本集简介
双语字幕
仅展示文本字幕，不包含中文音频；想边听边看，请使用 Bayt 播客 App。
你正在观看TVPN。
You're watching TVPN.
今天是2026年3月5日，星期四。
Today is Thursday, 03/05/2026.
我们正在TVPN超穹顶现场直播，
We are live from the TVPN UltraDome, the
科技的圣殿。
temple Of technology.
堡垒金融，资本之都。
Fortress finance The capital capital.
资本。
Capital.
让我告诉你关于ramp.com的事。
Let me tell you about ramp.com.
时间就是金钱。
Time is money.
同时节省两者。
Save both.
轻松使用企业卡片、账单支付、会计以及其他诸多功能，全部集中在一个地方。
Easy use corporate cards, bill pay, accounting, and a whole lot more all in one place.
我们有我们
We have We
病了。
are sick.
约翰和我都病得很严重。
John and I are both very sick.
厌倦了不录播客。
Sick of not podcasting.
我们开始吧。
Let's go.
我们确实病得很严重。
We are actually very sick.
我们
We are
病得很重。
deathly ill.
但是
But
这就是观看每日节目的妙处。
That's the beauty of watching A daily show.
直播。
Going live.
你不会被我传染，因为你隔着屏幕。
There's no risk of me getting you sick because you're on the other side of a screen.
屏幕时间。
Screen time.
这就是每日节目的妙处。
That's the beauty of a daily show.
你从来不会
There's no you don't ever
休息两天。
Two days off.
休病假。
To have sick days.
是的。
Yeah.
没有休息日。
There's no days off.
因为没有休息日。
Because there's no days off.
今天我们有一场精彩的阵容。
And we have a fantastic lineup today.
我们有马克·古尔曼加入。
We have Mark Gurman joining.
我们来调出线性排列的节目单。
Let's pull up the linear lineup.
线性推动了现代软件开发的系统。
Linear forces the system for modern software development.
70% 使用线性平台的企业工作空间都在使用智能代理。
70% of enterprise workspace on linear are using agents.
我们邀请了 Axios 的 Dan Primack 来分析 IPO 市场。
We have Dan Primack coming on from Axios to break down IPO market.
风险投资领域现在发生了什么？风险投资回报情况如何？
What's going on in venture capital, what's going on with venture capital returns.
我们有来自 Plaid 的 Zach 今天要来吗？
We have do do we have Zach from, from, Plaid coming on today?
那太棒了。
That'd be awesome.
如果是的话，我们还会邀请 Nominal 的 Cameron McCord 现场连线，他们刚刚完成了由创始人基金领投的巨额融资。
If so, we have Cameron McCord from Nominal, massive unicorn round that just happened with Founders Fund coming on, in person live.
当然，马克斯。
Of course, Max
马克斯亲自到场。
Max in person.
广阔的空间。
Vast space.
非常感谢您今天收听。
Well, thank you so much for tuning in today.
说到Plaid，Plaid为您用于消费、储蓄、借贷和投资的应用程序提供支持。
Speaking of Plaid, Plaid powers the apps you use to spend, save, borrow, and invest.
通过AI安全连接银行账户，实现资金转移、防范欺诈并改善贷款服务。
Securing connecting bank accounts to move money, fight fraud, and improve lending now with AI.
嘿，约翰。
Hey, John.
是的。
Why Yes.
没人提到吗
Is no one talking
关于丹尼尔·格罗斯。
About Daniel Gross.
丹尼尔·格罗斯。
Daniel Gross.
真的没人
No one literally no one is
在谈论
talking about
丹尼尔·格罗斯。
Daniel Gross.
他没有。
He no.
他应该家喻户晓才对。
He should be a household name.
你应该打个出租车，他们应该会说，哦，你确实如此。
He you should get into the a taxi, and they should be, oh, did you yeah.
你有没有注意到丹尼尔·格罗斯的AGI交易有多精准
Did you see how how incredibly dialed Daniel Gross's AGI trades
从2024年1月14日开始。
were From January 14.
2024年。
2024.
是的。
Yes.
这是一个非常棒的网站。
This is a great great website.
它到底是什么？
It's what is it?
danielgross.com/agi，经典的Times New Roman衬线字体，12号，纯粹用基础HTML直接呈现出来。
Danielgross.com/agi the the classic Times New Roman, serif font, 12, just, just hammering it out in the in the vanilla HTML.
不需要样式。
No need for styles.
不需要Bootstrap模板。
No need for bootstrap templates.
那个CS的Tailwind是什么？
What's that CS tailwind?
他做这个根本不需要Tailwind。
He didn't need tailwind for this.
他只是直接用Markdown或HTML写的。
He just, wrote it in probably Markdown or HTML directly.
很多预言都实现了。
And a lot of it has come true.
这很有趣，因为很多问题都是以开放式的提问形式提出的。
It's interesting because a lot of them are framed as just, like, open ended questions.
但如果你认真思考一下，如果你相信AGI，然后逐一分析这些问题，你就会看到过去两年里Exactly发生了什么。
But if you think about you know, if you believe in AGI and then you go through the questions, like you will see exactly what happened over the last two years.
这在很大程度上一直是情境感知的核心论点。
And this has been the underpinning thesis of situational awareness in many ways.
丹尼尔·格罗斯是该基金的支柱。
Daniel Gross was an anchor of the fund.
当然，这
That, of course, is
还有，这一点在今天尤其相关，因为我来读一下引言。
the And it's particularly relevant today because I'll read the intro.
是的。
Yes.
这是2024年1月14日，两年前了。
This is 01/14/2024, over two years ago.
他说，我认为我们都可以同意，GPT-4在许多任务上已经达到人类水平的熟练度。
He says, I think we can all agree that GPT-four completes many tasks at human level proficiency.
它在某些奇怪的方面还不完美。
It is imperfect in odd ways.
它能像一名聪明的麻省理工本科生一样编写软件，却无法像初级行政助理那样完成基本的任务规划。
It can write software like a smart MIT undergrad, but can't do basic task planning like an entry level EA.
它能说所有语言，却几乎不会做数学。
It speaks all languages, but can barely do math.
假设进步不会停止，就像GPT-4比GPT-3更强一样，GPT-5已经能够实现基本的代理行为，我。
Suppose the progress doesn't stop, just like GPT-four was better than three, GPT-five is capable of basic agentic behavior, I.
E.
E.
能够接受任务，持续处理一段时间后返回结果。
Able to accept a task, work on it for a while and return results.
完美命中。
Nailed it.
当然，如今OpenAI已经发布了GPT-5.4，它正是这样做的，嗯。
And of course, today, OpenAI is has released GPT 5.4 which does exactly this Mhmm.
相当不错。
Quite well.
是的
Yeah.
他们的评价陆续出来了，没错。
Their reviews are coming in Yep.
而且评价非常好。
And they are quite quite good.
丹尼尔继续说，现在有一部分Upwork的任务只需少量电子就能完成。
Daniel continues, some modest fraction of Upwork tasks can now be done with a handful of electrons.
假设每个人都能雇佣这样的一个智能代理。
Suppose everyone has an agent like this they can hire.
假设每个人都能雇佣一千个这样的智能代理。
Suppose everyone has 1,000 agents like this they can hire.
在这个世界里，人们该做什么？
What does one do in a world like this?
所以
So
那我们一个一个来讨论。
So let's go through them one by one.
有十七、十八个问题，差不多吧。
There's seventeen, eighteen questions, something like that.
首先，让我告诉你关于Labelbox的强化学习环境、语音、机器人、EDOX和专家人类数据。
First, let me tell you about Labelbox reinforcement learning environments, voice, robotics, EDOX, expert human data.
Labelbox是全球领先AI团队背后的数据工厂。
Labelbox is the data factory behind the world's leading AI teams.
他以一个简单的问题开场。
He kicks it off with an easy question.
在后AGI世界中，价值会集中在哪儿？
In a post AGI world, where does the value accrue?
当时人们正在对此进行辩论。
The people were debating this at the time.
应用层、基础模型层和基础设施层。
Application layer versus versus foundation model layer versus infrastructure layer.
价值肯定已经累积
Value has definitely accrued
到了这一层。
to this layer.
我的意思是，即使在AGI之后，约翰的新发型也在付费
I mean, you're even post AGI, John's new haircut is is paying
格雷厄姆写了一整篇文章，论述品牌为何重要。
Graham wrote a whole piece about how brands are important.
我们明天会讨论这个问题，但他将其与手表行业的石英危机相提并论，并对后AGI时代的价值分配提出了一系列见解。
We'll we'll we'll cover that tomorrow, but he he compares it to the courts crisis in the watch world and has a bunch of thoughts about value in the post AGI era.
但回顾过去两年，价值显然已经累积到了基础设施层。
But just looking back in the last two years, value has clearly accrued to the infrastructure layer.
这意味着芯片、封装、电力等等。
So that means chips, packaging, power, etcetera.
而这大体上就是情境感知的交易。
And this is the situational awareness trade by and large.
NVIDIA 基本上捕获了人工智能热潮中超过100%的利润，因为人工智能领域的许多其他公司都在亏损。
NVIDIA basically captures and and did for a while more than 100% of the profits from the AI boom because so many of the other companies in the AI space were losing money.
所以基础实验室都在亏损。
And so the Foundation Labs are losing money.
NVIDIA 的利润率从大约30%上升到了60%的毛利，这一点在市值上得到了体现。
NVIDIA's profit margin went from like 30% to 60% gross margins, and the it was born out in market cap.
自丹尼尔撰写这篇文章以来，NVIDIA 的市值增加了3.2万亿美元。
So NVIDIA added 3,200,000,000,000.0 in market cap from when Daniel wrote this piece.
我记得听过他在Strikelery上与本·汤普森和纳特·弗里德曼对话时说，是的。
And I remember listening to him on Strikelery talk to Ben Thompson alongside Nat Friedman and say, yeah.
基于ChatGPT，NVIDIA 似乎有点被低估了。
Based on based on ChatGPT, NVIDIA seems sort of undervalued.
我当时想，哦，既然他们在本·汤普森的Strikeckery播客上这么说，那肯定已经被市场定价了。
And I was like, oh, well, if they're saying it on Ben Thompson's Strikeckery podcast, it's obviously priced in.
人人都知道这一点。
Everyone knows this.
我完全错了。
And I was wildly wrong.
永远不要怀疑自己。
Never never doubt yourself.
但有趣的是，其他平台并没有看到三倍的增长。
But what's interesting is that the rest of the platforms did not see, you know, three x gains.
它们并没有增加3万亿美元的价值。
They did not add $3,000,000,000,000 in value.
从2024年1月到现在，微软只上涨了4%。
Microsoft, from from January 2024 to today is only up 4%.
亚马逊上涨了30%。
Amazon's up 30%.
而在私募市场中，你还有OpenAI、Anthropic和xAI。
And then you do have OpenAI, Anthropic, xAI in the private markets.
它们的收益巨大而惊人，堪称私募市场历史上增长最快的公司，彻底撼动了风险投资界。
The gains there, they've been huge and staggering in that it's like the fastest growing companies in private markets history, shaking the venture capital world.
你要么加入，要么出局。
You're either in or you're out.
这是一件大事。
It's a huge deal.
但如果你把它们都加起来
But you add them all
加起来达到1.4万亿，当你想到萨提亚的时候，真是疯狂。
up So to 1.4 crazy when you think it Satya Yeah.
在很多方面，它经历了一代人级别的高速增长。
In many ways went on such a generational run.
而你只增长了4%
And you got a 4%
突然上涨。
pop up.
如果你没有留意，确实如此。
If you hadn't if you hadn't looked Yes.
从股价来看，你会想，哦，它肯定涨了不少吧？
At the stock price at all, you would think, oh, it's gotta be up what?
是的。
Yeah.
4040%？
4040%?
现在，它涨了，是的。
Now, it's up Yeah.
4%。
4%.
于是，丹尼尔·格罗斯提出了下一个问题。
And so that was that was Daniel Gross's next question.
英伟达和微软会怎样？
What happens to NVIDIA and Microsoft?
当时，这两家公司是最值得关注的玩家，是最大、最符合AI趋势的公司之一。
These are the two interesting players at the time, the big some of the biggest companies, the most AI aligned.
英伟达的表现绝对出色。
NVIDIA absolutely crushed.
收入从2024财年的600亿美元增长到2026财年2月时翻了三倍。
Revenue tripled from 60,000,000,000 in fiscal year twenty twenty four to February in fiscal year twenty twenty six.
微软的主导地位则弱得多。
Microsoft has been far less dominant.
因此，Azure的增长实际上正在加速。
So Azure growth actually is accelerating.
其年增长率达到了40%，但股价仅上涨了4%，正如我们所说。
It's at 40% year over year, but the stock only returned 4%, as we said.
市场对萨提亚·纳德拉向投资者承诺的800亿美元AI资本支出进行了惩罚，因为每个人都想知道：我为什么不去投资英伟达？他们其实不需要投入那么多资金，因为他们是一家卓越的半导体设计公司。
The market punished the $80,000,000,000 in AI CapEx that Satya Nadella has been telling to investors because everyone's asking, well, I could be in NVIDIA, and they're not they don't need to really invest that much because they're a fabulous semiconductor design company.
他们的毛利率正在上升。
Their their gross margins are increasing.
他们正在源源不断地产生现金。
They're printing cash.
你说得对，好吧。
And you are saying, okay.
你得花800亿美元，但我们不知道什么时候才能从这笔投资中获得利润。
You gotta spend $80,000,000,000, and we don't know when we're gonna get the profits from that.
所以这里还有一个悬而未决的问题。
So there's an open question there.
因此，在涉及工具和铲子的生意时，你不想把自己绑定在某个初创公司或基础模型实验室身上。
So when it comes to the when it comes to the picks and shovels trade, you don't want to tie yourself up to an individual startup or a foundation model lab.
你只想持有那个最简单、价值最终会聚集其上的东西。
You just want to own the simplest thing that that that value will accrue to.
全程押注英伟达。
NVIDIA all the way.
他们是工具和铲子生意中明确的赢家。
They were the clear winner of the picks and shovels trade.
我认为微软的基础设施布局是个不错的决定。
Microsoft's infrastructure play, I think it's a good decision.
只是这些投资对股东来说尚未带来回报。
It's just the bets have yet to pay off for shareholders.
丹尼尔·格罗斯提出了一个更尖锐的问题：铜是否被定价错误？
Much more pointed question, Daniel Gross asked, is copper mispriced?
2024年1月，铜是否被定价错误？
Was miss was copper mispriced in January 2024?
答案是：是的，严重低估了。
The answer was, oh, yeah, majorly.
原材料的波动不像普通股票那样。
Now raw materials don't move like mean stocks.
实际上，2024年1月铜价为每磅3.75美元。
So the actual move, copper was $3.75 a pound in January 2024.
两年后，它涨至每磅6.61美元的历史高点。
Two years later, it went to an all time high of $6.61 a pound.
所以它甚至没有翻倍，但这种在我们已开采数百年的基础材料上的变动非常惊人，而这正是因为人工智能大量使用铜。
So it's not like it even doubled, but that type of move in a basic material that we've been mining for hundreds of years is remarkable, and that's because AI uses a lot of copper.
NVIDIA的GB200 NVL72服务器机架使用了超过5000根铜缆。
NVIDIA's g b 200 NVL 72 server rack uses over 5,000 copper cables.
虽然有72个GPU连接在一起，但你需要5000根铜线才能让它们协同工作。
There's 72 GPUs wired together, but you need 5,000 copper wires to get it all to work together.
如果把NVL72中所有的铜线拉直，从一端到另一端会长达两英里。
If you stretched out all of the copper wire in an NVL 72 from one side to another, it would go two miles long.
而这仅仅是一个服务器机架。
And and this is one server rack.
这并不是跨越整个数据中心的布线。
Like, this is, like it's not, like, going across the data center.
这并不是在把数据从AWS东海岸传到加利福尼亚。
This is not taking the data from AWS East to California.
这只是在那个单一服务器内部的布线。
This is within that one server.
把铜线拉直看看。
You stretch out the copper wire.
你这是在走两英里啊。
You're you're going two miles.
这数量太惊人了。
It's an incredible amount.
一个单体100兆瓦的数据中心——按现代标准来看只是个蚂蚁大小的数据中心——就需要大约3000吨铜。
A single 100 megawatt data center, which is a data center for ants by modern standards, needs around 3,000 tons of copper.
我觉得算下来差不多值五十万美元左右吧。
I think that's, like, half $1,000,000 or something like that when you multiply it all out.
或者不是。
Or no.
不是。
No.
几年内，全球数据中心每年将消耗五十万吨铜。
It's data centers broadly will be using half a million tons of copper annually in a few years.
有些人甚至说铜是新的石油，但其实也有很多东西都被称作‘新的石油’，尤其是在AI基础设施建设中。
And people are actually saying that copper is the new oil, but there are a bunch of things that are also the new oil, so in the in the AI build out.
所以这太复杂了。
So it's so complex.
到处都是瓶颈。
There's bottlenecks everywhere.
所以你得对这个说法打个折扣。
So you gotta take that with a grain of salt.
但如今铜确实
But copper today is certainly
看起来石油成了新的石油。
It's looking like oil is the new oil.
新的石油。
The new oil.
原油现在是多少？
The What is crude at?
80美元。
For 80.
80。
80.
在伊朗战争爆发之前，油价是70。
So it was at 70 before the the Iran war broke out.
当时有人预测油价会涨到100以上。
There were predictions that it would go to above a 100.
石油是另一种商品，它的走势与表情包股票或
Oil is another is another commodity that does not move in the same fashion as a meme stock or a
但今天除外。
Except for today.
但今天除外。
Except for today.
大幅上涨。
Big move Up
涨了8%。
8%.
但通常情况下，随着价格上升，企业会增加钻探，价格最终达到均衡。
But but typically as prices go up, the the the firms drill more and the prices reach equilibrium.
美国钻探公司并未急于增加石油产量，这一点今天登上了《华尔街日报》商业版的头版。
US drillers aren't rushing to increase oil production is on the front of The Wall Street Journal business section today.
中东正处在一场长期冲突的边缘，这场冲突可能将油价推至四年未见的高位。
The Middle East is on the cusp of a prolonged conflict that could push oil prices to heights not seen in four years.
目前，美国石油钻探公司选择按兵不动。
For now, American oil drillers are sitting this one out.
美国原油基准价格周三收于每桶74.66美元，为自去年六月以色列、伊朗和美国之间持续十二天的空袭以来的最高月度收盘价。
The US oil benchmark settled at $74.66 a barrel Wednesday, the highest front month settlement since the twelve day exchange of Israeli, Iranian, and American strikes last June.
但对于德克萨斯州的石油从业者来说，如果这场战争可能很快结束、原油价格随之下跌，那么增加昂贵的钻机并提高产量并没有意义。
But West Texas oilmen, it makes but to West Texas oilmen, it makes little sense to add expensive rigs and boost production when the war could be short lived and crude prices drop.
所以，这就是当前石油市场正在发生的情况。
So, that's what's happening in the oil markets.
总之，我们转到房地产领域。
Anyway, moving over to real estate.
丹尼尔·格罗斯问：旧金山是新的底特律吗？
Daniel Gross asked, is is San Francisco the new Detroit?
我不太确定他所说的‘新的底特律’指的是什么，是指当年作为汽车之城、制造惊人汽车的老底特律，还是指如今底特律已成空壳、昔日繁荣不再的样子？
And I am not exactly sure what he means by new new Detroit because new Detroit meaning like the old Detroit when it was Motor City and they were building amazing cars there or the New Detroit in the sense of like Detroit today is a hollowed out shell of what it used to be, a former boom town.
唯一明确的是，旧金山已经回来了。
The the one thing that's clear is that SF is back.
旧金山正在全面繁荣。
SF is completely booming.
办公楼空置率从36.9%下降至33.5%。
Office vacancy office vacancy fell from 36.9% to 33.5%.
OpenAI拥有百万平方英尺的办公空间。
OpenAI has a million square feet of offices.
Anthropic租下了一栋25层的塔楼。
Anthropic is a 25 story tower.
Sierra——一家应用层公司——在旧金山签署了30万平方英尺的办公空间。
Sierra, an application layer company, signed 300,000 square feet of office space in San Francisco.
2025年，湾区获得了78%的AI风险投资。
The Bay Area received 78% of AI venture capital in the 2025.
但这也有另一面。
And there is a flip side to this.
总体而言，旧金山的就业人数仍低于疫情前水平。
So overall employment in San Francisco is still down relative to the pre pandemic.
有些人离开了。
Some people left.
一些老牌公司招聘规模减少了。
Some legacy companies aren't hiring as much.
所有招聘都集中在AI初创公司区域，但房价依然坚挺。
All all the hiring is happening at the AI startup lab area, but housing prices remain strong.
这绝对不是一个空心化的外壳。
It's certainly not a hollow shell by any means.
如果你曾经去过旧金山，就会发现它比前几年更干净、更安全。
And if you've ever visited San Francisco, you can tell it's cleaner and safer than it has been in previous years.
所以，人工智能并没有彻底改变旧金山的格局，也没有变成开源或活动中心，科技也没有迁移到迈阿密、纽约、奥斯汀或洛杉矶，旧金山仍然是核心所在。
So AI overall was not a total rethinking of San Francisco, and it did not become open source or hub of activity and tech didn't move to Miami or New York Austin or Los Angeles, like San Francisco is still where it's at.
他接下来问的问题是，人工智能如何影响财富不平等？
The next question he asked, how does AI change wealth inequality?
现在下结论还为时过早。
It's sort of too soon to tell.
数据还不够清晰。
The data is not entirely clear.
数据并没有发生太大变化。
The data hasn't moved that much.
但有一些有趣的研究。
But there are some interesting studies.
国际货币基金组织在2025年发布了一份工作论文，指出人工智能可能减少工资不平等。
So the International Monetary Fund released a working paper in 2025 that said that AI could reduce wage inequality.
减少工资不平等。
Reduce wage inequality.
因此，人们的年收入可能会有所减少。
So the amount of money that people make on an annual basis could be, you know, reduced.
这怎么可能呢？
How is that possible?
这份论文中提到，高收入岗位，比如律师、高管、管理顾问的工作，将会比机械师、园丁、街头清洁工的工作更早被自动化。
Well, what they say in this paper is that high income tasks, the job of a lawyer, the job of an executive, the job of a management consultant those will be automated before the job of the machinist, the gardener, the street sweeper.
因此，高收入阶层的工资会下降，而低收入阶层的工资则保持稳定。
And so you will see wages at the high end fall, while wages at the low end remain stable.
而且，在低端，还有最低工资标准。
And also, at the bottom end, you have the minimum wage.
所以你会看到高收入工资下降，低收入工资保持不变，从而减少工资不平等。
And so you will see higher end wages go down, the bottom wages will stay the same, and that will reduce wage inequality.
另一方面，人工智能可能会加剧财富不平等，因为它将资本回报集中到了科技所有者手中。
On the other side, AI could worsen wealth inequality because it's concentrating capital returns into tech owners.
此外，经合组织发现，工资增长实际上在低技能职业中最为显著，比如装配工。
So the o e the OECD found that wage growth was actually the strongest in low skilled occupation, assemblers.
我甚至不知道还有这种工作，但如果你是装配工，你的工资已经上涨了11.6%。
I didn't even know that was a job, but if you're an assembler, you've seen your wages increase by 11.6%.
就是这样。
There we go.
你知道谁过得最艰难吗？
You know who's had it the hardest?
首席执行官。
CEOs.
那些高技能工作者，也就是首席执行官们，他们的工资只增长了2.7%。
The high skilled workers, those chief executives saw their wages increase by just 2.7%.
看来当CEO真不容易。
It's rough out there for a CEO apparently.
太残酷了。
Brutal.
这主要是因为最低工资上涨。
This is mostly because of minimum wage increases.
但总体而言，我们看到工资不平等在减少，而财富不平等却在加剧，因为目前股市高度集中。
But in general, we're seeing wage inequality decrease, but wealth inequality increase because there is incredible stock market concentration right now.
因此，美国七大科技公司——MAG七巨头，如今占标普500指数市值的32%，并在2025年贡献了42%的总回报。
So the MAG seven, the seven biggest tech companies in America, they now comprise 32% of the S and P 500 market cap, and they drove 42% of total returns in 2025.
所以，如果你投资了特斯拉、Meta、微软、苹果、谷歌、亚马逊等，还有英伟达。
So if you're invested in Tesla, Meta, Microsoft, Apple, Google, Amazon, etcetera, NVIDIA.
那你赚得不错。
You did very well.
你的财富增长了。
Your wealth increased.
但如果你没投资这些公司，而是持有一个广泛指数，或者一开始就没多少资本，那你就被甩在后面了。
But if you weren't in those and you had a broad index or you didn't have a lot of capital to begin with, you were left behind.
因此，这加剧了财富差距。
So that's increasing wealth.
保罗·格雷厄姆两天前分享说，现在公司增长得非常快。
Paul Graham shared two days ago, companies grow fast now.
这就是经济不平等加剧的原因，而不是某种阴险的政策转变。
That's the reason economic inequality is increasing, not some sinister policy shift.
对。
Yes.
他当时在强调Anthropic最近的增长。
He was highlighting Anthropics' recent growth.
他说，当公司快速增长时，会让创始人变得双倍富有。
He said, When companies grow fast, it makes founders doubly rich.
公司不仅更快地达到某个营收目标，而且在达到时价值更高，因为公司的价值是增长率的倍数。
The company not only hits a given revenue number sooner, but it is more valuable when it hits it because the value of the company will be a multiple of the growth rate.
不理解估值数学的人无法想象创始人会自然变得如此富有，而对创始人和投资者来说，这再明显不过了。
People who don't understand the math of valuations can't imagine that founders would get so rich naturally, whereas the founders and investors, it's the most obvious thing in the world.
这是科技界与政界之间存在巨大隔阂的原因之一。
This is one of the reasons there is such a disconnect between the tech world and politicians.
是的。
Yes.
所以，如果你想投资七巨头，或者不想投，去public.com看看。
So if you want to invest in the mag seven or you don't, over to public.com.
如今投资，必须认真对待。
Investing today is to take it seriously.
股票、期权、债券、加密货币、国债等，提供卓越的客户服务。
Stocks, options, bonds, crypto, treasuries, and more with great customer service.
正如保罗·格雷厄姆提到的，这在私募市场也同样成立。
So, this is also true in private markets, as Paul Graham mentioned.
AI初创公司的超级融资轮：OpenAI融资1100亿美元，Anthropic融资300亿美元。
AI startup mega rounds, a $110,000,000,000 for OpenAI, 30,000,000,000 for Anthropic.
这使得巨额私人财富集中在少数创始人和投资者手中。
This concentrates enormous private wealth among a small number of founders and investors.
因此，财富不平等，AI可能会改变它。
So wealth inequality, AI will probably change it.
至少，它会像以往的技术繁荣一样，带来同等程度的改变。
At the very least, it will change it as much as previous technology technological booms.
如果你是Instagram的创始人，你做得非常好，因为在移动技术这一波大繁荣中，突然间创造了大量财富。
If you were the founder of Instagram, you did very well because there was a lot of wealth creation all of a sudden in one major tech boom, mobile.
AI的范围和规模已经比移动技术更广更大了。
AI is is broader and bigger already than mobile.
接下来我们谈谈能源和数据中心。
So moving over to energy and data centers.
如果这真的变成了一场能源博弈，那该怎么做交易？
If it if it does become an energy game, what's the trade?
这确实变成了一场能源博弈，而交易策略基本上就是买入所有相关标的，因为每一种能源类资产都表现得非常好。
It did become an energy game, and the trade was buy everything, basically, because every energy thing basically did very well.
任何抓住这个机会的人都赚得盆满钵满。
Anyone who got the trade corrected very well.
Vistra的回报率达到321%。
Vistra returned 321%.
它是2024年标普500指数中表现第二好的股票。
It was the second best performing S and P 500 stock of 2024.
哇。
Wow.
你知道2024年谁超越了他们吗？
You know who beat them in 2024?
Palantir。
Palantir.
Palantir股价飙升，当时它已经入选标普500指数了。
Palantir mooned that they were already in the S and P 500.
他们表现非常出色。
They did very well.
但Vistra可以说是长期能源领域的赢家。
But, Vistra was sort of the the secular energy winner.
Constellation Energy在上市后规模扩大了三倍。
Constellation Energy tripled in size, after launch.
NRG能源公司，这个名字很棒，NRG，在2020年的上涨中上涨了95%。
NRG Energy, which is a great name, NRG, gained 95% in 2020 surge
12个月内上涨了700%，而且我认为他们甚至还没有产生任何能源？
700% in twelve months, and that's without, I think, even having any producing any energy?
没有。
No.
只是因为有这个想法
Just the idea that
他们将来会生产
they will produce
能源，是的。
energy Yeah.
总有一天。
Someday.
是的。
Yeah.
核电股一路飙升。
Nuclear went on a tear.
Oklo 更像是初创公司，还没有收入，或者说还没有产品。
Oklo is in more of, like, the startup pre revenue camp, I would say, or pre product camp.
但微软签署了一份价值160亿美元、为期20年的购电协议，以重启三哩岛核电站。
But Microsoft signed a $16,000,000,000 twenty year PPA to restart 3 Mile Island.
谷歌与Kairos Power签订了500兆瓦小型模块化反应堆的合同，而Meta则与多家核能供应商签订了总计6.6吉瓦的合约。
Google signed with Kairos Power for 500 megawatts of small modular reactors, and Meta contracted 6.6 gigawatts across multiple nuclear providers.
因此，从2024年AGI热潮初期开始，能源就成了一个绝佳的投资方向。
So energy was just such a great trade from the the the the beginning of the AGI boom in 2024.
在整个数据中心供应链的下一个阶段，哪些组件最难实现十倍扩张？
Across the next step across the entire data center supply chain, which components are hardest to scale up 10 x?
数据中心的‘芯片-晶圆-基板’是什么？
What is the chip on wafer on substrate of data center?
是的。
Yeah.
芯片-晶圆-基板，这正是台积电的核心技术。
So chip on wafer on substrate, that is TSMC's secret sauce.
这是限制TSMC两纳米芯片生产扩产的最关键因素。
That is the the the most gating factor in scaling TSMC's two nanometer chip production.
这使他们能够将HBM和GPU封装在同一个芯片上。
It's what allows them to package HBM and the GPU all on one chip.
TSMC表示，他们的产能早已被预订多年。
TSMC And reported that they were, like, sold out, like, years in advance.
这成了巨大的瓶颈。
It was a huge bottleneck.
在数据中心领域，最大的瓶颈可能是电力变压器。
In the data center world, the biggest bottleneck was probably power transformers.
我们听说，新电力变压器的交货周期在某些情况下超过三年，且供应缺口达30%。
We heard a lot about this lead time for new lead times for new power transformers reached over three years in some cases with a 30% supply shortfall.
高压断路器，我的天啊。
High voltage circuit break my in mom
公寓的托管？真的吗？
escrow on a condo Really?
而且交房的前提是开发方能获得变压器。
That and closing was contingent on the development getting a transformer.
她等了大概六周左右。
She waited like six weeks or something like that.
是的。
Yeah.
他们说，我们也不知道什么时候能拿到。
They were like, we don't know when we're gonna get it.
她最后干脆退出了。
And she ended up just backing out.
哇。
Wow.
是的。
Yeah.
就是因为需求太大了。
Just because there was there was so much demand.
展开剩余字幕（还有 480 条）
是的。
Yeah.
我的意思是，我们真的被这个打了个措手不及。
I mean, really got caught off we we really got caught off guard by this.
我的意思是，自2020年以来，变压器的成本上涨了150%。
I mean, transformers, the cost surged a 150% since 2020.
这是一种一百年前的技术。
It's a 100 year old technology.
它并不新。
It's not it's not new.
也不是什么疯狂的创新。
It's not crazy innovation.
一夜成名。
Overnight success.
一夜成名。
Overnight success.
但它成为了数据中心连接电网速度的瓶颈。
But it became the binding constraint on how fast data centers could connect to the grid.
如果你经营一家变压器公司，卖变压器，那就得上 Shopify。
If you are running a transformer company, you're selling transformers, gotta get on Shopify.
Shopify 是一个与你的业务共同成长的商业平台，让你能够在线、店内、移动设备、社交媒体、电商平台，以及现在通过 AI 代理，几秒钟内完成销售。
Shopify is a commerce platform that grows with your business and lets you sell in seconds online, in store, on mobile, on social, on marketplaces, and now with AI agents.
如果你在卖变压器，那就先去跑几个流程吧。
If you're selling transformers, start running some paths.
想象一下用 Shop Pay 购买一台变压器的体验。
Imagine checking out for a transformer with Shop Pay.
那将是一次无比棒的体验。
It'd be an incredible, incredible experience.
煤炭。
Coal.
煤炭被定价错误了吗？
Was coal mispriced?
有点吧，但远不如铜那么多。
Sort of, but not nearly as much as copper.
动力煤价格在2020年至2022年间下降了22%，但在2026年初略有回升。
Thermal coal prices actually declined 2020 22% in 2025, but they rebounded a little bit by early twenty twenty six.
煤炭股表现不错。
Coal stocks did fine.
Peabody Energy在十二个月内上涨了34%。
Peabody Energy gained 34% over twelve months.
Console Energy上涨了37%。
Console Energy, that was up 37%.
在运营方面，美国的燃煤发电。
On the operational side, US coal fired generation.
我喜欢你这么说：煤炭股表现不错。
I love how you're like, coal stocks did fine.
是的。
Yeah.
它们只上涨了3437%。
They were up only 3437%.
我的意思是，与其他翻倍甚至三倍上涨的相比，这简直就像，嗯，算了，好吧。
I mean, compared to everything else that's doubling and tripling, it's like, nah, you know, okay.
你知道，你之前还说，我是搞煤炭的。
You know, you were like, I'm the coal guy.
我相信煤炭才是那个东西，就像……
I I I believe coal is the thing like
Peabody Energy在过去六个月里上涨了73%。
Peabody Energy is up 73% in the last six months.
开玩笑的。
Just kidding.
才37%。
Only 37.
今年只上涨了161%。
Only up a 161% in year.
是的。
Yeah.
它只是表现得还不错。
Only It did only up It did fine.
过去五年里只上涨了728%。
Only up 728 percent in the past five years.
它表现得还不错。
It did it did fine.
只是不像铜那样令人兴奋，也不像核能那样令人兴奋。
It's just not as exciting as as copper, and it's not as exciting as as nuclear.
煤炭发电在美国整体增长了13%，在俄亥俄州增长了23%，在俄克拉荷马州增长了58%，那里是数据中心的热点地区。
Coal generation surged 13% overall in in America, 23% in Ohio, 58% in Oklahoma, where they're sort of data center hotspots.
因为那些地区有更多数据中心。
Like, there's more data centers in those in those areas.
所以向电网注入更多煤炭，有助于实现负载平衡。
So bringing more coal onto the grid allowed for, you know, more load balancing.
总之，各国。
Anyway, nations.
谁赢了，谁输了？
Who wins and loses?
猜猜谁赢了。
Take a wild guess at who won.
美国，宝贝。
America, baby.
美国赢了。
America won.
下一个问题。
Next question.
但这是真的。
But it's really true.
这些数据简直疯狂。
Like, the stats are crazy.
在过去两年的AI时代，美国确实是真正的主导赢家。
The United States is truly the dominant winner of the of the last two years in the AI era.
仅2024年，美国在私人AI投资上就达到了1090亿美元，如果算上亚马逊1100亿美元的资本支出指引，以及分别投入110亿给OpenAI、30亿给Anthropic、大量资金流向X AI，这个数字还要高得多。
$109,000,000,000 in private AI investment in 2024 alone, way more now if you look at 200,000,000,000 CapEx guide from AWS, 110 going to OpenAI, 30 going to Anthropic, tons of money going into x AI.
真的，资金正在源源不断地流向美国。
Like, the money is truly flowing in America.
2024年，中国在私人AI公司上的投资仅为93亿美元。
In 2024, China had just 9,300,000,000.0 invested in private AI companies.
自2013年以来，美国在AI领域的累计投资已达4700亿美元，超过其他所有国家的总和。
There's 470,000,000,000 cumulative since 2013, more than all other countries combined in America.
2024年，美国推出了40个重要的AI模型，而中国只有15个。
The US produced 40 notable AI models in 2024 versus China's 15.
不过，这场竞赛还没结束。
Now, the game's not over.
现在，许多国家都在进行大规模投资。
Lots of countries are making big investments now.
你不能忘记法国。
And you can't forget France.
法国投入了3000万欧元。
France is €30,000,000.
他们投入了3000万欧元。
They put in 30,000,000.
是的。
Yeah.
对。
Yeah.
这很严肃。
It's it's serious.
这很严肃。
It's serious.
让我告诉你关于纽约证券交易所的事。
Well, let me tell you about the New York Stock Exchange.
想改变世界？到纽约证券交易所融资吧。
Wanna change the world, raise capital at the New York Stock Exchange.
就这么做。
Just do it.
所以
So
下一个。
Next up.
印度。
India.
印度现在怎么样？
What's going on in India?
DG说，印度GDP出口中的2500亿美元本质上是GPT-4的令牌。
DG says 250,000,000,000 of India's GDP exports are essentially GPT 4 tokens.
那现在会发生什么？
What happens now?
约翰说，这已经开始成形了，但还为时过早。
John says, this is starting to take shape, but it's still early.
情境感知正在做一些与招聘下降相关的空头押注。
Situational awareness is placing some short bets which map to the decline in hiring.
我认为他们做空了印孚瑟斯，这是一家IT外包公司。
I think they were short Infosys, which is an IT outsourcer.
你需要一些软件被编写出来。
You need some software written.
你基本上只需要发送一个提示，然后就会得到代码，而这些模型在这方面非常擅长。
You will basically send a prompt and, basically, you would get back code, which is what these models are fantastic at.
因此，这确实意味着在近期乃至长期来看，竞争都异常激烈。
And so that does seem like some really, really steep competition in the near term, even in the long term.
所以我认为他有理由质疑这一点。
And so I think he's right to question that.
尽管我们目前在印度的整体招聘数据和菲律宾的整体招聘数据中还没有看到这种迹象，但在印度的IT出口领域已经出现了一些苗头。
Although we aren't seeing it in the overall India hiring data, the overall Philippines hiring data just yet, but it is showing up a little bit in India's IT export center.
因此，2024年和2025年，主要的印度IT公司合计裁员约58,000人。
So major Indian IT firms collectively in 2024 and 2025, they shrunk by 58,000 employees roughly.
这与2021年至2023年期间它们大量招聘的情况形成了鲜明对比。
And that's a dramatic reversal from the previous three years, 2021 to 2023 when they were staffing up.
那段时间，它们增加了36万名员工。
They added 360,000 employees.
所以是的。
So Yeah.
这里唯一的问题是，大型科技公司的情况在某种程度上是大型科技公司招聘的间接反映，因为这些印度公司所从事的大量工作实际上是由它们承接的。
The only thing here is that big tech this, like, is in some ways, like, proxy for big tech hiring because a lot of a lot of the labor that's happening at these firms is effectively offloaded to to to some of these firms in India.
所以某种程度上，这就像大型科技公司之前可能存在过度招聘。
So it's like, in some ways, it's like, if there was over hiring Mhmm.
在大型科技公司中，可能确实存在过度招聘的情况。
In big tech, there was could have been effectively over hiring here.
是的。
Yeah.
然后它在回撤。
And then it's re retracing.
是的。
Yeah.
有道理。
Makes sense.
而且，没错，你没提到这一点，但看看菲律宾的招聘活动也很有意思，其下滑幅度并没有人们想象的那么剧烈。
And, yeah, you didn't put it in here but it it's also interesting to look at the hiring activity in The Philippines which is not as dramatic to the downside as as one might think.
你什么意思？
What do you mean?
哦，只是。
Oh, just
当我们几周前查看数据时。
When we looked at the data, I think it was a couple weeks ago.
并不是说那里只是出现了大量职位削减。
Not like you there was just like job Yeah.
工作就像断崖式下跌。
Jobs, like, falling off of a cliff.
是的。
Yeah.
其实是相对稳定的。
Like, it's relatively stable.
对。
Yeah.
这些统计数据就像一圈圈的同心圆。
There's sort of these, like, concentric circles with all of these stats.
比如，看看Block公司，那里发生了什么？
Like, if you look at Block, what's happening at Block?
他们裁掉了50%的员工。
Like, well, they lost 50% of their workforce.
这是巨大的削减。
Like, that's a massive cut.
然后你看看整个科技行业的招聘情况，它正在放缓。
And then you look at tech hiring overall, and it's, like, slowing down.
你看看白领工作，我们并没有增加多少。
You look at white collar work, and it's, like, we're not adding a lot.
它可能正在缩小0.1%。
It's maybe shrinking by point 1%.
然后你再放眼整个美国经济，实际上它正在增长。
And then you zoom out to the American economy, and it's, like, it's actually growing.
所以你看到了这些层层递进的影响圈。
And so you have these, like, concentric circles of impacts.
因此，如果印度有一家从事IT外包的公司可能受到冲击，即使它是表现最差的，但管理最好的那一家或许能挺过去。
And so if there's, like, one company in India that's doing IT outsourcing that might be hurt, and then they're the least they're the worst performing, but the best managed one might be able to make it through.
然后你再放大视角，会发现整个印度经济状况还不错。
And then and then you zoom out and you're like, well, the overall Indian economy is doing okay.
而总体来看，全球经济也还不错。
And then overall, the global economy is doing okay.
所以这些影响会层层递进。
And so there's all these, like like, cascading effects.
我认为有趣的启示是，当像达里奥这样的人说，50%的白领工作可能会被自动化时。
And I think the the interesting takeaway was when when somebody like Dario says 50% of white collar work might be automated.
会是底层的那50%。
Like, it's gonna be the bottom 50%.
这并不是随机发生的。
It's not gonna be the it's not gonna be at random.
不是靠抛硬币来决定你是否能保住工作或是否从基层做起。
It's not gonna be flip a coin if you keep your job or entry level.
关键在于，你是否没有处于你所在领域的顶尖水平？
It's gonna be it's gonna be were were you not at the top of your were you not at the top of your game?
所以问题来了，如果我们正在取代软件工程师，那么软件工程师会去当机械师吗？
So the question is, if we are displacing software engineers, will software engineers become machinists?
这正是丹尼尔·格罗斯提出的问题。
This is what Daniel Gross asks.
他说，以往技术革命中的技能再培训的欧几里得距离是多少，而通用人工智能又如何与之相比？
He says, what is the Euclidean distance of reskilling in prior revolutions and how does AGI compare?
打字员变成了行政助理。
The typist became an executive assistant.
软件工程师能成为机械师吗？
Can the software engineer become a machinist?
但我们目前还没有看到这种情况。
So we're not seeing this yet.
软件工程师还没有立即转向蓝领工作，但数据显示，软件开发人员和程序员之间已经开始出现分化。
Software engineers aren't grabbing blue collar jobs just yet, but there is a divergence that's starting to show up in the data between software developers and programmers.
编程岗位正在减少，但软件工程岗位实际上在增长。
So programming jobs are in decline, but software engineering jobs are actually growing.
因此，我认为随着编码模型和智能系统允许在更高的抽象层次上构建系统，对AI工程师的需求增长了143%。
And so I think what's happening is that as coding models and agentic systems allow for building systems at higher levels of abstraction, demand for AI engineers has grown 143%.
所以，如果你要雇人帮你开发软件，你希望他们是原生AI思维的人，希望他们超越全栈开发——而全栈通常意味着你能用JavaScript做前端、用Python做后端。
And so if you're going to hire someone to help you build software, you want them to be like AI native, you want them to go beyond full stack or which typically meant you could do front end in JavaScript and back end in Python.
现在，全栈工程师或AI工程师意味着从提示词设计、产品开发、部署、运维到DevOps、数据库迁移，再到前端和后端，所有这些都将由智能体系统来处理。
And now full stack or AI engineer means, everything from prompt to design to product development to deployment to operations and DevOps to database migrations to front end and back end because all of that is gonna be handled by agentic systems.
因此，顶级科技公司的初级招聘人数下降了25%。
And so entry level hiring at top tech firms has dropped 25%.
可能是AI的原因，也可能是疫情后过度招聘的余波。
Maybe AI, maybe overhiring from COVID overhang.
我们拭目以待。
We'll see.
实习岗位发布减少了30%。
Internship postings fell 30%.
所以，你的选择要么是向上转型，成为所谓的‘无限思维的管理者’——我认为这是原话，也就是管理智能体的角色；要么，也许现在是时候去提升那些机械师的技能了。
So the decision is either, you know, reskill upwards to more of a manager of infinite minds, I think was the quote, but manager of agents role or, yes, you might it might be time to work on those machinist skills.
我对泰勒极为看好。
I I'm extremely bullish on Tyler.
我们曾亲眼见证他组装了美国第一台iPhone，而且他还是一个极其出色的软件开发者。
We watched him assemble the first iPhone ever in America, and he's also an incredible software developer.
他其实算不上程序员，因为他并不真正掌握所使用的编程语言，但这种说法不对。
Not really a programmer because he doesn't really know the programming languages that he employs, but Not true.
这不对。
That's not true.
就是这样。
There we go.
是的。
Yes.
我认为无论泰勒成为无限思维的管理者还是机械师，他都能成功。
I think Tyler makes it whether he becomes the the manager of infinite minds or the machinist.
我对那种浅蓝色的论点相当看好，那就是你会有机器人，但你只是会用，比如
I I I am pretty bullish on like the like the light blue color like thesis, which is that like you you'll have robotics, but you'll just have, like
好。
Yeah.
你会在机加工车间里，但只是用iPad操作。
So you'll be in the machine shop, but you're just using an iPad.
你其实并没有在做实际的工作。
You're not actually working on stuff.
是的。
Yeah.
我觉得我们会把你丢进一家耐克工厂。
I think we drop you in a Nike factory.
你两周内就能管起整个工厂。
You're running that place in, two weeks.
保证没问题。
Guaranteed.
我不是在开玩笑。
I'm not kidding.
我非常看好那些能够熟练使用所有工具、真正理解现代系统赋予他们的优势、而无需依赖旧有习惯的年轻人。
Like, I I I I'm so bullish on on young people who can use all the tools and and actually go and understand the leverage that they're getting from the modern systems natively and not need to fall back to old habits.
是的。
Yeah.
昨天有一篇关于高露洁的好期刊文章，是的。
There was a good journal article yesterday about the Colgate, like Yeah.
AI负责人。
Head of AI.
是的。
Yeah.
这非常有趣，因为我心想，哦，我觉得我完全可以胜任这份工作。
And it was very interesting because I was like, oh, this is like, I feel like I could do a pretty good job with this.
接在我后面。
Coming after me.
你就是说，是的，是的，是的。
You're like, yeah, yeah, yeah.
但你知道，他其实
But you know, he's
基本上就是在告诉所有人：是的，伙计们，你们真的得用好AI，不管那些随机的调查怎么说
just basically telling everyone like, yeah, guys, you actually gotta use AI like despite what these these random surveys might
这人一辈子就干过一份工作，做播客，却觉得自己能在大型牙膏市场称霸。
guy has had one job in his life working on a on a podcast and thinks he could go thinks he could go dominate in the big toothpaste space.
是的。
Yeah.
我喜欢这种自信。
I love the confidence.
我们是运动员。
We're athletes.
我们
We
明白了。
got it.
我们是企业运动员。
We we got corporate athlete.
你是个企业运动员。
You are a corporate athlete.
我们得请科莱特那个家伙来节目里，看看谁的表演更好。
We gotta get him we we gotta the Colgate guy on the show and see see whose takes are better.
泰勒·科斯格罗夫还是科莱特AI guy。
Tyler Cosgrove or or Colgate AI guy.
总之
Anyway
聊天室里说你不像是搞供应链的，泰勒。
The chat's saying you're not they don't think you're a supply chain guy, Tyler.
我觉得你能很快上手。
I think I think you could pick it up.
问问OM吧。
Ask ask the OM.
克劳德，我该怎么做？
Claude, what do I do?
是的。
Yeah.
别搞错了。
Make no mistakes.
为了克劳德。
For Claude.
对吧？
Right?
让我告诉你关于
Let me tell you about what
如果发生
happens if
如果我按下这个呢？
I hit if I hit this?
让我看看。
Let me see.
这个会接管吗？
Does that take does that take over?
Fin.ai，全球排名第一的客户服务AI代理。
Fin.ai, the number one AI agent for customer service.
如果你想让AI处理你的客户支持，请访问fin.ai。
If you want AI to handle your customer support, go to fin.ai.
电气化和流水线生产导致了高失业率，而新政包括公共事业振兴署——一项联邦项目，以巨额预算雇佣了八百五十万美国人。
Electrification and assembly lines lead to high unemployment and the new deal, including the Works Progress Administration, a federal project that employed eight and a half million Americans with a tremendous budget.
这会重复吗？
Does that repeat?
所以显然还没发生。
So clearly hasn't happened yet.
我们也没有看到高失业率，所以还得等等看。
We also haven't seen high unemployment yet, so we'll we'll see.
特朗普政府于2025年7月推出了美国AI行动计划。
The Trump administration did launch America's AI action plan in July 2025.
我们昨天在节目中采访了该AI行动计划的作者之一迪恩·鲍尔。
We talked to one of the authors of that AI action plan, Dean Ball, yesterday on the show.
他们还发布了关于人工智能教育和技能培训的行政命令。
And they issued some executive orders on AI education and skilled trades.
但劳工部和劳工部授予了8400万美元的学徒计划扩展拨款，但这就像法国的钱一样。
But the Department of Labor and the the Department of Labor awarded $84,000,000 in apprenticeship expansion grants, but that's like France money.
这点钱根本不足以产生实质性影响。
That's like not enough to do to really make an impact.
总体而言，美国的劳动力发展支出仅占GDP的0.1%，在经合组织国家中排名倒数第二。
Overall, US workforce development spending is at point 1% of GDP, which is second to last among OECD nations.
我想知道谁是倒数第一。
I wonder who's actually last.
我们应该查一下。
We should look that up.
无论如何，目前没有任何项目能接近公共事业振兴署或新政的规模。
Anyway, no program currently approaches the scale of the Works Progress Administration or the New Deal.
我认为，如果数据显示我们确实在大量裁员，这种情况会很快发生巨大变化。
I think this will change very, very quickly if the data comes through that we're actually shedding tons of jobs.
美国非常擅长创造新工作和印钞票，这整个反应都是针对西特里尼文章的。
America is very good at creating new jobs and and and running the printing press, and And this is the whole reaction to the Citrini article.
无论如何，他还提出了一个更哲学性的问题，这个问题其实没有详细的答案，但我很想知道你的看法，乔治迪。
Anyway, he also asks a more philosophical question that doesn't really have a detailed answer, but I'd love your take on it, Geordie, too.
终身学习值得投资吗？除了掌握技能的经济价值之外，它本身是否值得去做？
Is lifelong learning worth investing in, Something worth doing beyond the economic value of mastering the task?
这是一个非常抽象的问题。
It's a very abstract question.
这是一个非常个人化的问题。
It's a very personal question.
我倾向于认为值得。
I tend to think yes.
我认为在马斯洛需求层次中，你需要食物、住所、家人、朋友、影响力之类的。
I tend to think that on Maslow's hierarchy of needs, you you need food, shelter, family, friends, clout or whatever.
到了某个阶段，为了学习而学习一项技能，即使在一个你明白的世界里，也像去健身房一样有益。
At a certain point, learning a skill for the sake of learning that skill is edifying even in a world I, you know, it's like going to the gym.
比如，我听萨姆·阿尔特曼说过，今天出生的孩子永远不会比最聪明的AI系统更聪明。
Like, I was I was listening to Sam Altman talk about how kids will never kids that are born today will never be smarter than the smartest AI systems.
这很奇怪。
And that's weird.
这不一样。
That's different.
但我想到，几百年前，你确实可能是世界上最好的人选，比如搬运木材，除了牛之类的动物。
But I was thinking about like but like I was there was a time hundreds of years ago when you could actually be this the the the best thing in the world, the best option for like carrying lumber other than, like, maybe an ox or something.
但总的来说，在我们发明机械、发明汽车之前，你完全可以是跑得最快的人。
But, like, in general, like, before we created machinery, like, before we created the car, like, you could be the fastest person.
但一旦ZR1斧头问世，你就毫无胜算了。
And then as soon as as soon as the the the the the the z r one axe was released, you didn't stand a chance.
我不得不向我儿子解释过这一点。
And I had to explain this to my son at some point.
他对猎豹非常感兴趣。
Was he was very interested in cheetahs.
我告诉他，猎豹是跑得最快的动物。
And and I was telling him, like, cheetahs are the fastest animal.
他问：‘那它比车还快吗？’
And he was like, but is it faster than a car?
我说：‘根本没法比。',
And I was like, not even close.
坐在车里的我，轻松甩开那只猎豹。
Like like me in a car, I'm smoking that cheetah.
我比那只猎豹慢，但我快得多。
I am slower than that cheetah, but I'm way faster
是的。
Yeah.
在我的凯迪拉克里。
So in my Cadillac.
DG说，有些事值得去做，超越了掌握这项任务的经济价值。
DG says, something worth doing beyond the economic value of mastering the task.
是的。
Yes.
而我会这样反过来想：我生活中大部分的收获都来自于掌控自己、了解自己、认识自己，然后理解这个世界。
And the way I would kind of flip this, I feel like most of the gains in my life have come from mastering myself or understanding myself, learning about myself, and then understanding the world Yeah.
然后以某种方式将这两者结合起来。
And then combining those two things in a way.
并不是专注于掌握某一项技能。
It hasn't been about mastering any one Sure.
这个不错。
Specific This is good.
任务。
Task.
是的。
Yeah.
所以我仍然认为，重新理解自己、理解世界，以及这两者如何契合，能带来巨大的收获。
And so I still think there's tremendous tremendous gains to, again, under understanding yourself, understanding the world, how do these two things fit together.
所以我同意你的观点，不仅在于要掌握具体任务，更在于那些有价值的思想、概念以及技能的组合。
So I agree with you on just the on just the level of, like, there will be not just tasks to master, but, like, ideas and cons and and combinations of skills that will be valuable economically.
但即使在一个超智能AI世界中，任何任务掌握都毫无经济价值，我仍然认为投资于终身学习是值得的，因为这会让人获得成长。
But even in a in an ASI world where there is no economic value to any task mastery, I still think it is worth investing in lifelong learning because that will be edifying.
这会让人感到满足。
That will be satisfying.
是的。
Yeah.
这仅仅是一种享受。
It's just enjoyable.
就像成为一名画家。
It it it, you know, it's becoming a painter.
成为那种即使在相机和生成式AI已经存在的时代，依然纯粹从绘画风景的过程中获得快乐的人。
Becoming someone who who gets joy purely out of the process of painting the landscape when cameras exist, when generative AI exists.
即使结果没有经济价值，也远不如相机或生成式AI的成果，但坐在那里观察风景、绘画并最终完成作品，依然能带来成长和价值。
It can still be edifying and valuable to to sit there and look at a landscape and paint it and wind up with the result even if the result is not economically valuable or or anywhere near what you could get with technology like a camera or Gen AI.
所以非常抽象，但仍然是个有趣的思考话题。
So very abstract, but still something fun to think about.
在我们转向通货膨胀之前，让我跟你们说说CrowdStrike。
Before we move on to inflation, let me tell you about CrowdStrike.
你的业务是人工智能。
Your business is AI.
他们的业务是保障它的安全。
Their business is securing it.
CrowdStrike保障人工智能安全并阻止入侵。
CrowdStrike secures AI and stops breaches.
所以如果人工智能真的具有通缩性，我们该如何知道呢？
So if AI is truly deflationary, how would we know?
哪个图表或指标会最先显示出这种趋势？
What chart or metric would show it first?
你可以把人工智能API的价格看作这里最可能的答案。
And you could look at AI API pricing as probably the best answer here.
所以，你知道，现在AI能够完成某些任务了。
So, you know, there are certain tasks that AI can do now.
因此，这些任务不再受限于法规、人力输出或劳动力人口，而是受限于计算能力的供给。
And so those tasks have moved from being limited by regulation or human output or the population workforce to the compute allowed or compute allocated.
而计算成本正在下降。
And the cost of compute is falling.
以前电视很贵。
So TVs were expensive.
随着我们建造了越来越多、越来越自动化的工厂，能够生产出越来越便宜的电视，价格也就降了下来。
They got cheaper as we made more and more factories that were more and more automated that could produce ever cheaper TVs.
这就是为什么当你观察通胀趋势时，会看到医疗、教育这些高度受监管的领域。
And that's why when you look at the inflation trends, see health care, education, anything that's highly regulated.
能够通过律师资格考试的人数是有限的。
There are only so many lawyers that are allowed to pass the bar.
医生的数量也是有限的。
There are only many doctors.
因此，医疗成本一直在上涨，而电视则经历了全球最残酷的公司之间的激烈竞争。
And so health care has increased in cost, where TVs has been this knockout, drag out fight between the most ruthless companies all over the world.
电视的价格已经大幅下降。
And the price of TVs has fallen.
所以，如果你开始将更多这类工作交给AI，放在AGI曲线上，你就会看到通缩。
And so, if you start putting more of those jobs on the AI, on the AGI curve, you see deflation.
我们首先会在哪里看到这种现象？
Where should we see it first?
我们首先会在AI和API定价上看到。
We should see it in AI, API pricing.
GPT-4等同的推理成本已经暴跌。
And GPT-four equivalent inference costs have collapsed.
在2022年底，也就是GPT-4推出之前，价格是怎样的。
In late twenty twenty two, they were which is like right before GPT-four.
我认为GPT-4是在2023年1月发布的。
I think GPT-four came out January 2023.
2025年12月时，价格是每百万令牌20美元。
It was $20 per million tokens in December 2025.
同样的模型，现在每百万令牌只需0.04美元。
The same model, $0.04 0 per million tokens.
因此，这三年内成本下降了50倍。
So this is a 50x decline in three years.
我认为，根据你如何推理这些模型，你甚至可以用更低的成本获得更多的智能。
I think that you can even get more intelligence for cheaper depending on how you're inferencing these.
你可以将一些模型本地运行，此时能源成本还会更低。
You can run some of these models locally, and the energy cost is even cheaper at that point.
因此，这三年内50倍的成本下降，速度比个人电脑计算成本的下降还要快。
So this 50x decline in three years, this is falling faster than PC compute costs.
它下降的速度也快于互联网泡沫时期的带宽成本。
It's also falling faster than .com era bandwidth costs.
因此，这应该是下游服务通缩的领先指标。
And so this should be the leading indicator for downstream service deflation.
你举的任何一个例子，比如，我刚用ChatGPT检查了一下，我认识一个人说他特别喜欢ChatGPT，因为他的发动机故障灯亮了。
In any example that you give of like, oh, I was just able to use ChatGPT to check my I talked to somebody who said they love ChatGPT because their check engine light came on.
他拍了一张照片，ChatGPT就识别出车型，并告诉他需要做什么。
They were able to take a picture of it, and it identified the car, told them what they needed to do.
当他去维修时，他觉得自己不会被修车师傅宰了，因为他自己身边就像有个机械师在帮他撑腰、替他辩护。
And they felt like they when they went in for servicing, they weren't gonna get raked over the coals by their mechanic because they had, like, sort of a mechanic in their camp, in their fighting in their court.
所以这实际上相当于你身边有个自己的机械师，你可以自己动手，也可以去店里谈判，说：嘿，我知道这服务的合理价格是50美元，不是200，给我个公道价。
And so this should this should effectively act as like, okay, you have your own mechanic on your side and then you can do it yourself, you can also just go in and negotiate and say, hey, I know that this the fair price for this is $50, not 200, so give me the good price.
是的。
Yeah.
所以你会看到这种变化。
And so you should see that.
你会看到这种变化。
You should see that.
对。
Yeah.
你可以假设令牌的成本会越来越低，是的。
You can assume that tokens are gonna get cheaper Yep.
而且会无限地变得更高效。
And more effective indefinitely.
是的。
Yes.
但所有这些新产品发布，比如Open Claw，实际上正在扩大。
But all these new product launches, things like Open Claw are actually expanding.
令牌的数量。
The number of tokens.
是的。
Yes.
所以你看到的是通货紧缩，但同时也看到了杰文斯悖论带来的需求增长，这就是整体增长上升的原因。
So you are seeing deflation but then you're also seeing Jevan's Paradox and increase in demand and that's why overall growth is increasing.
那么，一个人应该如何，从历史上看，
So how should one Historically,
就像一个人用一个AI聊天，你只使用这么多。
it's like one person, one AI chat, you're only using so much.
然后你有了智能体产品。
Then you have agentic products.
你可以实际增加一个人能运行的智能体数量，产出会飙升。
Then you can actually multiply the number of agents that an individual can be running and the output goes through the roof.
是时候去加斯敦了，宝贝。
It's time to head to Gastown, baby.
所以他提出了一个后续问题：如果对知识产品的需求在生产成本下降的同时持续增长，人们该如何看待通缩？
So he asks this follow-up question, how should one think of deflation if demand for intellectual goods continues to grow as production costs go down?
AI API成本正在暴跌，但AI实验室的收入却在飙升。
AI API costs are plummeting, but AI lab revenues are skyrocketing.
所以价格下降，用量激增，总支出增加。
So prices fall, volume surges, total spend increases.
而真正有趣的是，我们正经历一场SaaS浩劫。
And the really interesting wrinkle is we're living through a SaaS pocalypse.
我们正经历着这场训练热潮，但SaaS供应商却像征收了一种AI税，我记得数据显示续约费用上涨了2037%。
We're living through this training article, but SaaS vendors are imposing like an AI tax of, I think the stats were somewhere between 2037% of renewals.
所以AI供应商，当你去续费时，他们会说：嘿，我们新增了AI功能。
So AI vendors, you you go for your renewal and they say, hey, we got AI features now.
你得再多付一点钱。
You got to pay a little bit more.
即使构建软件的底层成本趋近于零，SaaS的收入仍在持续增长。
SaaS revenues are still increasing and growing even as the underlying cost to build software approaches zero.
这与摩尔定律下计算技术的历史如出一辙。
This mirrors computing's history under Moore's Law.
每个单位的成本都在降低。
Each unit gets cheaper.
随着新应用场景的出现，总支出持续增长。
Total spending grows as new use cases emerge.
这正是经典的杰文斯悖论。
So classic Jevan's paradox.
让我跟你聊聊Figma。
Let me tell you about Figma.
无论你的想法从哪里开始，Figma都可能连接代码、Codex或草图，而Figma画布正是想法构建、连接并使产品成形的地方，用Figma确保方向正确。
No matter where your idea starts, Figma may clog code, codex, or or a sketch, the Figma canvas is where ideas build connect and products take shape, build in the right direction with Figma.
所以是地缘政治。
So geopolitics.
这很有趣，因为他把这个问题放在了地缘政治部分，但这个问题其实更像一个关于AI服务器建设的问题，他问的是：互连真的重要吗？
This was interesting because he put this question in the geopolitics section that feels much more like a server AI build out question, but he asks, does interconnect actually matter?
我当时非常困惑，因为互连通常指的是服务器之间的以太网线，或者NVLink，或者TPU在类似三环拓扑结构中所做的连接。
And I was very confused because interconnect typically refers to sort of like the Ethernet cables between servers or like NVLink or what TPU is doing, Ironwood, between, like, the three torus topology.
我认为这个问题的答案是：绝对重要。
And I think that the answer to that is, like, absolutely.
在大型GPU集群中，30%到50%的训练时间都花在了GPU之间的通信上，而不是实际的计算。
In large GPU clusters, 30 to 50% of training time is spent on inter GPU communication, not actually computation.
因此，关于TPU v7 Ironwood的所有炒作，都源于其三维环形拓扑技术，这种技术将9216个芯片连接在一起。
And so all the hype around TPU v7 Ironwood is due to their three d Taurus technology, the topology that connects 9,216 chips together.
NVIDIA的NVL 72可以将72个芯片连接在一起。
NVIDIA's NVL 72 connects 72 chips together.
互连至关重要。
Interconnect is incredibly important.
我在想，为什么互连会被放在地缘政治部分，是不是在暗示三星和SK海力士在韩国、台积电在台湾与美国之间的关联？
I'm wondering if the fact that interconnect is in the geopolitics section is maybe talking about the interconnections between Samsung and SK Hynix in South Korea and Taiwan TSMC and America.
这显然相关，我们会在下一个问题中深入探讨。
That's obviously relevant, we'll get into that in the next question.
但两者都极其重要。
But it is both are incredibly important.
所以，我认为无论你怎么理解这个问题，互连确实很重要。
So, I do think interconnect actually matters no matter where no matter how you read this question.
他在这里提出的一个更大、更引人注目的问题是：如果一个国家拥有更多能源，制程节点还重要吗？
The bigger question and the interesting, like, hot take question he has in here is, does node process matter if the country has more energy?
中国一直未能突破其半导体制造能力，其本土晶圆厂中芯国际（SMIC）进展受限。
So China has not been able to get their semiconductor manufacturing, their fab champion, SMIC.
他们有中芯国际，这是台积电的等效企业。
They have SMIC, which is the TSMC equivalent.
他们有华为，这是英伟达的等效企业。
They have Huawei, which is the NVIDIA equivalent.
他们还有盛美，这是他们版的阿斯麦。
And they have SME, which is their ASML equivalent.
他们未能将中芯国际复制到台湾半导体那样的水平。
They have not been able to clone TSMC to the level of Taiwan Semiconductor.
但他们能够生产14纳米的芯片。
But they can produce 14 nanometer chips.
所以问题在于，他们通常会把这些芯片用在各种随机的消费电子产品中，然后运送到这里。
And so the question is, usually they take those chips and they put them in just random, you know, consumer electronics goods that they ship over here.
所以，如果你去中国，想找一家封装厂或制造商为你生产产品，你说：是的，我想让它能连接互联网，或者我想让它内置一个扬声器。
So if you go to China and you want a co packer or manufacturer to build you something and you're like, yeah, want it to connect to the Internet or I want it to have a speaker inside.
他们会说：太好了。
They're like, oh, great.
我们会制造14纳米芯片，或者甚至更落后的工艺芯片。
We will we will fab a 14 mill 14 nanometer chip or or something even even less frontier than that.
但他们能否调动大量落后制程的产能，然后说：嘿，你知道吗？
But can they just marshal a ton of lagging edge capacity and then just say, hey, you know what?
我们有廉价的能源。
We have cheap energy.
我们在烧煤。
We're burning coal.
我们有核能。
We have nuclear.
我们有三峡大坝。
We have the 3 Gorges Dam.
我们有水力发电。
We have hydropower.
我们有如此多的免费能源。
We have so much free energy.
我们干脆在落后制程上多花十倍的能源。
Let's just spend 10 times as much energy on the lagging edge.
我们能这样实现通用人工智能吗？
Can we achieve AGI that way?
看起来不行。
And it seems like no.
似乎先进制程节点至关重要。
It seems like leading edge nodes are incredibly important.
你不能仅仅靠堆砌大量14纳米的落后制程芯片来解决这个问题，至少在当前的架构下不行。
You can't just throw a ton of lagging edge 14 nanometer chips at the problem, at least with current architectures.
这种情况可能会改变，但迄今为止，没有任何前沿模型是在比5纳米更落后的硬件上训练的。
Now, this might change, but no frontier model to date has been trained on hardware older than five nanometer.
所有领先的芯片，比如Blackwell、谷歌的TPUv7、AWS的Tranium-3，都采用台积电的4纳米或3纳米工艺。
All the leading chips, that's Blackwell, Google's TPUv7, AWS's Tranium-three, they use TSMC four nanometer or three nanometer process.
我认为，当我们深入研究Grok、Cerebras以及一些新出现的技术时，所有人都在说：我们希望使用最先进的制程节点。
And I think when we dig into Grok and Cerebras and some of the newer stuff that's coming, everyone is saying we want to be on the the most frontier node.
我们不希望在这一点上倒退。
We don't want to go backwards on that front.
因此，中国最努力的成果——华为的昇腾910C——采用中芯国际的7纳米级DUV工艺。
So China's best effort, Huawei's Ascend nine ten c, is on SMIC's seven nanometer class DUV process.
它不是极紫外光刻。
They're not it's extreme ultraviolet lithography.
它是一种深紫外光刻工艺。
They're a deep, deep ultraviolet lithography process.
这在推理方面具有竞争力，但在大规模训练时需要多得多的芯片和能源。
That's competitive for inference but requires dramatically more chips and energy for training at scale.
因此，通过忽视能源消耗来蛮力推进AI进展，似乎终将遇到经济瓶颈。
So brute forcing AI progress through disregard for energy consumption, it feels like it still hits economic walls at some point.
此外，还有一个问题：权重究竟有多重要？它们是核武器吗？
And then there's also the question of how valuable like, are the weights what are the nuclear weapons?
还是说真正的价值在于实际部署？
Or is it the actual deployment?
我刚才在听一场关于人工智能地缘政治影响的对话，你知道的，到底什么才是关键？
I was listening to this conversation about, you know, what are the geopolitical implications of AI?
你拥有天才，拥有2000智商的上帝模型，但你可能只能问它一个问题，这才是有价值的吗？
Is it that you have, okay, yes, you have the genius, you have the 2,000 IQ God model, but you can only really ask it like one question.
这真的有价值吗？
Is that what's valuable?
还是说你拥有一个由上千名天才组成的国家，部署在数据中心里，能够为敌方的每个人调动上万名网络安全专家？
Or is it you have a thousand, you know, country of geniuses in a data center and you're able to marshal effectively 10,000 cybersecurity experts for each person on the enemy team?
而对方只有上千人试图黑入你的手机，还有上千人不断试图黑入泰勒的系统。
And there's just a thousand people that are trying to hack your phone and a thousand people that are trying to hack Tyler constantly.
这其实才是推理的意义所在。
Is that is the inference, actually.
真正重要的是算力，是能源的来源，而不仅仅是训练。
What matters is the compute what where the power comes from as opposed to just the training.
训练当然重要，但推理也可能同样重要。
Training is obviously important, but inference might also be important.
所以这里有一个大问题。
So there's a big question there.
最后但同样重要的是，DG问：可能发生的台湾事件是什么？其主要预警信号又是什么？
And last but certainly not least, DG asks, What is the likely Taiwan event and what would be a leading indicator for it?
约翰认为，台湾封锁将是最大的触发因素，但台海紧张局势已经在升级。
John writes that Taiwan blockade would be the biggest trigger, but Taiwan Strait tensions are already escalating.
中国于2024年10月进行了‘联合利剑2024B’演习，围绕台湾展开协调军事行动。
China conducted Joint SWORD twenty twenty four b exercises in October 2024 surrounding Taiwan with coordinated military operations.
2025年12月，‘正义使命2025’行动部署了超过100架飞机，其中90架穿越了海峡中线，90艘舰艇和27枚从福建发射的火箭。
In December 2025, Justice Mission twenty twenty five deployed over a 100 aircraft, 90 crossing the median line, 13 warships and 27 rockets fired from Fujian.
有10枚火箭落在了台湾的毗连区。
10 rockets landed in Taiwan's contiguous zone.
毗连区。
Contiguous zone.
毗连。
Contiguous.
抱歉。
Sorry.
就在离岸12到24海里处。
Just 12 to 24 nautical miles offshore.
所以他们只是发射火箭，然后说：‘我们没打中陆地。’
So they're just like firing rockets and being like, they didn't hit land.
他们只是在你可能从海滩上看得见的地方发射，离岸10英里、12英里。
They're just you can see them from the beach probably, 10 miles offshore, 12 miles offshore.
我觉得那已经超出地平线了。
I think that's like just over the horizon.
是的。
Yep.
中国最近在谈论台湾时，将‘和平’与‘统一’分开了，这体现在他们2026至2035年的计划中。
China also recently separated Peaceful from reunification when talking about Taiwan and their twenty twenty six to two thousand and thirty five year plan.
这让你对事态可能的发展方向有所了解。
Gives you an idea of where things might be going.
台积电正在规划在亚利桑那州建设一座晶圆厂综合体，泰勒曾去参观过。
TSMC is planning ahead working on a fab complex in Arizona, which Tyler visited.
该工厂大规模投产后应能处理中国先进制程产能的30%，但目前处于微妙的平衡状态。
Should be able to handle 30% of total advanced China production at scale, but it's on a knife's edge.
如果你想大致了解这有多远，你一定见过洛杉矶海岸外的卡特琳娜岛。
If you wanna know roughly how how far that is, you've seen Catalina Island off the coast of LA.
是的。
Yes.
卡特琳娜岛我认为有26英里远。
Catalina Island's I think 26 miles.
所以想象一下，非常清晰可见。
And so imagine So very visible.
对。
Yeah.
你可以从长滩看到卡特琳娜岛，想象一枚导弹从你所在位置与卡特琳娜岛之间的中点处发射并落地。
You can see Catalina Island from Long Beach and imagine a missile coming and landing like halfway between where you are and Catalina Island.
疯狂。
Crazy.
确实如此。
Very Yeah.
问题是，伊朗及其冲突如何影响中国的思维方式。
The the question is how does Iran and the conflict there update China's kind of thinking
嗯。
Mhmm.
关于这个？
On this?
确实有。
There's Yeah.
一些情报账号分享说，中国可能在伊朗有特工，就像美国从乌克兰学习导弹能力一样，中国也在学习。
Some of the intel accounts have been sharing, who knows if it's true, that China actually has operatives in Iran, like, kind of learning in the same way that The US has learned Yeah.
特别是关于导弹能力方面。
From Ukraine, specifically, in in regards to missile capability.
真是疯狂的时光。
So wild, wild time.
而且，是的，自从这篇报道发布以来，感觉每一天都过去得飞快，但台湾事件的发生可能性却比前一天更低了。
And, yeah, it just feels like a day hasn't passed so far since this was published that Taiwan a Taiwan event feels less likely than it did the day before.
是的。
Yeah.
我认为这就是为什么帕尔默一直在强调封锁作为一个明确的红线，因为这是一个相对抽象的概念。
I think that's why Palmer has been trying to message around the blockade specifically as being like a line in the sand because that's somewhat of an abstract concept.
因为如果发生封锁，情况可能会非常糟糕。
Because if a blockade happens, it could be very bad.
它可能成为转折点的开端，但不会演变成热战。
It could be the start of a turning point, but it would not be a hot war.
不会出现地面部队。
It would not be boots on the ground.
也不会达到许多人会因此警醒的宣战级别。
It would not meet the level of, like, a declaration of war that many people would would wake up to.
所以他一直在传达这个意思：不。
And so he's been trying to message that, like, no.
封锁是更大事件的开端。
A blockade is the start of something bigger.
我们需要保持警惕，必须坚决阻止封锁的发生。
We need to be aware, and and we need to hold the line on on not allowing a blockade to happen.
无论如何，我们会继续关注这件事。
Anyway, we will continue following that.
但首先，让我跟你们说说 Railway。
But first, let me tell you about Railway.
Railway 是一个一体化的智能云服务平台。
Railway is the all in one intelligent cloud provider.
使用你最喜欢的工具部署网页应用、服务器、数据库等，Railway 会自动处理扩展、监控和安全问题。
Use your favorite agents to deploy web apps, servers, databases, and more while Railway automatically takes care of scaling, monitoring, and security.
我们转到 Mouse 的推特上看看。
Let's head over to Mouse Twitter.
Mouse Twitter 很重要。
Mouse Twitter is important.
Mouse Twitter 签到。
Mouse Twitter Check-in.
正处在上升势头。
Is on a roll.
来自 Mouse Twitter 的 Scott Wilson 说，一块帕尔马干酪是绝佳的随身零食。
Scott Wilson from Mouse Twitter says a block of Parmesan is an elite on the go snack.
这块八盎司的奶酪含有 80 克蛋白质和 64 克脂肪。
This eight ounce block has 80 grams of protein at 64 grams of fat.
我觉得帕尔马干酪根本不是那种适合直接啃的奶酪。
I don't Parmesan is like the worst cheese to just chomp on.
不同意。
Disagree.
你觉得不同意。
You think Disagree.
好吧。
Alright.
这太难了。
It's so hard.
我知道。
I know.
我觉得它可能会
I feel like it's gonna
我知道。
I know.
但还是吃点零食吧。
But but eat snack.
我同意。
I agree.
我自从大概那时候起就没再吃帕尔马干酪当零食了
I haven't I haven't been snacking on Parmesan since I was probably
你知道吗，这简直是一种高端零食。
You know, well, it's such an elite snack.
我从来没见你像老鼠一样啃一块帕尔马干酪。
I've never once caught you snacking like a mouse on a block of Parmesan cheese.
我不像老鼠那样吃东西。
I don't eat like a mouse.
口头偏好与实际偏好。
Stated preference versus revealed preference.
我不吃老鼠。
I don't eat a mouse.
假老鼠。
Fake mouse.
但小时候，我会吃一大堆大胡萝卜。
But as a kid, I'd have like a bunch of big carrots
是的。
Yeah.
还有一些帕尔马干酪块。
And some blocks of Parmesan.
好的。
Okay.
我就进入老鼠模式了。
And I'd just be going mouse mode.
老鼠模式。
Mouse mode.
这不错。
That's good.
让我跟你聊聊古斯托模式。
Let me tell you about Gusto mode.
这是一个为中小企业打造的、集薪酬、福利和人力资源于一体的统一平台。
The unified platform for payroll benefits and HR built to evolve its small and medium sized businesses.
所以他说是在Slack上，兄弟。
So he says it's on Slack, bro.
它在网盘里。
It's in the drive.
我只是把它放进了Notion，兄弟。
I just put it in the notion, bro.
我真正在Teams上发给你了。
I literally sent it to you on teams.
你检查过Airtable了吗，兄弟？
Did you check the Airtable, bro?
它在文件夹里，兄弟。
It's in the box, bro.
这是单点登录。
It's single sign on.
它在Okta上，兄弟。
It's on Okta, bro.
没有。
No.
你需要使用YubiKey。
You need the YubiKey.
检查一下你的资料，兄弟。
Check your deal, bro.
它在Gusto上。
It's on gusto.
检查一下Outlook，兄弟。
Check Outlook, bro.
让我跟你详细说说Okta。
Let me actually tell you about Okta.
Okta能为每个AI代理分配一个可信身份，让你在享受AI强大功能的同时规避风险。
Okta helps you assign every AI agent a trusted identity so you get the power of AI without without the the risk.
风险。
Risk.
保护每个代理。
Secure every agent.
保护任何代理。
Secure any agent.
今天早上。
This morning.
我喜欢这个段子。
I like that riff.
这是从‘It’s on Tubi’抄来的复制粘贴内容。
It's this is a this is a copy pasta from the it's on Tubi.
它在Hulu上。
It's on Hulu.
它在Zooby上。
It's on Zooby.
就像所有的流媒体平台一样。
Like the all the streaming services.
总之，今天早上发生了什么？
Anyway, what happened this morning?
今天早上，特朗普总统接受了一次电话采访，他说：‘我解雇了Anthropic。’
This morning, President Trump gave a phone interview in which he said, I fired Anthropic.
Anthropic现在有麻烦了，因为我像赶狗一样解雇了他们，而你根本不会解雇狗。
Anthropic is in trouble because I fired them like dogs, which You don't fire dogs.
别解雇狗。
Don't fire dogs.
我早就该把我的狗解雇了。
I would have fired my dog years ago.
它是最没用的狗。
He's the least effective dog.
是的。
Yeah.
他不是常说‘死得像条狗’吗？
Isn't the phrase he he died like a dog?
对。
Yes.
是的。
Yes.
他只是在套用这个说法。
He's just adapting that.
但是
But
他把Anthropic送到农场去了。
He sent anthropic to the farm.
不。
No.
如果他把它们比作狗，那狗可是人类最好的朋友。
If he's comparing them to dogs, it's man's best friend.
这是看涨的。
This is bullish.
他在说，他说它们是我的狗。
He's saying he's saying they're my dogs.
关于 Bayt 播客
Bayt 提供中文+原文双语音频和字幕，帮助你打破语言障碍，轻松听懂全球优质播客。