Meta的Muse AI助手能帮你省钱吗？

本期节目由Odoo赞助播出。

Support for the show comes from Odoo.

经营企业不应该像在软件小组项目里挣扎求生一样。

Running a business shouldn't feel like surviving a software group project.

一个应用管财务，另一个管库存，还有一个管销售，结果它们之间互不相通。

One app for accounting, another for inventory, another for sales, and somehow none of them talk to each other.

这就是Odoo的用武之地，一款集所有功能于一体的企业管理软件，把销售、财务、库存、营销等所有业务环节整合到一个强大的平台上。

That's where Odoo comes in, an all in one business management software that brings every part of your business together from sales and accounting to inventory and marketing all in one powerful platform.

没有混乱的集成。

No messy integrations.

不用在多个标签页之间来回切换。

No bouncing between tabs.

最重要的是，告别电子表格。

And best of all, no spreadsheets.

别再管理软件了，用一套统一的系统来管理你的企业吧。

Stop managing software and start managing your business with one unified system.

今天就可以去 odo.com/profg 免费试用。

Try for free today at odo.com/profg.

网址是 odoo.com/profg。

That's odoo.com/profg.

本期节目由 Northwest Registered Agent 赞助播出。

Support for the show comes from Northwest Registered Agent.

你的企业身份就是让企业看起来合法又专业的一切。

Your business identity is everything that makes your business legitimate and professional.

有了 Northwest Registered Agent，你不仅仅是成立一家公司，而是为隐私、信誉和成长打下完整的基础。

With Northwest Registered Agent, you don't just form a business, you start a complete foundation built for privacy, credibility, and growth.

这包括注册代理服务、公司地址、运营协议、域名、网站、专业邮箱、电话号码，还有内置的隐私保护。

That includes registered agent service, a business address, operating agreement, domain, website, professional email, phone number, and built in privacy.

换句话说，你的家庭住址、个人邮箱和电话号码都能保持私密。

In other words, your home address, personal email, and phone number stay private.

不用花几百甚至几千美元，在 Northwest 就能免费获得这些服务。

Don't pay hundreds or thousands of dollars for what you can get from Northwest for free.

访问 northwestregisteredagent.com/profgfree，开始使用免费资源来打造一番事业。

Visit northwestregisteredagent.com/profgfree and start using free resources to build something amazing.

在 northwestregisteredagent.com/profgfree 获取更多服务，选择 Northwest Registered Agent。

Get more with Northwest Registered Agent at northwestregisteredagent.com/profgfree.

你是宠物主人吗？

Are you a pet owner?

在美国，每六秒钟就有一位宠物主人收到超过一千美元的兽医账单。

Every six seconds, a pet owner in The US gets hit with a vet bill over a thousand dollars.

这总是让人措手不及。

Always an unwelcome surprise.

这时候 Fetch 就派上用场了。

That's where Fetch comes in.

Fetch 是为狗和猫提供的最全面的宠物保险。

Fetch is the most complete pet insurance for dogs and cats.

根据 consumeradvocate.org 的数据，你最多可以报销兽医账单的90%，在美国和加拿大的任何兽医诊所都适用，而且理赔最快两天就能到账。

According to consumeradvocate.org, you get paid back up to 90% of vet bills at any vet in The US and Canada with claims paid back in as little as two days.

现在就上fetchpet.com/save获取你的免费报价。

Go to fetchpet.com/save right now for your free quote.

网址是fetchpet.com/save。

That's fetchpet.com/save.

嘿，我是斯科特，想给你介绍Prof G Media的新节目，主持人是杰克·雷恩斯。

Hey, it's Scott, and I wanna introduce you to the new show from Prof G Media hosted by Jack Raines.

每周，杰克都会选取科技、金融或市场领域的一个大新闻，探讨它对你的生活和钱包意味着什么。

Each week, Jack takes a big story in tech, money, or markets, and asks what it means for your life and your wallet.

今天，他关注的是Meta新推出的AI助手Muse，以及当AI开始替我们做那些一直拖延的事情时会发生什么。

Today, he's looking at Meta's new AI assistant, Muse, and what happens when AI starts doing the things we've all been putting off.

下面有请杰克。

Here's Jack.

Meta刚刚发布了一款名为Muse的AI助手。

Meta just released an AI assistant called Muse.

可以免费试用，它的卖点是不只能回答你的问题。

It's free to try, and the pitch is that it can do more than just answer your questions.

只要你有权限访问正确的账户，它就能真正帮你做事。

You have access to the right accounts, and it can actually do things for you.

比如帮你找更便宜的价格、处理订阅，或者帮你搞定那些占用你一周时间的行政琐事。

Find a better price, deal with a subscription, or help sort through administrative tasks that eat up your week.

AI代理其实还挺新的。

AI agents are still pretty new.

它们大概从去年12月才开始变得实用，因为那时候AI模型在编程能力上有了明显的提升。

They really only became useful around December of last year as AI models had a noticeable step up in coding capabilities.

但直到最近，用AI代理通常意味着要买一台单独的电脑，花几个小时把各种应用和工作流程连起来，或者搭一个带数据库后端的云服务器。

But until recently, using an AI Agent often meant buying a separate computer and spending hours wiring together different apps and workflows, or hooking up some sort of cloud server with a database backend.

除了硅谷一小撮人或者那些沉迷于X的人，没人愿意这么干。

And nobody wanted to do that except a very small group of people in Silicon Valley or Addicted to X.

Muse实际上把所有这些复杂的设置步骤都简化掉了，给用户提供了一个开箱即用的“帮我在电脑上搞定一切”的工具。

Muse has actually abstracted away all of the complicated setup steps, giving users a do everything on my computer for me tool wrapped in a bow.

马克·扎克伯格的新赌注是，数百万普通人都准备好跳过这些麻烦，直接让一个助手帮他们把事情办妥。

Mark Zuckerberg's new big bet is that millions of ordinary people are ready to skip all of that and simply ask an assistant to get something done for them.

如果这个赌注真的能成，那会改变我们花时间和花钱的方式。

If that bet proves to be it will change how we spend our time and our money.

我是杰克·雷恩斯，早期风险投资基金Slow Ventures的投资人，Substack通讯《Young Money》的作者，也是新书《二十多岁时的财富与目标指南》的作者。

I'm Jack Raines, an investor at Early Stage Venture Capital Fund, Slow Ventures, writer of the substack newsletter Young Money, and author of the new book, A Field God to Wealth and Purpose in Your Twenties.

在这个节目里，我们会聊科技或金钱方面的大新闻，然后看看它对你的生活和钱包意味着什么。

In this show, we take a big story in tech or money and ask what it means for your life and your wallet.

那我们就开始吧。

So let's get into it.

Muse到底是什么？

What exactly is Muse?

根据Meta几周前的发布公告，Muse是一个个人AI代理。

Per Meta's launch announcement a few weeks ago, Muse is a personal AI agent.

它不只是回答问题，而是真的能帮你干活。

It doesn't just answer questions, it actually does the work.

它能帮人们跟上进度，把任务和项目从你手上接过去，还能把长期目标变成具体的行动计划。

It helps people stay on top of things, takes tasks and projects off their plate, and turns long term goals into action plans.

所以Meta在9月8号正式发布了Muse。

So Meta officially announced the Muse launch on September 8.

在过去的三个星期左右，Muse的下载量估计达到了340万次，App Store上的评分也超过了76000条。

And in the last three weeks or so, Muse has had an estimated 3,400,000 downloads, and over 76,000 ratings on the App Store.

作为对比，ChatGPT，目前最大的消费级AI应用，大概有1100万条评分。

For context, ChatGPT, which is still the biggest consumer AI app, has about 11,000,000 ratings.

而Claude，Anthropic的AI应用，评分大约在27万条左右。

And Claude, Anthropix AI app, is at around 270,000 ratings.

到目前为止，大多数面向消费者的AI应用其实都是聊天机器人。

So most of the consumer applications for AI so far have been a chatbot.

你打开ChatGPT，问它天气怎么样，或者让它帮你分析一个表格，或者干脆把它当谷歌用。

You open ChatGPT, you ask it for the weather, or to break down a spreadsheet, or you treat it like Google.

对超过90%的人来说，这就是2022年以来AI的样子。

And for more than 90% of people, that's kind of been what AI has been since 2022.

Meta推出了Muse，并把它宣传成一个随时在线的AI助手，能帮你干活。

Meta came out with Muse and is advertising it as an always on AI assistant that can do work.

当然，Muse 也可以当聊天机器人用，但你一登录 Muse 打开账户，它马上就让你连接日历、邮箱、iMessage 和备忘录应用。

Of course, Muse can be used as a chatbot, but as soon as you sign into Muse and open your account, they instantly want you to connect your calendar, your email, your iMessage, your notes app.

很明显，第一，他们想尽可能多地收集用户数据；第二，让 AI 能非常非常方便地读取和使用你连接的所有工具。

It's very obvious that one, they're trying to aggregate as much data as possible on their users, and two, make it very, very easy for the AI to both read from and use all the tools that you connect to it.

当然，Meta 通过 Facebook、Instagram、WhatsApp、Threads 以及浏览器 Cookie，已经掌握了很多关于你我的数据。

Of course, Meta already has a bunch of data on you and me through Facebook, Instagram, WhatsApp, threads, and just browser cookies.

真正有意思的是，有多少公司在 Muse 发布时就选择跟它合作。

What's been really interesting is how many companies have also partnered with Muse off the launch.

包括 PayPal、Shopify、Expedia、Stripe、Best Buy、Walmart、Instacart、American Eagle、Dick's Sporting Goods、Fanatics、Gap、Wayfair，还有很多其他公司。

We have PayPal, Shopify, Expedia, Stripe, Best Buy, Walmart, Instacart, American Eagle, Dick's Sporting Goods, Fanatics, Gap, Wayfair, and plenty of others.

值得注意的是，有一家公司没有跟 Muse 合作，那就是亚马逊。

Notably, one company who is not working with Muse is Amazon.

这个我们后面再讲。

We'll touch on that later.

它们还会连接你的通讯录、日历、邮箱、浏览器和 Google Drive。

They also connect to your contacts, your calendar, your email, your browser, your Google Drive.

像Granola这样的笔记应用。

Note taking apps like Granola.

你可以连接你的Spotify，甚至Peloton和OpenTable，最重要的是，连接Platt来管理你所有的金融账户。

You can connect your Spotify, even Peloton and OpenTable, and most importantly, Platt for all your financial accounts.

基本上，扎克伯格和他的团队已经让连接尽可能多的生活内容到Muse变得极其简单。

Basically, Zuckerberg and his team have made it as easy as possible to connect as much of your life as possible to Muse.

所以Muse可以看到你所有的数据，并且可以与你的所有账户互动。

So Muse can see all of your data and it can interact with all of your accounts.

但对我来说，Muse最有趣的地方在于，他们免费提供所有这些功能。

But the most interesting thing about Muse to me is that they're giving all this away for free.

ChatGPT确实有免费计划，但用的是较弱的模型，而且很快会达到使用限制。

ChatGPT has a free plan, sure, but it uses weaker models and you hit usage limits quickly.

马克·扎克伯格每周免费提供1亿个令牌，这算得上大量的计算资源。

Mark Zuckerberg is giving away a 100,000,000 tokens per week, which is a ton of compute.

对于重度用户，你可以每月支付20美元，获得每周5亿个令牌。

For power users, you can pay $20 a month for 500,000,000 tokens a week.

或者对于最高级别的用户，你可以每月支付100美元，获得每周30亿个令牌。

Or for max users, you can pay a $100 a month for 3,000,000,000 tokens per week.

但是，你懂的，从字里行间来看，他希望尽可能多的人尽可能多地使用Muse，而且他愿意为此补贴。

But, you know, reading between the lines, he wants as many people as possible using Muse as much as possible, and he's willing to subsidize that.

Anthropic和OpenAI都没有像Meta那样以每周令牌数的方式公布他们的免费额度。

Neither Anthropic nor OpenAI publish their free allowance numbers exactly the same way Meta does with the tokens per week.

但是作为Claude和ChatGPT的用户，我经常遇到使用限制。

But as a user of both Claude and ChatGPT, I've hit usage limits a lot.

你知道，你在两三个小时内发送数百条消息，然后在这两个平台上当天就被限速了。

You know, you send hundreds of messages over the course of one to two to three hours, and you get rate limited for the day on both platforms.

Anthropic、OpenAI和其他AI实验室对用户的使用限制更严格，是因为他们无法承担免费提供所有这些计算资源的成本。

The reason that Anthropic, OpenAI, and these other AI labs tend to have stricter usage limits for their users is they can't afford to just give away all this compute.

Meta拥有而OpenAI和Anthropic没有的是，Meta是一个庞大的广告业务。

Something that Meta has, that OpenAI and Anthropic don't have, is Meta is a massive advertising business.

仅在上个季度，Meta就实现了590亿美元的广告收入。

In the last quarter alone, Meta did $59,000,000,000 in advertising revenue.

其他所有业务加起来才15亿美元。

The entire rest of the business did 1,500,000,000.

Meta本质上是一家广告公司。

Meta is an advertising company.

Anthropic和OpenAI不是广告公司。

Anthropic and OpenAI are not advertising companies.

对，OpenAI确实开始涉足广告业务，也开始通过广告从免费用户那里赚钱。

Yes, OpenAI has started launching an ad business and they are starting to monetize their free user base through ads.

他们主要还是……

They are largely a...

一个AI消费业务。

An AI consumption business.

你要么付费订阅来使用更高级的模型，要么用API按token用量付费。

You either pay for a subscription plan to get access to higher models, or you're paying for an API plan where you're getting charged on your token usage.

他们只有在你使用模型的时候才能赚到钱。

They only get paid based off of how much you're using the models.

Meta的收入取决于你在Facebook和Instagram上刷多久。

Meta is getting paid by how much time you spend scrolling on Facebook and on Instagram.

所以，他们可以用这些广告收入来补贴他们推出新AI Agent平台的成本。

Therefore, they can use all that ad money to subsidize their ability to roll out a new AI Agent platform.

结果呢？

The result?

Meta能负担得起补贴需求，而其他AI公司却被迫限制用户的使用量。

Meta can afford to subsidize demand while other AI players are forced to rate limit their users.

到目前为止，股市似乎很买账。

And so far, the stock market seems to love it.

Meta的股价从Muse发布前一天每股613美元涨到了今天录制时的每股约720美元。

Meta's stock is up from $613 a share the day before Muse's launch to around $720 a share today at the time of this recording.

我的看法是，永远不要跟马克·扎克伯格对着干。

My take from this, never bet against Mark Zuckerberg.

三年前，扎克伯格因为全力押注元宇宙而被大家嘲笑。

Zuck was getting dunked on three years ago for going all in on the metaverse.

我是说，他真的把Facebook改名为Meta，因为元宇宙是他的下一个大赌注。

I mean, he literally changed the name of Facebook to Meta because the metaverse is the next big bet.

大家可能都见过那张扎克伯格在元宇宙或Horizon Worlds里的卡通形象图，三年前他们就是这么叫那个平台的。

Everybody's probably seen the picture of the cartoon image of Zuckerberg in the Metaverse or Horizon Worlds or whatever they were calling that platform three years ago.

当时网上到处都是梗，说Meta的股价跌了80%，而扎克伯格还在努力把元宇宙变成现实。

And there were all the memes going around of, oh, Meta's stock has collapsed 80% while Zuckerberg is trying to make the Metaverse a real thing.

元宇宙其实没怎么成功。

The metaverse didn't really work.

Meta最初在VR上的押注，比如Oculus和Quest，也一直没真正火起来。

Meta's original VR bets, like Oculus and Quest, never really worked.

但这都不重要。

But it doesn't matter.

扎克伯格一直在追逐他眼中的下一个大机会。

Zuckerberg keeps going after what he sees as the next big thing.

当AI在2022年成为下一个大热点时，他毫不犹豫地投入了。

And when AI became the next big thing in 2022, he didn't hold back.

他买下了Scale AI的大量股份。

He bought a big stake in Scale AI.

他让Scale AI的CEO亚历山大·王担任他们的AI负责人。

He made Scale AI's CEO Alexander Wang their head of AI.

在过去一年半里，Meta全力投入，积极构建自己的AI平台，训练自己的开源模型。

And over the last year and a half, Meta has aggressively gone all in on building out their own AI platform, training their own open source models.

现在推出Muse作为他们的AI代理。

Now launching Muse as their AI Agent.

不过，在我们录制这期节目时，Meta也刚刚宣布了一条新的面向企业的业务线。

However, as we record this, Meta also just announced a new business line to sell to the enterprise.

他们会向不同企业销售AI工具，包括Muse、一个编码工具，以及其顶级模型和针对特定业务的代理的访问权限。

They'll sell AI tools, including Muse, a coding tool, and access to its top models and business specific agents to different businesses.

前MongoDB CEO CJ Desai将负责这项业务，直接向马克·扎克伯格汇报。

Former MongoDB CEO, CJ Desai will run it, reporting directly to Mark Zuckerberg.

MongoDB还提供数据库软件，企业用它来存储和管理应用背后的数据，随着AI使用量的激增，这变得越来越重要。

MongoDB also makes database software that companies use to store and manage the data behind their apps, which has been increasingly important as AI usage has exploded.

扎克伯格的声明说，这项业务将利用到“很少有其他公司具备”的优势。

Zuckerberg's statement says the business will draw on strengths, quote, few other companies have.

模型、智能体、基础设施，以及多年积累的企业合作关系。

Models, agents, infrastructure, and years of relationships with businesses.

目前，还没有定价或运营细节。

As of right now, there are no pricing or operating details yet.

作为背景，Meta今年在AI基础设施上的投入超过1000亿美元。

For context, Meta is spending more than $100,000,000,000 on AI infrastructure this year.

7月份，Meta提高了支出指引的下限，投资者对扎克伯格的回答不满意，导致其股价盘后下跌10%。

In July, Meta's stock price fell 10% after hours when Meta raised the low end of its spending guidance, and investors didn't like Zuckerberg's answers.

在录制这段内容时，Meta的股价今天下跌了4%。

Meta's stock prices dropped 4% today at the time of this recording.

但这一切都表明，当扎克伯格想要大干一场时，他会全力以赴。

But all this goes to show that when Zuckerberg wants to take a big swing, he's all in.

很明显，目前他全力投入的方向就是AI。

And it's very obvious that the current swing is AI.

不过，企业这块就先聊到这儿吧。

But that's enough on the enterprise.

咱们还是回到AI代理和消费者这边。

Let's go back to focusing on AI agents and the consumer.

过去几周，是AI代理首次面向普通大众发布和推广的时期。

The last few weeks have been the first period in which AI agents have been released and marketed specifically to the average person.

几个月前，Open Claw火了一把，Cloud Code、Codecs和Cursor也都有过自己的高光时刻。

Several months ago, Open Claw had a moment, as did Cloud Code, Codecs, and Cursor.

但这些基本上都是AI编程助手，使用者主要是科技圈内人、硅谷的技术宅，还有那些整天泡在x.com上的重度网民。

But all of these were largely AI coding assistants that were being used by tech insiders, nerds in Silicon Valley, and the folks that are terminally online on x.com.

Open Claw在今年一月和二月的时候特别火。

Open Claw was all the rage back in January and February of this year.

基本上，OpenClaw就是一个跟实验室无关的AI助手。

Basically, OpenClaw was this non lab associated AI assistant.

咱们让OpenClaw跑起来吧。

Let's make OpenClaw work.

你得在另一台电脑上安装它，比如Mac Mini。

You had to install it on a separate computer like a Mac Mini.

你得下载几十个程序。

You had to download dozens of programs.

还有很多格式问题，你得进设置里确保它能访问你的邮箱，还有这个那个的。

There were a lot of formatting issues with having to go into the settings and making sure it had access to your email and this and that.

用户遇到的问题是OpenClaw实际上记不住他们告诉它的一些事情，他们还得调整数据库。

Users had issues with OpenCLaw not actually remembering certain things that they had told it and they had to tweak their databases.

基本上，普通用户或者普通消费者是不会花几个小时去设置OpenClaw，还要把它跟所有系统连起来。

Basically, the the normal user or the normal just consumer was not gonna spend the hours it took to get OpenCLL set up and actually wired them with all their systems.

他们也不会相信它不会被黑客攻击。

And they also weren't gonna trust that it wouldn't get hacked.

他们不会担心它遭受提示注入攻击。

They wouldn't suffer from a prompt injection.

他们不会担心它会意外泄露数据，或者买一些不该买的东西。

That it wouldn't accidentally leak their data or buy something that wasn't supposed to.

如果得花十个小时才能把OpenCLaw弄好，大多数人根本不会去弄。

If you had to spend ten hours to get your OpenCLaw working correctly, most people just weren't gonna do that.

Muse做得对的地方就是，上手特别特别简单。

The thing that Muse got right is the onboarding is so so simple.

五分钟之内，你就能把生活里的所有东西都接入Muse，然后直接问它问题，它就会帮你办事。

Within five minutes, you can have every piece of your life wired into Muse, and then you can just start asking it questions, and it'll do things for you.

有些批评者会说，不，我不想让马克·扎克伯格或者Muse或者Meta拿到我所有的数据。

Some critics would say, no, I don't wanna give Mark Zuckerberg or Muse or Meta access to all my data.

就我个人而言，我很喜欢精准广告。

Me personally, I love targeted ads.

我根本不在乎数据隐私。

I don't give a shit about data privacy.

我宁愿让马克·扎克伯格看到我所有的数据，然后通过Muse给我推送更好的广告、更有用的信息。

I would much rather Mark Zuckerberg see all my data and send me like better ads, more useful information through Muse.

我觉得他们比我自己搭建隐私系统、用OpenClaw，或者跟一个连SOC 2合规都搞不清楚的YC孵化初创公司合作，能更好地管理我的数据。

I would like to think that they can have a much better handle on my data than say, me building my own privacy systems and trying to use OpenClaw, or working with another like YC backed startup that can barely spell SOC two compliance.

我有一个比较激进的观点，就是我真的不太在乎我的个人数据。

One of my hotter takes is I really just don't care about my personal data that much.

如果我把数据交给一个更聪明的公司，然后因此得到更好的结果，那我肯定愿意这么做。

And if I give a much smarter company access to the data, and I get better outputs because of that, I'm gonna do that all day.

而且我觉得大多数消费者也有类似的想法，毕竟我们在Instagram和Facebook上花了那么多时间。

And I think most consumers feel pretty similarly given how much time we spend on Instagram and Facebook.

关闭广告定向投放几乎是一种很老派的想法。

It's almost like a very boomer mentality to turn off ad targeting.

我就喜欢定向广告。

Like, I love targeted ads.

那么，AI代理的广泛普及对AI领域意味着什么呢？

So what does this broad proliferation of AI agents mean for the AI landscape?

过去三年里，AI一直都只是代码而已。

For the last three years, AI has always just been code.

大语言模型擅长两件事：写作和编程。

LLMs are good at two things, writing and coding.

因为编程其实就是另一种形式的写作。

Because coding is really just another form of writing.

大语言模型的工作原理就是预测下一个词。

The way that LLMs work is just next token prediction.

它们会猜下一个词、字母或数字应该是什么，然后继续往下推。

They make a good guess on what the next word or letter or number should be and keep going from there.

编程在AI上普及得很快，原因之一是，我们有大量可验证的训练数据来训练这些编程模型。

Coding adoption took off really quickly with AI because one, we have a ton of verifiable training data needed to train these coding models.

原因之二是，用AI自动化编程有很大的经济潜力，因为现在我们的世界很大程度上依赖软件运行。

And two, there's a lot of economic upside in automating coding with AI because so much of our world runs on software now.

如果AI能比人类做得更好，让编程更快、更好、更高效，那就有很多钱可赚，因此也就有很多理由大力投资AI编程。

If you can make coding faster, better, and more efficient because AI can do it better than humans, there's a lot of money to be made, and therefore a lot of reasons to invest heavily in AI coding.

这就是为什么我们看到Anthropic在Cloud Code上加倍投入。

That's why we've seen Anthropic double down on Cloud Code.

这也是为什么我们看到OpenAI开始把Codecs作为他们的主打产品，而不是ChatGPT。

That's why we've seen OpenAI start pushing Codecs as their biggest thing rather than ChatGPT.

硅谷意识到，编码代理基本上可以在电脑上做任何事情。

Silicon Valley realized that coding agents could basically do anything on a computer.

在过去一年里，基本上从2025年第四季度Cloud Code变得好用以来，编码代理一直风靡一时。

And for the last year, basically since Cloud Code got good in Q4 twenty twenty five, coding agents were all the rage.

直到现在，我们还没有看到AI在消费者中的广泛使用。

What we have not seen until now is broad consumer use of AI.

例如，截至2026年2月，只有3%的美国银行客户在使用付费AI服务。

Only 3% of Bank of America customers, for example, were using paid AI services as of February 2026.

今年年初，AI能做的事情和大多数美国人认为AI能做的事情之间存在巨大差距。

There was a huge divergence in what AI was capable of doing at the beginning of this year, and what the majority of Americans thought AI could do.

Muse是前沿消费代理首次真正有机会实现突破性增长。

Muse is the first real chance for a cutting edge consumer agent to hit escape velocity.

我想指出的是，Muse并不是第一个面向消费者的AI代理平台。

I'd like to note that Muse is not the first AI Agent platform for consumers.

一家名为Instinct的新创公司在进入私人测试版七个月内就吸引了10万用户。

A new startup called Instinct hit 100,000 users within seven months of entering private beta.

Instinct是一个基于iMessage的AI代理，可以订机票、读邮件、根据你的日历发送跟进信息，现在它有一批狂热的追随者。

Instinct is an iMessage based AI Agent that can book flights, read your email, send follow ups based on your calendar, and it has a bit of a cult following right now.

而且有传言说他们正在以100亿美元的估值筹集更多资金。

And they're rumored to be raising more funding at a $10,000,000,000 valuation.

Grockbot是由SpaceX AI、Cursor以及现在埃隆·马斯克旗下所有公司联合推出的，它在8月发布一个月后就达到了每周41.8万用户。

Grockbot, which was launched by SpaceX AI, Cursor, the whole conglomeration of Elon Musk companies now, hit 418,000 weekly users one month after its August launch.

但Grockbot主要面向企业市场。

But Grockbot was geared toward the enterprise.

Instinct有一批狂热粉丝，但他们没有Meta那样的分发渠道。

Instinct has a cult following, but they don't have the same distribution that Meta has.

Muse在这两款应用面前遥遥领先，头几周就获得了340万次下载。

Muse is smoking both of these apps with 3,400,000 downloads in its first couple of weeks.

所以虽然Muse不是第一个消费者代理，但它是第一个同时受益于两点代理平台：一是广泛的分发渠道，二是Meta凭借其强大的广告业务能够免费提供它的能力。

So while Muse wasn't the first consumer agent, it is the first agent platform that benefits from both one, broad distribution, and two, Meta's ability to give it away for free thanks to its killer ad business.

越来越多的科技巨头加入消费者代理之战只是时间问题。

It's only a matter of time before more and more of the big tech players get involved in the consumer agent battle.

苹果很可能会推出自己的消费级智能助手。

Apple will probably have a consumer agent.

他们正在和Gemini合作，为新的Siri AI模型提供支持。

They're partnering with Gemini for the new Siri AI models.

谷歌也应该推出自己的Muse版本，毕竟他们有Gmail、日历和安卓系统。

Google should also have their own version of Muse, considering they have Gmail, they have your calendar, they have Android.

说实话，我很惊讶谷歌这次居然没抢在Muse前面。

It's actually shocking to me that Google didn't beat Muse to the punch on this.

AI智能助手既被商品化又向公众开放，这也带来了一些有趣的后果，比如在计算资源上的巨额投入。

There's also some interesting implications of AI agents both being commoditized and accessible to the public, in terms of insane spend on compute.

OpenAI和Anthropic现在必须上市了，因为私募市场基本上已经没钱可拿了。

OpenAI and Anthropic need to go public now, because the private markets are basically tapped out.

他们需要不断筹集更多资金来购买更多计算资源，以服务越来越多的客户。

They keep needing to raise more and more money to buy more and more compute, to keep servicing more and more customers.

他们目前并不盈利。

They aren't profitable.

他们现在处于烧钱模式。

They are in cash burn mode.

他们一直在筹集数千亿美元来继续补贴这种需求。

They've been raising hundreds of billions of dollars to continue to subsidize this demand.

他们在训练前沿模型上花了很多钱，但这些模型很快就被其他追赶上的模型商品化了。

They're spending a ton of money on training frontier models that quickly become commoditized as the other model catches up.

你会看到OpenAI发布消息说他们推出了一个打破所有基准的新模型。

You'll see OpenAI publish that they had a new model come out that breaks all these benchmarks.

然后Anthropic又发布一个模型，超过了他们。

Then Anthropic publishes another model that beats them.

接着还会有中国模型，性能达到95%，但价格只有5%。

And then you'll have a Chinese model that's 95% as good for 5% of the price.

所以这些公司通过新模型获得的任何优势，似乎在发布几周内就完全被商品化了。

So whatever edge these companies get with their new model seems to get fully commoditized within just a few weeks of launch.

Meta和谷歌有巨大优势，因为他们从广告业务中产生了大量自由现金流。

Meta and Google have massive advantages in that they generate a ton of free cash flow from their advertising businesses.

当你的主营业务能赚几百亿美元时，你就能免费送出几十亿美元的算力，只为了让用户用你的AI代理而不是别人的。

When you're making hundreds of billions of dollars with your main business line, you can then afford to give away billions of dollars of compute for free just to get users to use your AI Agent over theirs.

如果你用Muse，它一周不间断地为你工作，而你用ChatGPT时老是遇到限流，迟早你会选择那个不会出问题的AI工具。

If you're using Muse, and it works nonstop for you all week, and you're using ChatGPT, and you keep getting rate limited, sooner or later, you're just gonna choose the AI tool that doesn't break.

Muse发布时，最有趣的一点是看看哪些公司选择与它合作，哪些还没有。

One of the more interesting things about the Muse launch is seeing which companies did partner with Muse and which ones have not yet.

PayPal、Shopify、Expedia、Stripe、Best Buy、Walmart（这可是个大牌）、Instacart、American Eagle、Dick's Sporting Goods、Fanatics、Gap和Wayfair都已经与Muse合作了。

So PayPal, Shopify, Expedia, Stripe, Best Buy, Walmart, big one, Instacart, American Eagle, Dick's Sporting Goods, Fanatics, Gap, and Wayfair have all partnered with Muse.

你上去之后，他们有连接器，你可以直接插上，然后通过你的账户访问所有这些平台。

You go on, they have a connector, you can plug it in, and you can access all of those platforms through your accounts.

值得注意的是，亚马逊屏蔽了Muse。

Amazon notably blocked Muse.

几周前，很多用户试图用Muse在亚马逊上买东西，结果亚马逊直接屏蔽了Muse的访问权限，声称它违反了用户条款和条件。

A lot of users were using Muse to try to buy stuff on Amazon a couple of weeks ago, and Amazon went out and blocked Muse from being able to access the platform, claiming it was violating user term and conditions.

还扯了什么安全问题、确保数据隐私之类的，反正就是一堆废话。

There was something about like safety issues and making sure data privacy and blah blah blah.

他们不想让Muse进来搅乱他们的广告业务，这并不奇怪。

Like it's not a surprise they don't want Muse coming in and screwing up their advertising business.

而且他们也有机会继续打造自己的AI代理，这些代理拥有亚马逊的库存和用户的购买历史。

And they also have their own opportunity to keep building out their own AI agents that have Amazon's inventory, that have user purchase history.

如果你的Muse连接到你的亚马逊账户，获取了你的购买历史，那就有大量数据可以用来更精准地推荐商品或在代理广告中投放。

Where if your Muse is connected to your Amazon and it gets your purchase history, that's a lot of data where they can better target goods or in agent ads.

现在还不清楚那会是什么样子。

Not even sure what that would look like yet.

但亚马逊足够大，他们有筹码，不需要和Muse合作。

But Amazon's big enough, they have leverage, they don't have to partner with Muse.

看着挺有意思的。

It's something that is fun to watch.

就是看哪些公司对平台上的代理非常开放，哪些公司坚决反对。

It's just which companies are very open to agents being on platform, which companies are absolutely against it.

我们接下来要和我一个朋友JC聊聊，他因为用AI代理在Fort Charles订了一堆位子，被Resy封了号。

We're about to talk to one of my buddies, JC, who had gotten banned from Resy for having an AI Agent trying to make a bunch of reservations at Fort Charles.

与此同时，OpenTable 正在与 Muse 整合。

Meanwhile, OpenTable is integrating with Muse.

所以实际上没人知道这会怎么发展，但接下来六个月看看哪些公司非常支持代理上平台，哪些公司想尽办法阻止，会很有意思。

So nobody actually knows how this is gonna play out, but it'll be fun to watch over the next six months which companies are very pro Agent being on platform and which ones are trying to block it at all cost.

这位是 JC Bahr-de Stefano，Better Tomorrow Ventures 的冒险资本投资人。

Here's JC Bahr-de Stefano, adventure capital investor at Better Tomorrow Ventures.

我

I

被推荐使用 Instinct，它就像一个神秘兮兮、充满光环的代理，没人真正知道它到底是怎么回事。

got recommended to use Instinct, and it was this like mysterious kind of like aura filled agent where, like, nobody really knew what was going on with it.

它的网站非常非常简陋。

It had this website that was, like, very, very limited.

而且你知道，风投们总是想赶时髦，一有新工具出来就赶紧用。

And, you know, like, where VCs were, like, trying to be, like, cutting edge and use all of these tools when they come out.

所以我就开始用了。

So I started using it.

我第一次用它的时候，有个特别震惊的瞬间，就是我想约医生做年度体检。

And the first kind of, like, oh, shit moment that I had with it was I was trying to see my doctor do my annual physical.

我上次体检是去年八月，所以今年八月正好满一年。

And it would have been a year after my last physical would have been in August.

所以我七月就打电话给他们，问能不能八月去。

So I called them in July and said, can I come in in August?

他们说，我们下一个可预约的时间是一月。

And they said, the next appointment that we have is in January.

然后我就给Instinct发了个消息，几乎没给什么提示或指导，就说，你能想办法把我的医生预约提前吗？

And so I just texted instinct with, like, very, very little prompt or guidance and said, can you figure out a way to move up my doctor's appointment?

我发誓，大概十分钟后，我就接到医生办公室的电话，说，嘿。

And I swear to God, like, ten minutes later, I got a call from the doctor's office saying, hey.

我们看到了你的请求。

We saw your request.

刚好有人取消了预约。

Like, we just had a cancellation.

我们九月底有个预约。

We have an appointment in late September.

你想来吗？

Do you wanna come in?

我当时就震惊了。

And I was like, woah.

我简直不敢相信。

I was like, woah.

因为那感觉太真实了。

Because it was like real world.

这一切都在我不知情的情况下悄悄发生了。

It all happened in the background without me knowing.

后来我才发现，它基本上自己想办法进入了我的医生门户网站，因为它能从我的Gmail邮件里提取验证码来登录。

And then I figured out that it, like, basically figured out how to get into, like, my doctor's web portal because it can pull the OTP from the email when it's trying to sign in with my Gmail.

所以我就觉得，哇。

So I was like, wow.

这真是太神奇了。

This is this is this is pretty magical.

最后，我想，好吧。

Finally, I'm like, okay.

让我给它来个终极考验。

Let me give it the ultimate test.

比如，想办法订到我一直想去的纽约那家餐厅的位子。

Like, figuring out how to get a reservation at a restaurant in New York that I've always wanted to go to.

我就是在纽约市出生和长大的。

Like, I was born and raised in New York City.

大学毕业后大部分时间我都待在这里，但我还从来没去过Fort Charles。

I've been here most of the time post college, and I've still never been to Fort Charles.

而且，这可是最老牌的那家。

And, like, this is the OG.

一直以来，人们都在写机器人和脚本，想方设法订到Fort Charles的位子。

Like, this is like people have been writing bots and scripts to try to get reservations at Fort Charles forever.

你知道吗，人们一直在部署这些东西，然后在二手市场上卖掉它们。

People have, you know, have been deploying these things and then selling them on secondary marketplaces.

这本身就是一个完整的世界。

It's like a whole world in and of itself.

所以我就想，这是终极考验。

So I was like, this is the ultimate test.

于是我给它发了条短信，后来我看了下，提示词其实就是：帮我订一个查尔斯餐厅的位子。

So I texted it, and and the prompt was literally, because I reviewed it afterwards, just get me a Reset for, Charles.

然后在接下来的四五天里，它会时不时来问我进展，说，嘿。

And so then for the next four or five days, it would check-in with me, and it would say, hey.

等我进去的时候，全都订光了。

By the time I got in, it was all gone.

我明天再试试。

Like, I'm gonna try again tomorrow.

然后过了几天，我其实是想订个位子，跟我爸妈去吃晚饭。

And then a couple days later, I was like actually trying to get a reservation to go to dinner with my parents.

我一直想登录我的Resi账户，但就是登不进去。

And I kept trying to get into my Resi account, and I couldn't get in.

我当时就想，哎，这有点奇怪。

And I was like, oh, this is weird.

然后我琢磨了一下，心想，不知道Instinct是不是做了什么。

And then I thought about it, and I was like, I wonder if Instinct did anything.

我的Resi账户出事了。

Something happened with my Resi account.

于是我查了一下邮件，果然，大概一个小时前，我收到一封来自Resi的邮件，说：你违反了我们的服务条款。

And so I checked my email, and like, sure enough, like an hour before, I had an email from Resi saying, you violated our terms of service.

你已经被平台封禁了。

You've been banned from the platform.

然后我打开Instinct，问它。

And so I go into Instinct, and I asked it.

我说，你为了完成我给你的这个任务，到底做了什么？

I was like, what did you actually do to try to, like, accomplish this task that I gave you?

然后它回复说，哦，我一直在调用他们的API，你知道，每小时几百次，每天几千次，持续了大概一到两周。

And it comes back, and it's like, oh, I was calling their APIs, you know, multiple hundreds of times per hour, thousands of times per day over the course of, like, you know, one to two weeks.

它根本没有运用智能逻辑。

Like, it was not applying intelligent logic.

比如，这个时间段是取消预订的高峰期。

Like, this is the time of day when cancellations happen.

或者，我应该把精力集中在预订取消的时候。

Or like, I should, like, put all of my efforts into, when the reservation drops.

它只是整天不停地疯狂请求。

Like, it was just, like, pinging it constantly nonstop all day.

我看到之后就想，哦，难怪我被封了。

And so I saw that, and was like, oh, of course I got banned.

这完全不意外。

Like, that's not that's not surprising at all.

所以我发了条推文，说干得好，Instinct，还附上了那封邮件的截图。

And so I tweeted saying nice going instinct, and it was just like a screenshot of the email.

那封邮件最后获得了数百万的浏览量。

And that email ended up getting like like like millions and millions of views.

然后我还转发并评论了那条推文，说这就是它实际用来获取预订的方式。

And then I also like like quote tweeted that with the like, this is what it actually did to get the reservation.

我觉得这正好赶上了时机。

And I think it was it was just very timely.

大家都在讨论，如果我们都有这些智能代理，那实际会是什么样子、什么感觉？

Like everyone's talking about like, if we all have these agents, like what does it actually look like and feel like?

但有趣的是，看看谁在积极采用智能代理，谁在说不要这个。

But it is funny seeing like, who's leaning into agents, and who is saying like, no, like we don't want this.

现在市场上，尤其是在消费者界面方面，确实出现了明显的分化。

There's a real like bifurcation in the market of like just consumer interfaces right now.

我觉得到目前为止，我的观察是，这完全取决于商业模式。

I mean, like my observation on it so far is it it totally depends on the business model.

我记得就在上周，Muse 还在密集发布关于新集成的公告。

And, like, I think it was it was last week when Muse was just, like, banging out, like, announcements about new integrations.

他们接入了Shopify。

They had they had Shopify.

也接入了PayPal。

They had PayPal.

还有Expedia和Instacart，全都挤在几天之内，而亚马逊也出来说，不行。

They had Expedia, and they had Instacart all all all in the span of couple days when Amazon also came out and said, no.

你不能这么做。

You can't do this.

但你想一下这些商业模式，特别是Expedia和Instacart，它们可不想被那些能自己搞定一切的智能体给绕过去。

But, like, if you think about those business models, right, like Expedia and Instacart in particular, like, they don't wanna be disintermediated by agents that can go do all of these things without them being there.

所以，它们的动机很明显，如果自己不搞，智能体要么直接绕过它们，要么就跑去用Booking.com、Kayak或者DoorDash。

So, I mean, their incentive surely is like, if we don't do this, then the agents can either bypass us or go to booking.com or Kayak or DoorDash instead.

所以它们必须这么做，因为担心自己的商业模式会被颠覆。

And so we have to do this for fear of existential risk on our business model.

而PayPal和Shopify呢，它们根本不在乎。

And then PayPal and Shopify, they don't care.

他们只要有钱流动就能赚钱。

Like, they make money as long as money flows.

所以不管是代理做还是人做，都没关系。

So if an agent does it or a human does it, it doesn't matter.

对他们来说，蛋糕只会越来越大。

Like, the pie is gonna just grow for them.

所以他们只需要确保自己始终在交易和支付流程中，不管最终用户是谁。

And so they just need to make sure that they're in the transaction flow and the payment flow regardless of who who the end user ends up being.

但你再想想亚马逊和Resy，他们采取了不同的做法。

But then you think about, like, Amazon and Resi, which have kind of taken a different approach.

就像，如果我这里拥有库存，那我的……

It's like, if I own the inventory here, then what is my...

我是说，这尤其对Resy来说。

Like, I mean, this is particularly for Resy.

如果我拥有库存，那我有什么动力去做这件事呢？

If I own the inventory, then like, what is my incentive to do this?

我只有在活动量达到一定程度，经济上划算的时候，才会有动力这么做，因为那样我就能处理更多的预订。

Like, I am only incentivized to do this if we get to a point where there is so much activity that economically it makes sense because I'm gonna be able to do more reservations if I do that.

但在那之前，我并没有面临同样的生存威胁和风险。

But until then, I'm not under the same the same type of existential threat and risk.

真的会有人去打造一个以代理为核心的Resy吗？

Like, is someone really gonna build like an agent first resi?

我觉得不会，因为那不是这门生意的难点。

Like, I I I I don't think so because that's not the hard part about the business.

这门生意的难点在于和餐厅的关系，以及如何把他们吸引到平台上来。

The hard part about the business is the relationships with the restaurants and getting them onto the platform.

开发一个帮助人们预订的工具并不难。

It's not it's not hard to build a tool that helps people do, you know, reservations.

而亚马逊呢，他们不希望任何人这么做，因为他们的广告收入是

And then Amazon is like, I don't want anyone to do this because, you know, my like ad is

四分之一的广告收入。

a quarter in ad revenue.

他们需要的是眼球关注。

They they need the eyeballs.

十亿的收入。

Billion revenue.

他们需要眼球来看那些山寨的中国垃圾内容，这些内容排在正常内容上面。

They need the eyeballs to see the knock off like Chinese slop that goes above the regular stuff.

他们就是靠这个赚钱的。

That's how they make their money.

不过，哥们，谢谢你上节目。

But dude, thanks for coming on the show.

聊得很开心。

Good to chat.

谢谢邀请我。

Thanks having me.

是的，哥们。

Yeah, dude.

继续用你的智能体搞点有意思的东西。

Keep doing weird shit with your agents.

好的，兄弟。

Alright, man.

你也是。

You too.

不错。

Sweet.

回头见，JC。

See you JC.

我们休息一下，马上回来。

We'll be right back after a quick break.

本期节目由Odoo赞助播出。

Support for the show comes from Odoo.

如果你的公司靠五个不同的应用、十二个浏览器标签页，还有一个谁都不敢碰的电子表格来运转，那可能是时候试试Odoo了。

If your business runs on five different apps, 12 browser tabs, and one spreadsheet that everyone's afraid to touch, it's probably time for Odoo.

Odoo 是一款一体化的企业管理软件，能把公司各个部分整合到一个强大且易于使用的平台上。

Odoo is an all in one business management software that connects every part of your business into one powerful, easy to use platform.

这样就不用浪费时间在互不相通的系统之间切换，整个公司可以实时协同工作。

So instead of wasting time switching between disconnected systems, your entire business works together in real time.

团队效率更高，数据更准确，你也能真正专注于发展业务。

Your team moves faster, your data stays accurate, and you can actually focus on growing your business.

让一个统一的系统来管理你的整个公司。

Let one unified system run your entire business.

从第一个商机到最终收款，所有环节都集中在一个地方。

From the first opportunity to the final payment, everything together in one place.

无论你是小企业还是管理大型业务，Odoo 都提供一个灵活的、能随你成长的平台。

Whether you're a small business or managing a large operation, Odoo gives you one flexible platform built to grow with you.

今天就到 odoo.com/profg 免费试用吧。

Try today for free at odoo.com/profg.

网址是 odoo.com/profg。

That's odoo.com/profg.