Anthropic 的 Claude“做梦”：你需要了解的内容

今天节目里先来说说，Anthropic给Claude加上了‘做梦’的能力，这样它就能在夜间自主迭代升级了。

Today on the show Anthropic has taught Claude how to dream, so overnight it can improve itself.

还有安全研究公司Red Access刚发布的一份报告：目前有数千款基于Lovable、Repla和Base 44开发的应用正在泄露医疗与企业数据；另外谷歌已经悄悄叫停了Mariner项目，距离谷歌IO大会开幕只剩两周时间。

Red Access, a security research firm, just published a report that thousands of apps built on Lovable, Repla, and Base 44 are leaking medical and corporate data right now, and Google has quietly killed Project Mariner just two weeks before Google IO.

Anthropic最近在云算力合作方面势头猛得不得了，先是和谷歌云达成合作，又和SpaceX敲定了一项300兆瓦的算力供应协议，未来还有可能进一步和SpaceX合作，付钱让后者把轨道数据中心送上太空。

Anthropic is on an absolute tear with cloud compute deals, doing one with Google Cloud and also with SpaceX for 300 MW with a possible future deal where they'll be paying SpaceX to put orbital data centers into space.

谷歌IO大会马上就要开了，咱们先来讲讲和谷歌相关的这些新闻。

Let's get into some of the Google stories first with Google IO just around the corner.

他们最近在做很多调整，感觉都是悄悄上线功能，没怎么大张旗鼓宣传，但这些改动其实影响特别大。

They are making a bunch of changes, and it feels like they ship things and aren't really making huge announcements, but they have some big some big impact.

比如说谷歌的Mariner，本质上就是带智能代理功能的浏览器，和Claude的那个办公助理功能挺像的，对吧？

So Google Mariner, which is the basically agentic browser, it was kind of like a Claude coworker, right?

它原本的设计就是接管你的浏览器，自动点击页面、填写表单，还能帮你订机票酒店，各种杂活都能处理。

It was supposed to take control of your browser and click and fill out forms and book travel for you blah blah blah.

现在Claude办公助理做的那些事，和它原本的规划一模一样。

Kind of everything that Claude Cowork is doing.

谷歌已经把这个项目砍掉了。

Google has shut it down.

谷歌根本没为这事发什么大公告，不过Mariner项目的所有技术都会整合进Gemini智能体，并入整个智能助手生态系统里。

There wasn't, like, a big announcement about it, but all of the tech that was in Project Mariner is getting folded into the Gemini agent and kind of the whole agent assistant ecosystem.

我觉得真实原因是，Mariner这个项目从单位经济效益上来讲根本行不通。

What I think is actually happening here is that Mariner basically didn't make u unit economics sense.

这类浏览器智能体需要不断截图，还要对页面的DOM结构做视觉解析，成本非常高。

The browser agent that can take you know, it's taking screenshots and it's visually parsing the DOM.

这些工具的运行速度特别慢。

They're very slow.

成本还非常高。

They're very expensive.

而Codex和Claude Code这类命令行智能体，其实功能完全一样，但成本要低得多。

And command line agents like Codex and Claude Code have basically they they basically do the exact same thing, but it's a way lower cost.

所以我觉得谷歌也看到了这个情况，才做出了终止项目的决定。

And so I think Google saw that, and they kinda made this call.

有意思的地方在于，Claude Cowork还有Codex实现的不少功能，其实都和这个思路差不多。

Now what's interesting is that Claude Cowork and and presumably what some of the things that Codex does, they do a lot of the same things.

就拿我用Claude Cowork的经历来说，有时候我会跟它说，嘿。

So when I use Claude Cowork, you know, sometimes I'll say, hey.

你去打开那个Audacity音频编辑器，控制我的屏幕，帮我把这段音频处理一下，它真的会自己一步步点完所有操作。

Go over to, you know, go over to, like, Audacity, my audio editor, take control of my screen, and do this, you know, this chain of fixes on the audio, and it will go and literally click through.

要完成这些操作，它得不停截图、调整参数，能做到的事真的挺厉害的。

And in order to do that, it's gotta take screenshots, it's gotta make adjustments, and it can do some pretty incredible stuff.

我补充一点，最近有一份报告指出，这种“电脑操控”模式的算力消耗，大约是直接用Claude Code处理、或者让AI直接用文字回答问题的85倍。

What I will say is there was a recent report that came out, and it said that doing that, which is called computer use, is about 85 times as compute intensive as doing something through just directly through Claude code or, you know, asking it questions and it kinda giving you a text response.

所以这种模式虽然实用性强得多，但算力成本足足翻了85倍。

So this obviously is way more useful, but 85 times more compute intensive.

这么看谷歌把这个项目停掉，把相关技术整合到Gemini里，我一点都不意外。

So I'm not shocked that Google kind of, pulled this out and stuck it over into Gemini.

而且我觉得谷歌也在参考其他同类智能体的发展路线，意识到这个方向才是未来的趋势。

I also think that they're seeing how a lot of these other agents are doing it and and kind of realizing this might be the direction.

比如Claude Cowork其实做了一件很聪明的事，就连Claude Code也是，大多数时候我让它办事，它根本不需要用截图那种纯电脑操作的方式。

Now Claude Cowork, for example, is actually doing something very clever, even Claude Code, where most of the time when I ask it for something, it doesn't need to do the the strictly computer use screenshot tasks.

不过我猜它在后台还是会偶尔截几张图。

Although they do, I guess, kinda take some screenshots in the background as they go.

它很多操作都是想找到最直接用代码解决的办法。

A lot of what it does, it tries to figure out the most code way to do it.

所以如果我让它去CapCut里剪视频，它不会去打开CapCut应用，反而我试过让它直接操作CapCut的文件，效果更好。

So if I tell it to go edit a video in CapCut, instead of opening up the CapCut app, I've had better success with it going to the CapCut file.

它打开文件后，直接修改文件里的代码来实现我想要的编辑效果。

It opens up the file and just edits the code on the file to make the edits that I want.

这方法听起来有点像机器人在干活，但真的特别管用，效率高得多。

It's kind of a crazy robot way to do things, but it actually works really well, it's way more efficient.

我觉得谷歌可能会朝这个方向走一点，不过挺有意思的是，这个功能居然被停掉了。

So I think Google might move in that direction a little bit, but it's interesting that that got, shut down.

另外一件有意思的事是，谷歌刚推出了一款Fitbit Air。

The other thing that's interesting is that Google just launched a 100 Fitbit Air.

这是四年来我们首款全新的Fitbit硬件。

This is, the first Fitbit hardware that we've had in four years.

它没有屏幕，主要和你搭载Gemini的健康应用配对，甚至还内置了一个每月10美元的AI教练。

It's screenless, and it basically pairs with your Gemini powered health app, and they even have a $10 a month AI coach that is kind of built into this.

接下来是Red Access。

Next up, we have Red Access.

这是一家安全研究公司，他们发现成千上万款用Vibe编写的App正在泄露用户数据。

This is a security research firm, and they have found thousands of vibe coded apps that are leaking customer data.

具体来说，这些App泄露的包括企业数据、健康信息等各种不该外泄的内容。

And specifically, they're doing things like, corporate data and health data, all sorts of stuff you wouldn't want to have leaked.

Red Access是一家以色列的安全研究机构。

Red Access is an Israeli security forum.

他们扫描了38万个基于Level、Replit、Base44和Netlify构建的公开应用，其中大约5000个完全没有身份验证。

They basically scanned 380,000 publicly accessible apps built on Level, Replit, and Base 44, and then Netlify, and I think about 5,000 of those had no authentication.

也就是说，大约40%，也就是两千个应用，把敏感数据暴露给了任何能撞上这些链接的人。

So about 40% of those or about 2,000 apps were exposing sensitive data to anyone who happened to find the URL.

那都是些什么数据？

So what kind of data?

就像我之前说的，有个儿童长期护理机构里，医生和患者之间的对话也被泄露了。

Like I mentioned, there was conversations between doctors and patients at a long term care facility for children.

讽刺的是，他们还发现了一家安全公司的应急响应数据。

Ironically, they also found a security company's incident response data.

他们找到一个学校的应用，里面存着课程录音和学生档案。

They found a school app with lesson recordings and student records.

还有一个旅行规划指南，详细记录了客户的个人信息。

They found a vacation planning guide that detailed some of the customer's details.

他们还发现了客户聊天机器人的日志，可以直接看到所有客户的对话内容。

They had a customer chatbot logs that they were able to actually just see all of the conversations the customers were having.

内部的公司战略演示文稿也被暴露了。

They had internal corporate strategy decks.

只要知道网址，任何人都能点进去把这些内容全看完。

And basically anyone with their URL could go and read all of that.

现在这些平台，特别是Replit，正在对这份报告做出回应，他说：你看。

The platforms right now, Replit specifically, is kinda pushing back on this report, and he's saying, look.

你知道，Red Access只给了他们24小时就直接找媒体了，这话说得也没错。

You know, Red Access only gave them twenty four hours before going to the press, and, like, he's I guess, it's kind of true.

但他觉得，你们为啥不给我们多点时间，先把问题修好再公开呢？

So but he but he's like, why didn't you guys give us more time to, like, fix it before you went live with it?

这正是让我觉得不爽的地方。

So that's kind of what I was upset about.

我有个Base44的账号，我记得确实收到过他们的一封邮件，说在某个日期之前创建的所有项目都自动设为私有了。

I have a Base forty four account, and I think I actually saw an email from them saying something like all of the all projects on Base 44 made before a certain date were automatically made private.

如果你想重新设为公开，当然可以，但得先完成一些安全设置。

If you wanna go make it public again, you can go do it, but you gotta do some security stuff.

我觉得他们其实已经在处理了，但这个问题确实是大家都遇到过的。

So I think they are kind of working on it, but this is definitely an issue that they all encountered.

我觉得AI帮你搭一个能跑的App，和AI帮你搭一个不会泄露数据的安全App，完全是两码事。

And I think there's a huge difference between AI building you a working app and AI building you a secure app that isn't gonna leak your data.

所以，这正是现在许多公司接下来要应对的关键阶段。

So that's kind of the next big phase that a lot of these companies are having to work on.

我知道Lovable最近就发生过类似的安全事件，所有2025年11月之前创建的项目都存在漏洞，之后的就没问题。

And I think some of I know specifically with Lovable there was another kind of security incident just like this not too long ago, and it was like every lovable project that was made before, you know, November 2025 had issues, but everything after didn't.

有时候，问题出在那些老旧的项目上。

And so sometimes it's kind of the older legacy projects.

如果用户没及时更新安全设置，目前最好的处理方式就是把项目重新设为私有，因为现在新建项目的安全性已经提升了不少。

It seems like the best course of action if the users haven't updated the security, because right now the security is getting a little bit better when you're building stuff, is they just make them private again.

如果你想重新公开，就得自己去修复这些安全问题。

And if you wanna make it public, you gotta go and fix those security things.

vibe编程的世界，真是越来越魔幻了。

So really wild times with the world of vibe coding.

接下来，Anthropic正在教Claude做梦。

Next up, Anthropic is teaching Claude to dream.

他们管这叫‘做梦’，其实是个特别酷的功能——Claude在空闲时，比如晚上或会话间隙，会自动启动这个‘做梦’机制，回顾最多一百次之前的会话内容。

Now it's kind of what they're they're naming it dreaming, but it's a really cool feat a really cool functionality where Claude code, is gonna be able to basically in its idle downtime, overnight or between sessions, dreaming, this new feature called dreaming is gonna kick on, and it is a process that the agent basically reviews up to a 100 of its previous sessions.

它会寻找那些在实时操作中遗漏的模式，然后识别出反复出现的错误，并据此重新组织它的长期记忆。

It's gonna look for patterns that it missed in real time, and then it's gonna identify reoccurring mistakes, and it will restructure its long term memory store based off of that.

对吧？

Right?

比如说，我经常让它去修改某些文件、网站或者代码库，而且都有特定的方式。

So, like, if you I actually have a whole bunch of things where I'll, like, tell it to go and edit certain files or certain websites or certain code bases a certain way.

我有时候会一边看它的日志，然后发现：哦，这个我试过了。

I I kinda read the logs along as it's going sometimes, and it's like, oh, I've tried this.

哦，这个不行。

Oh, that didn't work.

那我换一个试试。

Let me try this.

啊，这个成功了。

Oh, this worked.

有时候呢，同样的问题一再出现，但我也不太在意。

And sometimes it's, yeah, the same problem over and over again, but I'm like, whatever.

它自己会搞定的。

It'll figure it out.

我不想坐那儿一遍遍修改我的提示词。

I don't wanna sit there and imperfection my prompt.

就算这样不太好，我也直接把想法扔给它，它迟早能弄明白，然后我就继续做我的事了。

Maybe that's bad then, but I'll just shoot off what I want it to get down, and it will get it down eventually, and I'll move on with my life.

这真的太棒了。

This is really awesome.

首先，它会更高效。

It's going to be, number one, more efficient.

其次，节省的token更少，第三，速度更快。

Number two, it's gonna use less tokens, And number three, faster.

速度这东西，我觉得人人都能认同，毕竟谁不希望有巨大提升呢？

And speed is, I think, one of the things that everyone can get behind as, you know, making some massive improvements.

这跟训练没关系。

And it's not training.

这其实就是记忆的整合。

It's just memory consolidation.

所以，Anthropic 的意思是，这跟人睡觉时大脑做的事情很像。

So what, Anthropic is kind of saying is that this is kind of the AI version of what your brain actually does during sleep.

他们在内部测试中发现，那个叫‘相关结果’的功能——用独立的评估代理来打分任务完成度——能把任务成功率提升多达10个百分点。

In their own internal testing, the related outcomes feature, which uses a separate grader agent to score task completion, improved task success rate by up to 10 points over their standard prompting loop.

现在，他们新推出的‘结果’功能、‘梦境’机制，还有多代理协同调度，都已经在云平台上开放公测了。

So outcomes that they this new feature they have called outcomes plus this new dreaming plus multi agent orchestration are all now in public beta on the cloud platform.

我觉得这些代理功能被严重低估了，他们推出来的这些进展真让人兴奋，其他公司，比如谷歌和OpenAI，恐怕得加把劲才能跟上Anthropic的步伐。

I think this is some majorly underrated agent features that are really excited that they've been shipping, and just every other company, Google and OpenAI, really have all their work cut out to try to keep up with what Anthropic is is putting out there.

除此之外，Anthropic还爆出另一个大消息：他们刚刚宣布，未来五年将向谷歌云投入2000亿美元。

In addition to this, another story for Anthropic is that the information just broke this, and basically Anthropic just committed $200,000,000,000 to Google Cloud over the next five years.

谷歌将专门划出5吉瓦的服务器算力，全部给Anthropic用。

Google is dedicating five gigawatts of server capacity, and this is exclusively for Anthropic.

Anthropic和博通合作的新TPU算力，预计2027年就能上线。

And the new TPU capacity from the Broadcom partnership that Anthropic just has is gonna come online in 2027.

我觉得这两项加起来，占了谷歌公开的云收入待办清单的四成以上，这规模太惊人了。

And I think between those two, that accounts for more than 40% of Google's disclosed cloud revenue backlog, which is massive.

很明显，市场已经注意到了这一点。

Basically, think the market noticed this.

消息一出， Alphabet 在盘后交易中的股价一度超过了英伟达的市值。

Alphabet's after hours price briefly pushed past NVIDIA's market cap on the news.

这可是好几年来头一回发生。

So this is the first time in years that that has actually happened.

Alphabet 当时的市值大约是 4.6 万亿美元。

Alphabet was at about $4,600,000,000,000.

英伟达是 4.8 万亿，但那会儿 Alphabet 短暂地超过了它。

NVIDIA is at 4.8, And they, you know, they were able to kind of bump above Anthropic for a little bit.

说到 Anthropic 和谷歌的关系，我觉得得公平点说，Alphabet 直接向 Anthropic 投入了高达 400 亿美元。

When it comes to Anthropic's relationship with Google, I will say I'd like to be fair, Alphabet is putting up to $40,000,000,000 into Anthropic directly.

换句话说，谷歌其实是在付钱让 Anthropic 在自家平台上开发，而 Anthropic 又反过来向谷歌购买算力。

So Google's kind of paying Anthropic to build on Google, and then and then Anthropic is gonna go and pay Google for compute.

这机制还挺有意思，400亿美元直接投给Anthropic，然后Anthropic又承诺回投2000亿美元。

And it's kind of interesting the way this works because 40,000,000,000 goes directly into Anthropic, then Anthropic makes a $200,000,000,000 commitment back.

看起来像是正向回报。

So it looks like positive ROI.

但这其实也是一个承诺。

That's also a commitment.

不是真金白银，而是分五年到位，所以你得把2000亿除以五来算。

It's not actual cash, and it's over five years, so you gotta, you know, divide 200 by five.

这里面有些微妙的地方，表面上看像是正向回报，前提是Anthropic能持续增长。

So there is some interesting things at play where it basically feels like positive ROI and assuming Anthropic continues to grow.

这一切都说得通，除非哪天泡沫破裂，或者出什么意外，这些东西就全悬了。

Like, it all makes sense unless some sort of bubble pops or something bad happens, then a lot of this stuff gets put into limbo.

当然，这不过是华尔街在AI领域最爱的那种循环资金流，挺耐人寻味的。

And, of course, this is just more of the kind of circular cash flow that Wall Street loves in the in the whole AI market, which is interesting.

好了。

Alright.

如果你已经在为ChattypuTea、Claude、Gemini、Grok，还有像Eleven Labs这样的音频工具或者某些图像模型付费，我特别希望你能看看我创办的创业公司，叫aibox.ai。

If you are already paying for ChattypuTea and Claude and Gemini and Grok and maybe, like, 11 Labs for audio or any of the image models, I would love for you to check out my startup, which is called aibox.ai.

你没必要付十几份不同的订阅费。

You don't have to pay a a dozen different subscriptions.

每月只要8.99美元，就能一次性访问80个不同的AI模型。

It is $8.99 a month, and you get access to 80 different AI models all in one platform.

你都能用上OpenAI、Claude、Gemini、Grok，还有Eleven Labs的音频功能。

You get everything from OpenAI, from Claude, from Gemini, from Grok, Eleven Labs for audio.

还能用上Google的VO three做视频生成，以及其他一大堆超强的模型。

You get Google VO three for video generation, tons of other incredible models.

每个月只要9美元，所有功能都整合在一个平台上。

It's only $9 a month, and it's consolidates everything into one place.

不用再开一堆账户，登录一堆不同的平台了。

You don't have to have tons of subscriptions and tons of different platforms to log into.

希望它能帮你省下大量时间和金钱。

I hope it saves you a ton of time and a ton of money.

我就是靠这个把所有功能都整合到一个平台里的。

It is how I am making it so everything is all consolidated into one place.

我特别喜欢Anthropic。

I love Anthropic.

我平时写东西、思考问题都用它，但它不支持生成图片。

It's what I use for basically all my thinking and writing stuff, but it doesn't have image generators.

它也不支持音频功能。

It doesn't have audio.

更别提视频了。

It doesn't have video.

所以视频我用的是Google。

So I use video with Google.

图片生成我用的是OpenAI。

I use image with OpenAI.

音频部分我用的是Eleven Labs。

I use audio with Lovin Labs.

为了把所有这些整合到一起，我全靠OpenAI来搞定。

And in order to get all of those into one place, I use OpenAI to to do all of it.

你可以在简介里找到 aibox.ai 的链接。

So you can see the link in the description to aibox.ai.

明白了。

Okay.

Anthropic 刚刚宣布和SpaceX达成了一项重大合作。

Anthropic has announced a massive deal with SpaceX.

他们将包下SpaceX的Colossus 1数据中心的所有算力。

They're gonna be locking up the full compute capacity of SpaceX's Colossus 1 data center.

这个数据中心在孟菲斯。

It's in Memphis.

当初刚建的时候，这地方还挺出名，也挺有争议的——他们直接找了一块水泥地，堆满服务器，拉来一堆发电机和移动冷却设备，转眼就搭起来了。

It kind of was, like, famous and controversial for a while because when they first built it, they just grabbed, like, a random slab of concrete and filled it with servers and got a bunch of power generators and portable cooling units, they just stood this thing up super fast.

有人抱怨说，那些发电机排出的废气特别多。

And people complained that the generators had, like, a lot of fumes and stuff.

我觉得他们现在应该已经解决掉那些问题了，至少我是这么希望的。

So I think they got a lot of that crap figured out by now, or at least I hope.

有了这些资源，Anthropic 能拿到大约 300 兆瓦的电力，还有大约 22 万个 NVIDIA 的 GPU。

And with all of that though, Anthropic's get is gonna get about 300 megawatts, and they're gonna get about 220,000 GPUs NVIDIA GPUs.

他们用的有 H100、H200，还有一大批密集部署的 GB200。

So there's h one hundreds, h two hundreds, and then they have a really dense deployment of g b two hundreds.

Anthropic 自己的博客提到，因为他们有能力这么做——我记得 Dario 在旧金山的活动上也提过类似的话，反正就是这么回事。

Anthropic's own blog post about all of this said, basically, because they were able to and I think Dario was, like, on stage at San Francisco and kind of announced it as well and kind of said the same thing, whatever.

这事是在 Twitter 上公布的。

It was on Twitter.

但他们主要的意思是，由于这次合作，立刻将 Pro 和 Max 用户的 Claude Code 五小时使用限额翻倍。

But, basically, what they're saying is they are doubling Claude code's five hour rate limit for pro and max users immediately because of this.

我个人特别高兴，因为我老是被限流，都快想开多个 Claude Max 账号了。

And personally, I'm thrilled because I keep getting rate limited, and I'm tempted to make multiple Claude Max subscription accounts.

一个账号 200 美元，这下我再也不用担心了。

It's a $200 one, so that doesn't happen to me anymore.

但不管怎样，他们因为这个大动作，把所有限制都翻倍了。

But in any case, they're doubling all of that because of this big deal.

他们还彻底取消了高峰时段的限速，那之前真是让人受不了。

They're also removing peak hour throttling entirely, which was something that was so brutal.

不过话说回来，他们也没别的办法。

But, I mean, they had no other option.

对吧？

Right?

他们手头的算力就那么多。

They they just had a certain amount of compute.

实在无能为力。

There's nothing they can do.

现在有了这个，总算能取消这个限制了。

So with this, they'll be able to remove that.

所以这对他们来说，算是一次巨大的升级了。

And so this is kind of a massive upgrade for them.

我还看到埃隆发了一条关于这件事的推文。

I also read a tweet that Elon put out about this.

他 basically 说，他花了很多时间和 Anthropic 的管理层以及一些员工交流，感觉他们身上一点恶意都没有。

He basically said that he spent a bunch of time with the leadership at Anthropic and some of the people there and talked to them, and he feels like he doesn't have any evil vibes from them.

所以他觉得 Claude 对人类整体来说会是好事。

So, he thinks that Claude will be generally good for humanity.

不管怎样，这些事情的运作方式真的挺有意思的。

Anyways, it's really funny the the way some of this stuff works.

但与此同时，彭博社还提到一个特别有意思的事：Anthropic 似乎对在太空中建设数吉瓦的算力设施很感兴趣。

But with all of that, another really interesting aspect that Bloomberg was writing about is that SpaceX and Anthropic have apparently had some interest from Anthropic's side in developing multiple gigawatts of compute capacity in space.

我不确定这是否就是他们和 SpaceX 达成合作的原因，还是说他们只是想搞更多算力，而这是目前最可行的方案。

And I'm not sure if that's why they made this deal with, you know, SpaceX or if that was just them trying to get more compute and this was the best way for them to do it.

我也听到不少人说，你看啊。

I'll also say from a lot of people were saying, oh, look.

这简直就像是 X AI 走向衰落的标志。

This is kind of like exit the the downfall of x AI.

如果他们真的把算力卖给了Anthropic，那他们现在就不会比Anthropic有更多的算力了。

They don't have more compute capacity than Anthropic anymore if they're literally selling it to Anthropic.

埃隆最近接受了不少采访，说他们要建一个超级巨大的吉加级数据中心，听说他们已经有一个叫Colossus Two的了。

Elon has had a bunch of, you know, recent press and stuff talking about how they are going to be building this massive giga Colossus, you know, data center thing, and I think they already have one called Colossus two.

所以他在帖子里提到：你看。

So in his post about it, he said, look.

我们已经有Colossus Two了，规模非常大。

We already have Colossus two, which is massive.

我们现在就是用它来做训练的。

That's what we're training on now.

所以Colossus One我们其实用不上了，干脆送人吧——这真有意思，真正投资算力的公司，居然把资源租给竞争对手。

So we don't really need the Colossus one, and we're happy to just go give it to somebody else, which is it's just fascinating that the companies that are investing in the actual compute are renting it to their competitors.

微软和谷歌那边也出现过类似情况，难怪OpenAI这么拼命砸钱建数据中心和算力设施。

We're seeing this with Microsoft and Google, and evidently this is why OpenAI wants to spend so much money building out data centers and compute.

而且我总觉得，这一切都是埃隆·马斯克在报复萨姆·阿尔特曼，他故意把最强劲的对手送上更多算力，好让对方在法庭上跟他较劲。

Also, can't help but feel like all of this is Elon Musk trying to get back at Sam Altman, so he's just giving his biggest you know, Sam's biggest competitor more compute to compete with him as he's going through court, you know, against Sam Altman.

摩根大通的一位策略师发了一份报告，被CNBC报道了，大意是警告说，Anthropic和SpaceX即将上市，可能会抽走原本支撑标普500指数的那部分AI资金——这些钱之前都流进了谷歌、英伟达这些顶级科技公司，推动了整个AI板块的上涨。

One Bank of America strategist put out a note that CNBC picked up, and, basically, it was warning that the upcoming Anthropic and SpaceX IPOs are gonna pull money out of the same, basically, AI trade that's been holding up the entire S and P 500 from all of the top, you know, Google, Nvidia, all of these top tech firms, and all of the AI money that's being made that's lifting up the S and P 500.

Anthropic在五月初的私募估值达到了1.2万亿美元，这是CoinDesk的说法。

Anthropic's private market valuation hit $1,200,000,000,000 in early May, if that's, you know, according to CoinDesk.

据说他们的IPO计划定在六月进行。

Their IPO is now reportedly gonna be coming in June.

这可不只是一个算力交易的问题。

So this isn't just a compute deal.

这其实还是一场发生在上市前的布局。

It's also kind of a pre IPO thing that's happening right now.

所以如果你是个开发者，从这些事里能得出的一个关键点就是：Claude Code的使用限额现在翻倍了。

So if you're a builder, one of the takeaways I would get from all of this is that Claude Code's rate limit is now double.

这太棒了。

That's amazing.

而且这周Opus推理的速度也会明显快很多。

And also, Opus Inference is gonna feel a lot faster this week.

如果你是投资者，记得关注Anthropic的IPO申报文件。

If you're an investor, watch the Anthropic IPO filings.

摩根大通对市场流动性影响的提醒，是首家大型银行把AI上市潮视为标普500整体市场的流动性风险，这点非常值得关注。

The Bank of America kind of, you know, drain on the market warning is the first time that a major bank has framed the AI IPO wave as a liquidity risk to the broader S and P, so that's very interesting.

好了，今天节目就到这里。

But that's it for the show today.

非常感谢你的收听。

Thank you so much for tuning in.

如果你喜欢这期节目，别忘了在你听播客的平台给个评分或评论。

If you enjoyed this episode, make sure to go leave a rating or review wherever you get your podcasts.

这能帮节目在算法里获得更多曝光，让像你这样的精彩听众更容易发现我们。

It helps the show out in the algorithm a ton to be found by more amazing people like yourself.

在Spotify上，你可以在‘关于’页面操作。

So over on Spotify, it's the about tab.

在Apple上，你可以直接留言。

You can leave a comment over on Apple.

我都看过了。

I read them all.

我都非常感谢。

I appreciate them all.

一如既往，如果你想一次性访问80多个AI模型，记得去aibox.ai看看，每月只需8.99美元。

And as always, make sure to go check out aibox.ai if you wanna get access to over 80 different AI models in one place for $8.99 a month.

简介里有链接，我们下一期再见。

There's a link in the description, and I'll catch you guys all in the next episode.