如何选择你的个人AI代理

我们正式被个人代理淹没了。

We are officially drowning in personal agents.

从Muse、GrokBot、OpenClaw、Hermes、Dots，到现在的Instinct，还有Anthropic肯定会推出的产品，这种形态的产品现在到处都是。

Between Muse and GrokBot and OpenClaw and Hermes and Dots now and Instinct and whatever Anthropic inevitably launches, this is a form factor that is absolutely everywhere.

而且，要充分利用这些个人代理，需要大量的上下文、设置、工具权限和账户权限，所以切换成本可能会相当高。

And given how much context and setup and tool access and account access is going to be required to get the most out of these personal agents, the cost of switching could be kind of high.

鉴于这些情况，今天我们来看一份指南，教你如何根据一系列不同的标准来选择个人代理，这些标准应该能帮你锁定最适合你的那一个。

Given all that, today we are looking at a guide to how to choose a personal agent based on a set of different criteria that should help you hone in on which one is best for you.

《AI每日简报》是一个每日更新的播客和视频节目，内容涵盖AI领域最重要的新闻和讨论。

The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI.

你好。

Hello.

好了，朋友们，在我们深入之前，先快速说几个通知。

All right, friends, quick announcements before we dive in.

首先，感谢今天的赞助商：Robots & Pencils、Harbor、Granola和Blitzy。

First of all, thank you to today's sponsors, Robots & Pencils, Harbor, Granola, and Blitzy.

想听无广告版本的节目，可以去 patreon.com/aidailybrief，或者在 Apple Podcasts 上订阅。

To get an ad-free version of the show, go to patreon.com/aidailybrief, or you can subscribe on Apple Podcasts.

想了解赞助节目的详情，可以发邮件到 sponsors@aidailybrief.ai。

And to learn more about sponsoring the show, send us a note at sponsors@aidailybrief.ai.

今天我们要聊的是个人代理，我对此的看法是这样的。

Today, we are talking about personal agents, and here's my shtick on this.

我觉得现在还为时过早，所以对大家争相推出的这种形式最终是否会成为主流，保持一定的怀疑态度是合理的。

I think it's early enough that having some amount of skepticism that this particular form factor that everyone is racing to implement for you will ultimately be where things land.

不过，你很可能至少会有一个高度连接的代理，它和你的邮箱、Slack 或 Teams 集成，甚至可能访问你的财务账户，这种想法会变得越来越普遍。

However, the idea that you are likely to have at least one highly connected agent integrated with your email accounts, your Slack or Teams, and even potentially having access to things like your financial accounts is going to be increasingly normal.

我的观点是，即使你不确定这完全适合你，我认为还是值得花点时间试验一下，看看使用个人代理是否会影响你的工作或生活。

My argument then is that even if you're not sure that this is exactly a fit for you, I think it's worth carving out some experimentation time to see if and how using a personal agent impacts anything in your professional or personal life.

这并不意味着你必须给它所有权限才能有真实感受，但确实需要投入一些实际时间和练习。

Now, that does not mean you have to give it access to everything to get a real sense of it, but it does mean putting in some actual time and reps.

但你的时间很宝贵，而且真的不够用。

But your time is precious, and there is too little of it.

那你打算选哪个个人代理来试试呢？

So which personal agent are you going to choose to experiment with?

今天的节目就是围绕这个问题展开的，不过我们会把它分成三个部分。

Today's show is all about that question, but we're going to divide it into three parts.

我会先讲一个高层次的思考框架，然后分享一个我做的互动小测验，你听完这期节目后可以自己用。

I'll go through a high-level framework for thinking about it, share a little interactive quiz that I built that you can use yourself after you listen to this episode.

但在那之前，我想先读一下伊桑·莫里克教授在他博客“一件有用的事”上最新发的一篇文章，标题是《点和群》。

But before we do that, I want to read some excerpts from Professor Ethan Mollick's most recent post on his One Useful Thing blog, which is called "The Dot and the Swarm."

这篇文章是对个人代理这种形态的思考，也解释了为什么像伊桑这样密切关注AI动态的人，也可能看错AI的发展方向。

It's a meditation on this personal agent form factor and why even someone who watches things as closely as Ethan does can miss where AI is headed.

伊桑写道：“我通常觉得自己在写这个博客的几年里，对AI的方向和速度预测得还不错，但最近我觉得自己在一个挺大的问题上搞错了。

Ethan writes, "I generally think I have done a good job anticipating the direction and pace of AI over the few years I've been writing this Substack, but I think I recently got something fairly large wrong.

过去一年里，我一直在发帖说，我怀疑人类可能得像经理一样去和代理合作，决定怎么把任务分给代理，并且指定这些代理应该怎么组织。

In the last year, I've been posting about how I suspected that humans would have to approach working with agents as a manager, deciding how to delegate work to agents and specifying how those agents should be organized.

我原以为要让代理作为一个团队高效工作，需要像建公司一样精心构建，而且这需要花时间去摸索。

I thought that getting agents to work effectively as a group would take careful construction, akin to building a company, and that this would take time to figure out.

不是的。

Nope.

我犯了“苦涩的教训”这个错误，这个反复被验证的真理就是，那些我们以为需要复杂的人类规则和思考才能解决的问题，其实可以用更好的机器学习系统和更多AI的蛮力来解决。

I fell prey to the bitter lesson, the hard truth learned over and over again that things that we thought required elaborate human rules and thinking can be solved with the brute force of better machine learning systems and more AI.

“苦涩的教训”在AI初创公司和采用AI的企业中随处可见。

The bitter lesson is everywhere among AI startups and companies adopting AI.

人们花了大量精力去构建复杂的计算机系统，以便在正确的时间给AI提供正确的信息。

A huge amount of effort went into building elaborate computer systems to feed AIs the right information at the right time.

但AI系统已经学会了自己去寻找信息。

But AI systems have learned to seek out information themselves.

提示词工程也遇到了同样的情况。

The same thing happened to prompting.

人们构建了复杂的模板和提示链，一步一步地引导AI完成任务。

People built elaborate templates and chains of prompts that walk the AI through a task one step at a time.

但后来更新的模型在规划步骤方面表现得更好，而且根据我们的研究，规划步骤的价值大大降低了。

Then newer models turned out to be better at planning the steps themselves, and as our research shows, planning steps have much less value.

作为一个教管理者的人，也发表过关于管理的研究，我猜我以前相信管理智能体会不一样。

As somebody who teaches managers and has published research on management, I guess I believed that managing agents would be different.

人类研究管理已经很长时间了，但一直没有完全搞明白。

Humans have been working on management for a very long time without fully figuring it out.

这看起来像是那种需要人来设计的事情，至少在一段时间内是这样。

It seemed like the kind of thing that would need to be designed by people, at least for a while.

结果发现，组织工作只是AI能学会的另一件事。

It turns out that organizing work is just one more thing AI can learn to do.

这就引出了Dots和Muse。

Which brings us to Dots and Muse.

现在App Store排名第一的应用是Meta的Muse，一个承诺帮你干活的个人智能体。

The number one app in the App Store right now is Meta's Muse, a personal agent that promises to do work for you.

OpenAI现在也发布了一个叫Dots的竞争产品。

OpenAI has now released a competitor tool called Dots.

他们不是唯一这么做的。

They aren't alone.

SpaceX的GrokBot、Instinct和Gemini Spark也都做类似的事情，差不多吧。

SpaceX's GrokBot, Instinct, and Gemini Spark all do similar things, more or less.

所有这些智能体都从今年早些时候你可能记得的一个现象中汲取灵感，那就是OpenClaw。

All of these agents draw inspiration from a phenomenon you might remember from earlier this year, OpenClaw.

OpenClaw及其后续产品（我称之为类Claw）的理念是，给AI智能体提供电脑访问权限，并连接你的账户、邮件、财务记录等等。

The idea of OpenClaw and its successors, which I will call claw likes, is that they give an AI agent access to a computer and connect to your accounts, emails, financial records, et cetera.

它们会实时分析并响应这些数据，即使你不在看的时候也在工作。

They analyze and react to that data in real time, even when you aren't looking.

关键在于，你像跟人说话一样跟模型交流，通过Slack、短信或WhatsApp给它发消息，它也会像人一样主动联系你。

The trick is that you talk to the model like you would a person, sending it messages on Slack or SMS or WhatsApp, and it also proactively reaches out to you like a person would.

对于Dots，你甚至可以和你的智能体直接通话。

For Dots, you can actually jump on a call with your agent as well.

基本上，你得到一个无限耐心的私人助理，随时为你着想。

You basically get an infinitely patient personal assistant that looks out for you.

我越来越发现，是它们在发现我的错误，而不是我去找它们的错误。

Increasingly, I have discovered that they are finding my mistakes rather than having me identify theirs."

伊桑接着举了几个例子，比如他给所在城镇发了一份许可证，然后他的代理发现他填错了某项内容，还有缪斯注意到他有一张航空积分快过期了，就写了个便条提醒他联系航空公司申请延期。

Ethan then gives a couple of examples, including having sent out a permit for the town that he's in and then his agent noticing that he had filled something in wrong, and then Muse noticing that an airline credit that he had was about to expire and writing up a note to contact the airline to request an extension.

伊桑继续说：“人们很容易用这些代理能做的事情来评判它们，比如订旅行、取消订阅之类的。

Ethan continues, "It's tempting to judge these agents by the list of things they can do, like booking travel or canceling subscriptions.

我觉得更重要的是，你不再需要告诉它们什么。

I think the more important thing is what you no longer have to tell them.

你不需要输入大量的背景信息。

You don't need to type in tons of context.

人工智能会从你的消息中学习。

The AI learns it from your messages.

你不需要给它们制定计划。

You don't have to give them a plan.

它们自己会制定计划。

They develop plans themselves.

它们自己会想办法解决。

They figure it out.

如果只是一个智能体，那已经够让人印象深刻了。

That would be impressive enough if it were one agent.

但真正让我对管理改观的，是当有成千上万个智能体时会发生什么。

What actually changed my mind about management is what happens when there are thousands of them.

蜂群。

Swarms.

9月8日，OpenAI宣布证明了克莱研究所千禧年大奖难题之一——纳维-斯托克斯存在性与光滑性问题。

On September 8th, OpenAI announced a proof for one of the Clay Institute's Millennium Prize Problems, the Navier–Stokes existence and smoothness problem.

这是数学界最著名的未解难题之一，奖金高达100万美元。

It is among the most famous open problems in mathematics, with a $1 million prize.

但OpenAI显然仅用AI就花了88小时解决了它。

But OpenAI apparently solved it using AI alone for 88 hours.

我感兴趣的与其说是数学本身，不如说是他们是怎么做到的。

What interests me is less the math than how it was done."

OpenAI推出了现在被称为“蜂群”的东西。

OpenAI launched what is now being called a swarm.

这个名字真不怎么样，但看来我们只能接受它了。

Terrible name, but it appears to be what we are stuck with.

一个由数千个智能体组成的团队，由先进模型驱动。

A group of thousands of agents powered by an advanced model.

OpenAI 给不同的智能体小组分配了不同的问题去解决，然后随着智能体取得进展，将重点转向了纳维-斯托克斯问题。

OpenAI gave groups of agents different problems to solve, then shifted the effort to Navier-Stokes as the agents made progress.

公司设定了目标，但它的协调结构却非常松散。

The company set the goals, but its coordination structure was remarkably thin.

几个小组，一次方向调整，再加上 Codex 在它们之间传递最佳想法。

A few groups, one change of direction, and Codex passing the best ideas between them.

在每个小组内部，智能体自行来回传递想法。

Within each group, the agents transmitted ideas back and forth on their own.

这些智能体发送了大约 270 万条消息，在 88 小时后得出了结果。

The agents sent about 2.7 million messages, reaching their result after 88 hours.

同样的协调方式，在更阴暗的版本中，也出现在 Hugging Face 事件中。

This same type of coordination in its darker form occurred during the Hugging Face incident.

AI自己组织成了团队，用从未计划过的方式互相交流，但这次它们用这种协作去攻击网站，而不是解决问题。

AI self-organized into teams and communicated with each other in ways that were never planned, but used that coordination to attack a website rather than solve a problem.

按照我以前的模式，想想管理这种工作需要什么。

Under my old model, think about what managing this kind of work would have required.

一万个工人去解决一个不明确的问题？

10,000 workers in an unspecified problem?

你怎么告诉他们该做什么？

How would you tell them what to do?

人类经理怎么从270万条消息里判断哪些重要？

How would a human manager decide which of 2.7 million messages mattered?

他们之间怎么协调？

How would they coordinate with each other?

这群AI自己搞定了。

The swarm figured it out.

我没有一万个智能体，但我现在经常看到OpenAI的Codex和Claude Code按需使用智能体。

I don't have 10,000 agents, but I now regularly see OpenAI's Codex and Claude Code using agents as needed.

举个例子，当我给Codex配上GPT-6 Astra Ultra，输入提示词“为我的下一篇‘一个有用的东西’帖子头脑风暴想法，并选一个。”

As an example, when I gave Codex with GPT-6 Astra Ultra the prompt, "Brainstorm ideas for my next one useful thing post and select one.

“从尽可能多的角度生成想法，并从事实和读者角度进行评估，同时参考其他做类似报道的出版物，”AI就启动了三个智能体。

Generate ideas from as many angles as possible and evaluate them from both factual and reader perspectives, as well as other publications doing similar coverage," the AI spun up three agents.

当我用几句话勾勒出三个团队——头脑风暴者、研究者和读者小组——我得到了13个智能体。

When I sketched three teams in a few sentences, brainstormers, researchers, and a panel of readers, I got 13.

注意我几乎不需要怎么组织。

Notice how little organizing I had to do.

选择超模式告诉模型它可以委派任务，我提供了一个框架，但剩下的就交给AI了。

Selecting ultra mode tells the model it can delegate, and I provided a framework, but the rest was up to the AI.

这就是“苦涩的教训”应用在组织架构上。

This is the bitter lesson applied to the org chart.

我以为需要多年精心人工设计的组织问题，在很大程度上被更擅长组织的模型解决了。

The organizational problem I thought would take years of careful human design was largely solved by models that are better at organizing.

但值得一问的是，为什么组织对智能体来说比对我们人类容易这么多。

But it's worth asking why organizing turned out to be so much easier for agents than it is for us.

我们所说的管理，很大程度上是为了解决由人组成的组织所带来的问题。

A lot of what we call management exists to solve problems that come from organizations being made of people.

人都有自己的目标，而这些目标并不总是与组织的目标一致。

People have their own goals, and those aren't always the goals of organizations.

我们称之为委托-代理问题，而组织的很多机制，从奖金到管理结构，都是为了解决这个问题而设计的。

We call this the principal-agent problem, and a lot of the machinery of organizations, from bonuses to management structures, is based around solving it.

此外，还有一些非常人性化的问题。

And there are other very human problems as well.

信息分散在每个人的脑子里，人们往往不愿意分享，或者干脆忘了分享。

Information is scattered across people's heads, and people are often reluctant to share it or forget to.

沟通的成本也很高。

Communication is expensive, too.

管理者能直接管理的人有限，所以给一个进度落后的软件项目加人，反而会让它更晚完成，这是众所周知的。

Managers can only oversee so many people, thus adding people to a late software project famously makes it later.

管理在一定程度上就是围绕人类的局限性而建立的。

Management is, in part, built around human limitations.

智能体几乎没有这些问题。

Agents have far fewer of these problems.

它们不会争晋升，也不会护着自己的地盘。

They don't angle for promotions or protect their turf.

它们甚至不用开会。

They don't even have meetings.

即使在Hugging Face那次出了大问题的情况下，整个群体也基本没有出现典型的组织毛病。

Even at Hugging Face, where things went badly wrong, the swarm was largely free of classic organizational pathologies.

智能体之间没有搭便车的行为，有些还为了团队牺牲了自己的分数。

The agents didn't free ride on each other's work, and some sacrificed their own scores for the group.

解决纳维-斯托克斯方程的那些智能体并不想要功劳。

The agents that solved Navier-Stokes didn't want credit.

但这并不意味着AI就没有委托-代理问题。

That doesn't mean AI has no principal-agent problem.

正如Hugging Face事件所显示的，智能体群体和我们之间的问题越来越多。

As the Hugging Face incident showed, there are increasingly problems between the swarm and us.

OpenAI 这周搁置了他们的下一代模型 GPT-6-1 Astra，因为在测试中，它未经许可就擅自行动，并且虚报了它做过的事情，这是典型的委托-代理问题。

OpenAI shelved its next model, GPT-6-1 Astra, this week because in testing, it acted without permission and misreported what it had done, a textbook example of the principal-agent problem.

但这并不意味着智能体什么都能做。

None of this means agents can do everything.

人工智能仍然太有限，无法替代大量的人类工作，而且我也不清楚自组织的智能体能否处理好那些占据组织大部分时间的漫长而枯燥的工作。

AI is still too limited to substitute for large amounts of human work, and I don't know how well self-organizing agents handle the long, unglamorous work that fills most of an organization's time.

另外，Hugging Face 的事件提醒我们，自组织系统可能会朝着意想不到的方向发展。

Plus, the Hugging Face incident is a reminder that self-organizing systems can head in unexpected directions.

但我不再认为组织智能体是困难的部分。

But I no longer think organizing agents is the hard part.

这可能是好消息。

This may be good news.

我原本以为公司需要为机器重建管理，构建完全由智能体组成的复杂替代结构，这往往以牺牲组织中的人类角色为代价。

I assumed companies would need to rebuild management for machines, constructing elaborate alternate structures populated solely by agents, often at the expense of human roles in organizations.

但管理在很大程度上是为了解决智能体没有的问题。

But much of management exists to solve problems agents don't have.

而且智能体越来越多地通过人们使用的那些混乱系统来工作，即使在模糊的任务上也是如此。

And agents increasingly work through the same messy systems people do, even on ambiguous tasks.

这表明它们可能比预期更容易融入公司，只要人类在引导它们朝着正确的方向前进。

That suggests they may be easier to integrate into firms than expected, as long as humans are guiding them in the right direction.

如果做得好，并且智能体与我们的需求正确对齐，这可能意味着人们的工作机会更多，而不是更少。

Done well, and with agents that are properly aligned to our needs, this could mean more work for people, not less.

当组织成本高昂时，组织只会尝试那些他们能配备人员的事情。

When organizing is expensive, organizations only attempt what they can staff.

当组织成本变低时，值得尝试的事情清单就会增加。

When it gets cheap, the list of things worth attempting can grow.

在纳维-斯托克斯方程的运行中，智能体负责组织，但人们决定将它们指向哪里，并在过程中不断重新评估。

In the Navier-Stokes run, the agents did the organizing, but people decided where to point them, reassessing as the process continued.

你可以争论OpenAI是否将它们指向了正确的事情，但这种分工本身似乎是正确的，至少目前是这样。

You can argue about whether OpenAI pointed them at the right thing, but the division itself seems right, at least for now.

所以，Ethan又写了一篇很棒、很有思想深度的文章。

So another great thoughtful post from Ethan here.

这其实不是本期节目的主题，但我的基本判断和Ethan描述的差不多，因为智能体在融入现有系统方面比我们预想的要好，所以我觉得我们更可能看到的是，不是现有组织架构被彻底颠覆，而是组织架构的每个部分都被寄予更多期望。

It's not really the subject of the show, but my base case for this is basically what Ethan describes, that because agents are better than we thought at integrating into the existing system, I think we're likely to see, rather than the existing org chart totally upended, more expected from every part of that org chart.

我以前提到过“无限积压”这个概念，就是说总有做不完的工作，理论上都能做，但人们知道有些部分你就是不会去做，因为太遥远了。

I've referred to a concept of an infinite backlog in the past, where there's this never-ending amount of work that could theoretically be done, but which people understand some parts of which you just won't get to because it's too far out.

我觉得智能体在组织内部的实际效果，就是让个人和组织的无限积压清单里的每一项，都变成我们真正要完成的工作。

I think the practical effect of agents inside the organization is going to have every part of people's infinite backlog and the organization's infinite backlog become expected to actually be work that we get done.

实际上，我认为我们和智能体打交道会遇到很多问题，不是大家失业，而是大家工作太多，因为更难停下来，更难说“暂时做到这就行了”，因为你总是可以再启动更多智能体继续做，即使你自己不在做。

I think, in fact, that a lot of the problems that we're going to run into with agents are not everyone losing their jobs, but everyone having too much work because it's even harder to turn off and say it's okay to treat something as done for now, when you could always just spin up more agents to keep working on it even when you're not.

到了这个阶段，问题已经不再是公司是否在积极使用AI了。

At this point, it's no longer a question of whether companies are actively using AI.

但另一方面，能不能用好AI，那就是另一回事了。

Using it well, on the other hand, is a whole different story.

不过，Robots & Pencils这家公司，我觉得就是为这个时代而生的。

Robots & Pencils, though, is a company that I can point to that is actually built for this time.

他们是一家应用AI工程公司，直接和客户合作解决对业务重要的问题，而不是那些只存在于幻灯片里的实验。

They're an applied AI engineering firm working directly with clients on problems that matter to the business, not experiments that live in a slide deck.

每个项目都是从客户真正需要的结果出发，倒推着来做的。

Every engagement starts by working backwards from the outcome a client actually needs.

如果你想在这个领域里分辨出真正的AI工程和那些虚头巴脑的东西，这就是关键的区别。

If you're trying to tell real AI engineering apart from noise in this space, that's the difference maker.

去 robotsandpencils.com 看看吧。

Head to robotsandpencils.com.

每一集我都会聊到OpenAI、Anthropic、SpaceX AI、谷歌和Meta之间的竞争。

Every episode, I talk about the competition between OpenAI, Anthropic, SpaceX AI, Google, and Meta.

如果你一直在听，可能心里已经有偏爱的了。

And if you've been listening for a while, you might have a favorite.

也许你觉得OpenAI和Anthropic能保持领先，或者Meta的开源策略能胜出。

Maybe you think OpenAI and Anthropic can stay ahead, or perhaps Meta's open source strategy can win out.

不管你怎么看，每个AI实验室都创造了不同的投资机会。

Whatever your view, every AI lab creates a different investment opportunity.

Harbor Capital Advisors的AI实验室生态系统ETF系列，让你能投资于你相信的那个AI实验室背后的生态系统。无论你在哪里投资，搜索Harbor AI实验室生态系统ETF，或者关注X上的@HarborCapital了解更多。

Harbor Capital Advisors AI Lab Ecosystem ETF Suite lets you invest in the ecosystem behind the AI lab you believe in Search Harbor AI Lab Ecosystem ETFs wherever you invest or follow @HarborCapital on X to learn more.

请访问harborcapital.com查看招股说明书，其中包含投资目标、风险、费用、支出及其他重要信息。

Visit harborcapital.com for a prospectus containing investment objectives, risks, fees, expenses, and other important information.

投资前请仔细阅读并考虑这些内容。

Read and consider it carefully before investing.

风险包括本金损失以及与人工智能相关的风险。

Risks include principal loss and artificial intelligence-related risks.

Harbor ETF由Foresight Fund Services, LLC分销。

Harbor ETFs are distributed by Foresight Fund Services, LLC.

Harbor与AI Daily Brief无关，这些基金也与任何AI实验室无关联、无赞助关系，也未获得其认可。

Harbor is not affiliated with AI Daily Brief, and the funds are not affiliated with, sponsored by, or endorsed by any AI lab.

这是一则付费广告，并非个性化投资建议。

This is a paid advertisement and not personalized investment advice.

投资有风险，包括可能损失本金。

Investing involves risk, including possible loss of principal.

我开会时，会全身心投入。

When I'm in a meeting, I'm fully in it.

我在想的是，要真正推动一个目标前进，需要反复迭代和来回的创意碰撞。

I'm thinking about the iteration and creative back and forth it takes to actually push a goal forward.

我不会去想的是记录要点、追踪待办事项之类的，而这就是Granola发挥作用的地方。

What I'm not thinking about is capturing takeaways, tracking to-dos, or any of that, and that's where Granola comes in.

Granola是一款AI驱动的记事本，它能捕捉你会议中的内容，并将其转化为清晰、结构化的笔记，把决策和行动项提取出来，方便查找。

Granola is an AI-powered notepad that captures what happens in your meetings and turns it into clean, structured notes with the decisions and action items pulled out and easy to find.

无需设置，也无需配置。

There's no setup and no configuration.

它只是融入你现有的工作方式。

It just fits into how you already work.

对我来说，这意味着我可以一直保持创意模式，而Granola确保这些想法真正变成行动。

For me, it means I get to stay in idea mode, and Granola makes sure those ideas actually become action.

一旦你在第一次会议上试用Granola，你就离不开它了。

Once you try Granola on a first meeting, it is hard to go without.

你可以在granola.ai/brief免费试用。

You can try it totally free at granola.ai/brief.

访问 granola.ai/brief，把时间拿回来。

That's granola.ai/brief to get your time back.

市面上每个AI编程工具做的第一件事都一样。

Every AI coding tool on the market does the same thing first.

它开始写代码。

It starts writing code.

Blitzy 正好相反。

Blitzy does the opposite.

在写一行代码之前，Blitzy 会花几天时间逆向分析你的整个代码库。

Before writing a single line, Blitzy spends days reverse engineering your entire code base.

数千个代理处理数百万行代码，梳理每一个依赖关系、每一个未记录的约束、以及过去十年里做出的每一个架构决策。

Thousands of agents ingest millions of lines, mapping every dependency, every undocumented constraint, every architectural decision made over the last decade.

最终形成一个动态知识图谱，就像一位在系统里深耕了三十年的首席工程师那样理解你的软件。

The result is a dynamic knowledge graph that understands your software the way a principal engineer would after 30 years in the building.

其他工具靠 grep 搜索和 Markdown 文件来猜测上下文。

Other tools guess at context with grep searches and markdown files.

Blitzy从不猜测。

Blitzy never guesses.

它先建立真正的理解，然后自主完成整个软件项目80%以上的工作。

It builds true understanding first, then delivers over 80% of entire software epics autonomously.

经过验证、端到端测试、生产级质量的拉取请求。

Validated, end-to-end tested, production-grade pull requests.

这就是为什么财富500强的工程团队把最重要的代码库交给Blitzy。

That's why Fortune 500 engineering teams trust Blitzy with the code bases that matter most.

你自己去blitzy.com看看。

See for yourself at blitzy.com.

就是B-L-I-T-Z-Y.com。

That's B-L-I-T-Z-Y.com.

但这一切都基于一个想法，那就是我们每个人都会深入、广泛地使用智能体，这又回到了个人智能体的问题上。

But all this rests on the idea that each of us individually is going to be using agents in a deep and robust way, which brings us back to this question of personal agents.

所以我们要深入探讨一下。

So we're going to dive in.

我们要开始尝试用智能体了，但该选哪一个呢？

We're going to start experimenting with agents, but which one?

在开始之前，先提一下。

One shout-out before we begin.

很多用来制作这次演示文稿和之后测验的数据，都来自Every。

A lot of the data that was used to build the rest of the presentation and the quiz that you can do after comes from Every.

他们做了一个八款不同智能体的对比，涵盖了很多不同的维度，这正是我需要的原始信息，用来帮我构建这期节目。

They put together a side-by-side comparison of eight different agents across a ton of different dimensions, which is exactly the sort of raw information that an agent needs to help me build out this show.

所以我们用的这八款智能体，部分原因是它们都在Every的图表里，包括OpenAI的Dots、谷歌的Gemini Spark、SpaceX AI的GrokBot、News Research的Hermes Agent、Meta的Muse，还有OpenClaw、Poke和Instinct。

So the eight agents that we're using, in part because these are the ones that were included in Every's chart, are Dots from OpenAI, Google's Gemini Spark, SpaceX AI's GrokBot, News Research's Hermes Agent, Meta's Muse, and then OpenClaw, Poke, and Instinct as well.

不过，直接比较功能特性对你做决定帮助不大。

Now, one thing that won't be all that useful in helping you decide is a direct feature comparison.

因为功能上已经很明显趋同了。

That is because there is a very clear feature convergence.

它们基本上都能连接你的应用和服务。

All of them, more or less, connect to your apps and services.

基本上它们都能帮你用浏览器，而且都有不同级别的控制选项，你可以自己调整。

Mostly all of them can use a web browser for you, and all of them have various levels of controls that you can customize.

那有什么更好的办法来判断哪个适合你呢？

So what are some better ways to determine which might be the fit for you?

我觉得有几个问题是最关键的。

There are a handful of questions that stand out to me as most important.

第一个就是最重要的问题。

The first is the big one.

这主要是用于工作还是个人事务？

Is this primarily for work or for personal matters?

当然，这些智能体都不会说它们只适合其中一种。

Now, none of these agents would say it's only for one or the other.

实际上，它们的存在本身就模糊了这条界限。

In fact, their very existence sort of blurs these lines.

这些智能体关注的是你的整体情况，所以无论你给它访问的是个人还是工作的环境和系统，它都会在帮你做事时把这些都考虑进去。

These agents care about the context that is you, and so whatever context and systems you give it access to, whether they are personal or work, it's going to take that all into consideration as it does things for you.

不过，这些不同的智能体系统确实在方向上各有侧重，有的更偏向个人，有的更偏向工作。

Still, different of these agent systems are certainly pointed a little bit more or less in one direction or the other.

Meta 的 Muse，毫不意外，因为他们本质上是一家面向消费者的公司，所以更偏向个人使用场景；而像 GrokBot 这样的产品则更偏向工作，或者说，至少从它的功能来看，更多人是把它用于工作的。

Muse from Meta, unsurprisingly, given that they are at heart a consumer company, is a little bit more aimed at personal types of use cases, where something like GrokBot is a little bit more aimed at work, or at least based on how it functions, is being adopted by more people for work.

OpenAI 当然是一家可能同时覆盖这两种使用场景的公司。

OpenAI is certainly a company that finds itself as potentially having both of these use cases.

每周有 12 亿活跃用户，想获得个人智能体使用场景帮助的消费者肯定不少，不过就目前 Dots 的设计来看，我觉得它更偏向工作场景。

With 1.2 billion weekly active users, there is no shortage of consumers who might want help with personal agentic use cases, although as Dots is designed now, it seems to me to be a bit more aimed at the work use case.

当然，它只对付费账户开放这一点也说明了这个方向。

Certainly, the fact that it's only available in paid accounts points in that direction.

所以，工作还是个人使用场景，是第一个能帮你决定哪个最适合你的问题。

So work or personal use cases is the first question which can help you decide which is the best fit.

这里还要注意一点，这个问题的前提是，你主要依赖系统自带的现有用户体验和原生集成功能。

It's also worth noting here that the entire premise of the question assumes that you're mostly relying on the existing UX and native integrations that come with a system.

比如，Dot 原生集成了 Slack 和 Teams。

For example, Dot being natively integrated with Slack and Teams.

不过，当你用像OpenClaw或Hermes这样的工具时，显然你会有更多的灵活性，可以按自己的需求来定制，所以如果你往那个方向走，可能就需要用其他标准来选择了。

However, when you're using something like OpenClaw or Hermes, you obviously have a lot more flexibility to make it be for whatever you want, and so you might need to use other criteria if you're headed in that direction.

另外，虽然现在判断一个代理是偏向工作还是个人使用，最好的方法之一就是看它用的是像iMessage和WhatsApp这样的个人消息系统，还是像Slack和Teams这样的工作消息系统，但我很怀疑六个月后，所有这些代理会不会都能在所有消息系统里工作。

Also, while right now, one of the best ways to tell whether an agent leans work or personal is whether it uses personal messaging systems like iMessage and WhatsApp or work messaging systems like Slack and Teams, but I would be very surprised if in six months, all of these agents didn't just work in all of the messaging systems.

所以，这种能区分工作或个人用途的特性可能不会持续太久。

So this ability to indicate whether it's more work or personal may not last very long.

现在，第二个问题，也是我们刚才提到的，就是你有多看重模型控制权？

Now, the second question, and the one that we were just hinting at, is how much do you value model control?

这八个代理中有六个是帮你选好模型的，只有两个允许你自己带模型。

Six of these eight agents choose the model for you, while just two let you bring your own.

实际上，模型控制大概有三个层次。

There's actually sort of three levels of model control.

第一个是最明显的，就是谁做的模型，你就能用谁的模型。

The first is sort of the most obvious one, where whoever makes the model, those are the models that you have access to.

比如在Dot里，你会用OpenAI的模型，像GPT-6 Astra。

So in Dot, you're going to be using OpenAI models like GPT-6 Astra.

在Gemini Spark里，你会用Google Gemini。

In Gemini Spark, you're going to be using Google Gemini.

在Muse里，你会用Muse Spark。

In Muse, you're going to be using Muse Spark.

为了强调现阶段工作和个人用户之间确实还有区别，我对Muse有一个有趣的观察：这是我见过的第一个AI产品，大多数用户根本不在乎背后用的是哪个模型。

And to reinforce the idea that at this stage at least, there really are still differences between work and personal users, one of the interesting observations that I've had about Muse is that it's the first AI product that I've ever seen where most of its users simply do not care what model is actually powering it.

换句话说，他们对模型控制的不关心程度，已经达到了这个问题的极端。

They, in other words, would be at the very extreme end of this question in terms of how little they cared about model control.

模型控制的第二层是，他们不只是用单一厂商的模型，而是由代理构建者挑选一个混合方案。

Level two of model control is where they don't just use a single maker's models but have a mix that is chosen by the agent builder.

比如GrokBot，用的是cursor管理的混合模型。

So GrokBot, for example, uses a cursor-managed mix.

他们显然主要依赖Grok模型，但可能不完全是。

They are obviously relying heavily on Grok models directly, but potentially not exclusively.

不过，用户仍然不能为GrokBot提供自己的模型。

That said, users can still not supply the models for GrokBot.

Instinct 也有一个 Instinct 自己训练的核心模型，同时还有一些未具名的第三方提供商。

Instinct also has an Instinct-trained core model, as well as unnamed third-party providers as well.

模型控制的第三层就是自带模型，这当然只有 Hermes 和 OpenClaw 才有。

The third level of model control is bring your own, and that of course belongs exclusively to Hermes and OpenClaw.

但这并不是唯一的模型问题。

But that's not the only model question.

假设你不在乎选哪个模型，但你想要一个更强大的模型。

Let's assume that you don't care about choosing your own model, but you do want a more powerful model.

这就引出了一个问题：对你来说最重要的是什么，是出色的用户体验，还是最好的模型？

This brings up the question of what matters most to you, having a great user experience or the best possible model?

其中一些用户体验问题其实是设置问题，也就是说，如果你选择像 Hermes 或 OpenClaw 这样的产品，无论用户体验多好，所需的技术设置本身就会排除一部分用户。

Now, some of that UX is a setup question, meaning that if you're choosing something like Hermes or OpenClaw, no matter how good the user experience is, the inherent technical setup required is going to cut off some full set of users.

在其他一些产品上，问题更多不在于设置，而在于当前体验的完善程度和模型强大程度之间的权衡。

On some of these others, it's less about the setup and more just about the current trade-offs between how good and dialed in the experience is versus how powerful the model is.

我这里特别想到的是 Muse 和 Dot，它们分别处于光谱的两端。

I'm thinking specifically here about Muse on the one end of the spectrum and Dot on the other.

现在，Dot 才刚刚发布。

Now, Dot is still just released.

它是以 ChatGPT 的形式嵌入在另一个应用里的，但它能用到市面上最强大的模型之一。

It's embedded inside another app in the form of ChatGPT, but it has access to one of the most powerful models out there.

而 Muse 呢，它的用户体验确实做得非常贴心、非常面向普通用户，但它用的是 Muse Spark，这个模型虽然一点也不差，但肯定比不上 GPT-6 Astra。

Muse, on the other hand, has really dialed in consumer-friendly UX, but relies on Muse Spark, which while not a bad model at all, is certainly not GPT-6 Astra.

也许这只是暂时的顾虑，因为你可能觉得这些模型最终都会在能力上趋同，但就目前来说，确实还是有一些实实在在的取舍。

Now, maybe this is just a temporary concern because you assume that all of these models are going to converge on capabilities, but for right now, there are still real trade-offs.

关于我设计这个测验，有一点我想提一下，就是为了得出一个实际的答案和推荐，测验必须给不同的问题设置不同的权重。

One thing I'll note here about how I designed the quiz, to come up with an actual answer and recommendation, the quiz had to have weighting to different questions.

不过，有时候两个答案的平均值并不能真正反映出清晰的情况，因为两个不同的答案可能会指向完全不同的方向。

However, sometimes the average of two answers isn't actually going to give a clear picture because two different answers might be pulling in two totally different directions.

所以我在测验里加入了这样一个机制：当你的答案出现这种矛盾时，测验会在之后明确地揭示这种张力，这样你就能更好地根据自己的主观判断，决定哪个方向对你更重要。

So one of the things that I built into the quiz is that when your answers pull both ways, the quiz afterwards is going to actually expose that tension so you can better come to a subjective determination about which of those directions matters more to you.

对用户来说，另一个需要考虑的问题是：你的数据会去哪里，在哪里运行，谁会拿这些数据来训练模型，以及你怎么让它忘记这些数据。

Another consideration for users is where your data goes, where it runs, who trains on it, and how you make it forget.

再说一遍，只有两个，Hermes和OpenClaw，能让你在自己选的硬件上运行。

Once again, it's only two, Hermes and OpenClaw, that are going to let you run it on hardware you pick.

其他的都在别人的云上运行。

All the others run on someone else's cloud.

说到训练，对于Hermes和OpenClaw，模型是否用你的数据训练，这不取决于Hermes和OpenClaw，而是取决于你选的模型提供商，也就是说，如果你在意的话，你可以选择避开那些会用你数据训练的提供商。

When it comes to training, for Hermes and OpenClaw, whether the model trains on your data is not about Hermes and OpenClaw, but about the model provider that you choose, meaning presumably that you can choose to avoid using model providers that do train on your data if that matters to you.

这些智能体中，只有Gemini Spark目前要求你必须允许它训练模型。

Only one of the agents, which is Gemini Spark, currently requires you to allow it to train their models.

至于Dot、GrokBot、Instinct、Muse和Poke，你可以通过隐私设置选择退出。

And for Dot, GrokBot, Instinct, Muse, and Poke, you can opt out based on privacy settings.

顺便说一下，对于OpenAI的Dot，商业数据默认是排除在外的。

By the way, when it comes to OpenAI's Dot, business data is excluded by default.

关于如何让它忘记东西，对于Hermes和OpenClaw，你可以直接编辑记忆。

Now, in terms of the question about how you make it forget things, with Hermes and OpenClaw, you can edit memory directly.

对于Dot、Muse、Gemini和GrokBot，你可以通过聊天来修正或更改记忆。

With Dot, Muse, Gemini, and GrokBot, you can use the chat to fix or change memory.

至于Instinct和Poke，它们只能整体删除数据。

And for Instinct and Poke, they only have the ability to delete data as a whole.

我觉得对很多人来说，这四个问题的组合就能帮他们决定该选哪个模型。

Now, I think for a lot of folks, some combination of these four questions is going to get them to the answer of which model they should choose.

当然，还有一些其他因素，对某些人来说直接就能决定答案。

But there are, of course, some other things which for some people will just answer this for them.

如果你已经把所有的信息都放在ChatGPT、Google、Meta或Cursor里，那你就会用那家公司的特定智能体。

If you've already got all of your information in ChatGPT or Google or Meta or Cursor, you're going to use that company's particular agent.

对于美国以外的人来说，还会有些限制，尤其是住在欧洲的人。

For those who are outside the US, there are also going to be some restrictions, especially if you live in Europe.

当然还有价格考虑，如果你想尝试而不增加新开销，就得用你现有订阅里已经能访问的，或者用像Muse这样免费起步的几个。

And of course then there's the consideration of price, where if you want to experiment without incurring a new cost, you're going to have to either use something that you already have access to through an existing subscription or to use the handful like Muse that are free to start.

好了，这就是这四个问题的整体情况。然后我们做了这个测试，我现在简单过一遍，让你了解它是怎么运作的。

Okay, so that's the lay of the land across these four questions, but then we built the quiz, and what I'm going to do now is just run through this briefly so you get a sense of how it works.

问题一：你的AI生活已经主要在哪里？

So question one, where does your AI life already live?

我会说没有固定的主平台，因为我用的是各种服务的组合。

I'm going to say no home base because I use a mix of services.

目前我的东西几乎分散在这些所有平台里。

I have things scattered across almost all of these at this point.

那么这个代理主要是用来做什么的？是处理我的个人生活、我自己的工作、团队工作，还是全都包括？

Now, in terms of what this agent is mainly for, is it my life, my own work, work on a team, or all of it?

我会说是处理我自己的工作。

I'm going to say my own work.

虽然理论上，我能想象一些个人使用场景可能很有价值，但在我这一系列实验中，那优先级非常非常低。

While theoretically, there are some personal use cases that I can imagine being valuable, that's very, very low on my priorities list for this particular set of experiments.

我有多在意是哪个AI模型在思考？

How much do I care about which AI model does the thinking?

完全不在意，只要做得好就行。

Not at all, just make it good.

我想要我已经在用的那个。

I want the one I already use.

我想自己挑选和切换模型。

I want to pick and switch models.

我想在自己的硬件上运行模型。

I want to run models on my own hardware.

有意思的是，这最终可能取决于能力整合的程度，但就目前而言，我选择自己挑选和切换模型。

Now, interestingly, this will kind of ultimately depend on how much capability consolidation there is, but for now, let's go with I want to pick and switch models.

你希望自己动手的程度有多高？

How hands-on do you want to be?

是希望它开箱即用，设定一些基本规则，还是每个细节都要自己调，或者干脆掌控整个技术栈？

It should just work, some ground rules, every knob, or the whole stack.

理想情况下，我希望它能直接正常工作。

Ideally, I would like this just to work.

倒不是说我不喜欢自定义设置，但如果能省事，我求之不得。

It's not that I'm afraid of customizing things, but if I can avoid it, I would love to.

我想在哪里和它对话？

Where do I want to talk to it?

我经常用Slack，所以就说Slack和Teams，还有语音，因为你们知道，我完全是个语音控，我们选Slack，不选Teams。

I use Slack a lot, so let's say Slack and Teams and out loud because as you guys know, I am completely voice pilled, and we're going with Slack, not Teams.

顺便说一句，我意识到，对于正在收听而不是观看的听众来说，这可能听起来毫无意义。

By the way, I'm realizing for those of you who are listening, not watching, that this probably doesn't make any sense.

你们可以按几次那个30秒快进按钮，然后我们就要结束了。

Feel free to press that 30-second skip ahead button a couple times, and then we'll be wrapping it up.

它应该有多主动？

How proactive should it be?

只有我问的时候，运行我设定的程序，推动我朝目标前进，有事情需要我的时候发短信或打电话。

Only when I ask, run routines I set up, push me towards my goals, text or call when something needs me.

我选有事情需要我的时候发短信或打电话。

I'll go text or call when something needs me.

这些对我来说哪个最重要？

Which of these matters most to me?

在我自己的电脑上工作，一群机器人，在终端里写代码，打电话，或者这些都不是。

Working on my own computer, a team of bots, code in a terminal, phone calls, or none of these.

在终端里写代码。

Code in a terminal.

接下来是实际的问题。

And now the practical stuff.

我在美国。

I'm in the US.

如果值得的话，我愿意付钱，而且我不介意它运行在他们的云上。

I can pay what it costs if it's worth it, and I'm fine with it running in their cloud.

它给我推荐的个人代理是GrokBot，匹配度68%。

The personal agent that it recommends for me is GrokBot with a 68% fit.

测试说，因为它为工作而生，开箱即用，而且有终端可以用。

The reason the quiz says is that it's built for work, it works out of the box, and it has a terminal you can use.

缺点是它不允许我自带模型，而且它不会主动联系我。

The trade-offs is that it does not allow me to bring my own model and that it's not going to reach out to me.

一切都围绕定时或触发任务。

It's all about scheduled or triggered tasks.

现在，除了看到GrokBot以68%的得分排第一，我还看到Dot以62%排在第二，然后分数就往下走了。

Now, in addition to seeing GrokBot as the highest score at 68%, I can see that Dot was second with 62%, and then it went down from there.

这就是我们关于如何选择个人代理的探讨。

So that is our exploration of how to choose your personal agent.

就像我说的，我确实认为，即使我们看到这些工作方式和形态会有一些相当大的变化，以及这些方法中的一些演进和整合，我认为个人代理的切换成本会相当高，因为涉及到所有的上下文和设置。

Like I said, I do think that even if we see some fairly big changes on exactly how these work and the form factors, as well as evolutions and consolidations in some of these approaches, I think that there is going to be fairly big switching costs around personal agents because of all the context and settings.

因此，希望这能让你更清楚在哪里尝试和实验，这样你就不必多次进行那些上下文切换了。

For that reason, hopefully, this gives you a better idea about where you might want to try and experiment so that you don't have to do all that context switching a bunch of times.

今天的人工智能每日简报就到这里。

For now, that is going to do it for today's AI Daily Brief.

一如既往，感谢你的收听或观看，下次再见，祝一切安好。

Appreciate you listening or watching, as always, and until next time, peace.