直播：Anthropic联合创始人谈人工智能与就业

本周《NPR政治播客》精彩内容。

This week on the NPR Politics Podcast.

几十年来，南方贫困法律中心一直追踪甚至渗透仇恨组织，但司法部现在指控他们资助这些活动的方式构成了银行欺诈。

For decades, the Southern Poverty Law Center has tracked and even infiltrated hate groups, but the justice department now alleges the way they funded that work amounted to bank fraud.

这到底是公正的追责，还是特朗普司法部针对政治对手的又一例证？

Is it an honest pursuit of justice or just the latest example of the Trump DOJ targeting the president's political opponents?

本周别忘了收听《NPR政治播客》。

Listen this week to the NPR politics podcast.

这是来自NPR的《金钱星球》。

This is Planet Money from NPR.

大家好。

Hey, everyone.

过去几周，《金钱星球》团队一直在全国各地跑书展巡讲。

For the past couple weeks, the Planet Money team has been on the road on book tour.

今晚，现场直播来自纽约的《金钱星球》新书发布会。

Tonight, live from New York, the Planet Money book launch.

我现在在旧金山。

I'm in San Francisco.

这儿怎么着？

What's going

啊？

on?

太兴奋了，终于到波士顿了。

So excited to be in Boston.

大家好。

Hello, everyone.

谢谢各位。

Thank you

那咱们也跟西雅图打个招呼吧？

so say hello, Seattle?

书展巡演还能这么玩？

Is that a thing we get to do on book tour?

是的。

Yeah.

对。

Yeah.

大家好。

Hello.

我们一直在全美各地巡回做现场演出，为《星球金钱》新书造势。

We have been crisscrossing America doing live shows to help promote the new Planet Money book.

我们去过纽约、华盛顿、波士顿、旧金山、西雅图、洛杉矶。

We've been in New York, DC, Boston, San Francisco, Seattle, Los Angeles.

这种机会其实挺难得的，我们平时很少能这样。

And this is a rare thing that we don't get to do very often.

平时很难有机会跟这么多听众面对面交流。

We don't get to meet and talk with so many of you.

在华盛顿的时候，我们跟观众玩了个游戏，来演示‘群体智慧’这个概念。

In DC, we played this game with the audience to demonstrate a concept known as the wisdom of the crowds.

这个观点是说，如果你让一大群人猜某样东西的价格，比如一幅画，他们的平均猜测值往往会非常接近真实价格。

This is the idea that if you ask a large group of people to guess the price of something, like a piece of art, the average will actually be pretty close to correct.

你们都猜这幅画值……得了吧，各位。

You all guessed that this piece of art is worth oh, come on, guys.

六十九万亿美元？

$693,000,000,000,000?

我觉得这不太对。

I don't think that's right.

这研究也不是完美的。

It wasn't a perfect study.

在纽约，我们的主持人阿曼达·奥伦奇克和玛丽·奇尔兹回答了观众的问题，讲我们是怎么选题的，哪些故事报，哪些不报。

In New York, our hosts Amanda Orenczyk and Mary Childs answered audience questions about how we choose what stories to report or not report.

我翻了翻原本打算做的那些节目清单，结果其中一个居然就叫‘可卡因’。

It was weird to look at the list of episodes that I thought I was gonna make, and one of them was just cocaine.

我们还有一个老梗，就是互相大喊‘毒品’，因为我们老想做一期‘毒品是怎么合法化的’。

We also have a running thread where we just yell drugs at each other because we keep wanting to make a how does a drug become legal episode.

是的。

Yes.

但律师们说不行。

But the lawyers have said no.

每到一站，我们都特别开心能遇到这么多从听众变成读者的朋友。

And at every stop, we have gotten the true delight of meeting so many of our listeners turned readers.

其中有些人，感觉简直该自己写本经济学书了。

Some of whom it seems like maybe should be writing their own economics books.

我叫瑞安，很高兴你们聊到了货运航运，因为我特别想听听你们对《琼斯法案》的看法——对不太了解海运的人，我简单解释一下。

My name is Ryan, and I'm really glad you were talking about cargo shipping because I'd like to hear your thoughts on the Jones Act, which was briefly and for the long for those who aren't, like, maritime shipping nerds, I'll just contextualize briefly.

《琼斯法案》是

The Jones Act is

一个条款。

a subsection.

对。

Yeah.

其实，你真该说说。

Actually, please do.

替我们把活儿干了。

Do our job for us.

我太喜欢这个了。

I love this.

这次书展巡讲我们还剩几站，但显然不可能去遍美国每个城市。

Now we do have a few stops left on this book tour, but obviously, we can't make it to every city in America.

其实连大多数城市都去不了。

Can't actually make it to most of them, really.

所以我们就想，不是每个人都能来参加书展。

And so we thought not everyone can come to the book tour.

不如我们把书展带到每个人身边。

Let's bring the book tour to everyone.

大家好，欢迎收听《星球金钱》。

Hello and welcome to Planet Money.

你好，欢迎来到《星球金钱》。

Hello and welcome to Planet Money.

你好，欢迎来到《星球金钱》。

Hello and welcome to Planet Money.

哇，太棒了。

Woah, wow.

这不错。

That's good.

你好，欢迎来到《星球金钱》。

Hello and welcome to Planet Money.

我是艾丽卡·巴里斯。

I'm Erica Barris.

在巡演的每个城市，我们都会采访一些特别嘉宾。

In every city on the tour, we've been doing these interviews with special guests.

他们有的是我们节目的常客，比如经济学家拉吉·切蒂和艾米莉·奥斯特，也有些是新闻人物，比如Anthropic的联合创始人杰克·克拉克。

They've been Planet Money regulars like economists Raj Chetty and Emily Oster, and some people in the news like the cofounder of Anthropic, Jack Clark.

到2027年4月，AI系统应该能完成一个人需要花150个小时才能做完的任务。

By April 2027, AI systems should be able to do tasks that might take a person a hundred and fifty hours.

那到底是什么任务啊？

Now that what is that?

这差不多是一个月的工作量，这意味着经济上得发生点离谱的变化。

That's almost a month's worth of work, which it requires strange things to happen in the economy.

对。

Yeah.

但具体是啥任务啊？你脑子里想的是哪种任务？

But, like, what task is it doing that's gonna like, what task are you thinking of?

你最近一次花了一个月才做完的工作是什么？

What's the last bit of work that took you a month?

给你写这些问题花的时间，比我预想的多太多了。

Writing these questions for you took more time than I wanted.

是啊。

Yeah.

对。

Yeah.

还有像经济学家达里尔·法韦尔这样的人，他们正努力帮普通人理解极其复杂的房地产市场。

And also people like economist Daryl Fairweather, who's trying to help regular people understand the massively complex housing market.

我觉得用‘抢椅子’来比喻房地产市场挺合适的。

So I think a good metaphor for the housing market is musical chairs.

那些已经买房的人，都稳稳坐在自己的椅子上。

Like, the people who already own houses are sitting in their chairs.

每到春季楼市活跃的时候，就有人站起来，换到另一把椅子上。

And whenever there's a new spring housing market, some people get up and they move and they select a new chair.

如果椅子总数不增加，想让首次购房者顺利进场，根本没那么容易。

And if you're not adding more chairs in, you can't get first time homebuyers into the market so easily.

今天这期节目，是我们线下读书会的两站——旧金山和西雅图，我们会聊聊AI焦虑，以及你——没错，就是你——如何成为解决住房危机的关键。

Today on the show, two stops from our live book tour, San Francisco and Seattle, where we'll talk about AI anxiety and how you, yes, you, could be the key to fixing the housing crisis.

本节目由第一资本赞助。

This message comes from Capital One.

Capital One 提供零费用、无最低余额的支票账户。

Capital One offers checking accounts with no fees or minimums.

你钱包里装的是啥？

What's in your wallet?

详情请见官网条款。

Terms apply.

了解更多请访问 capital1.com/bank。

See capital1.com/bank for details.

Capital One NA，FDIC成员。

Capital One NA, member FDIC.

首先我们来到旧金山，见到了 Anthropic 的联合创始人杰克·克拉克。

So first stop, San Francisco, where we got to talk with the cofounder of Anthropic, Jack Clark.

你知道，Anthropic 的 AI 模型，比如 Claude 和最新的 Claude Mythos，正引发一堆棘手的问题：AI 到底能做什么，不能做什么。

As you know, Anthropic and their AI models, Claude and now Claude Mythos, are raising all sorts of thorny questions about what AI can do and what it can't.

我们当然也打算问问杰克这些事。

And we were, of course, going to ask Jack about all of that.

但我们想先聊聊他创办Anthropic之前几年在网上发表的一系列短篇故事，这些故事能让我们看到他对未来的思考方式。

But we wanted to start with some writing he published online several years before starting Anthropic, a series of short stories that give some perspective into how he thinks about the future.

接下来是他和《Planet Money》主持人肯尼·马隆的对话。

Here he is with Planet Money host, Kenny Malone.

把这些称作科幻小说合适吗？

Is it fair to call these science fiction?

称它们为科幻小说是完全合适的。

It is fair to call them science fiction.

我想给大家简单感受一下里面的内容。

I'd like to give people a little taste of what's in here.

我想让你读一段特别有《Planet Money》风格的，讲的是2025年，我们会原谅你，因为你当时有点儿想错了——其实那句有点儿不对劲。

So the one I want you to read that feels very Planet Money, it imagines in the year 2025 we'll forgive you because you're it was a little you're a little wrong on that one.

但在2025年，你设想了一家在中国的工厂。

But in the year 2025, you imagine a factory in China.

这条生产线由机器人负责，而AI则在背后管理着这些机器人。

It is overseen by a robot factory line, and AI is overseeing the robots.

然后AI还会试着想象出一些全新的产品类型。

And then AI is also trying to imagine new types of products.

如果你只想读一读的话——

And then if you just wanna read

一旦发现新的候选产品，它就会调用一整套优化过的专家系统来评估，要是得分够高，就会在高精度模拟环境中测试这个产品。

When new candidate products are found, it runs for a series of tuned expert systems, and if it gets a high enough score, simulates the product in a high fidelity simulation.

如果能通过这些测试，系统就会生成一个实物样品，用无人机空运给附近的真人焦点小组，要是他们觉得满意，产品就会放到一个类似eBay的拍卖平台上，供工厂成千上万的分销商竞拍。

If it passes these tests, then a candidate product is produced, airlifted by drone to a nearby human focus group, and if it satisfies their criteria, is sold on an eBay like auction site frequented by the factory's thousands of distributors.

接着进入竞价环节，几天或几周后，就会收到市场反馈，看这产品卖得怎么样。

A bidding process takes place and in a few days or weeks, data comes back about how the product succeeds in the market.

要是它表现比现有产品更好，工厂就会把更多资源投入到这种风格的新产品研发上，这条即兴生产线也会开始探索这些新产品背后的潜在可能性。

If it does better than the existing product, more parts of the factory are dedicated to creating new products in this style, and the improvisation line begins exploring the latent space of the new products.

明白了。

Okay.

所以我们想象的世界是这样的：AI不仅能控制制造产品的机器人，还能自己创新、设计产品，把样品送到人类焦点小组，再卖给普通人。

So, we imagine a world where AI is able to both control the robots that make the product, innovate and design the products, ship them off to human focus groups, and then sell them to humans.

那我问你个问题。

So here's my question for you.

在这整个故事里，人类到底在哪儿？

Where are we the humans in this story?

我们是不是在田野里跳舞，晒着太阳吃椰枣？

Are we off in a field dancing, eating dates in the sun?

更重要的是，我们没参与生产，哪来的钱买这些产品？

And maybe more importantly, how do we have money to buy the products that we didn't help make?

这可是个正经问题。

Sort of a serious question.

这问题问得好。

That's a good it's a good question.

我觉得，如果AI真发展到大家想的那样，咱们得彻底重新想想资本主义到底是怎么运作的。

So I think I think if AI goes as far as people think, you actually need to re reconceptualize how how capitalism in the largest possible sense works.

要是真到了一个全是机器自产自销、人只负责买东西的世界，那人们总得有钱才行吧。

I think if you end up in a world where you have a closed loop production system with just machine to machine to machine to machine, and then people buy stuff, people need money.

早就成定局了。

Well established.

说的就是这个。

That is the thing.

没错。

Yep.

所以你得对机器人和AI公司大幅征税，还得想办法把钱重新分配出去。

So you need to tax the robots and AI companies significantly, and you need to somehow find a way to reallocate Yeah.

把机器经济里的钱，转到人类经济里来。

Money from this machine economy to the human economy.

明白了。

Okay.

我读这本书的时候想到一个问题：普通人能理解这个吗？

Now one of the things that occurred to me in reading this book is can people get this, by the way?

这事儿真能成吗？

Is this like a thing?

你可以在 jackclark.net 免费拿到电子版，不过我还没想好怎么在这儿卖纸质书。

You can get it for free online at jackclark.net, but I haven't figured out how to sell sell the book here yet.

每次我一打算出本正经的书，就老被别的事岔开，上回就是刚开始搞 Anthropic，忙起来了。

Each time I try and make an actual book to sell, I get sidetracked, and the last time was when I was starting Anthropic and when I got busy.

Surely 肯定会出你的书，杰克。

Surely, will publish your book, Jack.

我我我

I'm I'm I'm

我记在待办清单上了。

It's on my list.

这事一直搁我心上，挥之不去。

It plagues me.

我有个 Google 文档，空闲时候就盯着它发呆。

I have a Google Doc that I stare at on my free moments.

我做过不少关于图书出版的报道，觉得你这种人，出版社肯定抢着要。

Having gone through a series of reporting on on books and how they get made, I suspect you'd be an attractive candidate to a publisher.

我读你这本书的时候想到一点，科幻作家本质上就是在想象未来，预测未来。

One of the things that did occur to me in reading your book is that science fiction writers, by definition, are imagining a future, they're predicting a future.

但我觉得你可能是唯一一个同时在塑造未来的人——我们所有人现在生活的世界，正是你创造的技术所构建的。

But you may be genuinely the only one that I can think of who is simultaneously shaping the future, and we are all living in the world that your technology is creating.

所以我真的在想，这会不会改变我读这本书的方式？

And so I do genuinely wonder if that changes how I should read this book.

我是不是该把你的文字当成蓝图、操作指南、警钟，或者如何生活在杰克·克拉克未来世界的生存建议？

Like, should I think of your writing as a blueprint, as instructions, as a warning, as advice on how to live in Jack Clark's future?

是的。

Yes.

有一支叫Jawbreaker的乐队，他们有句歌词是：我的虚构比我的真实更有力。

There's there's a there's a band Jawbreaker, and they have a a lyric which says, my fiction beats the hell out of my truth.

我的通讯稿大部分是在写AI研究论文。

And so, the newsletter is mostly me writing about AI research papers.

但那些虚构故事，才是我真正试图面对自己在这其中的道德与责任问题的地方。

But the fiction stories are where I'm basically trying to grapple with what's happening in my moral and ethical responsibility in it.

所以你可以读这些故事，我觉得它们就像我从这个有点吓人的AI实验室里扔出去的漂流瓶，而我本人正是这个实验室里的核心人物。

So you can read them, and I view it as like messages in a bottle I'm trying to throw out of this semi frightening AI lab, which I'm I'm a principal character in.

你其实是在提醒我们，但又不是在提醒我们。

Are you and so you are warning us, and you're warning us, you're not warning us.

难怪你刚才说‘是的’。

I guess that's why you said yes.

这真让人兴奋。

It's exciting.

真的吗？有吗？

And it's Is is is it?

如果真算兴奋的话，我觉得这……

It's very if it's exciting, I think it's

特别让人兴奋。

very exciting.

大家都觉得兴奋吗？

Is everyone excited?

这大概就是我预料中的回应了，

That's that's about the response I expected, to

说实话。

be honest.

那些缺陷还挺诡异的。

Eerie of flaws.

是的。

Yeah.

但现在已经不一样了。

But it's changed.

对吧？

Right?

随之而来的，也有一丝不安。

And with that comes some amount of trepidation.

对吧？

Right?

是的。

Yes.

对我来说， personally 感觉特别忐忑。

For me, a lot of trepidation personally.

可能只是我这样想。

That may just be me.

我不确定。

I don't know.

我们今天早些时候聊过。

We talked earlier today.

这互动真是挺有意思的。

This is a very funny interaction.

我今天早些时候简短地见了杰克。

So Jack I met Jack briefly early today.

这简直是让人想赶紧撤的原因。

This is the greatest reason to run out of room.

他跟我说，我得赶紧跑出去做个预测。

He's like, I must run and make some predictions.

我当时就想，这到底在搞什么名堂？

And I was like, what is happening?

你后来停了一下，说你之前做了一连串预测，都会在九月应验，而且全都准得离谱，对吧。

And you you paused to briefly explain that you had made a series of predictions essentially that would come true in September, and they had they had done very well Yeah.

然后你又得去搞新的一轮。

And you had to go make another set.

那你那些预测都是啥？

So what were your predictions?

行。

Okay.

有两个。

There's two

现在跟我们说说吧。

Share them with us now.

这两条在Polymarket上能看到吗？

Two interesting see any of these on Polymarket though?

我们知道这些预测的来源。

We know where they came from.

其中一个，我觉得，如果你看时间线，到4月27号，AI系统能完成人类要花150小时才能做完的任务，这其实并不离谱。

Well, one, I think, again, if you do the time thing, it doesn't seem wild that by April 27, AI systems should be able to do tasks that might take a person a hundred and fifty hours.

那这到底是个啥任务？

Now that what is that?

这差不多是一个月的活儿了，得经济上发生点特别的事儿才行。

That's almost a month's worth of work, which it requires strange things to happen in the economy.

对。

Yeah.

但具体是啥任务啊？你脑子里想的是哪种活儿？

But, like, what task is it doing that's gonna like, what task are you thinking of?

你最近一次花了一个月才干完的活儿，是啥？

What's the last bit of work that took you a month?

给你写这些问题花的时间比我预想的还多。

Writing these questions for you took more time than I wanted.

是啊。

Yeah.

对。

Yeah.

就是这种任务。

So tasks like this.

没错。

Yeah.

不是。

No.

我是说，那种可能让人花超长时间的事，比如设计复杂的电路、做需要大量阅读和交叉引用的深入研究项目，或者开发一个软件。

I mean, tasks that might take a person a really, really long time might be designing a complicated piece of circuitry, doing a very very involved research project involving reading and cross referencing many things, building a piece of software.

我们现在花大价钱请顶尖人才做的那些工作，恰恰正是AI正在慢慢渗透进来的领域。

The the kinds of tasks which we currently pay extremely talented people large amounts of money to do are exactly the type of tasks which AI systems are kind of creeping into.

所以这带来了一些影响。

So, that has some implications.

对。

Yeah.

我们之前还预测过，你知道吗？Anthropic 里大部分代码都是由像 Claude 这样的 AI 写的，当时在场的我们都觉得，到四月份，可能 100% 的代码都得靠 AI 来写，于是我们就问自己：那我们到底在干嘛？

The other prediction we made was, you know, we've talked about how the majority of the code at Anthropic is written by AI systems like Claude, well, all of us in the room thought by April, maybe a 100% of the code is, which led to us say, what are we doing?

于是我们搞了个‘代码评审小组’，大家围在一起，分析、挑刺 Claude 写的代码，确认它是不是对的。

And, I think that we invented a guild system where we would sit around analyzing and critiquing the code that Claude writes and verifying that it's correct.

我真搞不清楚，这到底是像高级俱乐部一样挺好玩的工作，还是说，这其实反映了某种更奇怪的现状。

And I can't work out if this is like a really fun country club style job or if it paints some different picture.

没错。

Right.

你接着说吧。

You're gonna tell me.

我也不太清楚，杰克。

I don't know, Jack.

我在这俩情绪之间来回摇摆，一会儿吓坏，一会儿又挺乐观的，不知道你们是不是也这样？

I find it all I waffle between horrified and optimistic, and I find I'm I'm just is that is that kind of where everyone else is?

对。

Yeah.

没错。

Yeah.

你们在哪干这个？

Where do you do that?

我总觉得，我们正面对着一种百年未遇的巨大变革，真不知道我们会怎么应对。

I I find myself staring into the the face of of change that we haven't experienced in perhaps a century, and I I don't know how we'll respond.

一百年前，C.

A hundred years ago, C.

S.

S.

刘易斯写过一本书叫《地狱来信》，讲的是地狱里一个恶魔，专门给上级——也就是个大恶魔上司——写信，汇报他们怎么试图腐蚀地球上的人。

Lewis wrote a book called The Screwtape Letters, which is about a demon in hell whose job is writing letters to their boss, a senior demon, a manager demon, about their attempts to corrupt a person on earth.

他们有一封信里写道：好消息。

And they have a letter where they say, wonderful news.

我的人类刚发明了一种叫电话的东西。

My human has just got an invention called the telephone.

有了电话，每当他独处、本该反省自己、靠近上帝的时候，就会拿起电话找朋友聊天，反而离地狱更近了。

And the telephone means that whenever my person is alone and would otherwise introspect on themselves and be drawn closer to God, they will instead pick up the phone and call their friends, which brings them closer to hell.

那这说法，哪些对、哪些错呢？

Now, what what's what does that get right and wrong?

对的是，像电话这样的东西，确实彻底改变了人与人之间的相处方式。

What it gets right is that things like phones actually massively changed how people relate to one another.

没错。

Yes.

但我并不觉得，有了电话，我们就注定下地狱了。

But, I don't think that we're all like doomed to hell because we have phones.

嗯，这个嘛。

Well.

嗯，

Well.

我现在意识到我举的例子不太合适，但至少它点出了问题的核心。

I realize now I picked the wrong example, but but it gets at the shape of this.

对吧？

Right?

我们会经历巨大的变化，而我们现在正盯着这些变化，努力思考它对人与人之间关系意味着什么，它一定会改变我们彼此相处的方式。

Like, we are gonna experience immense change, and we're looking at it, and we're trying to think through what it means for how we relate to us as people, and it will change how we re relate to one another.

是的。

Yeah.

我想暂时换个话题，聊聊你的写作，因为你最近上了不少新闻。

I wanna shift gears from your writing for a second because you made a lot of news lately.

是吗？

Yeah?

我的意思是，你知道的，那还挺让人兴奋的。

I mean, you know, it was exciting.

我的意思是，你最近确实上了不少新闻，尤其是那个叫Mythos的新AI模型。

I mean, you've made a lot of news lately, but most recently, there's a new AI model called Mythos.

如果我说得不对，你尽管纠正我。

Please correct me if I get any of this wrong.

听说这个Mythos特别擅长搞渗透，要是有人想拿它干坏事，那可太厉害了。

So Mythos and that seems to be really good at hacking if someone wanted to use it for that.

所以你还没把这模型公开出来。

And so you have not made this model public yet.

你只给了大概四五十家公司，让他们提前做好防范，防着你造的这个工具。

You've given it to something like 40 or 50 companies to I guess gird themselves from the tool that you made.

对。

Yes.

明白了。

Okay.

这些公司包括苹果、谷歌和亚马逊这些大厂。

So that's companies like Apple and Google and Amazon.

据我们所知，目前唯一有报道说在用的金融机构是摩根大通，可能在测试。

And as far as we could tell, the only financial financial institution that I've seen reported is JPMorgan maybe has it to to test.

但我听说财长斯科特·贝桑特给你打过电话，想搞这个工具。

But I believe treasury secretary Scott Besant has called you asking for this.

所以第一个问题是，这些电话打得怎么样？

So question one is, how have these phone calls been?

就是说……

Like

我现在对各种奇怪的电话已经挺麻木了。

I have a high tolerance for weird phone calls at this point in time.

是的。

Yeah.

我觉得我们这些对话的核心是：好几年来我们一直在说，AI会越来越强，直到在大多数任务上超过人类。去年它先在编程上做到了，

I think the nature of the conversations we're having is we have been saying for years that AI systems would get better, and they would keep getting better until they got better than people at most tasks, and it started in coding last year.

所以当它在别的领域也超过人类时，别太惊讶。

Don't be surprised when they get better at other tasks as well.

没错。

Right.

但在电话里这么说可不太讨喜。

And that doesn't go down super well on the call.

所以你就得换个说法，问：那我们该怎么办？

So then you frame it as, well, what do we do about this?

对。

Yeah.

我觉得问题在于，这并不是什么特殊模型。

And I think the challenge is this isn't a special model.

对吧？

Right?

这代表的是普遍情况。

This is representative.

这可不是你们那个黑客模型，对吧？

This isn't your hacking model, is it?

这就是

This is

不是我们平时用的Claude。

not it's our regular Claude.

它特别擅长搞黑客攻击。

And it's just it's really good at hacking.

现在Claude居然挺会黑客操作的。

Claude turns out to be good at hacking now

这挺酷的。

as That's cool.

你瞧。

It Look.

我在看呢。

I'm look.

真厉害。

That's cool.

我知道。

I I know.

你看啊。

I it's look.

人都是复杂的。

We contain multitudes.

我们完全可以既乐观又害怕。

We can be optimistic and terrified at the same time.

但关键是，以后会冒出很多这样的系统，世界会慢慢适应的。

But what but the what it means is there will be many systems like this, and the world is going to adapt to this.

对。

Yeah.

我们有机会用它让这个世界变得更安全。

We have a chance to use it to make much much of the world much more secure.

但与此同时，我们也得面对这些突然出现的新能力。

But at the same time, we now have these new capabilities in the world we have to contend with.

真的只有这一家金融机构参与其中，让其他人测试自己的系统吗？

There was is it true there's just the one financial institution in the mix of people see like getting to test their own systems?

我们推出了Forti，之后一直在不断扩展。

Or We announced Forti and we've been expanding it a bunch over time.

明白了。

Okay.

我不确定到底是只限一家，还是已经扩了，但目前确实有越来越多的人在加入。

So I don't know exactly if it's just for one or if we've expanded it already, but more more people are being added all the time right now.

我是不是得天天盯着银行账户，怕钱被偷偷转走？

Do I need to be checking my banking account to see if money is going away?

不用。

No.

好。

Okay.

除非是你自己在花钱。

Unless you're spending it.

我是有可能会花。

I mean, I might be.

旧金山可是花钱的好地方。

San Francisco is a great city for spending money.

吃的真的不错。

It's the food is good.

现在你们免费提供这个工具给公司，但我心里总有点嘀咕。

So right now, you're offering this tool to companies for free, but there is this feeling in the back of my mind.

就是说，

Like,

你们既造了矛又造了盾，那以后要是想收费，怎么办？

you have invented both the weapon and the shield, and so what happens when you wanna charge for this?

这感觉有点像在说：要是你们公司出事了，那可别怪我们，谁叫你们不买我们的产品呢。

It does feel a little bit like, hey, you know, be ashamed if something happens to that company of yours if you don't pay for our products.

是的。

Yeah.

我觉得，那种更像黑手党式的商业策略，很难长久撑下去。

I I we we're we think that business strategies which tend more towards mafia than anything else are probably not long term resilient.

所以，你不会这么做吧？

So, won't

我们确实不会那样做。

be doing that.

我觉得，未来AI的某些部分必须真正变成一种公共事业——比如网络防御能力，应该按成本价提供，甚至零利润。

I do think that the shape of this in the future is is parts of AI need to become actually a true utility where you would expect things like cyber defense capabilities to be something that you provide like at cost or at the cost of it costs you to provide you with no margin.

要是我的财务官听见我说这话，估计得吓一跳，但未来就是这样：当这些能力对社会至关重要时，你最终得这么干。

I'm sorry if my financial officer is listening to me say that, but it's the it's the shape of where you end up in the future where there are all these capabilities that matter for society.

你得让这些能力广泛普及，不能向公众收费，否则就会形成一个特别糟糕的激励机制。

You need to proliferate those into society without charging society for it or you end up in a in a really bad incentive structure.

明白了。

Okay.

你现在有两个孩子，刚休完产假回来。

You now have two two children who just got back from parental leave.

是的。

Yep.

所以你也在为他们设想一个未来。

So you're imagining a future a future for them as well.

对。

Mhmm.

那我想知道，你为他们设想的未来是什么样的？

And so I am wondering what is the future you're imagining for them?

你觉得他们会如何融入你正在创造的世界？

How do you think that they will fit into the world that you are creating?

你会优先培养他们哪些能力？

And what skills will you prioritize for them?

我的想法是，如果我们做对了，未来会更好——人们能花更多时间陪伴彼此，而不是被孤立，远离家人和朋友。

I mean, I imagine if we get this right, a better future where people get to spend more time, more time with one another and less time being kind of atomized and driven away from their families and friends.

这就要求你真正解决分发的问题，

This requires you to actually solve the distribution problem,

你指的是我本人？

which Me, personally?

对。

Yes.

就是你。

You.

恭喜你。

Congratulations.

咱们一起在解决经济问题。

We're solving economics together.

但要实现这一点，结合我之前说的倡导，你得彻底重新思考教育，从根本上颠覆它。

But how you get there along with the advocacy I've talked about is actually thinking you're gonna need completely rethink education and sort of upend it.

你得从根本上改变教育的运作方式，还得教会人们变得更富有好奇心。

You're gonna need to upend how how education works in pretty fundamental ways, and you're going to have to teach people to be much more curious.

我特别注意到，我孩子总是不停地问我问题。

I am struck by how my child constantly asked me questions.

一天被问五十遍同样的问题，确实挺让人抓狂的。

It's it's obviously sometimes very enraging when you get asked the same question 50 times a day.

但有时候又特别棒，比如你就会想，骆驼为啥长驼峰啊？这类问题多有意思。

But sometimes it's amazing, like, you know, why why are camels why do they have humps or anything like this?

我也被问过这种问题，当时我就懵了：我哪知道啊。

I got that one too and I was like, I don't know.

可长大以后，你那问问题的本事，早被死记硬背的教育、朝九晚五的工作，还有朋友一句‘天呐，杰克，别再问这种怪问题了’给磨没了。

But but as you go into being an adult, you get kind of the the the skill of asking questions beaten out of you by by rote learning and and working in regular jobs and your friends saying, for heaven's sake, Jack, like, stop asking these insane sounding questions.

但AI能回答这些问题，帮我们把小时候那种好奇心一直保留到成年，这真的特别开阔眼界，也特别美好。

AI actually can answer these questions for you and can allow us to maintain that kind of childlike curiosity into adulthood in a way that I think is, like, very mind expanding and and wonderful.

我觉得这会把世界带向特别激动人心的方向——每个人都能像孩子一样充满创造力，而且成年后还能继续用、被鼓励去用。

And I think that's gonna lead the world to, like, really exciting places where everyone has the the inventiveness that you have as a child, but you can still use it and are encouraged to use it more as an adult.

这就是我乐观的看法。

That's my optimistic story.

对。

Yes.

我想用你前言里的最后一句话来收尾，我真的推荐大家去读一读。

I do wanna end with the line the final line from your preface, which I do recommend reading.

不过你选的那句结尾话是：祝我们所有人好运。

But the final line you have chosen is good luck to us all.

对。

Yes.

那我们就停在这儿吧。

And I think we'll end on that one.

杰克，真的太感谢你了。

Jack, thank you so much.

非常感谢。

Thanks very much.

多谢了。

Appreciate it.

杰克·克拉克，Anthropic联合创始人，与《Planet Money》主持人马龙的对话。

Jack Clark, the cofounder of Anthropic, talking with Planet Money host, Malone.

稍微澄清一下，刚才提到的《地狱来信》那个引用基本没错。

A quick clarification, that reference to the Screwtape letters was basically correct.

在那本书里，C.S.刘易斯确实写过，噪音作为一种大规模的干扰，会分散人的注意力。

In it, c s Lewis does write about noise writ large as a distraction to man.

他谈科技产品弊端——包括无线收音机——是在另一封私人信件里，不是这本书。

It was in a different private letter that he talked about the downsides of gadgetry, including the wireless radio.

广告后回来，聊聊当亚马逊的行为经济学家是什么体验。

After the break, what it's like to be a behavioral economist for Amazon.

这条广告来自富国银行。

This message comes from Capital One.

富国银行的支票账户，零手续费，也不设最低余额。

Capital One offers checking accounts with no fees or minimums.

你的钱包里装的是什么？

What's in your wallet?

具体条款请见官网。

Terms apply.

详情请访问 capital1.com/bank。

See capital1.com/bank for details.

Capital One NA，FDIC成员。

Capital One NA, member FDIC.

欢迎回来。

Welcome back.

接下来再放一段我之前采访的片段。

We have another interview from the tour I'm gonna play you part of.

这次采访的是达莉尔·费尔韦瑟。

It's with Daryl Fairweather.

达莉尔是行为经济学家，也是《恨透了这个游戏：生活、爱情与工作的经济潜规则》一书的作者。

Daryl is a behavioral economist and author of the book, Hate the Game, Economic Cheek Codes for Life, Love, and Work.

我们的书里多次引用了她的话，尤其是关于住房的部分。

She's also quoted many times in our book, especially on the topic of housing.

去目录里查一下。

Check the index.