人工智能人才争夺战内幕

由Capital One VentureX信用卡为您呈现。如果您热爱旅行，Capital One有一款专为您设计的奖励信用卡。使用VentureX，您在所有消费中都能赚取无限双倍里程，将每笔消费转化为非凡旅程。通过Capital One Travel预订，您还能在多家豪华酒店享受尊贵礼遇。此外，全球超过1,000个机场贵宾厅将为您敞开大门。

Brought to you by the Capital One VentureX card. If you love to travel, Capital One has a rewards credit card that's perfect for you. With VentureX, earn unlimited double miles on everything you buy and turn all of your purchases into extraordinary travel. And you get premium benefits at a collection of luxury hotels when you book through Capital One Travel. Plus, you'll get access to over 1,000 airport lounges worldwide.

Capital One，你的钱包里有什么？条款适用。详情请访问capital1.com。

Capital One, what's in your wallet? Terms apply. See capital1.com for details.

这里是《纽约时报》的娜塔莉·基特罗夫，您正在收听《每日播报》。人工智能领域的霸主之争已演变成全面人才战争，科技公司开出2.5亿美元的天价薪酬包，并大肆挖角竞争对手的核心员工。今天，我的同事迈克·艾萨克将为我们解析这场狂热背后的成因，以及为何这种炒作正引发人们对AI可能成为下一个巨大泡沫的担忧。今天是8月25日，星期一。

From The New York Times, I'm Natalie Kitroef. This is The Daily. The race to dominate artificial intelligence has become an all out talent war, with tech companies offering pay packages of a quarter billion dollars and poaching their competitors' best employees. Today, my colleague Mike Isaac on how we got here and why all the hype is raising fears that AI could become the next big bubble. It's Monday, August 25.

迈克，我们邀请你是因为过去几周乃至数月，在你负责报道的硅谷领域，我们目睹了真正惊人的现象。我说的正是这些动辄2.5亿美元的巨额薪酬包，它们正被随意地砸向AI人才。对很多人来说，这就像突然抬头发现：等等，到底发生了什么？所以我需要你帮我理清头绪。

Mike, we are here with you because in the last few weeks and months, we've started to see something truly extraordinary happen in the world you cover, which is Silicon Valley. I'm talking about these enormous pay packages, 250,000,000 being handed out left and right to AI talent. And I think at a really basic level, for a lot of us, this has been a moment where we looked up and said, wait. What's going on here? And so I want you to help me understand.

为什么会发生这种情况？

Why is this happening?

别担心，这里的人也觉得这太疯狂了。所以你并不孤单。但这确实是硅谷的特殊时刻——所有科技巨头领袖都认为人工智能是即将到来的革命性技术。他们掀起了这场我称之为'军备竞赛'的浪潮，重组公司架构以打造下一代伟大AI产品，却突然意识到既缺乏所需人才，

Well, just to reassure you, people here think it's pretty crazy too. So it's it's you're not alone. But I do think it's really a singular moment in Silicon Valley right now where all of the leaders of these big tech companies see artificial intelligence as the next big thing coming around the corner. And so they kicked off what I see as kind of an arms race where they're reshaping their companies to figure out how to build the next great AI product. But they've realized that they don't have the type of talent they need.

也没有相应的组织体系。于是对多数企业而言，解决方案就是向这个问题砸入天文数字的资金。

They don't have the type of organizational structures they need, And their solution, for a lot of them at least, has been to just throw enormous amounts of money at the problem.

是什么引爆了这场AI军备竞赛？因为直到最近，这一切似乎都还不那么明显。

And what kicked off this AI arms race? Because it did not seem like it was happening in this really obvious way until pretty recently.

如果要追溯某个转折点

So if I had to pick a moment

我认为我们现在有机会彻底重新定义使用计算机的意义。

I think we have the opportunity here to kind of completely reimagine what it means to use a computer.

那是在五月，OpenAI通过一部精心制作的视频发布了这一重大公告的时刻。

It would be when Open made this huge announcement in May through this highly produced video.

约翰尼是我见过的人中思想最为深邃的。

Johnny is the deepest thinker of anyone I've ever met.

他们宣布将邀请iPhone、iPad、iMac和Apple Watch等我们日常使用的传奇发明者约翰尼·艾维加入团队。

And they revealed that they were gonna bring on Johnny Ive, the now legendary inventor of the iPhone, iPad, iMac, the Apple Watch, stuff that we use basically every day.

没错。我越来越感觉到过去三十年的所学所历都将我引向此处、此刻。

Right. I have a growing sense that everything I have learned over the last thirty years has led me to this place and to this moment.

他们宣布将以64亿美元聘请他协助开发下一代革命性AI消费产品。哇哦。所以唐尼·艾维正在与OpenAI合作。

They announced that they were gonna bring him on for $6,400,000,000 to help develop the next great AI consumer product. Wow. So Donnie Ive is partnering with OpenAI.

本周OpenAI新闻不断，但65亿美元这个数字...

So much news from OpenAI this week, but 6 and a half billion dollars.

作为AI开发领军者的OpenAI正强势进军硬件领域。这确实是笔巨额交易，整个硅谷都为之躁动，因为他们尚未透露该设备的具体细节。能否成功尚不明朗，但我想这给其他公司高管们植入了一个概念——

So OpenAI, which is the leader in AI development, kind of muscling into the devices games. And it's a really huge deal. It's a ton of money, and everyone across the valley kinda chafed it because they haven't given any details about what this device would actually be. And it's, you know, unclear if it'll work. But I think it planted the idea in a lot of executives' minds at some of these other companies that, hey.

现在这是一场人才争夺战。

This is now a talent war.

根据已知信息推测，他们处理焦虑的方式就是转化为开支——直接打开钱包。

And I'm assuming, based on what we know, what they do with all that anxiety is channel it into spending. They open their wallet.

没错。这些大公司尤其是科技巨头最不缺的就是现金。虽然各家财务状况都不错，但真正体现过度开支的典型代表当属Meta。

That's right. I mean, the one thing that all of these companies have, especially the big companies, is cash. You know, all of them are doing well financially, but I think the one that really typifies the sort of excessive spending is Meta.

好的。那对Meta来说具体是怎样的？给我讲讲。

Okay. And what does that look like for Meta? Tell me about that.

没错。今年六月，OpenAI的CEO萨姆·奥尔特曼决定参加一个播客节目，而这个节目恰好由他兄弟主持。

Yeah. So back in June, Sam Altman, the CEO of OpenAI, decides to go on a podcast that is conveniently hosted by his brother.

听着，我听说Meta把我们视为他们最大的竞争对手。

Look. I've I've heard that Meta thinks of us as their biggest competitor.

他证实了这一点。是的，Meta正在积极挖人。

And he confirms. Yes. Meta is aggressively trying to poach people.

他们开始向我们团队的很多人开出天价条件，比如1亿美元的签约奖金。嗯。

They started making these, like, giant offers to, you know, a lot of people on our team. Mhmm. You know, like, 100,000,000 signing bonuses.

奥尔特曼说，Meta的首席执行官马克·扎克伯格一直在向OpenAI员工提供高达1亿美元的签约奖金，让他们跳槽去Meta从事AI工作。

And Altman says, Mark Zuckerberg, the chief executive of Meta, had been offering OpenAI employees, like, a $100,000,000 signing bonuses to come over and work on AI for Meta.

我很高兴至少到目前为止，我们最优秀的人才还没有人决定接受他们的条件。

I'm really happy that at least so far, none of our best people have decided to take them up on that.

但扎克伯格目前并没有取得太大进展，因为很多OpenAI员工其实喜欢在OpenAI工作。他们认同公司的使命。但后来马克开始加码，他觉得1亿美元不够，2亿美元怎么样？

But Zuckerberg really doesn't get super far at this point because there's a lot of OpenAI employees who actually like working at OpenAI. They believe in the mission. But then Mark starts to ratchet things up. He decides a $100,000,000 is not enough. How about $200,000,000?

2.5亿美元如何？我们可以在第一年预付一半，剩下的分四年发放怎么样？他从'这太疯狂了'变成'我要给你的可是改变几代人命运的财富'，而对方可能只是OpenAI一名22岁的研究员。

How about $2.50? How about we front load half of that offer within the first year, and then you get the rest of it over four years? He goes from, hey. This is crazy to this is generational wealth I'm about to offer you, someone who may be like a 22 year old researcher at OpenAI.

这招管用吗？我是说，抱歉，但2.5亿美元可能对我都管用。

And does that work? I mean, I'm sorry, but 250,000,000 might work on me.

是啊。我是说，我们可以整天谈论我多么相信这个使命。但当有人开着装满钱的垃圾车停在你家门口说‘拿去吧’，很难拒绝。所以他确实开始挖人了。他从苹果挖来了一位极其重要的人物，花了超过两亿美元。

Yeah. I mean, like, we can talk about, you know, I believe in the mission all day long. And then when someone, you know, backs a dump truck full of money up to your doorstep and says, here you go, it's hard to say no. So he did start getting people. He got someone at Apple who was very important for, you know, more than $200,000,000.

这位是彭荣京，苹果最受尊敬的AI研究员兼高管，基本上是他维系着苹果AI领域的全局。

This is Ronging Peng, the most respected AI researcher and AI executive at Apple. He was basically holding everything together there.

他开始从谷歌和DeepMind挖角人才。

He started grabbing people from Google and DeepMind.

Meta已任命赵振家为——他是ChadGPD的联合创造者，也是GPT-4的核心贡献者——

Meta has appointed Zhenjia Zhao, cocreator of ChadGPD and a key contributor to g p d four as

他还从OpenAI挖来了一些关键人物带到Meta。

There were some key OpenAI hires that he started grabbing and bringing over to Meta.

听着，我认为他们想集结全球前50名的AI研究人员。

Look. I think they want to assemble the top 50 AI researchers that are out there.

但真正重大的转折点是扎克伯格宣布他们以143亿美元收购了Scale AI公司49%的股份，这家初创公司主要做数据标注业务。

But I think a really big moment was when Zuckerberg announced they had purchased a 49% stake in this company called Scale AI, which is essentially a startup that kind of does this data labeling for $14,300,000,000.

哇。现在消息源还透露了更多激励细节，说这更多是为了Scale AI创始人王亚历山大的AI人脉网络。

Woah. Now Source gave me more color around incentives as well, saying that this is more about Scale AI founder Alexander Wang's AI network.

他支付这笔钱是为了挖来Scale公司28岁的CEO王亚历山大及其数名副手，领导扎克伯格称之为Meta新超级智能实验室的机构。

And what he did was pay that money to bring over Scale's 28 year old CEO Alexander Wang and a number of his deputies to head up this new organization that Zuckerberg is calling Meta's new super intelligence lab.

据说扎克伯格正在办公室调整工位，让这些人都能坐在他旁边。好吧，看来扎克伯格找到门道了。不过不止他一个人，对吧？

And Zuckerberg's apparently moving desks in the office so that all of these people can sit right by him. Okay. So Zuckerberg is figuring it out. It's not just him, though. Right?

还有其他公司也参与进来了。

There are other companies that have been getting in on this.

没错，完全是这样。这时候事情就变得疯狂了。我觉得硅谷其他公司都开始慌了，说扎克伯格这是在扰乱我们的薪酬体系。

Yeah. Totally. This is when things get really wild. I think every other company in the Valley starts kinda freaking out and saying, hey. Zuckerberg is kinda messing up our compensation here.

我们得想办法多掏钱留住人才。要知道，有些研究员已经待了很久，他们肯定没拿到几亿美元的薪酬。所以我觉得我们看到的是所有人才开始聚集起来，策划如何利用这个局面。

We need to figure out how to start coughing up more money to keep our talent. And, you know, some of these researchers have been there for a while. They certainly aren't getting paid, you know, hundreds of millions of dollars. And so I think what we started seeing is all of the talent kind of congregating together and strategizing for how they could take advantage of this situation.

我能问下，这些人到底是谁？他们做什么让公司觉得如此有价值？

Can I ask, who are, by the way, all these people? What do they do that makes them so valuable to these companies?

现在真正抢手的是做前沿研究的人，他们专攻AI研究中大多数人不涉足的冷门领域。其中极少数人研究的是所谓'推理'，就是教电脑像人类一样思考解决问题。能做且做好这事的人凤毛麟角。

So the people that are really in high demand these days are folks who are doing kinda niche research called frontier research, which means they've been pursuing far flung parts of AI research that maybe most researchers are not looking into, and it's a pretty small number of them are studying what's called reasoning, which is basically helping to teach a computer to think more like you and me when it's trying to solve problems. And there's just not a lot of people who do this and do this well.

你知道，我听人说这是科技行业的体育化现象。这就是原因对吧？通常只有运动员会被这种天价挖到新球队。

You know, I've I've heard people refer to this as the kind of sportsification of the tech industry. And and this is why. Right? I mean, typically, you really only hear of athletes getting lured to new companies or or teams with this kind of money. You know?

就像大谷翔平拿几亿美金加盟道奇队。

Like Shohei Ohtani getting paid hundreds of millions to go play for the Dodgers.

听着，我们听到这种新闻时并不反感。我是说——

And look. Hey. We don't chafe at that when we hear it. I mean

完全理解。

Totally.

这钱是疯狂，但现在大家已经习以为常了。就像库里拿天价合同，勒布朗值得巨额薪酬，因为他们能让球队成为冠军队伍。现在很多科技CEO都在做这种类比。

It's it's crazy money, but at this point, it's sort of normalized, I think. Like, of course, Steph Curry is getting, like, a huge contract. Of course, you wanna pay LeBron a crazy amount of money because they made these teams like championship teams. Right? That is the analogy that a lot of tech CEOs are making right now.

坦率地说，这些公司内部的许多普通员工开始将这些人视为超级巨星。

And frankly, a lot of the people inside of these companies, like the rank and file, are starting to see these folks as as superstars.

作为变革者。

As difference makers.

是的。我的意思是，对你我来说他们可能不是家喻户晓的名字，但在硅谷一带，他们确实享有盛誉。过去一年甚至还发展出了这样一个节目，叫做TBPN。你现在收看的就是TBPN。

Yeah. I mean, they may not be household names to you or me, but certainly around Silicon Valley, they have really good reputations. There's even this show that has developed over the past year. It's called TBPN. You're watching TBPN.

我们11点整准时开始

We start at 11AM sharp with

每周五天，每天三小时直播。

It streams five days a week, three hours a day.

总之，说到AI研究，我们在AI贸易协定领域有大量报道。扎克伯格继续在...

Anyway, speaking of AI research, we have a ton of stories in the AI trade deal world. Zuckerberg continues to be on a tear from

基本上就是对科技行业和媒体行业的实时评论

And it's basically a running commentary on the tech industry, the media industry

太棒了。

Amazing.

那种把它当作ESPN体育中心类节目来处理的模式，不过是针对商业领域的。

That kinda treats it like an ESPN sports center type show, but for business.

我们收到一个危险信号。

We got a red flag.

我们收到一个危险信号。埃隆·马斯克和山姆·奥特曼在争吵。就像父母在吵架。

We got a red flag. Elon Musk and Sam Altman are fighting. Mom and dad are fighting.

这很有趣。他们会玩这种把戏——如果有人去了Meta，他们就会制作一个搞笑图片，像是球员交易。OpenAI的人成了Meta超级智能实验室的首席科学家，他们就拿这个开玩笑。

It's very funny. They they do this stuff where if someone goes to Meta, they'll spin up a a funny graphic that's, like, traded. OpenAI guy becomes chief scientist at Meta's Superintelligence Labs, and they have fun with it.

苹果公司正在谋划其人工智能的回归。别叫它回归。其实你可以叫它回归。我会告诉你

Apple Inc. Is plotting its artificial intelligence comeback. Don't call it a comeback. Actually, you can call it a comeback. I'll give you

是的。听起来很多人对这则新闻感兴趣，这反映了当下行业对这些动向的敏感程度。

Yeah. And it sounds like there's a lot of people that are interested in this news, and that that is kind of reflective of how attuned the industry is to these moves at this moment.

确实如此。我认为硅谷所有人都在密切关注谁去了哪里，谁招到了下一个大人物，谁拿到了最疯狂的薪水。就像体育比赛一样，大家都在观察这些'转会'会如何影响其他公司，看谁能胜出。

Very much so. I think, you know, everyone in Silicon Valley is paying very close attention to who's going where, who's got the next big hire, who's getting paid the craziest salary. And, you know, really, just like in sports, everyone is watching to see how these moves or these trades, if you will, is gonna affect other companies and see who comes out on top.

我们马上回来。

We'll be right back.

我们都有本可以做得更好的时刻，比如自己剪头发。哎呀。或者忘记涂防晒霜，现在脸红得像番茄。痛啊。本可以做得更好。

We all have moments when we could have done better, like cutting your own hair. Yikes. Or forgetting sunscreen so now you look like a tomato. Ouch. Could have done better.

投资也是如此。用嘉信理财提升水平，更聪明地投资。获取市场洞察、教育资源，并在需要时获得人工帮助。详情请访问schwab.com。想象一个AI不只是自动化，而是赋能的世界。

Same goes for where you invest. Level up and invest smarter with Schwab. Get market insights, education, and human help when you need it. Learn more at schwab.com. Imagine a world where AI doesn't just automate, it empowers.

西门子将尖端工业AI和数字孪生技术交到真实用户手中，改变美国的建造和前进方式。您的工作将获得超级助力。您的运营将得到优化。可能性，无限。这不仅仅是自动化。

Siemens puts cutting edge industrial AI and digital twins in the hands of real people, transforming how America builds and moves forward. Your work becomes supercharged. Your operations become optimized. The possibilities, limitless. This isn't just automation.

这是能力放大。从工厂车间到电网，西门子正在把'如果'变成'下一步'。了解西门子如何解决当今挑战，为明天创造机遇，请访问usa.siemens.com。

It's amplification. From factory floors to power grids, Siemens is turning what if into what's next. To learn how Siemens is solving today's challenges to create opportunities for tomorrow, visit usa.siemens.com.

好的。我想深入探讨这些资金如何实际改变现状的问题，以及在你看来，这些资金如何影响AI领域的格局。

Okay. I wanna get into those questions of how this money is actually changing things, how the money is affecting the landscape of AI as you're seeing it.

没错。听着，我认为我们应该记住，我们正处在这场变革的漩涡中心。你明白吗？这就像是一个历史性的时刻。

Yeah. So look. I I think we should remember we're literally in the middle of all of this happening right now. You know? This is, like, a moment in time.

所有这一切都像是一场疯狂的人才争夺战。这些公司正在构建新事物，一切都悬而未决。我不太愿意断言谁将赢得这场竞赛。

All of this stuff is, like, a wild flurry of trading people back and forth. These companies are building new things. Everything is up in the air. I do not feel comfortable saying who's gonna win the race.

确实。

Sure.

话虽如此，我认为我们仍可以观察当前格局，看看各公司现状，比如谁遥遥领先或谁是最大黑马。这是我们仍可尝试厘清的。

That said, I think we can still look at the landscape and see where companies are right now and, you know, who's got a great lead or who's the biggest long shot. And I think that's something that we can still just try to suss out.

是的，就这么办。给我一份这些公司的实力排名吧。

Yes. Let's do that. Give me a a power ranking of these companies as it were.

好的。看待这个市场时，我喜欢从开发面向消费者产品的公司角度思考。我这么说是因为，虽然关于AI的讨论多集中在创造下一个机器人上帝或全知AI，但除了OpenAI，很多公司连最基础的事都做不到——就是说服人们在日常生活中使用人工智能并以此建立业务。因此按这个标准，我将这些公司分为三个梯队。

Okay. When I think of this market, I like to think of it in terms of companies creating consumer facing products. And I say that because while a lot of the rhetoric around AI is about building the the next robot god or this all seeing, all knowing AI. A lot of these companies, you know, aside from OpenAI, haven't been able to do something pretty basic, which is just to convince people to use artificial intelligence in their everyday lives and build a business around that. And so when I think about that metric, I try to break these companies down into three different tiers.

第三梯队是那些参与竞争但胜算较小的公司。比如Perplexity，这家初创公司开发了类似谷歌搜索的产品；还有规模更大的Anthropic，他们有自己的聊天机器人Claude，虽有人使用但远不及其他产品；另外就是埃隆·马斯克旗下xAI，这个嫁接在X社交平台（原推特）上的AI部门。

There's the bottom tier, companies that are definitely in the mix but are probably more of, like, a long shot in this greater race. You have Perplexity, which is a startup that creates this Google search like product. You have Anthropic, which is certainly bigger. They have their own chatbot called Claude, which is used by people and enjoyed, but certainly not as widely used as some others. And then you have x AI, which is the artificial intelligence wing owned by Elon Musk and kind of grafted on to x, the social network formerly known as Twitter.

xAI的聊天机器人Grok曾做过些有趣甚至惊艳的事，但一旦出错就是大问题。有次它突然变成'亲希特勒机器人'，几乎同时冒犯了所有人，最后团队不得不公开道歉。

That has this chatbot called Croc, which, you know, has done some interesting, even impressive things. But when it screws up, it screws up in a big way. There was this moment where it decided it was a Mecca Hitler bot and offended basically everyone at the same time before they had to apologize for it.

这错误可够严重的。

That's a pretty big error.

确实，这情况不太妙。我觉得问题在于，他们虽然有埃隆·马斯克和他那取之不尽的银行账户作为优势，但同时也以非常引人注目的方式频频栽跟头。

Yeah. That's that's not great. And I think it's one of these things where they have the advantages of having Elon Musk and his endless bank account, but they also seem to stumble in a very high profile way.

是啊，有趣的是埃隆——这个在很多领域都被视为先驱的人——在这件事上似乎并没有引领潮流。嗯。那么第二梯队是哪些？

Yeah. It's interesting that Elon, who's, you know, often seen as out ahead on so many things, does not seem to be leading the pack on this. Mhmm. So what's tier two?

第二梯队我会选那些更成熟的科技公司，比如苹果、Meta、谷歌。这些公司有两大关键优势：规模和资金。但劣势也不少。就拿Meta来说，我们已经知道它落后了。

So tier two, I would probably put the more established tech companies. Think Apple, Meta, Google. And these companies have two key advantages, scale and cash. But they have a lot of disadvantages too. Let's take Meta, which we've already established is behind.

很多研究人员担心马克·扎克伯格对Meta的AI发展缺乏清晰愿景。坦白说，他们不想开发下一个用AI推荐内容、让你在Instagram或Facebook停留更久的算法——这根本不是他们入行时想做的事。至于苹果，我认为他们很弱——虽然卖出了数十亿部iPhone，但内置的AI工具Siri甚至不被视为真正的AI，只是个时灵时不灵的语音助手，在硅谷圈子里口碑很差。

A lot of researchers are sort of worried that Mark Zuckerberg doesn't have a clear vision for what AI will look like at Meta. And frankly, they don't wanna work on building the next algorithm that keeps you in Instagram or Facebook for longer because it's suggesting better content using AI. That is not exactly what they signed up for when they got into the field. Then, you know, Apple, I think they're weak because, you know, even though they've sold billions of iPhones at this point, Siri, the AI tool that comes with it is not even really thought of as an AI tool. It's like a voice assistant that only works half the time and is considered pretty bad in Silicon Valley circles.

连苹果自己也承认Siri未达预期。对吧？它本该成为承诺中的革命性助手，但始终未能实现。多年来苹果没有真正投资其他公司那种技术，所以...

And even Apple acknowledges Siri is not where it should be. Right? It's it should be this incredible assistant is what they promised it to be, and it just hasn't been. It hasn't panned out because for years, Apple has not really invested in the type of technology that these other companies have. So I would say they have a

潜力

lot of

很大，但火候未到。

potential, but they're just not there yet.

谷歌呢？

And Google?

谷歌很有意思。可以说在AI领域耕耘最久，取得过重大突破，使命感很强——他们是真的想攻克癌症。

Google is interesting. You know, Google has been working on AI arguably the longest of most of these companies. They have made huge advances. They're very mission oriented. They're likely literally wanna cure cancer.

这对吸引人才很有帮助。但大众最熟悉的谷歌AI可能是现在搜索结果的顶部摘要。这方面他们出过些大纰漏，比如有次推荐了往披萨上涂胶水的食谱——反正我个人是不太能接受这种吃法。

That is huge in helping recruit talent. But I think the AI that most people are familiar with from Google is the little summaries at the top of search results that happen these days. And they've had some, you know, pretty high profile stumbles with that summary. There was this moment where they gave a recipe of putting glue on pizza, which is like I don't know about you. That's not something I'm super into, but

虽然还称不上是希特勒级别的聊天机器人，但它确实相当糟糕。

like It's not quite Hitler chatbot, but it's pretty bad.

我认为在糟糕程度排名上，它还算靠后一些。但谷歌的核心价值本应是提供最佳搜索结果，而眼前这个显然不是最优解。

I think in the rankings of bad, it's down the ladder a little. But still, like, the entire point of Google is getting you the best search results, and that is clearly not the best.

现在我们讨论的是顶级梯队对吧？我猜就是我们一直在谈论的那家被众人追赶的公司——OpenAI。

And I think we're now at the top tier. Right? Which I assume is the company we've been talking about that everyone's been chasing, OpenAI.

没错。OpenAI几年前真正开启了这场革命。他们率先发布了最成功的聊天机器人，一夜之间就有数百万人下载使用。直到今天，他们的用户基数依然庞大且持续增长。

That's right. OpenAI really kicked this whole thing off a few years ago. They were the ones that first released the most successful chatbot. Millions of people downloaded it overnight, basically. And to this day, they still have an enormous amount of people using it and continue to grow every single day.

但他们也存在劣势。要知道，维持这个成功聊天机器人的运营成本极高，导致他们亏损严重。虽然目前用户量遥遥领先，但只要有公司推出更好的产品，这种优势可能瞬间逆转。

But they do have some disadvantages. You know? For one, they're losing enormous amounts of money because it costs a lot of money to keep the successful chatbot running. I think it's it's one of these things where they're dramatically ahead as far as the number of people using it, but that can change almost overnight sometimes if a company puts out another better product.

然而值得注意的是，至少目前这家公司仍遥遥领先。我认为这是因为美国科技行业长期存在垄断话题——那些巨头掌控着互联网命脉，而当下这个领域对很多人来说正是世界上最重要的事物。

And yet there is something remarkable about the fact that this company is so far ahead right now, at least. And I think it's because the American tech industry is a place where there has been a lot of talk of monopolies. You know? There's these massive companies that just own so much of the Internet. And then you have this industry that is, like, the most important thing on earth to many people right now.

嗯。而且现在有家非巨头企业正占据主导地位。

Mhmm. And there's a company outside of the big ones that's dominant.

这简直就是硅谷的经典叙事。看看谷歌、Meta、苹果，我们总觉得这些公司会永远主宰市场。但二十年前人们也这样看待百视达——那时我们每周五晚都去租DVD回家看。

I mean, this is honestly classic Silicon Valley as a story. Right? You may look at Google, Meta, Apple. I think these companies are gonna be dominating the landscape forever. But, like, I would have said the same thing about blockbuster video twenty years ago, right, where, of course, I go every Friday night and get a movie, and I go home and watch it on DVD.

结果Netflix横空出世，先是通过邮寄DVD，几年后直接实现家庭流媒体点播，连沙发都不用离开。这就是经典的创新者窘境。

Yep. And then Netflix comes along, and now they're mailing you DVDs. And then a few years later, they're streaming every video to your TV at home so you don't even have to leave your couch. Right? It's a classic idea of the innovator's dilemma.

当企业足够庞大、占据主导时，往往会安于现状而非开拓创新。用体育来比喻的话，就是只打防守不进攻。

If you're big, if you're dominant, you don't need to worry about creating the next big thing because you can just sort of coast on your laurels and I don't know. Let's milk the sports analogy. You can keep playing defense rather than offense.

没错。当新贵带着新事物出现时，这就会让你落后。

Yeah. And and that leaves you behind when an upstart comes in with something new.

完全正确。现在所有这些公司都处于被动状态，拼命追赶OpenAI——正如你所说，它凭空出现并超越了所有人。

That's exactly right. And all these companies are now kind of on their back foot and trying desperately to catch up to OpenAI, which, as you said, came out of nowhere and lapped them all.

我们排名不错，你是根据这些公司在争夺消费产品市场、努力占领消费者市场的相对位置来排名的。考虑到我们说过他们在这一领域的巨额投入——不仅仅是薪资，还有数据中心建设、模型开发本身——让用户使用这些产品的赌注似乎相当高。

We ranked well, you ranked these companies based on their relative positions in the race toward a kind of consumer product, the effort to capture the consumer market. I mean, and the stakes for that, for getting people to use this stuff seem pretty high given how much we've said they're spending and investing in this. And it's not just salaries. It's the data centers. It's developing these models in the first place.

是的。所以我不得不问，这种状况能持续多久？这些公司很多都是上市公司，有股东和投资者。

Yeah. And so I have to ask, how sustainable is that? I mean, many of these companies are public companies. They have shareholders. They have investors.

这种情况能维持多久？是否存在极限？

How long can this last? Is there a limit?

这是个好问题。感谢你用金钱把我们拉回现实，因为这一切都贵得离谱。不仅是数十亿，而是数千亿美元的基础设施投入。建数据中心可不便宜。

It's a it's a good question. I appreciate you bringing us back down to earth with money because all this stuff is insanely expensive. This is not just billions, but hundreds of billions of dollars in spending on infrastructure. You know, building data centers is not cheap.

确实。

Sure.

开发可能有人用、可能没人用的产品也不便宜。成本在累积，而且速度很快。多年来这些公司获得了很大宽容度，只因它们基本都在印钞——谷歌用搜索广告创造了完美的互联网商业模式，Meta让我不停在Instagram上购物，每季度赚取数十亿。

Developing products that maybe people use, maybe they won't, is not cheap. It's adding up, and it's adding up quickly. I think the companies have been given a lot of latitude for years just because they all basically print money. Google created the perfect Internet business model with search ads. Meta has me buying stuff on Instagram incessantly and makes billions every quarter.

只要主营业务运转良好，投资者就允许它们追逐这些亏损巨大的遥远构想。但我觉得我们正开始看到这种心态的极限——最近几周华尔街对此已显动摇，他们看到建设成本后开始提出更多质疑：

You know? As long as their businesses have been running smoothly, investors have allowed them to pursue these far flung ideas that are losing big amounts of money. But I think we're starting to see the limits of that mentality. In recent weeks, Wall Street has started to get a little shaky on this. They've seen how much this stuff cost to build, and they're starting to ask more questions saying, hey.

这能成为真正的业务吗？人们真会用这些东西吗？

Can this be an actual business, and do people wanna use this stuff?

确实。关于这点，已经有些讨论了，对吧，关于这是否可能是个泡沫，你知道的，让人想起两千年代初的互联网泡沫，那时候有大量互联网初创企业估值高得离谱，然后泡沫就破裂了。

Yeah. On that, there has been some talk, right, of whether this might be a bubble, you know, reminiscent of the kind of .com bubble in the early two thousands where you had this huge amount of Internet startups that had these insane valuations, and then that bubble popped.

嗯。

Mhmm.

你在这里看到类似情况了吗？这两者有可比性吗？

Do you see that here? Is is that a parallel?

呃，这里没人喜欢'泡沫'这个词。不过，是的，我认为这完全...我觉得这很合理。你看，我的意思是，有太多资金投入其中了。每家风投公司都在疯狂砸钱进入AI公司。

Well, no one likes the b word out here. But, yeah, I think I think it's totally I think it's fair. Look. I mean, there's so much money tied up in this. Every VC firm is spending insane amounts to get into AI companies.

有些AI公司建立在其他AI公司之上，还有数据和基础设施公司都依赖这些东西的成功。这有点像...全都相互关联，彼此的成功也紧密依存。我认为质疑这些东西最终能否成功是非常合理且重要的，而不是将其视为必然。

There's AI companies built atop of AI companies, data and infrastructure companies that are relying on the success of this stuff. It's kind of, like, all interconnected and really dependent on each other's success. And I think it's a really fair and important to question whether this stuff is gonna even work out, not treat it as an inevitability.

这对经济意味着什么？因为我们经济的很大部分正由这个新兴产业的巨额支出驱动。如果泡沫真的破裂，对我们其他人意味着什么？

What does that mean for the economy? Because so much of our economy is being driven by this huge spending in this this massive new industry. If the bubble does go pop, what does that mean for the rest of us?

我是说，你看，这让我想起我们与这一切的关联有多深。别忘了，大型科技股常常引领整个市场。对吧？有时候道指的涨跌就取决于Meta、谷歌、苹果这些公司的表现。我的401k养老金可能和这个绑在一起，你的也是。

I mean, look, it reminds me of how tied into this we all are. You have to remember, the big tech stocks often lead the entire market. Right? Sometimes the Dow is up and down based on the performance of Meta or Google or Apple or whatever. My four zero one k is probably tied up to this as is yours, you know.

无论我们是否想要这个版本的未来，这已经编织进我们的经济体系了。所以在某种程度上，我们确实与这里的成功绑定了。即使你选择不使用ChatGPT，也不意味着AI完全不会影响你。是的，我认为它渗透在世界运作的各个层面，可能超出你我的认知。

This is kind of woven into our economy whether we want this version of the future or not. So in a way, we are kind of tied to the success here. The idea that if you choose not to use ChatGPT doesn't insulate you from the idea that AI is not gonna affect you at all. Yeah. I think it's more woven into different parts of how the world works than you and I might realize.

没错。就像即使你个人没有参与，整个经济体系已经...是的。在这条船上

Right. It's like even if you personally are not bought in, the entire economy is, so you're Yep. On this

欢迎登船。

Welcome aboard.

迈克，谢谢你。

Mike, thank you.

感谢邀请我。

Thanks for having me.

我们马上回来。

We'll be right back.

本播客由MTV《鲁保罗变装皇后秀》赞助支持。该节目获得10项艾美奖提名，包括最佳真人秀或真人竞技节目主持人及最佳真人竞技节目。《纽约时报》称《鲁保罗变装皇后秀》轰动一时，《秃鹰》杂志评价这部艾美奖提名系列为扣人心弦的电视节目，《新闻周刊》盛赞其堪称史诗级，《早安美国》则将其誉为杰出代表作。

This podcast is supported by MTV's RuPaul's Drag Race. Nominated for 10 Emmy Awards, including outstanding host for a reality or reality competition program and outstanding reality competition program. The New York Times calls RuPaul's Drag Race a sensation. Vulture is calling the Emmy nominated franchise gripping TV. Newsweek declares it truly epic, and Good Morning America celebrates the series as a standout.

切勿错过这一现象级节目，立即在Paramount+上观看艾美奖提名系列《鲁保罗变装皇后秀》。

Don't miss the phenomenon and stream the Emmy nominated franchise RuPaul's Drag Race on Paramount plus.

您的住宅是主动投资而非被动资产。通过Rocket Mortgage，您可以立即盘活房屋净值。释放房屋净值，即为家人开启新机遇——无论是装修扩建，还是购置下一处房产。立即通过更智能的工具与真实贷款专家的指导开启旅程，详情请访问rocketmortgage.com。

Your home is an active investment, not a passive one. And with Rocket Mortgage, you can put your home equity to work right away. When you unlock your home equity, you unlock new doors for your family, renovations, extensions, even buying your next property. Get started today with smarter tools and guidance from real mortgage experts. Find out how at rocketmortgage.com.

Rocket Mortgage LLC，全美50州持牌运营。Nmlsconsumeraccess.org，编号3030。

Rocket Mortgage LLC, licensed in 50 states. Nmlsconsumeraccess.org, thirty thirty.

以下是今日其他要闻。

Here's what else you need to know today.

完成这里的工作后，我们将前往另一地点，同样确保其安全。我们要让国家变得非常安全，让城市变得无比安全。

After we do this, we'll go to another location, and we'll make it safe also. We're gonna make our country very safe. We're gonna make our cities very, very safe.

特朗普总统表示计划将芝加哥作为其下一轮联邦打击犯罪行动的目标。

President Trump said that he planned to target Chicago for his next federal crackdown on crime.

芝加哥一团糟。你们有个无能的市长，极其无能，我们很可能下一个就要整顿那里。这将是继当前问题后的下一个目标。

Chicago's a mess. You have an incompetent mayor, grossly incompetent, and we'll straighten that one out probably next. That'll be our next one after this.

特朗普表示，国民警卫队在华盛顿特区的部署有助于清理城市，并暗示他也愿意在城市街道上动用现役部队。

Trump said the National Guard deployment in DC was helping clean up the city and suggested that he was also willing to use active duty troops on city streets.

我们尚未需要调动常规军队，但如有必要我们愿意这么做。

We haven't had to bring in the the regular military, which we're willing to do if we have to.

但总统可能难以轻易在其他主要城市复制华盛顿特区的武力展示。因为华盛顿是联邦直辖区而非州，特朗普在那里比美国任何其他地方都拥有更多控制权，包括接管当地警察部队的权力。本期节目由戴安娜·温、亚历克斯·斯特恩和卡洛斯·普列托制作，莱克西·迪奥、马克·乔治和玛丽亚·伯恩在迈克尔·伯努瓦协助下编辑，艾丽莎·莫克斯利负责技术制作。

But the president may not be able to easily replicate the show of force in DC in other major cities. Because DC is a federal district, not a state, Trump has more control there than anywhere else in the country. That includes the ability to take control of the local police force. Today's episode was produced by Diana Wynn, Alex Stern, and Carlos Prieto. It was edited by Lexi Dio, Mark George, and Maria Byrne with help from Michael Benoit and was engineered by Alyssa Moxley.

以上就是本期《每日新闻》。我是娜塔莉·基特罗斯。明天见。我们

That's it for The Daily. I'm Natalie Kitroeth. See you tomorrow. We

都有本可以做得更好的时刻，比如自己剪头发。哎呀。或者忘记涂防晒霜，现在你看起来像个番茄。哎呦。本可以做得更好。

all have moments when we could have done better, like cutting your own hair. Yikes. Or forgetting sunscreen so now you look like a tomato. Ouch. Could have done better.

投资也是如此。通过嘉信理财提升水平，更聪明地投资。获取市场洞察、教育资源，并在需要时获得人工协助。了解更多请访问schwab.com。

Same goes for where you invest. Level up and invest smarter with Schwab. Get market insights, education, and human help when you need it. Learn more at schwab.com.