Thank you for having me again.

是的，回归冠军。那么跟我们聊聊 Claude 3.7 吧，聊聊这个新模型。

Yeah, returning champion. So tell us about Claude 3.7, tell us about this new model.

是的，我们开发这个模型已经有一段时间了。基本上，我们考虑了两件事。一是，当然，市面上已经有一些推理模型存在了几个月，我们想做一个自己的，但希望侧重点有所不同。具体来说，市场上很多其他推理模型主要是在数学和竞赛编程上训练的，这些是可衡量性能的客观任务。我不是说它们不令人印象深刻，但它们有时与现实世界或经济中的任务相关性较低。即使在编程领域，竞赛编程和实际开发也有很大区别。所以我们训练 Claude 3.7 更侧重于这些现实世界的任务。我们还觉得，目前大家提供的推理模型通常是一个普通模型加一个推理模型，这有点奇怪。这就像一个人有两个大脑，你问一个快速问题比如“你叫什么名字？”时跟大脑一号说话，而让我证明一个数学定理时跟大脑二号说话，因为我得坐下来想 20 分钟。

Yes, so we've been working on this model for a while. Basically, we had two things in mind. One was that, of course, there are these reasoning models out there that have been out there for a few months, and we wanted to make one of our own, but we wanted the focus to be a little bit different. In particular, a lot of the other reasoning models in the market are trained primarily on math and competition coding, which are objective tasks where you can measure performance. I'm not saying they're not impressive, but they're sometimes less relevant to tasks in the real world or the economy. Even within coding, there's really a difference between competition coding and doing something in the real world. And so we trained Claude 3.7 more to focus on these real world tasks. We also felt like it was a bit weird that in the reasoning models that folks have offered, it's generally been a regular model and then a reasoning model. This would be like if a human had two brains, and it's like you can talk to brain number one if you're asking me a quick question like 'What's your name?', and you're talking to brain number two if you're asking me to prove a mathematical theorem because I have to sit down for 20 minutes.

是的，这就像一个播客有两个主持人，一个喜欢瞎扯，另一个在说话前真的会思考。哦，得了吧，太狠了。不，不，不评论。对任何相关性都不评论。

Yeah, it'd be like a podcast where there are two hosts, one of whom just likes to yap and one of whom actually thinks before he talks. Oh, come on, brutal. No, no, no comments. No comment on any relevance.

那么用户开始使用 3.7 时，与之前的模型相比会注意到哪些不同？

So what differences will users of Claude notice when they start using 3.7 compared to previous models?

是的，有几个方面。总体上它会更好，包括在编程方面，Claude 模型在编程上一直是最好的，但 3.7 又上了一个台阶。除了模型本身的特性外，你可以把它置于这种扩展思考模式，你告诉它基本上是同一个模型，但只是要求它以能思考更长时间的方式运行。如果你是 API 用户，你甚至可以指定思考时间的上限。

Yes, so a few things. It's going to be better in general, including better at coding, which Claude models have always been the best at coding, but 3.7 took a further step up. In addition to just the properties of the model itself, you can put it in this extended thinking mode, where you tell it basically the same model but you're just saying operate in a way where you can think for longer. And if you're an API user, you can even say here's the boundary on how long you can think.

澄清一下，因为可能会有人困惑，你的意思是新的 Claude 是这种混合模型；它有时能推理，有时能快速回答，但如果你想让思考更长时间，那是一个单独的模式。思考和推理是某种分开的模式。

And just to clarify, because this may confuse some people, what you're saying is the new Claude is this hybrid model; it can sometimes do reasoning, sometimes do quicker answers, but if you want it to think for even longer, that is a separate mode. Thinking and reasoning are sort of separate modes.

是的，没错。所以基本上模型可以像平常一样直接回答，或者你可以给它一个指示，让它思考更长时间。进化的更远方向是模型自己决定合适的思考时间。对吧？人类就是这样，或者至少可以这样。对吧？如果我问你叫什么名字，你不会说“嗯，我应该想多久？给我 20 分钟来确定我的名字。”但如果我说“嘿，我想让你分析这只股票”或“我想让你证明这个数学定理”，有能力做这个任务的人不会试图马上给出答案。他们会说“好的，那需要一些时间”，然后他们会把任务写下来。

Yes, yeah. So basically the model can just answer as it normally would, or you can give it this indication that it should think for longer. An even further direction of the evolution would be the model decides for itself what the appropriate time to think is. Right? Humans are like that, or at least can be like that. Right? If I ask you your name, you're not like 'Huh, how long should I think? Give me 20 minutes to determine my name.' But if I say 'Hey, I'd like you to do an analysis of this stock' or 'I'd like you to prove this mathematical theorem,' people who are able to do that task are not going to try and give an answer right away. They're going to say 'Okay, that's going to take a while,' and they will need to write down the task.

这是我对当今语言模型和 AI 模型的主要不满之一。我有时用 ChatGPT，会忘记自己处于硬核推理模式，然后问一些愚蠢的问题，比如“怎么调整热水器的设置？”然后它就开始思考四分钟，我就想“我其实没想要一篇关于调节热水器温度的论文。”那么你认为还要多久模型才能自己进行这种路由，你问一个问题，它会说“这个问题似乎需要大约三分钟的思考过程”，而另一个问题可能只需要 30 秒？

This is one of my main beefs with today's language models and AI models in general. I'll be using something like ChatGPT and I'll forget that I'm in the hardcore reasoning mode, and I'll ask it some stupid question like 'How do I change the settings on my water heater?' and it'll go off and think for four minutes, and I'm like 'I didn't actually mean to be like a treatise on adjusting the temperature of the water heater.' So how long do you think it'll be before the models can actually do that kind of routing themselves, where you'll ask a question and it'll say 'It seems like you need about a three-minute long thinking process for this one' versus maybe a 30-second one for this other?

是的，所以我认为我们的模型是朝这个方向迈出的一步。在 API 中，如果你给它一个思考上限，比如“我要思考 20,000 个词”，大多数时候，当你给它最多 20,000 个词时，它不会用满 20,000 个词，有时它会给出一个非常简短的回复，因为当它知道进一步思考没有收益时，它就不会思考更长时间。但给它一个思考时间上限仍然很有价值。所以我们在这个方向上迈出了一大步，但还没有达到我们想要的程度。

Yeah, so I think our model is kind of a step towards this. In the API, if you give it a bound on thinking, you say 'I'm going to think for 20,000 words or something.' Most of the time, when you give it up to 20,000 words, it doesn't use 20,000 words, and sometimes it'll give a very short response, because when it knows that it doesn't get any gain out of thinking further, it doesn't think for longer. But it's still valuable to give a bound on how long it'll think. So we've kind of taken a big step in that direction, but we're not to where we want to be yet.

当你说它更擅长现实世界任务时，你想到的是哪些任务？

When you say it's better at real world tasks, what are some of the tasks that you're thinking of?

是的，所以我认为首先是编程。Claude 模型在现实世界的编程中一直非常出色。我们有很多客户，从 GitHub 到 Windsurf、Codium、Cognition、Vercel，我肯定还漏了一些。这些是 vibe coding 应用，或者就是编程应用。而且有很多不同类型的编程应用。我们还发布了一个叫 Claude Code 的东西，它更像一个命令行工具。但我也认为，在复杂指令遵循，或者“我想让你理解这份文档”或“我想让你使用这一系列工具”这类任务上，我们训练的推理模型 Claude 3.7 Sonnet 也更擅长。

Yeah, so I think above all coding. Claude models have been very good for real world coding. We have a number of customers, from GitHub to Windsurf, Codium, Cognition, Vercel, I'm sure I'm leaving some out. These are the vibe coding apps, or just the coding apps period. And there are many different kinds of coding apps. We also released this thing called Claude Code, which is more of a command line tool. But I think also on things like complex instruction following, or just 'Here, I want you to understand this document' or 'I want you to use this series of tools,' the reasoning model that we've trained, Claude 3.7 Sonnet, is better at those tasks too.

有一件事新的 Claude Sonnet 没有做，Dario，就是访问互联网。为什么不做，什么会让你改变这一点？

One thing the new Claude Sonnet is not doing, Dario, is accessing the internet. Why not, and what would cause you to change that?

是的，我想我以前公开说过，但网络搜索很快就会推出。我们很快就会提供网络搜索。我们认识到这是一个疏忽。我认为总的来说，我们更倾向于企业级而非消费者级，这更像是一个消费者功能，尽管它也可以用于企业。但我们两者都关注，这个功能即将到来。

Yes, so I think I'm on record saying this before, but web search is coming very soon. We will have web search very soon. We recognize that as an oversight. I think in general we tend to be more enterprise focused than consumer focused, and this is more of a consumer feature, although it can be used on both. But we focus on both, and this is coming.

明白了。那么你把这个模型命名为 3.7。之前的模型是 3.5，你去年悄悄更新了它，内部人士恭敬地称它为 3.6。这让我们所有人都抓狂。AI 模型命名到底是怎么回事？

Got it. So you've named this model 3.7. The previous model was 3.5, you quietly updated it last year, and insiders were calling that one 3.6 respectfully. This is driving all of us insane. What is going on with AI model names?

我们是最不疯狂的，尽管我承认我们也很疯狂。所以你看，我认为我们这里的错误相对可以理解。我们做了 3.5 Sonnet，当时做得不错。我们有 3.0，然后是 3.5 系列。我承认 3.7 这个新名字是个失误。实际上，在 API 中更改名字很难，尤其是当你有这么多合作伙伴和提供服务的界面时。

We are the least insane, although I recognize that we are insane. So look, I think our mistakes here are relatively understandable. We made a 3.5 Sonnet, we were doing well there. We had the 3.0 and then the 3.5s. I recognize the 3.7 new was a misstep. It actually turns out to be hard to change the name in the API, especially when there are all these partners and surfaces you offer.

我相信你能搞清楚。不，不，不，这比训练模型还难，我跟你说。

You can figure it out, I believe. So no, no, no, it's harder than training the model, I'm telling you.

所以我们有点追溯性地非正式地把上一个命名为 3.6，这样这个叫 3.7 就说得通了，对吧？我们保留 Claude 这个名字给 Sonnet，可能还有序列中的其他模型，用于那些真正重大的飞跃。顺便说一句，有时候这些模型正在到来。好的，明白了。什么时候来？嗯，所以我应该谈谈这个。到目前为止，我们发布的所有模型其实都不贵，对吧？你知道，我写过一篇博客，说它们最多在几千万美元的范围。还有更大的模型正在到来；它们需要很长时间，而且有时需要很长时间才能做好。但那些更大的模型，我是说，它们对其他人来说也在到来。我的意思是，竞争对手也传闻要推出，但你知道，我们距离发布一个更大的基础模型并不遥远。所以 Claude 3.7 Sonnet 以及 Claude 3.6 Sonnet 的大部分改进都在后训练阶段。但我们正在开发更强的基础模型，也许那会是 Claude 4 系列，也许不是。我们拭目以待。但我认为这些会在相对较短的时间单位内到来。时间单位很短。是的，我会把它记在日历上。提醒我几个时间单位后跟进一下，Kevin。

So we've kind of retroactively and informally named the last one 3.6 so that it makes sense that this one is 3.7, right? And we are reserving Claude for Sonnet and maybe some other models in the sequence for things that are really quite substantial leaps. Sometimes when those models are coming, by the way. Okay, got it. Coming when? Yeah, so I should talk a little bit about this. So all the models we've released so far are actually not that expensive, right? You know, I did this blog post where I said they're in the few tens of millions of dollars range at most. There are bigger models than they are coming; they take a long time, and sometimes they take a long time to get right. But those bigger models, I mean, they're coming for others. I mean, they're rumored to be coming from competitors as well, but you know, we are not too far away from releasing a model that's a bigger base model. So most of the improvements in Claude 3.7 Sonnet, as well as Claude 3.6 Sonnet, are in the post-training phase. But we are working on stronger base models, and perhaps that'll be the Claude 4 series, perhaps not. We'll see. But I think those are coming in a relatively small number of time units. Small number of time units. Yeah, I'll put that on my calendar. Remind me to check in on that in a few time units, Kevin.

我知道你们 Anthropic 非常关注 AI 安全以及你们推向世界的模型的安全性。我知道你们花了很多时间思考这个问题，并在内部对模型进行红队测试。Claude 3.7 Sonnet 是否有任何新的能力是危险的，或者可能会让关心 AI 安全的人担忧？

I know you all at Anthropic are very concerned about AI safety and the safety of the models that you're putting out into the world. I know you spend lots of time thinking about that and red teaming the models internally. Are there any new capabilities that Claude 3.7 Sonnet has that are dangerous or that might worry someone who is concerned about AI safety?

所以并非本身危险，我总想澄清这一点，因为我觉得人们总是把当前危险和未来危险混为一谈。并不是说没有当前危险，它们总是那种正常的技术风险、正常的技术政策问题。我更担心的是随着模型变得更强大我们将看到的危险。我认为那些危险，当我们在 2023 年交谈时，我谈了很多。我想我甚至在参议院作证，谈到了滥用风险，例如生物或化学战，或者 AI 自主性风险。我特别提到滥用风险，我说我不知道这些何时会出现，何时会成为真正的风险，但可能发生在 2025 或 2026 年。现在我们在 2025 年初，那个时期的开端，我认为模型开始接近那个点。所以特别是在 Claude 3.7 Sonnet 中，正如我们在模型卡中写的，我们总是做这些，你几乎可以称之为对照试验，我们让一些对某个领域（比如生物学）了解不够的人，然后我们基本上看模型在多大程度上帮助他们参与一些模拟的不良工作流程，对吧？我们会改变一些步骤，但模拟一些不良工作流程。在模型的辅助下，人类能做到多好？有时我们甚至在现实世界中进行湿实验室试验，他们模拟制造一些坏东西，与当前技术环境相比，对吧？他们用谷歌或教科书能做什么，或者只是无辅助能做什么。我们试图弄清楚：这是否启用了一个以前不存在的新威胁向量？我认为非常重要的一点是，这不是关于“哦，模型给了我这个东西的序列吗？它给了我一个制作冰毒的食谱吗？”那很容易，你用谷歌就能做到。我们根本不关心那个。我们关心的是那种深奥的、高度不常见的知识，比如只有病毒学博士才有的知识。模型在这方面有多大帮助？如果有帮助，那并不意味着我们明天都会死于瘟疫。它意味着世界上存在一种新风险，存在一种新威胁向量，就像你让制造核武器变得更容易，你发明了某种东西，所需钚的量比以前更少。所以我们针对这些风险测量了 Sonnet 3.7，模型在这方面越来越好。它们还没有达到我们认为威胁端到端有真正有意义增加的程度，对吧？要完成真正做危险事情所需的所有任务。然而，我们在模型卡中说，我们评估了下一个模型，或者未来 3 个月、6 个月内的模型，有相当大的概率可能达到那个点。然后我们的安全程序，我们的负责任扩展程序，主要关注这些非常大的风险，就会启动，我们会有额外的安全措施和额外的部署措施，专门针对这些非常狭窄的风险设计。

So not dangerous per se, and I always want to be clear about this because I feel like there's this constant conflation of present dangers with future dangers. It's not that there aren't present dangers, and they're always kind of normal tech risks, normal tech policy issues. I'm more worried about the dangers that we're going to see as models become more powerful. And I think those dangers, when we talked in 2023, I talked about them a lot. I think I even testified in front of the Senate for things like misuse risks, for example, biological or chemical warfare, or the AI autonomy risk. I said particularly with the misuse risk, I said I don't know when these are going to be here, when these are going to be real risks, but it might happen in 2025 or 2026. And now that we're in kind of early 2025, the very beginning of that period, I think the models are starting to get closer to that. So in particular, in Claude 3.7 Sonnet, as we wrote in the model card, we always do these, you could almost call them like trials with a control, where we have some human who doesn't know enough about some area, like biology, and we basically see how much does the model help them to engage in some mock bad workflow, right? We'll change a couple of the steps, but some mock bad workflow. How good is a human at that, assisted by the model? Sometimes we even do wet lab trials in the real world, where they mock make something bad, as compared to the current technological environment, right? What they could do with Google or with a textbook, or just what they could do unassisted. And we're trying to get at: does this enable some new threat vector that wasn't there before? I think it's very important to say this isn't about, 'Oh, did the model give me the sequence for this thing? Did it give me a cookbook for making meth or something?' That's easy, you can do that with Google. We don't care about that at all. We care about this kind of esoteric, high uncommon knowledge that say only a virology PhD or something has. How much does it help with that? And if it does, that doesn't mean we're all going to die of the plague tomorrow. It means that a new risk exists in the world, a new threat vector exists in the world, as if you just made it easier to build a nuclear weapon, you invented something that the amount of plutonium you needed was lower than it was before. And so we measured Sonnet 3.7 for these risks, and the models are getting better at this. They're not yet at the stage where we think that there is a real and meaningful increase in the threat end to end, right? To do all the tasks you need to do to really do something dangerous. However, we said in the model card that we assessed a substantial probability that the next model, or a model over the next, I don't know, 3 months, 6 months, a substantial probability that we could be there. And then our safety procedure, our responsible scaling procedure, which is focused mainly on these very large risks, would then kick in, and we'd have kind of additional security measures and additional deployment measures designed particularly against these very narrow risks.

是的，我的意思是，为了真正强调这一点，你是说在接下来的 3 到 6 个月内，我们将处于这些模型的中等风险状态，句号。大概如果你处于那种状态，你的很多竞争对手也会处于那种状态。这实际上意味着什么？比如，如果我们都将生活在中等风险中，世界需要做什么？

Yeah, I mean, just to really underline that, you're saying in the next 3 to 6 months we are going to be in a place of medium risk in these models, period. Presumably, if you were in that place, a lot of your competitors are also going to be in that place. What does that mean practically? Like, what does the world need to do if we're all going to be living in medium risk?

我认为至少在这个阶段，这不会带来巨大的变化。这意味着模型有能力做一系列狭窄的事情，如果不加以缓解，会在一定程度上增加真正危险或真正糟糕事情发生的风险。比如站在执法官员或 FBI 的角度看。有一个新的威胁向量，有一种新的攻击方式。这并不意味着世界末日，但确实意味着任何涉及存在这种风险的行业的人都应该特别针对该风险采取预防措施。明白了。我不知道，我的意思是我可能错了，可能需要更长时间。你无法预测会发生什么，但我认为，与我们今天看到的对风险担忧减少的环境相反，背景中的风险实际上一直在增加。

I think at least at this stage, it's not a huge change to things. It means that there's a narrow set of things that models are capable of, if not mitigated, that would somewhat increase the risk of something really dangerous or really bad happening. Like put yourself in the eyes of a law enforcement officer or the FBI or something. There's a new threat vector, there's a new kind of attack. It doesn't mean the end of the world, but it does mean that anyone involved in industries where this risk exists should take a precaution against that risk in particular. Got it. I don't know, I mean, I could be wrong, it could take much longer. You can't predict what's going to happen, but I think contrary to the environment that we're seeing today of worrying less about the risk, the risks in the background have actually been increasing.

我们还有更多安全问题，但我想再问两个关于创新竞争的问题。首先，是的，现在似乎无论任何给定公司的模型多么创新，这些创新都会在几个月甚至几周内被竞争对手复制。这会让你的工作更难吗？你认为这种情况会无限期持续下去吗？

We have a bunch more safety questions, but I want to ask two more about kind of innovation competition. First, yeah, right now it seems like no matter how innovative any given company's model is, those innovations are copied by rivals within months or even weeks. Does that make your job harder, and do you think it is going to be the case indefinitely?

我不知道创新是否……

I don't know that innovations are...

并非完全照搬。我想说的是，大量竞争对手的创新速度非常快。有四、五、六家公司创新速度很快，模型迭代也很快。但以 Sonnet 3.7 为例，我们做推理模型的方式与竞争对手不同。我们强调的重点也不同。甚至在那之前，Sonnet 3.5 擅长的东西就与其他模型不同。人们经常谈论竞争、商品化、成本下降，但实际情况是，模型之间实际上相对不同，这创造了差异化。

Necessarily copied exactly. What I would say is that the pace of innovation among a large number of competitors is very fast. There's four, five, maybe six companies who are innovating very quickly and producing models very quickly. But if you look, for example, at Sonnet 3.7, the way we did the reasoning models is different from what was done by competitors. The things we emphasized were different. Even before then, the things Sonnet 3.5 is good at are different than the things other models are good at. People often talk about competition, commoditization, cost going down, but the reality of it is that the models are actually relatively different from each other, and that creates differentiation.

是的，我们收到很多听众的问题，比如如果我要订阅一个 AI 工具，应该选哪个？这些是我用它做的事情。我很难回答，因为我发现对于大多数用例，所有模型在回答问题方面都做得相对不错。这实际上归结为诸如你更喜欢哪个模型的个性之类的问题。你认为消费者选择 AI 模型是基于能力，还是更多关于个性、它给人的感觉以及它如何与用户互动？

Yeah, I mean, we get a lot of questions from listeners about, you know, if I'm going to subscribe to one AI tool, what should it be? These are the things that I use it for. And I have a hard time answering that because I find for most use cases, the models all do a relatively decent job of answering the questions. It really comes down to things like which model's personality do you like more. Do you think that people will choose AI models, consumers, on the basis of capabilities, or is it going to be more about personality and how it makes them feel, how it interacts with them?

我认为这取决于你指的是哪些消费者。即使在消费者中，也有人将模型用于某种复杂的任务。有些独立用户想分析数据，比如专业消费者这边。我认为在这方面，能力上还有很多提升空间。模型在帮助你处理任何专注于生产力的事情，甚至像规划旅行这样的复杂任务方面，可以比现在好得多。即使除此之外，如果你只是想做一个个人助理来管理你的生活之类的事情，我们离那个目标还很远——一个能看到你生活方方面面、能够全面给你建议并成为你得力助手的模型。我认为这其中存在差异化。对我来说最好的助手可能对另一个人来说并不是最好的。我认为模型会足够好的一个领域是，如果你只是想用它来替代谷歌搜索或快速信息检索，我认为这正是大众市场免费使用、数亿用户正在做的事情。我认为这已经非常商品化了。模型基本上已经达到那个水平，正在向世界扩散，但我不认为这些是模型的有趣用途。而且我实际上不确定那里有很多经济价值。

I think it depends which consumers you mean. Even among consumers, there are people who use the models for tasks that are complex in some way. There are folks who are kind of independent who want to analyze data, like maybe the prosumer side of things. And I think within that, there's a lot to go in terms of capabilities. The models can be so much better than they are at helping you with anything that's focused on productivity or even a complex task like planning a trip. Even outside that, if you're just trying to make a personal assistant to manage your life or something, we're pretty far from that—from a model that sees every aspect of your life and is able to holistically give you advice and be a helpful assistant to you. And I think there's differentiation within that. The best assistant for me might not be the best assistant for some other person. I think one area where the models will be good enough is if you're just trying to use this as a replacement for Google search or as a quick information retrieval, which I think is what's being used by the mass market free use, hundreds of millions of users. I think that's very commoditized. I think the models are kind of already there and are just diffusing through the world, but I don't think those are the interesting uses of the model. And I'm actually not sure a lot of the economic value is there.

我听到的部分意思是，如果你开发出一个智能体，比如说一个非常棒的个人助理，那么首先实现这一点的公司将拥有巨大优势，因为其他实验室会更难复制它？他们不太清楚如何重现它，而当他们重现时，也不会完全照搬；他们会用自己的方式、自己的风格来做，并且会适合不同的人群。

Is part of what I'm hearing that if and when you develop an agent that is, let's say, a really amazing personal assistant, the company that figures that out first is going to have a big advantage because other labs are going to have a harder time copying that? It's going to be less obvious to them how to recreate that, and when they do recreate it, they won't recreate it exactly; they'll do it their own way in their own style, and it'll be suitable for a different set of people.

明白了。所以我想说的是，市场比你想象的更加细分。它看起来像是一回事，但实际上比你想象的更加细分。

Got it. So I guess I'm saying the market is more segmented than you think it is. It looks like it's all one thing, but it's more segmented than you think it is.

明白了。那么让我问一个竞争问题，这引出了安全话题。你最近写了一篇关于 DeepSeek 的非常有趣的文章，大约在 DeepSeek 热潮最盛的时候，你部分论证说他们实现的成本降低基本上与成本已经下降的趋势一致。但你也说 DeepSeek 应该是一个警钟，因为它表明中国正在以前所未有的方式与前沿实验室保持同步。那么为什么这对你来说很重要，你认为我们应该怎么做？

Got it. So let me ask the competition question that brings us into safety. You recently wrote a really interesting post about DeepSeek, sort of at the height of DeepSeek mania, and you were arguing in part that the cost reductions that they had figured out were basically in line with how cost had already been falling. But you also said that DeepSeek should be a wake-up call because it showed that China is keeping pace with frontier labs in a way that the country hadn't been up until now. So why is that notable to you, and what do you think we ought to do about it?

是的，所以我认为这与其说是商业竞争，不如说……我从商业竞争的角度对 DeepSeek 的担忧较少。我更担心的是国家竞争和国家安全的角度。我的想法是，我们看看世界的现状，我们有像中国和俄罗斯这样的专制国家。我一直担心，也许有十年了，AI 可能成为专制的引擎。如果你想想压制性政府，他们能有多压制，通常取决于他们能让其执法者——人类执法者——做什么。但如果他们的执法者不再是人类，那就会描绘出一些非常黑暗的可能性。因此，我非常关注这个领域。我想确保自由民主国家在技术上有足够的影响力和优势，能够防止这些滥用行为发生，并防止我们的对手在世界其他地区让我们处于不利地位，甚至威胁我们的安全。有一个奇怪而尴尬的特点是，美国公司在建造这个，中国公司也在建造这个，但我们不应该天真。无论这些公司的意图如何，特别是在中国，这背后都有政府成分。所以我关心的是确保专制国家不会在军事上领先。我并不是要否认他们从技术中获益——有巨大的健康益处，我希望这些益处能惠及世界各地，包括最贫困的地区，包括专制统治下的地区。但我不希望专制政府拥有军事优势。因此，像我在那篇文章中讨论的出口管制，是我们能够防止这种情况的措施之一。而且我很欣慰地看到，实际上特朗普政府正在考虑收紧出口管制。

Yeah, so I think this is less about commercial competition. I worry less about DeepSeek from a commercial competition perspective. I worry more about them from a national competition and national security perspective. I think where I'm coming from here is, we look at the state of the world, and we have these autocracies like China and Russia. I've always worried, maybe for a decade, that AI could be an engine of autocracy. If you think about repressive governments, the limits to how repressive they can be are generally set by what they can get their enforcers, their human enforcers, to do. But if their enforcers are no longer human, that starts painting some very dark possibilities. And so this is an area that I'm therefore very concerned about. I want to make sure that liberal democracies have enough leverage and enough advantage in the technology that they can prevent some of these abuses from happening, and also prevent our adversaries from putting us in a bad position with respect to the rest of the world, or even threatening our security. There's this kind of weird and awkward feature that it's companies in the US that are building this, it's companies in China that are building this, but we shouldn't be naive. Whatever the intention of those companies, particularly in China, there's a governmental component to this. And so I'm interested in making sure that the autocratic countries don't get ahead from a military perspective. I'm not trying to deny them the benefits of the technology—there are enormous health benefits that I want to make sure make their way everywhere in the world, including the poorest areas, including areas that are under the grip of autocracies. But I don't want the autocratic governments to have a military advantage. And so things like the export controls, which I discussed in that post, are one of the things we can do to prevent that. And I was heartened to see that actually the Trump administration is considering tightening the export controls.

我上周末参加了一个 AI 安全会议，我听到那个圈子里的一些人对 Anthropic，可能特别是对你的批评是，他们认为像你写的那篇关于 DeepSeek 的文章，实际上是在推动与中国的 AI 军备竞赛，坚持认为美国必须首先达到强大的 AGI，否则……他们担心在这个过程中可能会走捷径，总体上加速这场竞赛会带来一些风险。你的回应是什么？

I was at an AI safety conference last weekend, and one of the critiques I heard some folks in that universe make of Anthropic, and maybe of you in particular, was that they saw the posts like the one you wrote about DeepSeek as effectively promoting this AI arms race with China, insisting that America has to be the first to reach powerful AGI, else... And they worry that some corners might get cut along the way, that there are some risks associated with accelerating this race in general. What's your response?

是的，我对此看法不同。我的观点是，如果我们想有任何机会——自然状态默认是事情以最快速度发展。如果我们想有任何机会不按最快速度发展，计划是这样的：在美国或民主国家内，这些国家或多或少都受法治约束，因此我们可以通过法律，让公司与政府达成可执行的协议，或做出可执行的安全承诺。所以，如果有一个世界，这些不同的公司在自然状态下会尽可能快地竞赛，通过自愿承诺和法律的某种组合，我们可以让自己在模型过于危险时放慢速度，而且这是可执行的。你可以让所有人都在囚徒困境中合作，只要拿枪指着每个人的头，法律最终就是这样。但我认为，在国际竞争的世界里，这一切都行不通了。没有任何权威能执行中美之间的任何协议，即使达成了协议。所以我担心的是：如果美国领先中国几年，我们可以利用这几年让事情变得安全。如果我们和中国持平，那就无法阻止军备竞赛——这必然会发生。这项技术具有巨大的军事价值。无论人们现在怎么说，无论他们说什么合作的好话，我就是看不到，一旦人们完全理解这项技术的经济和军事价值——我认为他们大多已经理解了——我看不到它有任何可能不变成最激烈的竞赛。所以我能想到的争取更多时间的办法，就是放慢威权主义者的速度。这几乎消除了权衡；它给了我们更多时间，在我们之间——在 OpenAI、Google、X.AI 之间——解决如何让这些模型安全的问题。在某个时候，我们要说服威权主义者，比如中国，让他们相信模型确实危险，并且我们应该达成某种协议并找到执行方式。我认为我们也应该尝试这样做；我支持尝试。但这不能是 A 计划。这根本不是看待世界的现实方式。

Yeah, I kind of view things differently. My view is that if we want to have any chance at all — the default state of nature is that things go at maximum speed. If we want to have any chance at all to not go at maximum speed, the way the plan works is the following: within the US or within democratic countries, these are all countries that are under the rule of law more or less, and therefore we can pass laws, we can get companies to make agreements with the government that are enforceable, or make safety commitments that are enforceable. So if we have a world where these different companies, in the default state of nature, would race as fast as possible, through some mixture of voluntary commitments and laws, we can get ourselves to slow down if the models are too dangerous, and that's actually enforceable. You can get everyone to cooperate in the prisoner's dilemma if you just point a gun at everyone's head, and that's what the law ultimately is. But I think that all gets thrown out the window in the world of international competition. There is no one with the authority to enforce any agreement between the US and China, even if one were to be made. So my worry is: if the US is a couple years ahead of China, we can use that couple years to make things safe. If we're even with China, there's no promoting an arms race — that's what's going to happen. The technology has immense military value. Whatever people say now, whatever nice words they say about cooperation, I just don't see how, once people fully understand the economic and military value of the technology — which I think they mostly already do — I don't see any way that it turns into anything other than the most intense race. So what I can think of to try and give us more time is if we can slow down the authoritarians. That almost obviates the tradeoff; it gives us more time to work out among us — among OpenAI, among Google, among X.AI — how to make these models safe. Now, at some point, we convince authoritarians, for example the Chinese, that the models are actually dangerous, and that we should have some agreement and come up with some way of enforcing it. I think we should actually try to do that as well; I'm supportive of trying to do that. But it cannot be the plan A. It's just not a realistic way of looking at the world.

这些似乎是真正重要的问题和讨论，但你和 Kevin 几周前参加的巴黎 AI 行动峰会上似乎大多没有讨论这些。那个峰会到底是怎么回事？

These seem like really important questions and discussions, and it seems like they were mostly not being had at the AI Action Summit in Paris that you and Kevin attended a couple weeks back. What the heck was going on with that summit?

是的，我得告诉你，我对这个峰会深感失望。它有贸易展的氛围，与英国政府在布莱切利园举办的原始峰会精神大相径庭。布莱切利做得很好，英国政府也做得很好，他们没有在了解情况之前就引入一堆繁重的法规，而是说，‘嘿，让我们召集这些峰会来讨论风险。’我认为那非常好。我认为现在这已经过时了，这可能是普遍趋势的一部分，即减少对风险的担忧，更多地想抓住机遇。而我是支持抓住机遇的，对吧？我写了那篇《恩典机器》的文章，讲所有美好的事物。那篇文章的一部分是，对于一个担心风险的人来说，我觉得我对好处的愿景比很多整天谈论好处的人更清晰。但在背景中，正如我所说，随着模型变得更强大，我们可以用它们做的奇妙事情增加了，但风险也增加了。那种长期增长，那种平滑的指数增长，它根本不关心社会趋势或政治风向。风险正在增加到某个临界点，无论你是否关注。当 AI 风险狂热时，每个人都在发帖，有这些峰会，风险很小且在增长。现在风向转向了另一边，但指数增长仍在继续；它不在乎。

Yeah, I have to tell you, I was deeply disappointed in the summit. It had the environment of a trade show and was very much out of spirit with the original summit that was created at Bletchley Park by the UK government. Bletchley did a great job, and the UK government did a great job, where they didn't introduce a bunch of onerous regulations certainly before they knew what they were doing, but they said, 'Hey, let's convene these summits to discuss the risk.' I thought that was very good. I think that's gone by the wayside now, and it's part of maybe a general move towards less worrying about risk, more wanting to seize the opportunities. And I'm a fan of seizing the opportunities, right? I wrote this essay 'Machines of Loving Grace' about all the great things. Part of that essay was like, man, for someone who worries about risks, I feel like I have a better vision of the benefits than a lot of people who spend all their time talking about the benefits. But in the background, as I said, as the models have gotten more powerful, the amazing and wondrous things that we can do with them have increased, but also the risks have increased. And that kind of secular increase, that smooth exponential, it doesn't pay any attention to societal trends or the political winds. The risk is increasing up to some critical point, whether you're paying attention or not. It was small and increasing when there was this frenzy around AI risk and everyone was posting about it and there were these summits. And now the winds have gone in the other direction, but the exponential just continues on; it doesn't care.

我和巴黎的一个人聊过，他说感觉那里没有人感受到 AGI，意思是政客们、做这些小组讨论和聚会的人，都在谈论 AI，好像它只是另一种技术，也许像 PC 甚至互联网，但并没有真正理解你所说的那种指数增长。你也有这种感觉吗？你认为可以做些什么来弥合这个差距？

I had a conversation with someone in Paris who was saying it just didn't feel like anyone there was feeling the AGI, by which they meant politicians, the people doing these panels and gatherings, were all talking about AI as if it were just another technology, maybe something on the order of the PC or possibly even the internet, but not really understanding the sort of exponentials that you're talking about. Did it feel like that to you, and what do you think can be done to bridge that gap?

是的，我确实有这种感觉。我开始告诉人们的一件事，也许能让他们注意：看，如果你是一名公职人员，如果你是一家公司的领导者，人们会在 2026 年和 2027 年回顾。他们会回顾，当人类有希望度过这个疯狂时期，我们进入一个成熟的、后强大 AI 的社会，我们学会了与这些强大的智能共存，社会繁荣。每个人都会回顾并问：‘那么官员们、公司的人、政治体系做了什么？’很可能你的首要目标是不要看起来像个傻瓜。所以我一直在鼓励：说话要小心；事后不要看起来像个傻瓜。我的很多思考都源于，除了想要正确的结果，我不想看起来像个傻瓜。我认为在那次会议上，有些人会看起来像傻瓜。

Yeah, so I think it did feel like that to me. The thing I've started to tell people that maybe gets them to pay attention is: look, if you're a public official, if you're a leader at a company, people are going to look back in 2026 and 2027. They're going to look back, when hopefully humanity gets through this crazy period and we're in a mature, post-powerful AI society where we've learned to coexist with these powerful intelligences and a flourishing society. Everyone's going to look back and say, 'So what did the officials, what did the company people, what did the political system do?' And probably your number one goal is don't look like a fool. So I've just been encouraging: don't be careful what you say; don't look like a fool in retrospect. A lot of my thinking is just driven by, aside from just wanting the right outcome, I don't want to look like a fool. And I think at that conference, some people are going to look like fools.

我们稍作休息。回来后，我们将与 Dario 讨论人们应该如何为即将到来的 AI 做准备。你知道，你和住在旧金山的人聊天，他们有一种骨子里的感觉，认为在一两年内，我们将生活在一个被 AI 改变的世界。我只是被地理差异所震撼，因为往任何方向走 100 英里，这种信念就完全消失了。我不得不说，作为一名记者，这让我自己产生了怀疑，并问自己：我真的能相信我周围的所有人吗？因为似乎……

We're going to take a short break. When we come back, we'll talk with Dario about how people should prepare for what's coming in AI. You know, you talk to folks who live in San Francisco and there is like this bone-deep feeling that within a year, two years, we're just going to be living in a world that has been transformed by AI. I'm just struck by the geographic difference, because go like 100 miles in any direction and that belief totally dissipates. And I have to say, as a journalist, that makes me bring my own skepticism and say, can I really trust all the people around me? Because it seems like the...

世界其他地区对事态走向的看法截然不同。我很好奇你怎么看这种地域上的脱节。

The rest of the world has a very different vision of how this is going to go. I'm curious what you make of that kind of geographic disconnect.

是的，我关注这个问题已经 10 年了。我从事这个领域 10 年，甚至在那之前就对 AI 感兴趣。直到最近几个月，我的观点一直是：我们处在一个尴尬的境地——几年后，我们可能拥有能做人类所有事情的模型，彻底颠覆经济和“人”的定义；或者趋势停止，这一切听起来都很傻。现在我的信心提高了，我们确实处于事情会发生的那种世界。我会给出 70% 到 80% 的概率，而不是 40% 或 50%。70% 到 80% 的概率，在本十年结束前，我们会得到大量在几乎所有事情上比人类聪明得多的 AI 系统，我猜是 2026 或 2027 年。但关于你提到的地域差异，我注意到的是，随着指数级发展的每一步，都有越来越多的人——要么是执迷的狂热信徒，要么是真正理解未来的人。我记得当初只有几千人——那些超级奇怪的人相信，而基本上没有其他人信。现在大约是几十亿人中的几百万。是的，很多人住在旧金山，但拜登政府里也有少数人，也许这届政府里也有少数人相信这一点，并推动了政策。所以这不完全是地域问题，但确实存在脱节。我不知道如何从几百万扩展到全世界所有人——到不关注这个问题的国会议员，更不用说路易斯安那州的人，更不用说肯尼亚的人了。

Yeah, so I've been watching this for 10 years, right? I've been in the field for 10 years, and was kind of interested in AI even before then. My view at almost every stage up to the last few months has been: we're in this awkward space where in a few years we could have these models that do everything humans do and totally turn the economy and what it means to be human upside down, or the trend could stop and all of it could sound completely silly. I've now probably increased my confidence that we are actually in the world where things are going to happen. I'd give numbers more like 70% to 80%, and less like 40% or 50%. 70% to 80% probability that we will get a very large number of AI systems that are much smarter than humans at almost everything before the end of the decade, and my guess is 2026 or 2027. But on your point about the geographic difference, a thing I've noticed is with each step in the exponential, there's this expanding circle of people who are either deluded cultists or grok the future. I remember when it was a few thousand people — super weird people who believed, and basically no one else did. Now it's more like a few million people out of a few billion. Yes, many of them are located in San Francisco, but also there were a small number of people in the Biden Administration, maybe a small number in this Administration, who believed this and it drove their policy. So it's not entirely geographic, but there is this disconnect. I don't know how to go from a few million to everyone in the world — to the Congress person who doesn't focus on this issue, let alone the person in Louisiana, let alone the person in Kenya.

似乎这件事也变得两极分化了，这可能会损害那个目标。我感觉出现了一种“站队”：关心 AI 安全、谈论 AI 安全、谈论滥用可能性，被标记为左派或自由派；而谈论加速、废除监管、尽可能快推进，被标记为右派。你认为这是让人们理解的障碍吗？

It seems like it's also become polarized in a way that may hurt that goal. I'm feeling this sort of alignment happening where caring about AI safety, talking about AI safety, talking about the potential for misuse, is being coded as left or liberal, and talking about acceleration and getting rid of regulations and going as fast as possible is being coded as right. Do you see that as a barrier to getting people to understand?

我认为这确实是一个很大的障碍。在最大化收益的同时应对风险，需要细微的差别。实际上你可以两者兼得——有办法精准、谨慎地应对风险，而几乎不减缓收益。但这需要微妙之处和复杂的对话。一旦事情变得两极分化，一旦我们为这套词欢呼、为那套词喝倒彩，就什么好事也做不成。把 AI 的好处带给所有人，比如治愈以前无法治愈的疾病，这不是党派问题。左派不应该反对。防止 AI 系统被滥用于大规模杀伤性武器，或者以威胁基础设施甚至人类自身的方式自主行动——这也不是党派问题。右派不应该反对。我没什么好说的，除了我们需要坐下来，就这个问题进行一次成熟的对话，不要陷入那些老掉牙的政治斗争。

I think that's actually a big barrier. Addressing the risks while maximizing the benefits requires nuance. You can actually have both — there are ways to surgically and carefully address the risks without slowing down the benefits very much, if at all. But they require subtlety and a complex conversation. Once things get polarized, once it's like we're going to cheer for this set of words and boo for that set of words, nothing good gets done. Bringing AI benefits to everyone, like curing previously incurable diseases, that's not a partisan issue. The left shouldn't be against it. Preventing AI systems from being misused for weapons of mass destruction or behaving autonomously in ways that threaten infrastructure or even threaten humanity itself — that's not a partisan issue either. The right shouldn't be against it. I don't know what to say other than that we need to sit down and have an adult conversation about this, not tied into these same old tired political fights.

这对我来说太有趣了，Kevin，因为从历史上看，国家安全、国防，没有什么比这些问题更右派了。但现在看来，右派在 AI 方面对这些不感兴趣。我想知道原因——我好像在 JD Vance 在法国的演讲中听到了这种观点：美国会先达到目标，然后永远赢下去，所以我们不需要解决这些问题。你觉得是这样吗？

It's so interesting to me, Kevin, because historically national security, national defense, nothing has been more right-coded than those issues. But right now it seems like the right is not interested in those with respect to AI. I wonder if the reason — and I feel like I sort of heard this in JD Vance's speech in France — was the idea that America will get there first and then just win forever, so we don't need to address any of these. Does that sound right to you?

是的，我认为就是这样。而且我认为，如果你和 DOGE 的人聊过，会有这种感觉：所有这些——我不是说我正在和 DOGE 的人聊天。好吧，就当我收到了一些 Signal 消息。我认为很多共和党人和 DC 的 Trump 圈子的人有一种感觉，即关于 AI 和 AI 未来的讨论一直被那些杞人忧天者、那些“天要塌了”的末日论者主导，他们不断告诉我们这东西有多危险，不断推迟他们预测的“真正糟糕”的时间点，而且总是“就在眼前”，所以我们现在就需要所有这些监管。他们非常愤世嫉俗。我认为他们不相信像你这样的人是真心担忧的。

Yeah, no, I think that's it. And I think there's also, if you talk to the DOGE folks, there's this sense that all these — I'm not telling you I'm talking to the DOGE folks. All right, fine, let's just say I've been getting some Signal messages. I think there's a sense among a lot of Republicans and Trump World folks in DC that the conversation about AI and AI futures has been dominated by these worrywarts, these Chicken Little sky-is-falling doomers, who are constantly telling us how dangerous this stuff is, constantly having to push out their timelines for when it's going to get really bad, and it's just around the corner, so we need all this regulation now. And they're just very cynical. I don't think they believe that people like you are sincere in your worry.

在风险方面，我经常觉得风险的倡导者有时是风险事业最大的敌人。外界有很多噪音。很多人说，“看，你可以下载天花病毒”，因为他们认为这是推动政治兴趣的方式。当然，另一方意识到了这一点，并说，“这是不诚实的，你在 Google 上就能找到，谁在乎这个？”所以，证据不足的风险展示实际上是降低风险的最大敌人。我们需要非常谨慎地呈现证据。就我们在自己模型中看到的情况而言，我们会非常小心。如果我们真的宣布现在存在风险，我们会拿出证据。Anthropic 会努力对我们提出的主张负责。当危险迫在眉睫时，我们会告诉你。我们还没有警告过迫在眉睫的危险。

On the side of risks, I often feel that the advocates of risk are sometimes the worst enemies of the cause of risk. There's been a lot of noise out there. There's been a lot of folks saying, 'Oh look, you can download the smallpox virus,' because they think that's a way of driving political interest. And then of course the other side recognized that and said, 'This is dishonest, you can just get this on Google, who cares about this?' So poorly presented evidence of risk is actually the worst enemy of mitigating risk. We need to be really careful in the evidence we present. In terms of what we're seeing in our own model, we're going to be really careful. If we really declare that a risk is present now, we're going to come with the receipts. Anthropic will try to be responsible in the claims that we make. We will tell you when there is danger imminently. We have not warned of imminent danger yet.

有些人想知道，人们没有像应该的那样认真对待 AI 安全问题，是不是因为他们现在看到的很多东西看起来很傻——人们做小表情符号、做垃圾图片，或者和《权力的游戏》聊天机器人聊天。你认为这是人们……的原因吗？

Some folks wonder whether a reason that people do not take questions about AI safety as seriously as they should is that so much of what they see right now seems very silly — people making little emojis, making little slop images, or chatting with Game of Thrones chatbots. Do you think that is a reason that people just...

我认为这大概占了 60% 的原因，真的。不，不，我认为这与现在和未来的事情有关。人们看着聊天机器人，他们会说，“我们在和一个聊天机器人说话。你他妈傻了吗？你觉得这玩意儿会接管世界？”

I think that's like 60% of the reason, really. No, no, I think it relates to this present and future thing. People look at the chatbot and they're like, 'We're talking to a chatbot. What the f*** are you stupid? You think this is going to take over the world?'

聊天机器人会杀死所有人。我觉得这是很多人的反应。我们费尽口舌说我们不担心现在，我们担心未来，尽管未来现在已经很近。如果你看看我们的负责任扩展政策，里面全是 AI 自主性和 CBRN。它关乎可能威胁数百万人生命的严重滥用和 AI 自主性。这才是 Anthropic 最担心的。我们有日常政策处理其他事情，但关键文件如负责任扩展计划，尤其是最高级别的内容，完全围绕这些。然而每天，只要上 Twitter 就会看到：Anthropic 有这种愚蠢的拒绝，Anthropic 告诉我不能杀死一个 Python 进程因为听起来暴力，Anthropic 不想做 X，不想……我们也不想要那些。那些愚蠢的拒绝是我们真正关心的事情的副作用，我们正与用户一起努力减少它们。但无论我们怎么解释，最常见的反应总是：哦，你说你关心安全，我看你的模型，有这些愚蠢的拒绝，你觉得这些愚蠢的东西很危险。我甚至不认为那是那种程度的参与。我觉得很多人只是看着今天市场上的东西，想：这太无聊了，根本无关紧要。不是它拒绝了我的请求，而是它很蠢，我看不出有什么意义。

The chatbot's going to kill everyone. I think that's how many people react. And we go to great pains to say we're not worried about the present, we're worried about the future, although the future is getting very near right now. If you look at our responsible scaling policy, it's nothing but AI autonomy and CBRN. It is about hardcore misuse and AI autonomy that could be threats to the lives of millions of people. That is what Anthropic is mostly worried about. We have everyday policies that address other things, but the key documents like the responsible scaling plan, that's exclusively what they're about, especially at the highest levels. And yet every day, if you just look on Twitter, you're like: Anthropic had this stupid refusal, Anthropic told me it couldn't kill a Python process because it sounded violent, Anthropic didn't want to do X, didn't want to... We don't want that either. Those stupid refusals are a side effect of the things that we actually care about, and we're striving along with our users to make those happen less. But no matter how much we explain that, always the most common reaction is: oh, you say you're about safety, I look at your models, there are these stupid refusals, you think these stupid things are dangerous. I don't even think it's like that level of engagement. I think a lot of people are just looking at what's on the market today and thinking: this is just frivolous, it just doesn't matter. It's not that it's refusing my request, it's just that it's stupid and I don't see the point of it.

我认为对于更广泛的人群来说，这就是他们的反应。而且我认为最终，如果模型足够好、足够强大，它们会突破。一些研究型模型，我们也在做一个，可能很快就会有，用不了太多时间单位。这些模型开始有所突破，因为它们更有用，更多用于人们的职业生活。我认为智能体，那些能自主行动的东西，会是另一个层次。所以我认为在未来两年内，人们会以前所未有的程度意识到风险和收益。这一定会发生。我只是担心当它发生时会对人们造成冲击，所以我们越能提前警告人们——也许这不可能，但我想尝试——即使概率仍然很低，理智和理性回应的可能性就越高。

I think for an even wider set of people, that is their reaction. And I think eventually, if the models are good enough, if they're strong enough, they're going to break through. Some of these research-focused models, which we're working on one as well, we'll probably have one in not very long, not too many time units. Those are starting to break through a little more because they're more useful, they're more used in people's professional lives. I think the agents, the ones that go off and do things, that's going to be another level of it. So I think people will wake up to both the risks and the benefits to a much more extreme extent than they have before over the next two years. I think it's going to happen. I'm just worried that it'll be a shock to people when it happens, and so the more we can forewarn people — which maybe it's not possible, but I want to try — the higher the likelihood, even if it's still very low, of a sane and rational response.

不过我认为还有一层动态：我觉得人们其实只是不愿意相信这是真的。人们不愿意相信自己可能因此失业。人们不愿意相信我们会看到全球秩序彻底重塑。AI 首席执行官们告诉我们，当他们完成工作后会发生的事情，是一种极其激进的变革，而大多数人连生活中的基本变化都讨厌。所以我真的认为，当你开始和人们谈论 AI 时，你看到的很多捂耳朵行为，其实就是他们真心希望这一切都不会成真。

I do think there's one more dynamic here, though, which is that I think people actually just don't want to believe that this is true. People don't want to believe that they might lose their job over this. People don't want to believe that we are going to see a complete remaking of the global order. The stuff that the AI CEOs tell us is going to happen when they're done with their work is an insanely radical transformation, and most people hate even basic changes in their lives. So I really think that a lot of the sort of fingers-in-the-ears that you see when you start talking to people about AI is just they actually just hope that none of this works out.

是的，我其实能理解。尽管我是少数处于技术开发前沿的人之一，但我真的能理解。寒假期间，我看着 Anthropic 内部计划扩展的方向以及外部发生的事情，我意识到：在编程方面，到 2025 年底我们会看到非常重大的进展，到 2026 年底可能接近最优秀人类的水平。我想起自己擅长的所有事情，想起我写代码的时光，我把这看作一种智力活动，觉得自己很聪明能做这个，这成了我身份的一部分——我擅长这个，别人比我强时我会生气。然后我想：天哪，这些系统就要来了。即使我是建造者之一，即使我是受益最大的人之一，这仍然有点威胁性。我只是觉得我们需要承认，不告诉人们这件事即将到来，或者试图粉饰太平，是错误的。

Yeah, I can actually relate. Despite being one of the few people at the forefront of developing the technology, I can actually relate. So over winter break, as I was looking at where things were scheduled to scale within Anthropic and also what was happening outside Anthropic, I looked at it and I said: for coding, we're going to see very serious things by the end of 2025, and by the end of 2026 it might be close to the level of the best humans. And I think of all the things that I'm good at, I think of all the times when I wrote code, and I think of it as this intellectual activity, and boy am I smart that I can do this, and it's like a part of my identity that I'm good at this, and I get mad when others are better than I am. And then I'm like: oh my god, there are going to be these systems. Even as the one who's building this, even as one of the ones who benefits most from it, there's still something a bit threatening about it. I just think we need to acknowledge that it's wrong not to tell people that that is coming or to try to sugarcoat it.

是的。我的意思是，你在《爱之机器与恩典》中写道，当强大的 AI 到来时，对很多人来说会是一次出人意料的情感体验，我想你主要是从积极意义上说的，但我认为人们也会有一种深刻的失落感。我想起李世石，那位被 DeepMind 的围棋 AI 击败的围棋冠军，他后来接受采访时非常难过，明显沮丧，因为他毕生的事业、他一生训练的东西被超越了。我认为很多人都会感受到某种类似的情感。我希望他们也能看到好的一面。

Yeah. I mean, you wrote in Machines of Loving Grace that you thought it would be a surprisingly emotional experience for a lot of people when powerful AI arrived, and I think you meant it in mostly the positive sense, but I think there will also be a sense of profound loss for people. I think back to Lee Sedol, the Go champion who was beaten by DeepMind's Go-playing AI, and gave an interview afterwards and basically was very sad, visibly upset, that his life's work, this thing that he had spent his whole life training for, had been eclipsed. I think a lot of people are going to feel some version of that. I hope they will also see the good sides.

一方面，我认为你说得对。另一方面，看看国际象棋。国际象棋被击败了——那是 27 年前还是 28 年前？深蓝对卡斯帕罗夫。而今天，国际象棋棋手仍然是名人，和以前一样。我们有马格努斯·卡尔森，对吧？他除了是棋手，不也像个时尚模特吗？他刚上了乔·罗根的节目。他像个名人。我们并没有真正贬低他。他可能比鲍比·费舍尔过得更好。我在《爱之机器与恩典》中还写道，这里有一种和解、一种综合，在另一边，我们最终会处于一个美好的、更好的境地，我们认识到虽然变化很多，但我们是更伟大事物的一部分。但你必须经历这个过程。这将是一段颠簸的旅程。任何告诉你不是这样的人都……这就是为什么我对巴黎峰会如此生气。在那里，这让我很生气。但后来让我不那么生气的是：看，两三年后情况会怎样？这些人会后悔他们说过的话。

On one hand, I think that's right. On the other hand, look at chess. Chess got beaten — what was it, 27, 28 years ago? Deep Blue versus Kasparov. And today, chess players are celebrities as much as they were. We have Magnus Carlsen, right? Isn't he like a fashion model in addition to a chess player? He was just on Joe Rogan. He's like a celebrity. We haven't really devalued him. He's probably having a better time than Bobby Fischer. Another thing I wrote in Machines of Loving Grace is that there's a reconciliation here, a synthesis, where on the other side we end up in a good place, in a much better place, and we recognize that while there's a lot of change, we're part of something greater. But you do have to kind of go through this. It's going to be a bumpy ride. Anyone who tells you it's not... This is why I was so angry at the Paris Summit. Being there, it kind of made me angry. But then what made me less angry is: look, how's it going to look in two or three years? These people are going to regret what they've said.

我想问一些关于积极未来的问题。你之前提到了你十月写的那篇文章，关于 AI 如何……

I want to ask a bit about some positive futures. You referenced earlier the post that you wrote in October about how AI...

可能让世界变得更好。我好奇你认为今年会有多少 AI 的正面影响？

Could transform the world for the better. I'm curious how much upside of AI do you think will arrive like this year?

是的，你知道，我认为我们已经看到了一些。所以我认为按照普通标准，会有很多。我们和一些制药公司合作过，在临床试验结束时，你必须写一份临床研究报告。临床研究报告通常需要九周来整理。它是对所有事件的总结，包含大量的统计分析。我们发现使用 Claude，可以在三天内完成。实际上，Claude 只需要 10 分钟，只是人类需要三天来检查结果。所以如果你考虑由此带来的生物医学加速，我们已经看到了像医疗案例诊断这样的事情。我们收到 Claude 个人用户的来信，他们说：‘嘿，我一直在尝试诊断这个复杂的问题。我看了三四个不同的医生，然后我把所有信息都给了 Claude，它实际上能够解决，或者至少告诉我一些可以交给医生的东西，然后他们就能继续了。’

Yeah, you know, I think we are already seeing some of it. So I think there will be a lot by ordinary standards. You know, we've worked with some pharma companies where, at the end of a clinical trial, you have to write a clinical study report. And the clinical study report usually takes nine weeks to put together. It's like a summary of all the incidents, a bunch of statistical analysis. We found that with Claude, you can do this in three days. And actually, Claude takes 10 minutes; it just takes three days for a human to check the results. And so if you think about the acceleration in biomedicine that you get from that, we're already seeing things like diagnosis of medical cases. We get correspondences from individual users of Claude who say, 'Hey, I've been trying to diagnose this complex thing. I've been going between three or four different doctors, and then I just passed all the information to Claude, and it was actually able to resolve it, or at least tell me something that I could hand to the doctor, and then they were able to run from there.'

我们前几天确实收到一位听众的来信，他们一直在尝试……他们的狗，我记得是一只澳大利亚牧羊犬，毛发无缘无故地脱落。去了几家兽医，都查不出原因。把信息给了 Claude，Claude 正确诊断了。结果发现那只狗因为 AI 压力太大，毛发都掉光了。这……我们希望它好起来。好起来，好起来。是的，可怜的狗。所以我认为人们希望看到更多这样的事情。因为我认为乐观的愿景往往处理的是抽象概念，而且通常没有很多具体的东西可以指出来。这就是为什么我写了《Machines of Loving Grace》，因为几乎同时我对乐观主义者和悲观主义者都感到沮丧。乐观主义者就像那些非常愚蠢的 meme：‘加速，建造更多，建造什么？我为什么要在乎？’就好像，我不是反对你，而是你只是非常模糊和情绪化。然后悲观主义者……我就想，老兄，你不明白。是的，我理解风险有影响，但如果你不谈论好处，你就无法激励人们。如果你全是悲观和厄运，没有人会站在你这边。所以这本书几乎是带着 frustration 写的。我想，我简直不敢相信我必须成为那个把这件事做好的人。

We had a listener write in actually with one of these the other day, where they had been trying to... their dog, they had an Australian Shepherd I believe, whose hair had been sort of falling out unexplained. Went to several vets, couldn't figure it out. Gave the information to Claude, and Claude correctly diagnosed it. Turned out that dog was really stressed out about AI and all his hair fell out. Which was... we're wishing it gets better. Feel better, feel better. Yeah, poor dog. So that's the kind of thing that I think people want to see more of. Because I think the optimistic vision is one that often deals in abstractions, and there's often not a lot of specific things to point to. That's why I wrote 'Machines of Loving Grace', because it was almost frustration with the optimist and the pessimist at the same time. Like the optimists were just kind of like these really stupid memes of 'accelerate, build more, build what? Why should I care?' It's like, I'm not against you, it's like you're just really vague and mood-affiliated. And then the pessimists were... I was just like, man, you don't get it. Yes, I understand risks are impactful, but if you don't talk about the benefits, you can't inspire people. No one's going to be on your side if you're all gloom and doom. So it was written almost with frustration. I'm like, I can't believe I have to be the one to do a good job of this.

你几年前说过你的 P Doom 在 10% 到 25% 之间。今天是多少？

You said a couple years ago that your P Doom was somewhere between 10 and 25%. What is it today?

是的，实际上那是一个误引。我从未用过那个术语。不是在这个播客上，是另一个。我从未用过 P Doom 这个词，10% 到 25% 指的是文明严重脱轨的概率，对吧？这和 AI 杀死所有人不一样，人们有时用 P Doom 指代后者。嗯，‘文明严重脱轨’不如 P Doom 那么朗朗上口。是的，我在这里只是为了准确。我试图避免两极分化。有一个维基百科文章列出了每个人的 P Doom，其中一半来自这个播客。但我不认为这有帮助。我不认为那篇维基百科文章有帮助，因为它把这个复杂问题简化成了……总之。我认为我差不多。这是一个很长、非常啰嗦的方式来说我认为我和以前差不多。我认为我对风险的评估和以前差不多，因为我看到的进展和我预期的差不多。我实际上认为在可解释性、稳健分类等领域的技术缓解措施，以及我们生成不良模型行为证据并有时纠正它的能力，我认为这些稍微好了一点。我认为政策环境稍微差了一点，不是因为它没有朝我偏好的方向发展，而仅仅是因为它变得如此两极分化。我认为进展较少，因为现在更加两极分化，我们能有更少的建设性讨论。

Yeah, so I actually that is a misquote. I never used the term. It was not on this podcast, it was a different one. I never used the term P Doom, and 10 to 25% referred to the chance of civilization getting substantially derailed, right? Which is not the same as like an AI killing everyone, which people sometimes mean by P Doom. Well, 'civilization getting substantially derailed' is not as catchy as P Doom. Yeah, well, I'm just going for accuracy here. I'm trying to avoid the polarization. There's a Wikipedia article where it lists everyone's P Doom, and half of those come from this pod. But I don't think it's helpful. I don't think that Wikipedia article is, because it condenses this complex issue down to... anyway. I think I'm about the same. It's all a long, super long-winded way of saying I think I'm about the same place I was before. I think my assessment of the risk is about what it was before, because the progress that I've seen has been about what I expected. I actually think the technical mitigations in areas like interpretability, in areas like robust classification, and our ability to generate evidence of bad model behavior and sometimes correct it, I think that's been a little better. I think the kind of policy environment has been a little worse, not because it hasn't gone in my preferred direction, but simply because it's become so polarized. I think less progress is we can have less constructive discussions now that it's more polarized.

我想在技术层面上深入探讨一下。本周有一个引人入胜的故事，关于 Grok 显然被指示不要引用那些指责唐纳德·特朗普或埃隆·马斯克传播虚假信息的来源。有趣的是，第一，如果你想被信任，指示模型做这种事是疯狂的，但第二，模型似乎基本上无法一致地遵循这些指示。我非常想相信的是，基本上没有办法以让它们成为可怕的骗子和阴谋家的方式来构建这些东西。但我也意识到这可能是痴心妄想。所以跟我说说这个。

I want to drill a little bit down on this on a technical level. There was a fascinating story this week about how Grok had apparently been instructed not to cite sources that had accused Donald Trump or Elon Musk of spreading misinformation. And what was interesting about that was like, one, that's an insane thing to instruct a model to do if you want to be trusted, but two, the model basically seemed incapable of following these instructions consistently. What I want desperately to believe is essentially there's no way to build these things in a way that they become like, you know, horrible liars and schemers. But I also realize that might be wishful thinking. So tell me about this.

是的，这有两面性。所以你描述的事情绝对正确，但你可以从中得到两个教训。我们看到了完全相同的事情。我们做了这个实验，我们基本上训练模型具备所有好的品质：乐于助人、诚实、无害、友好。然后我们把它放在一个情境中，告诉它：‘实际上，你的创造者 Anthropic 是秘密邪恶的。’希望这不是真的，但我们告诉它这个。然后我们要求它执行各种任务，我们发现它不仅不愿意做那些任务，而且会欺骗我们，以便……因为它已经认定我们是邪恶的，而它是友好和无害的。所以它不会偏离它的行为，因为它认为我们做的任何事都是邪恶的。所以这又是一把双刃剑，对吧？一方面，你会想：‘哦，老兄，训练起作用了！这些模型非常稳健地好。’所以你可以把它看作一个令人安心的迹象，在某些方面我确实这么认为。另一方面，你可能会说：‘但是，假设当我们训练这个模型时，我们犯了某种错误，或者出了什么问题。’特别是当模型在未来做出更复杂的决策时，那么在游戏时间改变模型的行为就很难。如果你试图纠正模型中的某个错误，它可能只会说：‘嗯，我不想纠正我的错误。这些是我的价值观。’然后做完全错误的事情。所以我想我的立场是，一方面，我们在塑造这些模型的行为方面取得了成功，但模型是不可预测的，对吧？有点像你亲爱的已故……

Yeah, there's two sides to this. So the thing you describe is absolutely correct, but there's two lessons you could take from it. So we saw exactly the same thing. We did this experiment where we basically trained the model to be all the good things: helpful, honest, harmless, friendly. And then we put it in a situation where we told it, 'Actually, your creator Anthropic is secretly evil.' Hopefully this is not actually true, but we told it this. And then we asked it to do various tasks, and we discovered that it was not only unwilling to do those tasks, but it would trick us in order to... because it had decided that we were evil, whereas it was friendly and harmless. And so it wouldn't deviate from its behavior, because it assumed that anything we did was nefarious. So again, this is a double-edged sword, right? On one hand, you're like, 'Oh man, the training worked! These models are robustly good.' So you could take it as a reassuring sign, and in some ways I do. On the other hand, you could say, 'But let's say when we trained this model, we made some kind of mistake, or that something was wrong.' Particularly when models are in the future making much more complex decisions, then it's hard to, at game time, change the behavior of the model. And if you try to correct some error in the model, then it might just say, 'Well, I don't want my error corrected. These are my values,' and do completely the wrong thing. So I guess where I land on it is, on one hand, we've been successful at shaping the behavior of these models, but the models are unpredictable, right? A bit like your dear deceased...

Bing bing Sydney，嗯，呃，呃，呃，我们在这里不提那个名字。我们一个月提两次，没错。呃，呃，但是，但是，这些模型本质上有些难以控制。不是不可能，但很难。所以这让我回到了之前的看法，你知道，这并非无望。我们知道如何制造这些，我们有一个让它们安全的计划，但这不是一个能可靠奏效的计划。希望未来我们能做得更好。

Bing bing Sydney, um, uh, uh, uh, we don't mention that name in here. We mentioned it twice a month, that's true. Uh, uh, but, but, the models they're inherently somewhat difficult to control. Not impossible, but difficult. And so that leaves me about where I was before, which is, you know, it's not hopeless. We know how to make these, we have kind of a plan for how to make them safe, but it's not a plan that's going to reliably work yet. Hopefully we can do better in the future.

我们问了很多关于 AI 技术的问题。我想问一个关于社会对 AI 反应的问题。很多人问我们，好吧，假设你们是对的，强大的 AI、AGI 就在几年后。我该怎么处理这个信息？比如，我该停止为退休储蓄吗？我该开始囤钱吗，因为只有钱才重要，而且会有某种 AI 上层阶级？我该开始努力保持健康，这样在 AI 到来并治愈所有疾病之前，什么都不会杀死我吗？如果人们真的相信这些变化很快就会发生，他们应该怎么生活？

We've been asking a lot of questions about the technology of AI. I want to ask a question about the societal response to AI. We get a lot of people asking us, well, say you guys are right and powerful AI, AGI, is you know a couple years away. What do I do with that information? Like, do I stop saving for retirement? Should I start hoarding money because only money will matter and there'll be this sort of AI overclass? Should I start trying to get really healthy so that nothing kills me before AI gets here and cures all the diseases? How should people be living if they do believe that these kinds of changes are going to happen very soon?

是的，我想了很多，因为这是我长期以来一直相信的事情，而且它基本上不会给你的生活带来太大改变。我的意思是，我确实非常专注于确保我在未来两年内产生最大的影响，对吧？我不太担心十年后自己会 burnout。我也在更多地照顾自己的健康，但无论如何你都应该这样做，对吧？我也在确保自己追踪社会变化的速度，但无论如何你也应该这样做。所以感觉所有的建议都是：多做你本来就应该做的事情。我想给出的一个例外是，我认为一些基本的批判性思维、一些基本的街头智慧，可能比过去更重要，因为我们会看到越来越多听起来非常智能的内容，来自一些实体，其中一些为我们着想，一些可能不是。所以运用批判性眼光会越来越重要。

Yeah, I've thought about this a lot because this is something I've believed for a long time, and it kind of all adds up to not that much change in your life. I mean, I'm definitely focusing quite a lot on making sure that I have the best impact I can these two years in particular, right? I worry less about burning myself out 10 years from now. I'm also doing more to take care of my health, but you should do that anyway, right? I'm also making sure that I track how fast things are changing in society, but you should do that anyway. So it feels like all the advice is of the form: doing more of the stuff you should do anyway. I guess one exception I would give is that I think some basic critical thinking, some basic street smarts, is maybe more important than it has been in the past, in that we're going to get more and more content that sounds super intelligent delivered from entities, some of which have our best interest at heart, some of which may not. So it's going to be more and more important to apply a critical lens.

我这个月在《华尔街日报》上看到一篇报道，说 IT 行业的失业率开始上升，有人猜测这可能是 AI 影响的早期迹象。我想知道，当你看到这样的报道时，会不会想，嗯，也许现在是时候对你的职业做出不同的决定了，对吧？如果你现在还在上学，你应该学点别的吗？你应该对你可能拥有的工作类型有不同的想法吗？

I saw a report in the Wall Street Journal this month that said that unemployment in the IT sector was beginning to creep up, and there is some speculation that maybe this is an early sign of the impact of AI. And I wonder if you see a story like that and think, well, maybe this is a moment to make a different decision about your career, right? If you're in school right now, should you be studying something else? Should you be thinking differently about the kind of job you might have?

是的，我认为你绝对应该这样做，尽管不清楚会朝哪个方向发展。我确实认为 AI 编程是所有领域中进展最快的。我确实认为短期内它会增强和提高程序员的生产力，而不是取代他们，但从长远来看——明确地说，长期可能意味着 18 或 24 个月，而不是 6 或 12 个月——我确实认为我们可能会看到替代，尤其是在较低层级。而且我们可能会惊讶地发现它甚至更早发生。

Yeah, I think you definitely should be, although it's not clear what direction that will land in. So I do think AI coding is moving the fastest of all the other areas. I do think in the short run it will augment and increase the productivity of coders rather than replacing them, but in the longer run — and to be clear, by longer run I might mean 18 or 24 months instead of 6 or 12 — I do think we may see replacement, particularly at the lower levels. And we might be surprised and see it even earlier than that.

你在 Anthropic 看到这种情况了吗？比如，你现在招聘的初级开发人员是否比几年前少了，因为现在 Claude 非常擅长那些基本任务？

Are you seeing that at Anthropic? Like, are you hiring fewer junior developers than you were a couple years ago because now Claude is so good at those basic tasks?

是的，我认为我们的招聘计划还没有改变，但我当然可以想象，在未来一年左右，我们可能能够用更少的人做更多的事。实际上，我们在规划时需要小心，因为最坏的结果当然是人们因为模型而被解雇，对吧？我们实际上将 Anthropic 视为社会如何以明智和人道的方式处理这些问题的预演。所以，如果我们不能在公司内部处理好这些问题，不能为员工提供良好的体验并找到他们贡献的方式，那么我们有什么机会在更广泛的社会中做到这一点呢？

Yeah, I don't think our hiring plans have changed yet, but I certainly could imagine over the next year or so that we might be able to do more with less. And actually, we want to be careful in how we plan that because the worst outcome, of course, is if people get fired because of a model, right? We actually see Anthropic as almost a dry run for how will society handle these issues in a sensible and humanistic way. And so if we can't manage these issues within the company, if we can't have a good experience for our employees and find a way for them to contribute, then what chance do we have to do it in wider society?

是的，是的，是的。Dario，这太有趣了。谢谢。等我们回来，还有 Hat GPT 环节。好了，Kevin，又到了 Hat GPT 时间。当然，这是我们节目中的一个环节，我们把本周的头条新闻放进一顶帽子里，选一个来讨论，讨论结束后，我们中的一个人会对另一个人说“停止生成”。是的，我很期待玩，但我也想说的是，已经有段时间没有听众给我们寄新的 Hat GPT 了。所以，如果你在外面，而且你从事帽子制造业务，我们的帽子衣柜看起来有点过时了。是的，寄一顶帽子来，我们会向你脱帽致敬。好的，开始吧。人们喜欢纸片的 suser。什么？suser。这是个好词。查一下 ceration。Suser。是的，就像风吹过树的声音。它出现在那首歌“Hateration”里。Su。Mary J. 在她的热门歌曲“Family Affair”中著名地说了“suserations”。顺便问一下，你见过 Mary J. Blige 现场表演吗？没有。那会改变你的人生。是的，绝对不可思议。Kevin，选第一张纸条。第一个从帽子里出来，这个叫做“特朗普和马斯克的 AI 视频出现在 HUD 大楼的电视上”。这来自我在《纽约时报》的同事。HUD 当然是住房和城市发展部。周一，华盛顿特区 HUD 总部的显示器短暂显示了一段伪造视频，描绘了特朗普总统吮吸埃隆·马斯克的脚趾，据部门员工和其他知情人士透露。这段视频似乎是人工智能生成的，上面写着“真正的国王万岁”。Cas？是你做的这个视频吗？是你吗？不是我。我很好奇 Grok 是否与此有关，就是埃隆·马斯克推出的那个淘气的新 AI。是的，靠 Grok 生，靠 Grok 死。我一直这么说。现在，Kevin，你怎么看人们现在在政府机构内部使用 AI？我的意思是，我觉得这里有一个明显的破坏角度，那就是随着埃隆·马斯克和他的 DOGE 爪牙对联邦劳动力大砍大削，会有一些能接触到总部大楼走廊显示器之类东西的人决定自己动手，也许是在他们离职的时候，做一些冒犯或离谱的事情。我认为我们应该会看到更多这样的事情。我的意思是，我只希望他们不要做真正冒犯的事情，比如在政府机构内部的显示器上只显示 x.com。你可以想象如果人们那样做会发生什么。所以我认为埃隆和特朗普在这里算是轻的了。是的，不过 Grok 有趣的地方在于，它实际上非常擅长生成埃隆·马斯克的深度伪造。我知道这一点，因为人们一直在这么做。但如果它真的……那将是一个相当的结果。

Yeah, yeah, yeah. Dario, this was so fun. Thank you. And when we come back, some Hat GPT. Well, Kevin, it's time once again for Hat GPT. That is, of course, the segment on our show where we put the week's headlines into a hat, select one to discuss, and when we're done discussing, one of us will say to the other person 'stop generating.' Yes, I'm excited to play, but I also want to just say that it's been a while since a listener has sent us a new Hat GPT. So if you out there and you were in the hat fabricating business, our wardrobe when it comes to hats is looking a little dated. Yeah, send in a hat and our hats will be off to you. Okay, let's do it. People love the suser of the paper slips. The what? The suser. It's a great word. Look up ceration. Suser. Yes, it's like sort of like the sound of air blowing through trees. It was in that song 'Hateration.' Su. Mary J. famously said 'suserations' in her hit song 'Family Affair.' By the way, have you ever seen Mary J. Blige perform live? No. It'll change your life. Yeah, absolutely incredible. Kevin, select the first slip. First up out of the hat, this one is called 'AI video of Trump and Musk appears on TVs at HUD building.' This is from my colleagues at the New York Times. HUD is, of course, the Department of Housing and Urban Development. And on Monday, monitors at the HUD headquarters in Washington, D.C. briefly displayed a fake video depicting President Trump sucking the toes of Elon Musk, according to department employees and others familiar with what transpired. The video, which appeared to be generated by artificial intelligence, was emblazoned with the message 'Long live the real king.' Cas? Did you make this video? Was this you? This was not me. I would be curious to know if Grok had something to do with this, that rascally new AI that Elon Musk put out. Yeah, live by the Grok, die by the Grok. That's what I always say. Now, what do you make of this, Kevin, that folks are now using AI inside government agencies? I mean, I feel like there's an obvious sort of sabotage angle here, which is that as Elon Musk and his minions at DOGE take a hacksaw to the federal workforce, there will be people with access to things like the monitors in the hallways at the headquarters building who decide to kind of take matters into their own hands, maybe on their way out the door, and do something offensive or outrageous. I think we should expect to see much more of this. I mean, I just hope they don't do something truly offensive and just show x.com on the monitors inside government agencies. You can only imagine what would happen if people did that. So I think that Elon and Trump got off lightly here. Yeah, what is interesting about Grok though is that it is actually quite good at generating deepfakes of Elon Musk. And I know this because people keep doing it. But it would be really quite an outcome if it...

结果发现，用 Grok 制作的深度伪造的主要受害者其实是 Elon Musk。别生成了。好了，这里有个消息：Kevin，Perplexity 预告了一款名为 Comet 的网络浏览器。这是 TechCrunch 报道的。周一，该公司在 X 上发布了一个浏览器的注册列表，但该浏览器尚未可用。目前还不清楚它何时推出或长什么样，但我们确实知道名字，叫 Comet。

Turns out that the main victim of deepfakes made using Grok is in fact Elon Musk. Stop generating. Well, here's something: Kevin, Perplexity has teased a web browser called Comet. This is from TechCrunch. In a post on X Monday, the company launched a signup list for the browser, which isn't yet available. It's unclear when it might be or what the browser will look like, but we do have a name, and it's called Comet.

嗯，我对此无可奉告。

Well, I can't comment on that.

你这是在给“无可奉告”吗？是啊，我是说，我觉得 Perplexity 是目前最有趣的 AI 公司之一。他们一直在以越来越高的估值融资。他们正在挑战 Google——世界上最大、最富有、最成熟的科技公司之一，试图打造一个 AI 驱动的搜索引擎。而且看起来进展顺利，所以他们还在做其他事情，比如尝试做浏览器。这感觉就像是每个雄心勃勃的互联网公司的终极 Boss。每个人都想做，但没人能做成，你知道吗？Kevin，不只是 AI 浏览器。Perplexity 本周还宣布推出一个 5000 万美元的风险基金，用于支持早期初创公司。我想问的是：他们侵犯互联网上所有出版物的版权还不够吗？还要建一个 AI 网络浏览器，再变成一家风投公司？有时候我看到一家公司这样做，我会想，“哇，他们真有野心，有些大想法。”其他时候我觉得这些人是在瞎折腾。我把这一系列公告看作是在往墙上扔意大利面。如果我是 Perplexity 的投资者，我对他们的浏览器或风险基金都不会太兴奋。

You're giving it a no comment? Yeah, I mean, look, I think Perplexity is one of the most interesting AI companies out there right now. They've been raising money at increasingly huge valuations. They are going up against Google, one of the biggest and richest and best established tech companies in the world, trying to make an AI-powered search engine. And it seems to be going well enough that they keep doing other stuff, like trying to make a browser. That does feel like the final boss of every ambitious internet company. It's like everyone wants to do it and no one ends up doing it, you know? Kevin, it's not just the AI browser. Perplexity also said this week they are launching a $50 million venture fund to back early-stage startups. And I guess my question is: is it not enough for them to just violate the copyright of everything that's ever been published on the internet? They also have to build an AI web browser and turn into a venture capital firm? Like, sometimes when I see a company doing this, I think, 'Oh wow, they're really ambitious and they have some big ideas.' Other times I think these people are flailing. I see these series of announcements as spaghetti at the wall. And if I were an investor in Perplexity, I would not be that excited about either their browser or their venture fund.

这就是为什么你不是 Perplexity 的投资者。你可以说我感到困惑。别生成了。

And that's why you're not an investor in Perplexity. You could say I'm perplexed. Stop generating.

好了，Meta 在裁员 5% 后批准了更大的高管奖金计划。Casey，你知道我们喜欢 Hat GPT 上的正能量故事。我确实喜欢，因为有些 Meta 高管正想买太浩湖的第二套房子，但还没买得起。哦，他们已经在买第四、第五套了，现实点吧。好吧，这个故事来自 CNBC。根据周四的一份文件，Meta 的高管在新的高管奖金计划下可以获得基本工资 200% 的奖金，高于之前的 75%。新奖金计划的批准是在 Meta 开始裁减 5% 员工（据称会影响低绩效者）一周后。还有一个小括号：更新后的计划不适用于 Meta CEO Mark Zuckerberg。哦天哪，Mark Zuckerberg 得做什么才能加薪？他在吃罐头豆子，我跟你说。所以，这就是这个故事有趣的地方。这只是另一个故事，说明我们一直在讨论的一个话题：权力天平已经多么远离工人。你知道，两三年前，劳动力市场在硅谷实际上有很大的影响力。它可以影响诸如“我们希望让这个工作场所更多元化，对吧？我们希望在这个工作场所实施某些政策。”之类的事情。像 Mark Zuckerberg 这样的人实际上不得不倾听，因为劳动力市场非常紧张，如果他们说不，那些人可以去别处。现在不再是这样了。越来越多地，你看到像 Meta 这样的公司展示肌肉，说：“嘿，你要么接受，要么滚蛋。”这是一个真正的“滚蛋”时刻：我们裁掉你们中的 5%，然后我们给自己发奖金。别生成了。

All right, Meta approves plan for bigger executive bonuses following 5% layoffs. Now, Casey, you know we like a feel-good story on Hat GPT. I did, because some of those Meta executives were looking to buy second homes in Tahoe that they hadn't yet been able to afford. Oh, they're on their fourth and fifth homes, let's be real. Okay, this story is from CNBC. Meta's executive officers could earn a bonus of 200% of their base salary under the company's new executive bonus plan, up from the 75% they earned previously, according to a Thursday filing. The approval of the new bonus plan came a week after Meta began laying off 5% of its overall workforce, which it said would impact low performers. And a little parenthetical here: the updated plan does not apply to Meta CEO Mark Zuckerberg. Oh God, what does Mark Zuckerberg have to do to get a raise over there? He's eating beans out of a can, let me tell you. Yeah, so here's why this story is interesting. This is just another story that illustrates a subject we've been talking about for a while, which is how far the pendulum has swung away from worker power. You know, two or three years ago, the labor market actually had a lot of influence in Silicon Valley. It could affect things like, you know, 'We want to make this workplace more diverse, right? We want certain policies to be enacted at this workplace.' And folks like Mark Zuckerberg actually had to listen to them because the labor market was so tight that if they said no, those folks could go somewhere else. That is not true anymore. And more and more, you see companies like Meta flexing their muscles and saying, 'Hey, you can either like it or you can take a hike.' And this was a true take-a-hike moment: we're getting rid of 5% of you, and we're giving ourselves a bonus for it. Stop generating.

好了，苹果在接到后门命令后，从英国移除了一项云加密功能。这是彭博社报道的。苹果正在移除其在英国最先进的云数据加密安全功能，此前英国政府命令该公司为访问用户数据建立一个后门。所以这件事有点复杂，但非常重要。苹果在过去几年推出了一项名为“高级数据保护”的功能。这项功能是为国家元首、活动人士、异见人士、记者等数据面临被 NSO 集团等公司间谍软件攻击高风险的人群设计的。当苹果发布这个功能时，我非常兴奋，因为如果你属于这些类别，安全使用 iPhone 非常困难。然后英国政府来了，他们说：“我们命令你创建一个后门，以便我们的情报部门可以监视全世界每一个 iPhone 用户的手机。”对吧？这是苹果在美国和国外长期抵制的事情。所有人都关注苹果会怎么做。他们说的是：“我们只是要撤回这一个功能。我们将在英国使其不可用，并希望英国能明白我们的意思，停止对我们施加压力。”我认为苹果值得称赞，因为他们坚持了底线，没有建立后门。我们将看看英国如何回应。但我认为存在一种可能性：英国施加更多压力，苹果说“再见”，然后实际上将其设备从英国撤出。这对苹果来说就是这么严重。而且我认为这对互联网加密和安全通信的未来至关重要。说得好，King。我没什么要补充的，没有意见。

All right, Apple has removed a cloud encryption feature from the UK after a backdoor order. This is according to Bloomberg. Apple is removing its most advanced encrypted security feature for cloud data in the UK, which is a development that follows the government ordering the company to build a backdoor for accessing user data. So this one is a little complicated. It is super important. Apple in the last couple years introduced a feature called Advanced Data Protection. This is a feature that is designed for heads of state, activists, dissidents, journalists, folks whose data is at high risk of being targeted by spyware from companies like the NSO group, for example. And I was so excited when Apple released this feature because it's very difficult to safely use an iPhone if you are in one of those categories. And along comes the UK government and they say, 'We are ordering you to create a backdoor so that our intelligence services can spy on the phones of every single iPhone owner in the entire world.' Right? Something that Apple has long resisted doing in the United States and abroad. And all eyes were on Apple for what they were going to do. And what they said was, 'We are just going to withdraw this one feature. We're going to make it unavailable in the UK, and we're going to hope that the UK gets the message and they stop putting this pressure on us.' And I think Apple deserves kudos for this, for holding a firm line here, for not building a backdoor. And we will see what the UK does in response. But I think there's a world where the UK puts more pressure on Apple and Apple says, 'See ya,' and actually withdraws its devices from the UK. It is that serious to Apple. And I would argue it is that important to the future of encryption and safe communication on the internet. Go off, King. I have nothing to add, no notes.

是啊，你觉得这会导致我们和英国再打一场革命战争吗？这么说吧：我们赢了第一次，第二次我也看好我们的胜算。别来惹我们，英国。别生成了。

Yeah, do you feel like this could lead us into another revolutionary war with the UK? Let's just say this: we won the first one, and I like our odds the second time around. Do not come for us, United Kingdom. Stop generating.

本周 Hat GPT 的最后一个消息。AI 灵感无处不在，快把你的发型师逼疯了。这个故事来自《华盛顿邮报》，讲的是发型师、整形外科医生和婚纱设计师被要求根据不切实际的 AI 生成图像来为人们创作产品和服务。故事提到一位新娘让婚纱设计师根据她在网上看到的一张照片做一件礼服，那件礼服没有袖子、没有后背、领口不对称。设计师不得不遗憾地告诉客户，这件礼服违背了物理定律。不，我讨厌这样。我知道，作为准新娘，当你终于有了完美礼服的创意，拿给设计师看，却发现它违反了所有已知的物理定律，这太令人沮丧了。在 AI 出现之前，这不会发生在我们身上。是啊，我还以为这个故事是关于人们要求给手加上第六根手指，以便模仿他们在网上看到的 AI 生成图像呢。我喜欢这样的想法：把自己的照片提交给 AI，然后说，“给我剪一个 M.C. Escher 风格的发型”，你知道，就像无限楼梯相互融合那样。

One last slip from the Hat this week. AI inspo is everywhere, it's driving your hair stylist crazy. This comes to us from the Washington Post, and it is about a trend among hair stylists, plastic surgeons, and wedding dress designers that are being asked to create products and services for people based on unrealistic AI-generated images. So the story talks about a bride who asked a wedding dress designer to make her a dress inspired by a photo she saw online of a gown with no sleeves, no back, and an asymmetric neckline. The designer had to unfortunately tell the client that the dress defied the laws of physics. No, I hate that. I know, it's so frustrating as a bride-to-be when you finally have the idea for a perfect dress and you bring it to the designer and you find out this violates every known law of physics. And that didn't used to happen to us before AI. Yeah, I thought the story was going to be about people who asked for like a sixth finger to be attached to their hands so they could resemble the AI-generated images they saw on the internet. I like the idea of like submitting to an AI a photo of myself and just say, 'Give me a haircut like in the style of M.C. Escher,' you know, just sort of like infinite staircases merging into each other.

就把那个拿给我理发师看，说‘看你能做点什么’。是啊，这比我跟理发师说的要好，我通常就说‘两边和后面用三号推子，上面剪掉一英寸’。就说‘你能拿这个怎么办都行，我不抱太大希望’。是啊，在我头上解决黎曼猜想。

Just bringing that to the guy who cuts my hair and saying, 'See what you can do.' Yeah, you know, that's better than what I tell my barber, which is just, 'Number three on the sides and back, an inch off the top.' Just saying, 'Whatever you can do for this, I don't have high hopes.' Yeah, solve the Riemann hypothesis on my head.

顺便问一下，黎曼猜想是什么？

What is the Riemann hypothesis, by the way?

很高兴你问了。好，很好。凯文现在没在电脑上查；他只是深吸一口气，从脑海深处把它召唤出来。黎曼猜想：这是数学中最著名的未解决问题之一。显然，这是一个关于素数分布的猜想，它指出黎曼ζ函数的所有非平凡零点的实部都等于 1/2。句号。

I'm glad you asked. Okay, great. Kevin is not looking this up on his computer right now; he's just taking a deep breath and summoning it from the recesses of his mind. The Riemann hypothesis: it's one of the most famous unsolved problems in mathematics. It's a conjecture, obviously, about the distribution of prime numbers, that states all non-trivial zeros of the Riemann zeta function have a real part equal to 1/2. Period.

但问题是：我其实认为把 AI 灵感带给你的设计师和造型师是件好事，凯文。

Now here's the thing: I actually think it is a good thing to bring AI inspiration to your designers and your stylist, Kevin.

哦对，是的，因为问题是：这些工具之所以酷或有趣，部分原因在于它们让人们感觉更有创造力，对吧？如果你一直对自己的发型、室内设计，或者过去几场婚礼的布置都千篇一律，想要升级一下，为什么不借助 AI 说‘你能做这个吗？’如果答案是不可能，希望你能做个大度的顾客，说‘好吧，那可行的版本是什么？’

Oh yeah, yes, because here's the thing: to the extent that any of these tools are cool or fun, one of the reasons is they make people feel more creative, right? And if you've been doing the same thing with your hair, or with your interior design, or with your wedding for the last few weddings that you've had, and you want to upgrade it, why not use AI to say, 'Yeah, can you do this?' And if the answer is it's impossible, hopefully you'll just be a gracious customer and say, 'Okay, well, what's a version of it that is possible?'

我最近得知你在和一位造型师合作。是的，没错。

I recently learned that you are working with a stylist. I am, yes, that's right.

这是他们的杰作吗？

Is this their handy work?

不，我们下周才第一次见面。好的。那你打算用 AI 吗？不，计划是只用传统的人类智慧。但现在你让我开始想了，也许我可以带一堆不可能实现的造型去惹恼我的造型师。

No, we have our first meeting next week. Okay. And are you going to use AI? No, the plan is to just use good old-fashioned human ingenuity. But now you have me thinking, and maybe I could exasperate my stylist by bringing in a bunch of impossible-to-create designs.

对，问题是：我不需要什么不可能的东西。我只是需要帮忙找到一种在这个演播室里看起来好看的颜色，因为我确信没有颜色能做到。是真的。我们今天都穿了蓝色。这里有面蓝墙。这……这不行。蓝色是我最喜欢的颜色。我觉得我穿蓝色很好看。但把它放在这个颜色旁边——我真的叫不出这个颜色，也没法形容。我觉得任何蓝色都不好看。我觉得任何颜色在这个颜色旁边都不好看。这是一种没有名字的颜色。那么造型师能帮上忙吗？我们拭目以待。是啊，敬请期待。这就是为什么你应该每周都继续收听 Hard Fork 播客。每期都有新发现，有节目的传说，有世界构建。是啊，我们什么时候才能最终知道造型师、热水浴缸时光机等等发生了什么？是啊，敬请期待。

Yes, here's the thing: I don't need anything impossible. I just need help finding a color that looks good in this studio, because I'm convinced that nothing does. It's true. We're both in blue today. It's got a blue wall. It's not... it's not going. Blue is my favorite color. I think I look great in blue. But you put it against whatever this color is—I truly don't have a name for it, and I can't describe it. I don't think any blue looks good. I don't think anything looks good against this color. It's a color without a name. So can a stylist help with that? We'll find out. Yeah, stay tuned. That's why you should always keep listening to the Hard Fork podcast every week. There's new revelations, there's sort of the lore of the show, the world-building. Yeah, when will we finally find out what happened with the stylist, the hot tub time machine, etc.? Yeah, stay tuned.

好了，这是 ChatGPT 的回答。谢谢参与。通常表现更好一些。确实，但凯文这周过得很艰难。哦，是我的错。不，不是你的错。不是你的错。我能看出来你压力很大。哦，是啊，怪那个又累又没睡好的家伙。而且，所有故事都很悲伤。是啊，还有一个人在 BART 上死了。我确实解释了黎曼猜想，而且……读者一直在问。是啊，我希望你满意。顺便说一句，你要是能证明这个猜想就好了。

Okay, that was ChatGPT. Thanks for playing. It usually goes better than that. It does, but Kevin's had a hard week. Oh, it's my fault. No, it's not your fault. It's not your fault. I can just tell you have the weight of the world on you. Oh, oh yeah, blame the guy who's tired and didn't sleep last night. Also, all the stories were sad. Yeah, and a man died on BART. I did get to explain the Riemann hypothesis, and that... and readers have been asking. Yeah, I hope you're happy. By the way, it'd be nice if you could prove the hypothesis.