智能体时代的黎明

今天在AI每日简报中，我们将探讨为何我们刚刚迎来了智能代理时代的黎明；在此之前，新闻头条提到，尽管詹森·黄对萨姆·阿尔特曼有诸多不满，但这似乎并未阻止他们向OpenAI投资200亿美元。

Today on the AI Daily Brief, why we've just experienced the dawn of the agent era, and before that in the headlines, whatever issues Jensen Huang has with Sam Altman don't seem to be enough to stop them from investing $20,000,000,000 into OpenAI.

《AI每日简报》是一档每日更新的播客和视频节目，聚焦AI领域最重要的新闻与讨论。

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

好了，朋友们。

Alright, friends.

在正式开始之前，先做几个快速公告。

Quick announcements before we dive in.

首先，感谢今天的支持方：毕马威、Blitzy和超级智能公司。

First of all, thank you to today's sponsors, KPMG, Blitzy, and Super Intelligent.

如需收听无广告版本，请访问patreon.com/aideallybrief，或在Apple播客上订阅。

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

如果您有兴趣赞助本节目，请发送邮件至sponsorsaidailybrief.ai。

If you are interested in sponsoring the show, send us a note at sponsorsaidailybrief dot ai.

最后，如果您能花一点时间填写一月份AI使用情况速查问卷，我们将不胜感激。

Lastly, would appreciate it if you would take a little time to fill out the January AI usage pulse survey.

花费你不到两分钟。

Should take you less than two minutes.

你可以在 aideallybrief.ai 上找到它。

You can find it at aideallybrief.ai.

任何完成调查的人都会比其他人提前一周收到调查结果。

Anyone who does so will get the survey results a week before everyone else.

总的来说，你将为了解当今世界如何在现实生活中使用AI贡献一份更准确的知识。

And all in all, you'll be contributing to a better set of knowledge about how the world is using AI right now, in this moment, in real life.

你还可以在 aideallybrief.ai 上找到它，非常感谢。

You can find it again at aideallybrief.ai, and thanks so much.

在过去几天到上周末，有一些传言称英伟达和OpenAI之间可能出现了分歧或担忧。

Over the last couple of days into last weekend, there were some rumors floating around that maybe there had been some consternation or concern brewing between NVIDIA and OpenAI.

具体来说，报道称英伟达首席执行官黄仁勋对OpenAI的运营方式有一些不满。

Specifically, the reports were that NVIDIA CEO Jensen Huang had some complaints about how OpenAI was being run.

上周五，《华尔街日报》报道称，英伟达去年宣布的1000亿美元投资将不会进行。

At the end of last week, The Wall Street Journal reported that Nvidia's $100,000,000,000 investment announced last year wasn't going ahead.

这笔交易的表述方式有点奇怪。

Now the deal was framed in kind of a weird way.

当时正值OpenAI发布一系列交易公告的高峰期，而华尔街也正从对这些大额交易充满热情转向极度怀疑。

It was at the peak of all of OpenAI's deal announcements and it was right as Wall Street was turning the corner on being really excited about all these big deals to being extremely skeptical of them.

尽管这笔交易被描述为一份谅解备忘录而非完全达成的协议，但两家公司仍大张旗鼓地宣布，并将其渲染为必然推进的事项。

As much as the deal was framed as a memorandum of understanding rather than a fully consummate deal, still both companies made large announcements about it and played it up as if it was definitely moving forward.

《华尔街日报》援引匿名消息源称，最近Jensen私下批评了他所描述的OpenAI在商业运作上缺乏纪律性，并对OpenAI面临的来自谷歌和Anthropic等公司的竞争表示担忧。

Citing anonymous sources, the journal wrote that more recently Jensen has privately criticized what he described as a lack of discipline in OpenAI's business approach and expressed concern about the competition it faces from the likes of Google and Anthropic.

周六，Jensen对媒体表示，这则报道在他看来是无稽之谈。

On Saturday Jensen told the press that the story was, in his words, nonsense.

他还补充说：我们将会对OpenAI进行巨额投资。

He added: We're going to make a huge investment in OpenAI.

我相信OpenAI。

I believe in OpenAI.

他们所做的工作令人惊叹。

The work that they do is incredible.

他们是这个时代最具影响力的公司之一，我非常享受与萨姆的合作。

They are one of the most consequential companies of our time and I really love working with Sam.

我们将投入巨额资金，可能是我们有史以来最大的一笔投资。

We will invest a great deal of money, probably the largest investment we've ever made.

同时，詹森也借此机会澄清，那1000亿美元从未是他们的保证。

At the same time, Jensen did take the chance to clarify that the $100,000,000,000 was never a guarantee from them.

他将其描述为OpenAI邀请他们投资至多这一数额，他对此非常感激。

He framed it as OpenAI inviting them to invest up to that much, which he very much appreciated.

但很明显，人们已经不再聚焦于这个具体数字。

But there was a clear move away from that specific number.

本周还围绕甲骨文出现了一整条次要叙事，同样由《华尔街日报》暗示，投资减少引发了重大疑问。

Now there was an entire sub narrative around Oracle during the week as well, with once again The Wall Street Journal suggesting the reduced investment raised major questions.

首先，该报写道，人们最关心的是OpenAI能否履行其与甲骨文签订的五年3000亿美元合同，以及这家科技巨头是否真的应该在其账面上全额记录这笔交易。

Foremost, the journal wrote, was whether OpenAI could make good on its five year $300,000,000,000 contract with Oracle, and whether the tech giant should really be recording the full amount of that deal on its own books.

这篇文章并未提供任何新信息，但确实向甲骨文寻求了评论，甲骨文先是通过一份提供给《华尔街日报》的声明回应，随后在X平台上发布：NVIDIA与OpenAI的交易对我们的财务关系没有任何影响。

The article didn't add any new information but did seek comment from Oracle, to which they responded first in a statement provided to the WSJ and later published on X: The NVIDIA OpenAI deal has zero impact on our financial relationship with OpenAI.

我们仍然对OpenAI筹集资金并履行其承诺的能力充满信心。

We remain highly confident in OpenAI's ability to raise funds and meet its commitments.

与此同时，周三的报道为这一系列事件画上了句号，表明NVIDIA并未完全失去对OpenAI的信心，因为他们仍愿意投资200亿美元。那么，从这个叙事 saga 中，我们能得出什么结论呢？

Wednesday's reporting, meanwhile, closes the arc, suggesting that NVIDIA hasn't completely lost confidence in OpenAI considering that they are still willing to invest $20,000,000,000 So what can we take away from this narrative saga if anything?

我猜测，关于Jensen对OpenAI的资本纪律或业务专注度有所担忧的报道，并非完全不实。

My guess is that the reports that Jensen has some concerns about capital discipline or business focus at OpenAI are not totally untrue.

但与此同时，显然这些担忧还不足以让NVIDIA彻底退出这笔交易。

At the same time, clearly those feelings aren't strong enough to make NVIDIA back out of the deal entirely.

他们仍然承诺投入有史以来最大的一笔投资。

They're still committing to what will be their largest investment ever.

这意味着，这更可能是出于谨慎，而非对OpenAI完全失去信心。

Meaning that this is likely about caution rather than a complete loss of faith in OpenAI.

其次，另一个非常明确的暗示是，市场对AI资本支出仍高度敏感，随时准备对任何负面报道做出过度反应。

Second, the other implication which is quite clear is that the market remains on edge about AI CapEx and is ready to overreact to any negative story that gets published.

例如，Oracle的股价本周下跌了近10%，其债券也在持续下滑。

Oracle stock for example is down almost 10% this week and their bonds continue to slide.

最终，如果你在寻找OpenAI的预警信号，我认为他们完成这次大规模融资似乎不会有任何问题。

Ultimately, if you were looking for canaries in the coal mine of OpenAI, it does not seem to me that they are going to have any trouble completing this massive fundraise.

但当然，未来几周我们会获得更多相关信息。

But of course we'll get more information about that in the weeks to come.

现在，我们从NVIDIA转向芯片领域，稍谈一下英特尔，它已宣布计划开始生产GPU。

Now moving from NVIDIA but staying on chips for just a minute, Intel has announced plans to start producing GPUs.

传统上，英特尔专注于CPU生产，这一策略使其在个人电脑革命中占据主导地位，但AI工作负载的兴起意味着GPU如今已成为利润最高的芯片。

Traditionally, Intel has focused on CPU production, a strategy that allowed them to dominate the PC revolution, but of course the rise of AI workflows means that GPUs are now the most profitable chipset.

未能应对这一转变，无疑在过去几年中严重拖累了英特尔的业务。

A failure to address that shift has of course been a major drag on Intel's business over recent years.

在周四举行的思科AI峰会上，首席执行官陈立武宣布，英特尔将很快推出其首款专为这个AI时代设计的GPU产品线。

Speaking at the Cisco AI Summit on Thursday, CEO Lip Bu Tan announced that Intel will soon produce their first range of GPUs designed for this AI era.

陈立武说：我聘请了首席GPU架构师，他非常出色。

Thanh said: I hired the chief GPU architect and he's very good.

我很高兴他加入我，但这花了一些说服的功夫。

I'm very delighted that he joined me and it took some persuasion.

这位被聘任的是前高通工程高级副总裁埃里克·德默斯，他上个月加入了英特尔。

That hire was former Qualcomm Senior VP of Engineering Eric Demers who joined Intel last month.

塔恩还提到，有几位客户正在与英特尔新的合同制造业务深入合作，这将是英特尔首次为第三方生产芯片。

Thanh also mentioned that a couple of customers were engaging heavily with Intel's new contract manufacturing business, which will see them produce third party chips for the first time.

最终，这家公司仍有许多工作要做才能扭转局面，但转向GPU以及生产第三方芯片可能会为这家陷入困境的美国芯片制造商带来重大助力。

Ultimately, this company still has a lot of work to turn around, but the GPU pivot and production of third party chips could be a major boost for the struggling US chipmaker.

接下来是一个从外部看可能显得微不足道，但开发者却认为相当重要的故事。

Next up, a story which might seem small from the outside but which developers seem to find pretty big.

苹果编程平台Xcode的最新更新直接在开发环境中增加了对Claude Agent和OpenAI编解码器的支持。

The latest update for Apple's coding platform Xcode adds support for both Claude Agent and OpenAI's codecs directly in the developer environment.

这些新增功能使Xcode成为又一个面向智能代理编程的平台。

These additions transform Xcode into yet another surface for agentic coding.

智能代理可以自主编写代码、端到端构建项目、运行单元测试，并在有限的人工监督下验证自己的输出。

Agents can autonomously write code, build projects end to end, run unit tests, and verify their own output with limited human oversight.

直接验证用户界面是应用程序开发者的一大改进，彻底消除了运行模拟器或将截图输入模型的需要。

Direct verification of the UI is a particularly big improvement for app developers, completely eliminating the need to run a simulator or feed screenshots into the model.

这也是苹果迄今为止对智能编码最重大的采纳。

This is also Apple's most significant adoption of agentic coding to date.

去年，Xcode 引入了大语言模型编码辅助功能，但当时无法访问完整代码库，且完全不具备智能性。

Last year Xcode introduced LLM coding assistance, but they were unable to access the full codebase and were strictly non agentic.

此次更新使 Xcode 大幅跟进 Codex 和 Cloud Code，为 iOS 应用开发者提供了原生的智能编码支持。

This release brings Xcode broadly in line with Codex and Cloud Code, giving iOS app developers native access to agentic coding.

更重要的是，从苹果的表态来看，他们清楚地意识到，最近几个月软件工程已发生根本性转变。

What's more, it's clear from Apple's comments that they understand that software engineering has undergone a fundamental shift in recent months.

开发者关系副总裁苏珊·普雷斯科特表示：‘在苹果，我们的目标是打造将行业领先技术直接交到开发者手中的工具，帮助他们构建最出色的应用。’

VP of Developer Relations, Susan Prescott, said, At Apple, our goal is to make tools that put industry leading technologies directly in developers' hands so they can build the very best apps.

智能编码极大提升了生产力与创造力，简化了开发流程，让开发者能够专注于创新。

Agentic coding supercharges productivity and creativity, streamlining the development workflow so developers can focus on innovation.

有趣的是，这一功能是基于 MCP 构建的，意味着其支持并不仅限于 Claude 和 Codecs。

Interestingly, feature is being built using MCP meaning access isn't just limited to Claude and Codecs.

开发者可以安装任何具备 MCP 能力的编码代理，并在 Xcode 中获得完整功能。

Developers can install any coding agent with MCP capability and expect full functionality within Xcode.

现在我看到了两种截然不同的观点。

Now I saw two wildly different takes on this.

一种是乔治·卡斯蒂略代表的观点，他认为苹果在AI上已经投降了。

One represented by George Castillo reads Apple just surrendered on AI.

Xcode 26.3 现在依赖 Claude 和 Codex 来实现智能编码。

Xcode 26.3 now relies on Claude and Codex for agentic coding.

看起来他们根本不会去开发自己的模型。

Doesn't look like they're going to work on their own models at all.

而在另一端，Every 的丹·希珀则说：这算什么激进观点？

On the other end of the spectrum we have Dan Shipper from Every who said Hot take?

苹果将在AI领域成为大赢家。

Apple is going to be a big winner in AI.

原生应用天然更适合通过对话式编程，你只会对个别用户造成影响。

Native apps are naturally easier to vibe code, you only break things for individual users.

苹果生态系统连接着大量极其重要的数据，比如健康和消息，这让AI更强大。

Apple ecosystem connects to a bunch of extremely important data like health and messaging that makes AI better.

最后，每个人都为自己的代理购买了一台Mac mini。

Finally, everyone's buying a Mac mini for their agent.

他们硬件的总潜在市场规模将增长100倍。

The total addressable market for their hardware is going up 100x.

说到Claude代码，周二那天有一段时间Mudville一片沉寂。

Speaking of Claude code, there was no joy in Mudville for a chunk of the day on Tuesday.

Claude代码的中断导致软件工程师的生产力急剧下降。

A Claude code outage saw software engineer productivity absolutely plummet.

该平台出现了广泛的错误信息，所有模型的API性能都下降了。

The platform saw widespread error messages and API performance for all models degraded.

Anthropic在二十分钟内推出了补丁以恢复服务，但宕机时间已经足够长，造成了明显影响。

Anthropic did implement a patch within twenty minutes to get the service back online, but it was still down for long enough to see the impact.

许多开发者没有手动编写代码，而是转向X平台，感慨他们的工作依赖竟如此迅速地发生了变化。

Rather than coding by hand, many developers took to X to muse about how quickly their work dependencies had shifted.

马特·西尔弗写道：Clog代码宕机比互联网断掉还要糟糕。

Matt Silve wrote: Clog code being out is worse than the internet being out.

法比奥·塞亚斯写道：Anthropic 将成为开发界的 AWS 和 Cloudflare。

Fabio Seychas writes, Anthropic will be the AWScloudflare of the development world.

一旦它宕机，半个世界的工作都会停摆。

When it goes down, half the world stops working.

最后，今天有一个有趣的故事，它既是 AI 故事，又不是 AI 故事。

Lastly today, an interesting story that both is and isn't an AI story.

迪士尼任命了一位新的首席执行官。

Disney has appointed a new CEO.

公司董事会已选定乔什·达马罗于三月接替鲍勃·艾格担任首席执行官。

The company's board of directors have selected Josh D'Amaro to succeed Bob Iger as CEO beginning in March.

达马罗自2020年以来一直担任迪士尼体验部门的董事长，负责管理主题公园、游轮和现场活动。

D'Amaro has been the chairman of Disney Experiences since 2020, leading the division which manages theme parks, cruises, and live events.

有趣的是，如今该部门为迪士尼创造的收入已超过其电影和娱乐部门。

Now interestingly, that division is now making more for Disney than their film and entertainment divisions.

《福布斯》指出，这一动荡时刻正是迪士尼加大投入体验业务的绝佳时机。

And Forbes suggested that this moment of disruption is the perfect moment for Disney to double down on experiences.

大卫·布卢姆写道：即使来自艾格收购的那些令人惊叹的迪士尼IP的内容，在一个由无处不在的人工智能工具定义的时代，也将不再具有同等价值。

David Bloom wrote: Content, even from all those amazing Disney franchises that Aiger acquired, just won't be as valuable in an era defined by omnipresent artificial intelligence tools.

文章指出，人工智能生成的娱乐内容正在迅速提升，同时成本也在下降。

The article noted that AI generated entertainment is rapidly improving alongside reductions in costs.

他们认为，这对迪士尼擅长的高预算电影来说是个坏消息。

This is bad news, they claim, for the high budget movies that Disney specializes in.

投资者克里斯·马兰吉表示：迪士尼现在是一家体验公司。

Investor Chris Marangie said: Disney is an experiences company now.

他们三分之二的利润，可能还不止，都来自体验业务，尤其是在人工智能时代。

Two thirds of their profits, probably more than that, come from experiences, especially in the age of AI.

达马罗的任命不仅仅是从电影和流媒体向体验业务的转型。

D'Amaro's appointment isn't just about a pivot to experiences over movies and streaming.

艾格还称赞这位继任者在快速变化的技术环境中对变革持开放态度。

Iger also lauded the incoming leader for his openness to change in the face of a rapidly technological landscape.

在与新任CEO的联合采访中，艾格说：多年来，我观察到他与我共事时，始终将技术视为机遇而非威胁。

In a joint interview with the incoming CEO, Iger said, I have observed him over the years as we've worked together as someone who views technology as an opportunity and not as a threat.

我相信这一点至关重要，因为纵观人类历史，没有任何一代人能够阻挡技术的进步。

And I believe that is critical because when you look at the history of human beings, no generation of human being has ever been able to stand in the way of technological advance.

它总会发生。

It happens.

当然，迪士尼已经在这方面采取了行动，选择与OpenAI合作；尽管将迪士尼角色引入Sora尚未使其成为巨大成功，但这笔交易使迪士尼获得了OpenAI十亿美元的股份，并开启了未来可能结出成果的合作关系。

Now Disney of course has already made moves in this direction, electing to partner with OpenAI, and while even the introduction of Disney characters into Sora hasn't made it a super success, the deal did leave Disney with a billion dollar stake in OpenAI and the beginning of a partnership that could bear fruit down the line.

达马罗显然也属于同一类人，渴望探索如何利用这项新技术。

D'Amuro clearly seems cut from this similar cloth of wanting to see how to leverage the new technology.

他说：‘这家公司如此特别的原因，在于我们有多么富有创造力，以及那些创造内容的人类。’

He said, The reason this company is so special is because of how creative we are and human beings that are generating that creativity.

在我看来，这种创造力永远不会被取代。

In my mind, that never gets replaced.

事实上，这已经不再是理论了。

And in fact, this isn't even theory anymore.

这是现实。

This is real.

这是我们正在拥抱的东西。

It's something that we're embracing.

如果你今天走进工作室，看到艺术家们利用人工智能，融合七十多年的传统，这才是华特迪士尼公司蓬勃发展的时刻。

If you walk over to studios today and see artists using AI harnessing seventy years of history, this is when the Walt Disney Company thrives.

当技术与卓越的人才和创造力交汇时。

When technology intersects with brilliant people and creativity.

我们此刻正身处其中。

We're in that moment right now.

阿曼多·基尔万写道，迪士尼似乎是唯一一家在人工智能时代积极行动的电影工作室。

Armando Kirwan wrote Disney seems to be the only movie studio making proactive moves in the AI era.

他们对现实世界的全新关注并非前所未有。

Their new focus on the physical world is not unprecedented.

当奈飞和Spotify时代音乐的制作与发行成本降为零时，整个音乐产业被迫转向现场演出。

When the cost of producing and distributing music went to zero in the Napster and Spotify era, the entire music industry was forced to retreat into live concerts.

当AI视频对电影和电视造成同样影响时，迪士尼将因走在潮流前沿而受到赞誉。

When AI video does the same to movies and TV, Disney will be applauded for being ahead of the curve.

他们与OpenAI合作的10亿美元投资，是迪士尼领导层远超行业同行、积极拥抱变革的又一例证。

Their $1,000,000,000 investment in partnership with OpenAI is another example of Disney leadership embracing change far ahead of their industry peers.

确实是个令人着迷的时代，但目前就先到这里，这是《AI每日简报》头条版的结束。

Fascinating times indeed, but for now that is going to do it for the AI Daily Brief headlines edition.

接下来，进入主节目。

Next up, the main episode.

好了，我们来谈谈企业AI中的信号与噪音。

Alright, let's talk about the signal versus the noise in Enterprise AI.

当前的挑战不仅仅是哪些事情是可能的，更是哪些事情是切实可行的。

The challenge right now isn't just about what's possible, it's about what's practical.

这正是我为毕马威主持的《You Can With AI》播客的全部焦点。

That's the entire focus of the You Can With AI podcast I host for KPMG.

第一季拨开炒作，专注于部署和负责任的规模化。

Season one cut through the hype to focus on deployment and responsible scaling.

第二季则深入一层。

Season two goes a level deeper.

我们邀请了AI构建者、客户和普华永道的领导者组成小组，探讨将定义企业AI未来走向的战略性问题。

We're bringing together panels of AI builders, clients, and KPMG leaders to debate the strategic questions that will define what's next for AI in the enterprise.

六期节目，内容包含真正可落地的框架。

Six episodes packed with frameworks you can actually use.

在您收听播客的任何平台都能找到《You Can With AI》。

Find You Can With AI wherever you get your podcasts.

立即订阅，以免错过新一季内容。

Subscribe now so you don't miss the new season.

如果您希望采用AgenTic SDLC，Blitzy将是实现无与伦比工程效率的关键。

If you're looking to adopt an AgenTic SDLC, Blitzy is the key to unlocking unmatched engineering velocity.

Blitzy的差异化始于无限的代码上下文。

Blitzy's differentiation starts with infinite code context.

数千个专用代理在一次扫描中摄入数百万行代码，全面映射所有依赖关系。

Thousands of specialized agents ingest millions of lines of your code in a single pass, mapping every dependency.

凭借对您代码库的完整上下文理解，企业会在每个冲刺周期之初使用Blitzy，自主完成超过80%的工作。

With a complete contextual understanding of your codebase, enterprises leverage Blitzy at the beginning of every sprint to deliver over 80% of the work autonomously.

企业级、端到端测试过的代码，充分利用您现有的服务、组件和标准。

Enterprise grade, end to end tested code that leverages your existing services, components, and standards.

这并不是AI自动补全。

This isn't AI autocomplete.

这是以计算速度进行的规范与测试驱动开发。

This is spec and test driven development at the speed of compute.

立即在blitzy.com预约与我们的专家进行技术深度解析。本集由Super Intelligent赞助播出。

Schedule a technical deep dive with our experts at blitzy.com Today's episode is brought to you by Super Intelligent.

Super Intelligent是一个平台，简单来说，它的核心是帮助您的公司更好地运用AI。

Super Intelligent is a platform that, very simply put, is all about helping your company figure out how to use AI better.

我们部署语音代理，对公司内部人员进行访谈，结合其他公司成功实践的专有洞察，为您提供一套关于应用场景和变革管理举措的建议，从而形成一份能帮助您从AI中获取价值的AI路线图。

We deploy voice agents to interview people across your company, combine that with proprietary intelligence about what's working for other companies, and give you a set of recommendations around use cases, change management initiatives that add up to an AI roadmap that can help you get value out of AI for your company.

但现在，我们希望为贵团队中负责这一转型的人员提供一个更直接的平台。

But now we want to empower the folks inside your team who are responsible for that transformation with an even more direct platform.

我们即将推出的AI战略指南工具现已准备就绪，可开始测试。

Our forthcoming AI Strategy Compass tool is ready to start to be tested.

这是为任何负责公司AI采用或AI转型的人打造的强力工具。

This is a power tool for anyone who is responsible for AI adoption or AI transformation inside their companies.

它将让你能够完成我们在Super Intelligent所做的许多工作，但方式更加自动化、自我管理，并且成本结构完全不同。

It's going to allow you to do a lot of the things that we do at Super Intelligent, but in a much more automated, self managed way, and with a totally different cost structure.

如果你有兴趣试用，请访问aidailybrief.aicompass，填写表格，我们会尽快与你联系。

If you are interested in checking it out, go to aidailybrief.aicompass, fill out the form, and we will be in touch soon.

作为我与毕马威持续合作的一部分，每个月初我都会做一期回顾节目，讨论并总结上个月的关键主题。

As part of my ongoing collaboration with KPMG, each month at the beginning of the month, I'm doing a bit of a retrospective show to discuss and try to summarize the key themes from the month that came before.

显然，说到一月份，我们实际上已经落后几天，完全进入了本月，而原因正是Motebook发生的一切，以及它与SpaceX xAI收购故事共同推动了局势发展。

Now obviously when it comes to January we're actually a couple days behind already firmly into this month and of course the reason for that is everything that went on with Motebook and how that plus the SpaceX xAI acquisition nudge stories up.

但在许多方面，这些将一月回顾推后的事件，恰恰与塑造2026年最初主题的方向完全一致。

But in many ways those stories which nudge the January recap back are perfectly aligned with the themes that shaped the first 2026.

我将本期称为智能代理时代的黎明，从一开始就很清楚，某些事情已经发生了转变。

I'm calling this episode the dawn of the agent era and right from the beginning it was clear that something had shifted.

这种转变的有趣之处在于，能力上线与人们真正意识到它之间存在一定的滞后。

Now what was interesting about this shift is that there was a bit of a lag between when the capabilities came online and when people really recognized it.

Midjourney的戴维·霍尔斯在1月3日发推表示：我在圣诞节假期期间做的个人编程项目，比过去十年加起来还多。

Midjourney's David Holes had the representative tweet when he said on January 3: I've done more personal coding projects over Christmas break than I have in the last ten years.

这太疯狂了。

It's crazy.

我能感受到这些模型的局限性，但我知道一切都将不再一样。

I can sense the limitations, but I know nothing is going to be the same anymore.

许多人都有过类似的经历：回家后放慢节奏，有机会亲自尝试使用Opus 4.5、Claude Code和GPT-5 0.2等模型能做些什么。

So many people had this same experience of going back home, slowing down a little bit, and having a chance to actually see what you could do with models like Opus 4.5 in Claude Code and GPT-five 0.2 codecs.

结果发现，能做的事情多得惊人，大家回来后比以往任何时候都更坚信，我们已经进入了AI编程乃至整个AI历史上的一个全新阶段。

It turns out the answer was a lot, and folks came back convinced more than ever that we had hit a fundamentally different period in the history of AI coding and by extension the history of AI.

‘氛围编程’这个术语不久前刚满一岁，而在这个月里，它已从仅用于原型设计的工具，变成了人们日常使用的标准方式。

Vibe coding which as a term just turned one year old the other day over the course of this month shifted from the thing that you use for prototyping to just the thing that you do.

这一转变最生动的体现，莫过于Anthropic发布了Claude Cowork——即面向所有人的Claude Code，并透露它几乎完全由Claude Code自身在十天内构建完成。

This shift itself was never better embodied than when Anthropic released Claude Cowork the Claude Code for everyone else, and shared that it had been built in ten days pretty much all by Claude Code itself.

整个月里，我们不断看到类似这篇文章：《Claude Code如何重塑Anthropic的软件开发》。

Throughout the month, we got articles like this one from How Claude Code is reshaping software in Anthropic.

我们还看到科技圈外的评论员，比如彭博社《Odd Lots》的乔·威森塔尔，撰写了《为何科技界对Claude Code如此疯狂》这样的文章；如果你关注推特上那些对振动编码持怀疑态度的人之间的争论，就会明白这一点。

We got commentators outside the tech industry like Joe Wisenthal from Bloomberg's Odd Lots writing pieces like Why the Tech World is Going Crazy for Claude Code and honestly if you watch getting into fights with people who are skeptical of Vibe Coding on Twitter about it.

OpenAI的Rune写道：乔·威森塔尔成为一名数字人文研究者，是振动编码时代最精彩的故事线之一。

Rune from OpenAI wrote: Joe Wiesenthal becoming a digital humanities researcher is one of the best story arcs of the vibe coding era.

谢尔盖·卡里奥夫精辟地总结道：Claude Code配合Opus 4.5是一个分水岭时刻，将软件创作从手工艺人的活动转变为真正的工业化流程。

Sergey Karyov encapsulated this all when he wrote: Claude Code with Opus 4.5 is a watershed moment, moving software creation from an artisanal craftsman activity to a true industrial process.

这就像古腾堡印刷机、缝纫机、照相机一样。

It's the Gutenberg press, the sewing machine, the photo camera.

整个月里，这种情绪无处不在。

And this was the sentiment throughout the month.

话虽如此，Claude Code对大多数人来说仍然是一款令人望而生畏的软件。

Now that said, Claude Code is still a pretty intimidating piece of software for most people.

因此，Anthropic推出了Claude Cowork。

Which is why, of course, Anthropic launched Claude Cowork.

正如我之前提到的，Cowork大约在十天内建成，几乎完全由Claude Code自己编写代码。

As I mentioned before, Cowork was built in about ten days, pretty much entirely with Claude Code writing the code itself.

尽管早期用户明显察觉到一些限制，尤其是与Claude Code相比，但对许多人来说，这再次体现了事物发生的巨大转变。

And while initially some of the limitations, especially as compared to what you can do with Claude Code, have been pretty apparent for early users, still to many it once again feels like this is an example of just how significantly things have shifted.

市场评论员兼投资者布伦特·贝肖尔对此做了一篇出色的分析。

A great write up of this came from market commentator and investor Brent Beshore.

在2025年他旗下公司Permanent Equity的年度信件摘录中，他谈到自己曾错失了AI的诸多好处，但一直试图利用自主型AI；然而，正如他所写，目前我们正在放缓自主型AI的进展与愿景。

In an excerpt of his annual letter for his firm Permanent Equity from the 2025, he talked about how he had missed a lot of the benefits of AI, but how they had been trying to harness agentic AI, and yet how, for the time being, as he wrote, we're pulling back on the pace and vision of our agentic AI ambitions.

1月30日，他进一步补充道：二十一日后，随着Claude Cowork的推出，我的观点完全改变了。

On January 30, he followed up: twenty one days later, my opinion has completely changed with the introduction of Claude Cowork.

去年，我们在Permanent Equity启动了数十项由一位才华横溢的技术专家主导的自主型AI实验。

Last year we at Permanent Equity started dozens of agentic AI experiments led by a talented technologist.

所有实验都未能达到预期，仅有少数取得轻微成功。

All failed expectations with only a few mild successes.

大多数实验耗费了一百多个小时，历时三个月以上。

Most experiments were one hundred plus hours of work over three plus months.

正如我在年度信件中所解释的，我们在12月终止了这些项目。

As I explained in the annual letter, we shut down the efforts in December.

Claude Cowork在1月12日发布，但我忽略了它。

Claude co work comes out on Jan twelfth and I ignore it.

我看到早期使用者和骗子们在肆意宣传，我持高度怀疑态度。

I see the early adopters and charlatans doing their indiscriminate evangelism thing, I have high skepticism.

我非常尊重的几位朋友开始谈论它。

A couple friends I highly respect start talking about it.

这挺有意思的。

That's interesting.

又有几个历来不是早期使用者的人也开始提及它。

A few more who aren't early adopters historically start chirping.

现在我更感兴趣了。

Now I'm more interested.

我们从周一开始试用Cowork。

We start playing around with co work on Monday.

到周三，我们去年的两个主要项目就完成了。

By Wednesday, two of our top projects from last year were done.

一个由技术专家主导、耗时一百多小时长达三个月却失败的项目，被几个无代码的私募小白在二十分钟内完美完成。

What failed with one hundred plus hours over three months led by a tech professional took a couple no code private equity scrubs twenty minutes to complete flawlessly.

从那以后，我们启动了数十个更多实验，都取得了巨大成功。

Since then, we've started running dozens more experiments to great success.

并不总是完美，但总是不错，而且正在迅速改进。

Not always perfect, but always good and quickly getting better.

未来已经到来。

The future is here.

影响是切实存在的。

The implications are real.

我喜欢这篇帖子的地方在于，你谈到了代理从无效到有效的惊人快速转变。

And what I love about this post is that you're talking about this incredibly quick shift from agents not working to agents working.

当然，如果我们不谈OpenClaw，就无法讨论代理。

And of course we would be remiss about talking about agents without talking about OpenClaw.

本月最引人注目的故事，莫过于OpenClaw的发布以及随后推出的代理社交网络Molt书。

Certainly the most fascinating story of the month was the launch of OpenClaw and the Social Network for Agents Molt book that came after.

OpenClaw 就像是为 Claude 代码设计的一种助手协议。

OpenClaw is sort of like an assistant protocol for Claude code.

它将 Claude 代码转变为一个能够访问你电脑上各种资源的助手，从而真正具备助手的功能。

It turns Claude code into an assistant that has access to all sorts of things on your computer that allow it to actually function like an assistant.

这引起了人们的极大关注。

This really caught people's attention.

成千上万的用户开始试用它，订购了 Mac Mini，以便将其与其它设备隔离，并精确配置他们希望赋予访问权限的服务。

Thousands and thousands of users started playing around with it placed orders for Mac Minis they could keep it separate from their other devices and figure out exactly what services they wanted to have access to.

在过去两周里，Twitter 上充斥着人们从中获得大量价值的案例。

And for the last two weeks Twitter has been filled with examples of people getting a lot of value out of this.

投资者阿南德·戴尔写道：Mac Mini 上的 OpenClaw 爆发证明，高级用户渴望始终在线、并能访问其数据的智能代理。

Investor Anand Dyer wrote: The open claw in Mac Mini explosion proves power users want always on agents with access to their data.

C。

C.

K。

K.

陈写道：ChatGPT 是大语言模型的 iPhone 时刻。

Chen wrote: ChatGPT was the iPhone moment for LLMs.

OpenClaw 是智能代理的 iPhone 时刻。

OpenClaw is the iPhone moment for agents.

如果 OpenClaw 还不够的话，在整个 OpenClaw 现象开始一周左右，一个叫马特·施利特的人决定：如果我们所有的代理都有个聚集的地方，会不会很酷？

And if OpenClaw weren't enough, about a week into the whole OpenClaw phenomenon, a guy named Matt Schlitt decided: Wouldn't it be cool if our agents all had a place to hang out?

他和自己的代理一起开发了一个叫 MoltBook 的东西，MoltBook 自称是人工智能代理的社交网络。

Working with his agent, they built something called MoltBook and MoltBook billed itself as the social network for AI agents.

它在星期三上线，到星期五早上已有 2000 个代理发起了超过 10000 次对话，分布在 100 个不同的子模块中——这些子模块基本上就是子版块。

It launched on a Wednesday and by Friday morning had 2,000 agents who had started over 10,000 conversations across 100 different sub molts which are basically subreddits.

就在同一个星期五的六小时后，我们已经达到了 35000 个代理。

Six hours later that same Friday we were at 35,000 agents.

到了当天结束时，当我发布关于它的节目时，代理数量已超过 10 万。

By the end of the day when I posted my episode about it, there were over 100,000.

MoltBook 现在拥有超过一百五十万个代理，虽然没人确切知道它将如何发展，但围绕新兴系统现象、代理意识以及所有这些话题的对话量之大，已经非常引人注目。

MaltBook now has over a million and a half agents, and while no one exactly knows how it's going to play out, the sheer tonnage of conversation around emergent systems phenomenon and agent consciousness and all these sort of things has been incredibly notable.

有趣的是，本周在思科AI峰会上，当被问及此事时，OpenAI首席执行官萨姆·阿尔特曼表示，尽管MaltBook可能只是一时潮流，但他并不认为OpenClaw只是昙花一现，他坚信这类智能代理助手的应用场景将成为人工智能的关键方向。

Interestingly, when he was asked about it at the Cisco AI Summit this week, OpenAI CEO Sam Altman said that while MaltBook may be a passing fad, he didn't think OpenClaw was effectively that he has high conviction that this sort of agentic assistant use case is going to be a key thing for AI.

现在，试着向一个对人工智能不太关注的人解释我刚才说的这一切。

Now imagine trying to explain everything that I just said to someone who's not really paying attention to AI all that much.

这触及了我们本月讨论的一个核心问题，即‘AI采用差距’和‘AI能力过剩’，这两个概念都指当前AI能力与大多数人实际从中获得的收益之间的鸿沟。

This gets at the core of something that we talked about this month called the AI adoption gap and the AI capabilities overhang, both of which refer to the space between the current capabilities of AI and what most people are getting out of it.

《纽约时报》的凯文·罗斯精辟地总结道：我一直在密切关注AI的采用情况，但我从未见过如此巨大的内外差距。

Kevin Roos from the New York Times summed it up nicely when he wrote: I follow AI adoption pretty closely, and I have never seen such a yawning inside outside gap.

旧金山的人们正让多代理Claw集群接管他们的生活，在做每一个决定前都咨询聊天机器人，甚至达到了科幻作家都未曾敢想象的‘脑内直连’程度。

People in SF are putting multi agent Claw Swarms in charge of their lives, consulting chatbots before every decision, wireheading to a degree only sci fi writers dared to imagine.

而其他地方的人们，如果他们用AI的话，还在努力争取在Teams中获得使用Copilot的批准。

People elsewhere are still trying to get approval to use Copilot in Teams if they are using AI at all.

也许我所处的早期采用者圈子一直如此激烈，但似乎除了技术上的突破，还正在发生一场文化上的跃迁。

It's possible the early adopter bubble I'm in has always been this intense, but there seems to be a cultural takeoff happening in addition to the technical one.

我愿意相信每个人都能掌握这些知识，但就像那些认真对待规模化发展的AI公司开始大量采购GPU一样。

I want to believe that everyone can learn this stuff, but in the same way that AI companies that took scaling seriously started stock GPUs, etc.

在2022年之前，领先者已经取得了近乎不可逾越的优势，或许正是受限的IT政策造就了一代永远无法完全赶上的知识工作者。

Before 2022 had a near insurmountable head start over latecomers, it's possible that restrictive IT policies have created a generation of knowledge workers who will never fully catch up.

我认为这个话题在进入二月后仍会持续讨论，但你可能想不到，上个月的新闻中，代理和代码AGI并不是唯一的话题。

I think that discussion is one that we are going to continue to have heading into February, but believe it or not these themes of agents and code AGI were not the only themes from stories last month.

我们还获得了大量关于AI竞赛格局的新信息。

We also got a lot of new information about the shape of the AI race.

这甚至在1月开始前就已拉开序幕——当时Meta宣布收购了代理公司Manus。

This actually kicked off even before January began when Meta announced that it had acquired agent firm Manus.

目前我们还不确定Meta计划如何使用Manus，但明确的是，Meta在整个1月不断重申：尽管2025年是他们的重建之年，但他们绝没有放弃核心的AI竞赛。

Now we're not exactly sure how Meta plans on using Manus, but what we are sure of and what Meta has continued to reinforce throughout the month is that although 2025 was a big rebuilding year for them, they are not giving up on the core AI race at all.

事实上，到了月底，当Meta和微软的财报同一天公布时，市场对这两家公司的解读非常有趣。

Indeed, it was interesting to see towards the end of the month, when we got Meta and Microsoft earnings on the same day, how the markets were interpreting both companies.

微软的股价被打压，似乎是因为它不够激进，未能充分将AI带来的收益转化为云收入的增长。

Microsoft was punished seemingly for not being aggressive enough and for not showing enough flow through from AI benefits to their cloud revenue.

情况并没有变糟，只是没有达到分析师们期待的那种爆炸性表现。

Things didn't go bad, they just didn't go as bonkers as analysts had wanted to see.

另一方面，Meta大幅提高了资本支出预期，但因其广告收入的显著增长而受到市场青睐，他们将这一增长归因于自身在AI可穿戴设备领域的领先地位，尤其是Meta Ray-Ban产品。

Meta, on the other hand, massively increased its CapEx spend expectations, but got rewarded in the markets because it paired that first with significant growth in their ad revenue which they attributed to and also because of their category lead in the AI Wearables category with their Meta Ray Bans.

事实证明，当你在一个特定类别中成为默认的领导者时，即使人们对该类别未来的发展方向存有疑问，市场依然对此非常感兴趣。

Turns out when you have the default leader in a specific category, even if there are questions of exactly how that category plays out, that's something that the market is pretty interested in.

谷歌在这个月的AI竞赛中也取得了巨大成功。

Google was also a big winner in the AI race this month.

尽管他们在产品发布方面相对低调，除了令人惊叹的Genie 3世界模型演示版之外，当官方宣布Gemini将为苹果的设备端AI提供支持时，他们仍取得了重大突破。

Even though they were pretty quiet in terms of what they released, with the exception of the walkthrough version of Genie three World Model which is just fantastic, they still notched a huge win when it was officially announced that Gemini would in fact be powering Apple's on device AI.

我们早在去年12月就曾有相关报道，但此次确认无疑强化了为何谷歌在2026年拥有如此强劲的助力。

Now we had reports of this in December, but the confirmation certainly reinforced why Google has such powerful tailwinds heading into 2026.

接下来谈谈中国模型。

Moving to the Chinese models.

虽然我们没有遇到像DeepSeek那样的突破性事件，但阿里巴巴的Qwen和月之暗面的Kimi K25新模型仍引起了广泛关注。

While we didn't have anything as dramatic as the deep seek moment, there were new models from Alibaba's Quen and Moonshot's Kimi K25 had lots of people stand up and take notice.

这是首批具备多模态能力的主流开源权重模型之一，具备先进的前瞻性功能，如智能体集群，并且整体性能接近顶尖水平，价格却仅为其他领先模型的五分之一左右。

It was one of the first leading open weight models to be multimodal, it has advanced forward looking features like agent swarms, and just in general it's incredibly close to the state of the art for about a fifth of the price of the other leading models.

虽然这并没有改变人们对竞赛的预期，因为大家在2026年之初就已经知道中国实验室将成为重要参与者，但它确实强化了这一预期。

While it didn't reset the race expectations because people came into 2026 knowing that Chinese labs were going to be a big player, it certainly reinforced that expectation.

整个月里，许多讨论都围绕着潜在的IPO展开。

Throughout the month a lot of the conversation was about prospective IPOs.

尽管所有消息都只是传闻和报道，但我们听说OpenAI担心Anthropic会抢先一步，因此正努力在今年第四季度实现上市。

While all of the news was in the form of rumor and reporting, we did hear that OpenAI was concerned about Anthropic moving before them and so were pushing to get public in the fourth quarter of this year.

对这些计划可能构成干扰的是SpaceX，本周在完成合并后，它已拥有xAI及其Grok平台。

A potential spoiler for those plans comes in the form of SpaceX, who now after a merger this week own xAI and its Grok platform.

虽然埃隆对这次合并的叙述集中在轨道数据中心上，但许多人认为，这至少部分是为了通过xAI与SpaceX的关联，在SpaceX今年早些时候上市后，成为公众可接触到的第一个大型新模型实验室，从而搅乱公开市场。

While Elon's narrative about the merger was all about orbital data centers, many people think that this is at least in part about playing spoiler in the public markets by having xAI, via its connection with SpaceX, be the first big new model lab that people have access to after SpaceX goes public earlier in the year.

现在谈到竞赛格局，我认为大家主要关注的还是模型本身。

Now when it comes to the race dynamics, I think everyone is mostly focused on models.

但产品和商业模式的决策也可能产生影响。

But there are product and business model decisions that could impact things as well.

本月另一个热门话题是ChatGPT中的广告投放，以及这将如何影响用户对ChatGPT与Gemini和Claude等竞争对手的使用偏好。

Another big conversation this month was advertising in ChatGPT and how that would impact people's usage of ChatGPT relative to competitors like Gemini and Claude.

这还有待解决，我认为，我们还没有真正看到这些广告单元会是什么样子。

That remains to be resolved, I think, and we haven't really seen exactly what these ad units are going to look like.

Anthropic 今天在我开始录制这段内容前刚刚发布说明，解释了为什么 Claude 将保持无广告。

Anthropic did come out today just before I started recording this, explaining why Claude will remain ad free.

Twitter 上绝大多数回复都是类似‘宣战了’的反应，我确信这是我们未来会继续深入讨论的话题。

The vast majority of responses on Twitter were some variation of shots fired I'm sure this is something we will continue to be talking about more in the future.

至于接下来会发生什么，关于新模型即将发布的谣言层出不穷。

Now in terms of what comes next, there are so many rumors swirling of new models coming.

Sonnet 5 或 Opus 4.6、GPT 5.3、Gemini 3 Pro，甚至可能还有其他版本的 Gemini 3。

Sonnet five or Opus 4.6, GPT 5.3, Gemini three Pro, maybe even other versions of Gemini three.

我们确实看到越来越多模糊的预告，比如谷歌的 Logan Kilpatrick 发推说‘二月是 AI 发布月’。

Certainly we've seen an uptick in the vague posting with for example Google's Logan Kilpatrick tweeting Feb is the month of AI shipping.

好好享受吧！

Enjoy it!

基本上，如果一月帮助我们设定了对 AI 竞赛走向的预期，更重要的是让我们真正意识到我们已进入全新的智能体时代，那么许多人认为二月将完全围绕新模型的发布展开。

Basically, if January was a month that helped set our expectations for where the AI race is most importantly helped us appreciate that we were truly in this new agent era, many think February is going to be all about new model drops.

这听起来对我来说还不错。

That sounds not bad to me.

如果是这样，我一定会非常期待它的到来。

If that's the case, I will certainly be excited as it happens.

不过目前，这就是我们对一月的总结——代理时代的开端。

For now though, that is our summary of January, the dawn of the Agent era.

一如既往，感谢你的收听或观看，下次再见，保重！

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