克劳德代码终结了AI泡沫

今天在AI每日简报中，我们聊聊Claude Code是如何终结AI泡沫的。

Today on the AI Daily Brief, how Claude Code killed the AI Bubble.

《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，先做个快速公告。

Quick announcements before we dive in.

首先，感谢今天的支持方：Assembly、Robots and Pencils、Blitzy 和 Super Intelligent。

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

要获取无广告版本的节目，请访问 patreon.com/aideallybrief，或在Apple播客订阅。

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

无广告版本每月仅需3美元。

Ad free is just $3 a month.

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

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

最后，当你访问aideallybrief.ai时，可以了解我们正在进行的所有项目，包括我在录制这段内容时还未上线的一个项目，但我明确承诺将在周日前发布。

And finally, while you are at aideallybrief.ai, can find out about all the various projects that we have going on, including one which as I am recording this, I have not pushed live yet, but which I am clearly committing to by Sunday.

这是对我们AI新年自主学习计划的后续，它将是一个新项目，旨在匹配OpenAI在3月31日前实现以智能体为核心的内部目标。

It's the follow-up to our AI New Year self directed learning program, and it's going to be a new program to match OpenAI's internal objective of agent first work by March 31.

最稳妥的方式是前往aideallybrief.ai查找链接，但我预计它也会出现在march31.ai或aidbtraining.com上。

The safest thing is to go to aideallybrief.ai to look for the link, but I assume it will also be on march31.ai or aidbtraining.com.

说完这个公告，我们进入今天的节目内容。

With that announcement out of the way, let's move on to today's episode.

这是一期周末特别节目，正如你们所知，这是长文阅读或深度思考类的节目。

So this is a weekend episode, which as you guys know is a long reads and or big think episode.

最近出现了一个非常有趣的话题，我觉得它极其迷人，完美地总结并收束了我们自2026年以来所讨论的所有内容。

And there is a really interesting theme that has taken hold that I think is so fascinating and a perfect encapsulation and capstone to everything we've been talking about throughout 2026 so far.

当然，周四我们短短二十分钟内接连发布了两个前沿模型：Anthropic的Opus 4.6和OpenAI的ChatGPT 5.3 Codex。

On Thursday, of course, we got two Frontier models within twenty minutes of each other: Anthropic's Opus 4.6 and OpenAI's ChatGPT 5.3 Codex.

这件事让很多人产生了共鸣。

Something about this clicked for people.

著名思想家泰勒·科文写道：今天将被记作某种转折点，虽然这种说法有些武断，但记者和历史学家如此呈现也无妨。

Prominent thinker Tyler Cowen wrote: Today will go down as some kind of turning point somewhat arbitrarily, but it is okay if journalists and historians have to present things in that manner.

内森·杨写道：如果你在旧金山街头漫步，是否感觉像新冠疫情初期那样，每个人心里想的都一目了然？

Nathan Young wrote, If you're walking around SF, does it feel like the early days of COVID where it's clear what's on everyone's mind?

韦恩在推特上说：有谁能告诉我，过去48小时究竟发生了什么具体的事，能解释为什么我看到了57,246条类似这种模糊的帖子？

Wayne on Twitter said, Can someone explain to me what concrete thing happened in the last forty eight hours that explains the fact that I've seen 57,246 vague posts like this one?

安迪·马斯利写道：我知道人人都说现在感觉像2020年2月，但确实感觉就像2020年2月。

Andy Massley wrote, I know everyone's saying it's feeling a lot like February 2020, but it is feeling a lot like February 2020.

那么，到底发生了什么？

So what is going on?

投资者王 Chow 简明扼要地表达了看法。

Investor Chow Wang put it simply.

他写道：我认为，AI的泡沫远没有我两个月前以为的那么严重。

He wrote: I think AI is much less of a bubble than I thought two months ago.

而且我认识的几乎所有在过去两个月使用过Claude和Codex的人，都有同样的感受。

And pretty much everyone I know who used Claude and Codex in the last two months feels that way.

简而言之，我们在2026年迄今为止所经历的，是一系列连锁性的认知转变。

In short, what we've experienced so far in 2026 is a set of cascading recognitions.

正如我们反复讨论过的，即便是最精通技术的AI用户，也需要在假期回家后，有时间和空间去真正理解像Opus 4.5和Codex 5.2这样的模型能力究竟有多大的不同。

As we've discussed ad nauseam, it took even the most enfranchised and technical AI users going home over the holidays and having some time and space to really understand just how different the capabilities of the models including Opus 4.5 and Codex 5.2 really were.

Cloud Code 当然成为了整合这些工具、从而改变你所能实现之事的封装工具。

Cloud Code, of course, became the harness encapsulation of using those things to transform what you can do.

当人们回来后，他们开始谈论自己在过去两周内提交的代码量，超过了之前一整年的总和。

When people came back, they started talking about how they had pushed more code in the last two weeks than they had done in the year before.

你开始看到叙事的转变，就连那些曾经认为‘氛围编程’只适用于原型设计的人，现在也开始意识到‘智能体编程’几乎适用于一切。

You started to see a shift in the narrative, where even the folks who had previously said Vibe coding is just for prototyping were now recognizing that agentic coding was kind of for everything.

Claude Cowork 发布了，其背后的团队透露，他们仅用十天就完成了开发，而且几乎完全使用Claude Code编写。

Claude Cowork came out, and the team behind it revealed that they had put it together in ten days, and it was basically exclusively coded using Claude Code.

Claude Cowork 作为这个故事中的一个转折点非常关键，因为它在月中发布，也正是从那时起，主流群体也开始关注这一趋势。

Now Claude Cowork was interesting as an inflection point in the story because it came out around the middle of the month, and that's when the mainstream started picking up on this story as well.

这不仅仅是因为他们正在使用Claude CoWork，尽管许多人确实如此。

It wasn't just that they were using Claude CoWork, although many of them were.

有些人甚至开始使用Claude Code，尽管它在技术上更具挑战性。

Some were even finding their way into Claude Code even though it's technically more challenging.

你开始看到商业和金融媒体（而非科技媒体）出现分析文章，讨论Claude Code的智能代理能力有何不同。

You started to see think pieces show up in business and finance publications away from technology about how different the agentic capability set was with Claude Code.

你也开始看到它对市场产生了影响。

And you started to see it have a market impact as well.

新的担忧不再是关于AI泡沫，而是有些人称之为'SaaSpocalypse'的现象：软件类股票出现广泛下跌，而其他类型的技术股票则未受影响，因为这些智能编码工具的兴起让人们开始质疑这些SaaS公司的价值和持久性。

The new concern started to be less about an AI bubble and more about what some dubbed the SaaSpocalypse: a broad based plunge specifically in software but not other types of technology stocks where the rise of these agentic coding tools had people really questioning how valuable and how durable the positioning of those SaaS companies was.

四点六和五点三版本的Codecs就是在这样的环境中推出的。

That is the environment into which four point six and five point three Codecs came.

而Semi Analysis正是在这样的背景下写下了他们最近的文章《Claude Code是转折点》。

And that's the environment in which Semi Analysis wrote their recent post, Claude Code is the inflection point.

因此，本集的长文部分将节选Semi Analysis团队的精彩内容，而非全文朗读，开头是一个相当深刻的统计数据。

So this will be the long reads portion of this episode, and we'll read not the whole thing, but a number of excerpts from the great team at Semi Analysis that starts with a fairly profound stat.

ClaudeCode于2025年3月发布（请注意，当时仅为研究预览版，仅在安德烈·卡帕西提出'vibe coding'概念约一个月后），如今已占GitHub公开提交代码的4%。

ClaudeCode, which was released less than a year ago in March 2025 (as a research preview mind you, just about one month after Andre Karpathy coined the term vibe coding) now represents 4% of GitHub public commits.

你可以从这张图表中看到，这种趋势正在加速。

And you can see in this chart that this is accelerating.

从十月开始，出现了病毒式增长。

There started to be viral growth around October.

到了一月，情况真正开始升温。

Then at the January, things really started to heat up.

这在一定程度上是因为Cloud Code的创作者Boris在Twitter上现身，并开始分享他如何使用它。

It came in part around Boris, the creator of Cloud Code, introducing himself on Twitter and starting to talk about how he used it.

但显然，这个月发生了许多事情，极大地提升了人们对Cloud Code的参与度。

But obviously, there has been a lot going on this month that has significantly increased the engagement with Cloud Code.

OpenCLaw Motebook。

OpenCLaw Motebook.

这一直是2026年至今的主旋律。

This has been the story of 2026 so far.

Semi Analysis的Dylan Patel继续说道：按照目前的趋势，我们相信到2026年，Claude Code将占所有每日提交量的20%以上。

Semi Analysis's Dylan Patel continues: At the current trajectory, we believe that Claude code will be 20% plus of all daily commits by the 2026.

你一眨眼的功夫，AI就已经吞噬了整个软件开发领域。

While you blinked, AI consumed all of software development.

让我们继续深入探讨更大的部分。

Let's continue on into the larger piece.

Semi Analysis 团队写道：我们认为 Claude Code 是 AI 代理的转折点，也是窥见 AI 未来运作方式的一扇窗口。

The semianalysis team writes: We believe that Claude code is the inflection point for AI agents and is a glimpse into the future of how AI will function.

它有望在 2026 年为 Anthropic 带来惊人的收入增长，使该实验室能够大幅超越 OpenAI。

It's set to drive exceptional revenue growth for Anthropic in 2026, enabling the lab to dramatically outgrow OpenAI.

他们认为，Anthropic 有望在接下来三年内增加与 OpenAI 相当的算力。

Anthropic, they argue, is on track to add as much power as OpenAI in the next three years.

随后，他们分享了一份逐栋追踪 Anthropic 和 OpenAI 的数据，并指出 Sam 的 AI 实验室明显因多个数据中心延误而受阻。

They then share a building by building tracker of Anthropic and OpenAI, and write Sam's AI Lab is notably suffering from multiple data center delays.

由于更多的算力意味着更多的收入，我们可以预测 ARR 增长，并直接比较 Anthropic 和 OpenAI。

And since more compute means more revenue, we could forecast ARR growth and compare Anthropic to OpenAI directly.

值得注意的是，他们继续写道：我们的预测显示，Anthropic 的季度 ARR 增量已经超过了 OpenAI。

Notably, they continue: Our forecast shows that Anthropic's quarterly ARR additions have overtaken OpenAI's.

Anthropic 每月新增的收入已经超过 OpenAI。

Anthropic is adding more revenue every month than OpenAI.

我们相信 Anthropic 的增长将受到算力的限制。

We believe Anthropic's growth will be constrained by compute.

下一节他们称之为 Claude Code 与智能体的未来。

The next section they call Claude Code and the Agentic Future.

他们写道，智能体将成为有机智能与人工智能交互的主要方式。

Agents they write will be the primary method of how organic intelligence interacts with artificial intelligence.

但 Claude Code 也展示了相反的一面，展现了智能体如何与人类互动。

But Claude Code is also a demonstration of the reverse, showing how agents interact with humans.

我们相信，人工智能的未来将在于令牌的编排，而不仅仅是按基础成本出售令牌。

We believe the future of AI will be about the orchestration of tokens, not just selling tokens at base cost.

以历史为鉴，我们认为 OpenAI 的 ChatGPT API 就像 Web 1.0 时代的请求与响应，TCP/IP 将用户连接到托管在互联网上的静态网站。

With history as a guide, we view the OpenAI ChatGPT API as the call and response of tokens akin to Web one point zero with TCPIP connecting users to static websites hosted on the Internet.

尽管 TCP/IP 是一项基础技术，但在 Web 2.0 时代转向动态网页的过程中，这一通信协议仅成为实现互联网功能的手段。

While TCPIP is a foundational technology, this communication protocol became just the means to the end of enabling the Internet during Web two point zero and the shift to dynamic web pages.

如今，互联网使用TCP/IP数据包来组织比静态网站大得多的信息集。

Today the Internet uses TCPIP packets to organize much larger sets of information than a static website.

协议很重要，但真正创造数万亿美元价值的是建立在这一协议之上的应用。

The protocol matters, but it was the applications built on top of this protocol that created trillions in value.

这就是Semi Analysis认为我们再次处于人工智能关键时刻的原因。

This is why Semi Analysis believes we are yet again at another critical moment in AI.

这一时刻甚至可能超越了2023年初ChatGPT带来的转折点。

One that matches, if not exceeds, the ChatGPT moment in early twenty twenty three.

每一个关键时刻都拓展了AI的能力：GPT-3证明了规模有效，Stable Diffusion展示了AI可以生成图像，ChatGPT证明了对智能的需求，DeepSeq证明了可以在更小规模上实现，而O1则表明你可以将模型扩展到更优的性能。

Each moment expanded what AI could do: GPT-three proved scale worked, stable diffusion showed AI could make images, ChatGPT proved demand for intelligence, DeepSeq proved that it could be done on a smaller scale, and O1 showed you that you could scale models to even better performance.

Studio Ghibli的病毒式传播只是采用的节点，而Claude Code则是在组织模型输出方面的新突破，使其变得更有意义。

The viral moments of Studio Ghibli are just adoption points, while Claude Code is a new breakthrough in the agentic layer of organizing model outputs into something more.

在描述Claude Code时，他们进一步表示：将Claude Code仅仅视为专注于代码的工具可能是错误的，它更应被视为Claude计算机。

Now in describing Claude Code, they continue: It might be incorrect to think of Claude Code only as focused on code, but rather as Claude Computer.

在完全访问你的计算机的情况下，Claude能够理解其环境、制定计划，并迭代完成该计划，同时始终接受用户的指导。

With full access to your computer, Claude can understand its environment, make a plan, and iteratively complete this plan, the whole time taking direction from the user.

Claude Code 的功能不止于编写代码，它是 AI 代理的最佳范例。

Claude Code does more than just code and is the best example of an AI agent.

你可以用自然语言与计算机交互，描述目标和结果，而不是实现细节。

You can interact with a computer with natural language to describe objectives and outcomes rather than implementation details.

给 Claude 提供一个输入，比如电子表格、代码库、网页链接，然后要求它完成某个目标。

Provide Claude an input such as a spreadsheet, a code base, a link to a webpage, and then ask it to achieve an objective.

接着它会制定计划、核实细节，然后执行。

It then makes a plan, verifies details, and then executes it.

这是对未来的一瞥，但如今在软件中已经真实存在。

It's a glimpse of the future, but it is also here today in software already.

你最喜欢的工程师正在‘氛围编码’。

Your favorite engineers are vibe coding.

一年前提出‘氛围编码’一词的安德烈·卡帕西，正在公开讨论这一范式转变，他特别表示：我已经注意到，自己正在缓慢地丧失手动编写代码的能力。

Andre Karpathy, who coined the term vibe coding one year ago, is openly discussing the phase shift, and specifically says, I've already noticed that I am slowly starting to atrophy my ability to write code manually.

生成代码和辨别代码——阅读代码——是大脑中不同的能力。

Generation, writing code, and discrimination, reading code are different capabilities in the brain.

Vercel的首席技术官马尔特·许贝尔声称，他现在的主要工作是告诉AI哪里做错了。

Malte Hubel, the CTO of Vercel, claims that his new primary job is to tell AI what it did wrong.

Node的创造者瑞安·达尔。

Ryan Dahl, creator of Node.

JS表示，人类编写代码的时代已经结束了。

Js, says the era of humans writing code is over.

Ruby on Rails的创造者大卫·亨尼默·汉森正经历一种预期中的怀旧情绪，在亲手编写代码的同时怀念着亲手写代码的日子。

David Henimer Hansen, creator of Ruby on Rails, is having some sort of anticipated nostalgia, reminiscing about writing code by hand while writing code by hand.

Claude Code的创造者博里斯·丘尼表示，我们几乎100%的代码都是由Claude Code和Opus 4.5编写的。

Boris Churney, creator of Claude Code, says that pretty much 100% of our code is written by Claude Code and Opus four point five.

就连林纳斯·托瓦兹也在进行氛围编码。

Even Linus Torvald's is vibe coding.

但这不仅仅是程序员的事，Semi Analysis描述了他们团队的不同成员如何以不同方式使用这一工具。

But it isn't just coders, from which Semi Analysis describes how the different members of their team all use this tool in different ways.

他们写道，数据中心模型团队每周需要审阅数百份文档。

They write that the data center model team needs to review hundreds of documents every week.

AI供应链团队需要检查包含数千项条目的物料清单。

The AI supply chain team needs to inspect BOMs with thousands of line items.

内存模型团队需要在现货市场价格飙升时近乎实时地构建预测。

The memory model team needs to build forecasts in near real time as spot market prices explode.

技术人员需要维护一个实时仪表板，这意味着总的来说，从监管申报到许可证，从规格表到文档，从配置到代码，我们与计算机的交互方式已经改变。

Technical staff needs to maintain a live dashboard, meaning in total as they write: from regulatory filing to permits, spec sheets to documentation, config to code, the way we interact with our computers has changed.

程序员将不再编写代码，而是请求他人代为完成任务。

Coders will stop doing code and rather request jobs to be done on their behalf.

Claude Code的神奇之处在于它能正常运行。

And the magic of Claude Code is that it just works.

许多著名程序员终于接受了这种新的氛围编程浪潮，并意识到编程本质上已接近一个被解决的问题，更适合由代理而非人类来支持。

Many famous coders are finally giving in to the new wave of vibe coding and now realizing that coding is effectively close to a solved problem that is better off supported by agents than humans.

竞争的核心正在转移。

The locus of competition is shifting.

对线性基准的痴迷，比如哪个模型最好，将显得过时，就像过去比较拨号上网和DSL速度一样。

Obsessions over linear benchmarks as to what model is best will look quaint, akin to how fast your dial up is compared to DSL.

速度和性能至关重要，模型是驱动代理的核心，但性能将通过生成网站所需的网络数据包净输出来衡量，而非数据包本身的质量。

Speed and performance matter, and the models are what power agents, but performance will be measured as the net output of packets to make a website, not the packet quality itself.

明天的网站功能将通过工具、记忆、子代理和验证循环进行协调，以创造成果，而不仅仅是回应。

The website features of tomorrow is going to be the orchestration through tools, memory, sub agents, and verification loops to create outcomes and not responses.

所有信息工作现在都可以由模型来处理。

And all information work is finally addressable by models.

如果你正在构建任何语音AI产品，你必须了解AssemblyAI。

If you're building anything with voice AI, you need to know about AssemblyAI.

他们打造了业内最出色的语音转文本和语音理解模型，这些模型是Granola、Dovetail、Ashby和Cluely等产品的幕后基础设施。

They've built the best speech to text and speech understanding models in the industry the quiet infrastructure behind products like Granola, Dovetail, Ashby, and Cluely.

正如我之前所说，语音是AI最重要的模态之一。

Now, as I've said before, voice is one of the most important modalities of AI.

它是人类最自然的交互方式，我认为这也是下一波创新的关键所在。

It's the most natural human interface, and I think it's a key part of where the next wave of innovation is going to happen.

AssemblyAI的模型在准确性和质量上领先业界，让你能够真正信赖你的产品所依赖的数据。

AssemblyAI's models lead the field in accuracy and quality so you can actually trust the data your product is built on.

他们的语音理解模型帮助您超越转录，自动发现洞察、识别说话者并提取关键时刻。

And their speech understanding models help you go beyond transcription, uncovering insights, identifying speakers, and surfacing key moments automatically.

它以开发者为先，无需签订合同，按使用量付费，且可轻松扩展。

It's developer first, no contracts, pay only for what you use, and scales effortlessly.

前往 AssemblyAI.com/brief，领取50美元免费额度，今天就开始构建您的语音AI产品。

Go to AssemblyAI dot com slash brief, grab $50 in free credits, and start building your Voice AI product today.

大多数公司并不缺乏创意。

Most companies don't struggle with ideas.

他们的问题在于如何将这些创意转化为真正创造价值的AI系统。

They struggle with turning them into real AI systems that deliver value.

Robots and Pencils 是一家致力于弥合这一差距的公司。

Robots and Pencils is a company built to close that gap.

他们设计并交付以生成式和代理式AI驱动的智能云原生系统，注重专注力、速度和明确的结果。

They design and deliver intelligent, cloud native systems powered by generative and agentic AI, with focus, speed, and clear outcomes.

Robots and Pencils 以小型、高影响力团队的形式开展工作。

Robots and Pencils works in small, high impact pods.

工程师、战略家、设计师和应用AI专家协同合作，消除从想法到落地过程中的不必要的障碍。

Engineers, strategists, designers, and applied AI specialists working together to move from idea to production without unnecessary friction.

借助他们的Identic加速平台Roboworks，团队能够快速交付切实成果，根据项目范围，首次上线时间可短至45天。

Powered by Roboworks, their Identic Acceleration Platform, teams deliver meaningful results including initial launches in as little as forty five days depending on scope.

如果你的组织准备加速前进、降低复杂性，并将AI愿景转化为实际成果，Robots and Pencils正是为此刻而生。

If your organization is ready to move faster, reduce complexity, and turn AI ambition into real results, Robots and Pencils is built for that moment.

立即访问 robotsandpencils.com/aideallybrief 开始对话。

Start the conversation at robotsandpencils.com/aideallybrief.

网址是 robotsandpencils.com/aideallybrief。

That's robotsandpencils.com/aideallybrief.

Robots and Pencils 的 Impact at Velocity Weekends 专为氛围编程而设。

Robots and Pencils Impact at Velocity Weekends are for vibe coding.

如今将热情项目变为现实从未如此简单，所以赶紧启动你最爱的氛围编程工具吧。

It has never been easier to bring a passion project to life, so go ahead and fire up your favorite vibe coding tool.

但周一即将到来，不知不觉间，你将面对一堆微服务、一个来自1970年代的遗留COBOL系统，以及一条贯穿你退休派对之后的工程路线图。

But Monday is coming, and before you know it, you'll be staring down a maze of microservices, a legacy COBAL system from the 1970s, and an engineering roadmap that will exist well past your retirement party.

这就是你需要 Blitzy 的原因，它是首个专为企业级代码库设计的自主软件开发平台。

That's why you need Blitzy, the first autonomous software development platform designed for enterprise scale codebases.

在每个冲刺周期开始时部署，让你的路线图提速 500%。

Deploy at the beginning of every sprint and tackle your roadmap 500% faster.

Blitzy 的智能代理会分析你的整个代码库，规划任务，并自主完成超过 80% 的工作。

Blitzy's agents ingest your entire codebase, plan the work, and deliver over 80% autonomously.

以计算速度交付经过验证、端到端测试、高品质的代码。

Validated, end to end tested, premium quality code at the speed of compute.

数月的工程工作被压缩至数天内完成。

Months of engineering compressed into days.

周末用 vibe coding 实现你的热情项目。

Vibe code your passion projects on the weekend.

周一把 Blitzy 带到工作中。

Bring Blitzy to work on Monday.

了解为什么财富 500 强企业信赖 Blitzy 来处理关键代码，访问 blitzy.com。

See why Fortune 500s trust Blitzy for the code that matters at blitzy.com.

这就是 blitzy.com。

That's blitzy.com.

今天的节目由我的公司超级智能赞助。

Today's episode is brought to you by my company, Super Intelligent.

到2026年，企业人工智能的关键主题之一，甚至可能是最关键的主题，将是您将人工智能代理部署到的基础设施究竟有多完善。

In 2026, one of the key themes in enterprise AI, if not the key theme, is going to be how good is the infrastructure into which you are putting AI in agents.

超级智能代理准备度审计专门帮助您明确：第一，人工智能代理在何处以及如何最大程度地提升您的业务影响；第二，您需要做哪些准备，才能让组织最有效地利用这些新获得的收益。

Superintelligence agent readiness audits are specifically designed to help you figure out: one) Where and how AI in agents can maximize business impact for you, and two) What you need to do to set up your organization to be best able to leverage those new gains.

如果您想真正利用人工智能和代理在今年不仅提升生产力，更能以可衡量的方式根本性改变您业务成果的能力，请访问 bsuper.ai。

If you want to truly take advantage of how AI and agents can not only enhance productivity, but actually fundamentally change outcomes in measurable ways in your business this year, go to bsuper.ai.

而这就是他们从那里延伸出的核心主题。

And this is really the big theme they pick up from there.

之所以这是一个转折点，并不仅仅因为编码能力，更因为这种能力所带来的一切。

That the reason that this is an inflection point moment is not just about coding capability, but about what that leads to.

他们继续说道：在2020年代软件工程热潮时期，编程曾是所有工作中最有价值的，程序员供不应求。

They continue: Coding was once the most valuable work of all, with programmers in hot demand during the 2020 era of software engineering.

如今，编码已成为代理式信息处理所带来颠覆的前沿阵地，而价值高达15万亿美元的信息工作经济正面临风险。

Coding is now a beachhead in terms of the disruption that agentic information processing has, and the larger $15,000,000,000,000 information work economy is now at risk.

全球有超过10亿信息工作者，约占全球劳动力的三分之一。

There are 1,000,000,000 plus information workers or roughly a third of the global workforce.

信息工作类别的每一个工作流程通常都相似，并共享一种已被Claude Code验证有效的流程：读取并摄入非结构化信息，思考并应用领域知识，编写并生成结构化输出，然后验证并对照标准进行检查。

Every single workflow in the information work category is often similar and shares a workflow that Claude Code proves works for software: read ingest unstructured information, think apply domain knowledge, Write, produce structured output And then verify, check against standards.

这涵盖了大多数信息工作者的很大一部分，包括研究工作。

This is large swaths of most information workers, including research.

如果代理能够取代软件，那么还有哪些劳动力群体是它们无法触及的呢？

And if agents can eat software, what labor pool can they not touch?

我们的观点是，相当多的领域都会受到影响。

Our view is quite a few.

随着Claude Code和Cowork的兴起，代理的总可服务市场规模远超大语言模型。

And with the rise of Claude Code and Cowork, the total addressable market of agents is much larger than LLMs.

鉴于编码这一杀手级应用场景，以及Claude Code和Cowork明显的可泛化性，这要求我们采用一种完全不同的评估逻辑。

Given the killer use case in coding and the clear generalizability of Claude Code and Cowork, this justifies a completely different calculus.

自动化大多数问答和信息获取是可行的，这将释放巨大的潜在价值。

Automating most call and response and information fetching is likely doable, and this opens the absolute dollars possible.

他们所说的真正让更大块的市场可供颠覆的原因，是更长的任务周期。

And what they say really makes larger parts of the pie available for disruption is longer task horizon.

一个代理在失败前能持续工作多久？

How long can an agent work before it fails its task?

数据显示，自主任务周期每四到七个月翻一番，到2025年将加速至每四个月左右翻一番。

Meter data shows autonomous task horizons doubling every four to seven months, accelerating to around every four months in 2425.

每次翻倍都解锁了更多整体市场潜力。

Each doubling unlocks more of the total pie.

在30秒时，你可以自动补全代码片段。

At 30, you can auto complete code snippets.

在4.8小时时，你可以重构一个模块。

At four point eight hours, you can refactor a module.

对于多日任务，你可以自动化整个审计流程。

Multi day tasks, you can automate an entire audit.

很明显，Anthropic 也看到了这一点。

And it's clear Anthropic sees this, too.

2026年1月12日，Anthropic 推出了 CoWork，即面向通用计算的云端代码工具。

On 01/12/2026, Anthropic launched CoWork, Cloud Code for General Computing.

四位工程师在十天内完成了它。

Four engineers built it in ten days.

大部分代码是由 Cloud Code 自己编写的。

Most of the code was written by Cloud Code itself.

架构相同：Claude Agent SDK、MCP、子代理。

Same architecture: Claude Agent SDK, MCP, Sub Agents.

它能从收据生成电子表格，按内容整理文件，并从零散的笔记中起草报告。

It creates spreadsheets from receipts, organizes files by content, and drafts reports from scattered notes.

这就是去掉终端、加上桌面版的 Claude Code。

It's Claude Code minus the terminal plus a desktop.

这展现了未来的图景：一个能理解你日常工作的上下文，并按需构建和生成信息处理的工具。

This is the glimpse of the future: a harness that understands the context of your day to day job or work and can build and generate information processing as needed.

与其从数据库下载报告后再创建图片，一个代理会为你生成一份比你自己在Excel中做得更好的格式化报告。

Instead of creating images from reports you download from your database, an agent will generate a report with better formatting than you could do yourself within Excel for you.

每当你需要获取有关销售配额等信息时，你的代理会从用户界面或API中提取数据，并代表你生成报告。

Whenever you need to gather information about, say, a sales quota, your agent will extract the information from a UI or API and generate the report for you on your behalf.

信息工作本身将像Cloud Code自动化软件工程一样被自动化。

Information work itself is going to be automated like Cloud Code has automated software engineering.

尽管今天它还不完美，但它显然能比大多数人类更快地处理、综合和格式化数据。

And while it's not perfect today, it clearly can generally process, synthesize, and format data faster than most humans can.

在某些领域，这一切的保真度更高、成本更低，优于普通员工。

This all comes at higher fidelity and lower cost than the average worker in some areas.

虽然会出现幻觉，但大多数现有系统中早已存在大量人为错误。

While there will be hallucinations, most systems already exist with many human led errors in the process.

如果信息以可行的保真度被处理并传递到下一步，这本身将大幅增加工作供给。

If the information is processed at a viable level of fidelity and then passed to the next step, this itself will massively increase the supply of work.

我们正处在一个任何人都可以输入这些代理工作流，运行一个在2000年代需要终身学习才能完成的多变量回归分析的时刻。

We are literally at the point where any individual could type into one of these agent workflows to run a multivariable regression that would have taken a lifetime of training in the 2000s.

Stack Overflow 2025 年开发者调查显示，84% 的程序员正在使用人工智能，这代表了采用的前沿水平。

The Stack Overflow 2025 Developer Survey has eighty four percent of coders using AI, and that is the bleeding edge of adoption.

只有 31% 的人使用编码代理。

Only thirty one percent use coding agents.

这意味着，这种渗透曲线在更广泛的信息工作领域仍处于早期阶段。

And that means that this penetration curve is early for broader waves of information work.

就像编码代理的渗透率曾经历的那样，更广泛的信息工作将迅速迎来人工智能的普及。

Just like the blink for coding agent penetration, broader information work will quickly see AI adoption.

现在，我们将要阅读的这篇文章的最后一部分是关于成本和市场影响的。

Now the last section of this piece that we're going to read is about cost and market impact.

他们还专门设置了一个关于竞争格局和谁领先的次要部分，但至少对这个节目来说，那不是重点。

They have a whole secondary section on competitive race and who's winning, but that's less the point at least for this show.

回到我们之前中断的地方，他们写道：工程工作一直以来都是信息工作的黄金标准，但随着质量终于跨越了关键阈值，程序员与其工具之间的关系已经发生了逆转。

Moving back to where we left off, they write: Now engineering has and always will be the gold standard information work, but as the quality has finally crossed over a critical threshold, the relationship between coders and their tools have flipped.

程序员实际上只是在利用一个黑箱来实现目标。

Coders are effectively just harnessing a black box to achieve outcomes.

这一切之所以成为可能，不仅是因为智能的品质提升，还因为令牌的智能成本大幅下降。

And that was all possible because not only the quality but the cost of the intelligence of tokens has fallen an amazing amount.

现在，一位使用Claude Code的开发者就能完成过去一个团队一个月的工作量。

One developer with Claude Code can now do what took a team a month.

企业已经开始行动了。

And Enterprise is already starting to move.

智能的大幅成本下降将重新定价每家信息公司可重复性工作的利润率。

The massive deflationary cost of intelligence is going to reprice every information company's margin for repeatable work.

埃森哲刚刚签署了一项协议，培训30,000名专业人士使用Claude，这是迄今为止最大的Claude Code部署。

Accenture just signed a deal to train 30,000 professionals on Claude, the largest Claude code deployments to date.

埃森哲将重点聚焦于金融服务、生命科学、医疗保健和公共部门。

Accenture will focus on financial services, life sciences, healthcare, the public sector.

这些领域都是信息自动化尚未充分开发的巨大市场。

Those are all huge untapped markets for information automation.

OpenAI刚刚宣布推出Frontier，专注于企业级应用。

OpenAI just announced Frontier focused on enterprise adoption.

企业软件无疑是智能成本大幅下降后的首个牺牲品。

Enterprise software has easily been the first casualty of the great cost decline of intelligence.

SaaS本身只是将工作流程的信息处理固化为代码。

SaaS itself is just crystallized information processing of workflows into code.

SaaS的三大切换成本——数据、工作流锁定和集成复杂性——都已在边缘部分被削弱。

The three motes of SaaS switching costs of data workflow lock in and integration complexity have all been partially eroded at the margins.

SaaS高达75%的毛利率看起来是一个巨大的机会，因为智能代理在系统间迁移数据时成本降低，代理本身不依赖人类导向的工作流程，而MCP集成也使整合变得容易得多。

The 75% gross margin of SaaS looks like a huge opportunity as agents migrate data between systems with lessened migration costs, agents themselves do not rely on human oriented workflows, and MCP integrations make integration much easier.

SaaS的各个方面都在降价，而利润率已成为人工智能的首个突破口。

Every aspect of SaaS is cheapening, and the margins have become the first opportunity of AI.

在我们看来，任何需要人类点击按钮、收集信息、再将其重新格式化到另一种媒介的工作，都面临巨大风险。

In our view, anything that has a human click buttons, gather information, reformat it into another medium, is a huge risk.

好了，这就是我们接下来要阅读的这篇论文的部分内容。

So, okay, that's the part of this essay that we're going to read.

当事情发展到关键地步时，这里的关键词是转折点。

And when push comes to shove, the key phrase here is inflection point.

过去一个月最重要的事情，不仅是那些最具话语权、技术素养最高的AI用户集体意识到我们已经到达了一个转折点。

What's important about the last month is not just that en masse the most enfranchised and highly technically literate AI users realized that we had reached an inflection point.

而是这种认知现在已经蔓延到了更广泛的人群中。

It's that that perception has now cascaded into the wider world.

真正让我对此产生深刻认识、并促使我想要做这期节目的，是前《大西洋月刊》作者、《富足》一书的合著者德里克·汤普森在周四发了一条推文：对我而言，过去三周内，AI是泡沫的概率显著下降，而我们实际上在推理能力和使用需求方面严重供给不足的概率则大幅上升。

What really crystallized this for me and what basically prompted me to want to do this show was when former Atlantic author and co author of Abundance, Derek Thompson, tweeted out on Thursday, For me, the odds that AI is a bubble declined significantly in the last three weeks, and the odds that we're actually quite underbuilt for the necessary levels of inference and usage went significantly up in that period.

我认为，未来两年内，AI将成为绝大多数白领工作者的主屏幕，而并行代理将在知识工作领域以近乎苏联式的规模被部署。

Basically, I think AI is going to become the home screen of a ludicrously high percentage of white collar workers in the next two years, and parallel agents will be deployed in the battlefield of knowledge work at downright Soviet levels.

《纽约时报》的凯文·罗斯转发了这条推文，并说：这就是为什么大家在寒假期间对Claude Code如此恐慌。

The New York Times' Kevin Roose reposted it and said, This is why everyone was freaking out about Claude Code over winter break.

一旦你看到代理能自主为你做事，你就会立刻明白，几乎所有基于计算机的工作都将以此方式完成。

Once you see an agent autonomously doing stuff for you, it's so instantly clear that roughly all computer based work will be done this way.

凯文继续说：这就是为什么我认真的AI政策建议是，让每一位国会议员都坐在房间里，拿着笔记本电脑，花三十分钟亲手建一个网站。

Kevin continued, This is why my serious AI policy proposal is to sit every member of Congress down in a room with laptops for thirty minutes and have them all build websites.

黛德丽·博萨，你可能还记得她在准备关于SaaSpocalypse的节目时，是CNBC的一名记者，她尝试自己编写一个monday.com的版本，并不指望真能做成什么。

Deidre Bossa, who you might remember in preparation for a show about the SaaSpocalypse as a reporter for CNBC, tried to code herself up a version of monday.com, not expecting to actually do anything.

大约一小时后，她就做出了一个完全可用的版本，并且某种程度上成了一个信徒。

About an hour later, she had a fully working version and kind of became a convert.

她对这一转变的描述让我觉得非常精准：在过去几个月里，按她的话说，AI 从‘说话’变成了‘做事’。

The way that she described this shift, which I thought was quite crisp, was that over the last couple of months, in her words, AI went from talking to doing.

并不是所有人都完全认同。

Not everyone fully agrees.

迈克·科托内转发了德里克的推文，并说：我基本同意这一点，但其中有一个重大假设，即这些白领员工所就职的组织是否真的有意愿整合这些工具。

Mike Cotone reposted Derek and said, I agree mostly with this, however, there's a big assumption contained within that the organizations these white collar workers are employed by actually have the appetite to integrate the tools.

要利用当前的能力，需要对现有流程和系统进行大量变革。

Lots of process and system change will need to be made with current capabilities.

我认为这甚至走得更远。

I think it goes even farther than that.

坦率地说，好好使用AI的价值已经大幅上升。

To put it bluntly, the value of using AI well has gone way up.

但学会好好使用AI的难度也大幅上升了。

But the difficulty of learning how to use AI well has also gone way up.

这使得企业固有的惯性障碍更加明显。

That makes the natural enterprise inertia barriers even more pronounced.

还有许多类似的反应，比如范·杰克逊写道：AI泡沫源于企业缺乏盈利能力且过度杠杆化，而非使用问题。

There's also plenty of reactions like this one from Van Jackson, who writes, The AI bubble is about lack of profitability in firms being over leveraged, not about usage.

每个人已经在无盈利的情况下使用AI了。

Everyone already uses AI unprofitably.

摧毁大部分劳动力，并强迫那些仍 clinging 工作的人使用AI，并不会改变任何事情。

Destroying most of the workforce and press ganging those still clinging to jobs into using AI changes nothing.

但至少在市场层面，这正是发生的变化。

But this, at least on the market side, is kind of what's shifted.

这一观点为泡沫讨论增添了一个有趣的转折点，也是德里克和周等人认为AI泡沫很可能出现的原因：对于普通人来说，AI泡沫的论点是，我们正在过度建设AI基础设施，而这些基础设施我们可能根本不需要，或者这些公司根本无力承担。

The interesting wrinkle that this adds to the bubble conversation, and the reason that folks like Derek and Chau are talking about why an AI bubble is likely, is that for the average person, the AI bubble argument was that we were overbuilding AI infrastructure that maybe we weren't even going to need, or that maybe these companies couldn't even pay for.

你希望它一直运行着。

You kind of want it running all the time.

事实上，你希望多个代理程序持续运行，以完成更多任务。

In fact, you want multiple agents running all the time to do more things.

多个代理持续运行意味着消耗更多的令牌。

Multiple agents running all the time means more tokens consumed.

正如伊桑·莫洛赫所言，现在代理能够完成长期经济可行的任务，因此需要更多的计算资源。

And that, as Ethan Moloch puts it, are going to need more compute now agents can complete long term economically viable tasks.

伊桑澄清，这并不意味着在融资计算资源方面不会出现财务问题，但他指出计算资源并未被过度建设。

Ethan clarifies this does not mean that there couldn't be some sort of financial issue with financing the compute, but does point to the idea that compute is not being overbuilt.

而这至少开始发生变化。

And that is what is at least starting to shift.

现在，若认为这一点已经完全反映在公开市场上，那就过于夸大了。

Now, it would be way overblown to argue that this has fully found its way into public markets.

但你已经开始看到这种情况发生，而且令人头晕目眩，因为没人知道这些信号综合起来究竟意味着什么。

But you're starting to see it happen, and it's kind of head spinning, as no one knows what all these signals taken together should mean in aggregate.

塞布·K总结了这种困惑：今天突然形成的共识是，AI的起飞正在迅速且出人意料地加速，但谷歌、微软、亚马逊、脸书、Palantir、博通和英伟达的股票在过去五天内均下跌了约10%。

Seb K sums up the confusion: Sudden smart consensus today is that the AI takeoff is rapidly and surprisingly accelerating, but stocks for Google, Microsoft, Amazon, Facebook, Palantir, Broadcom, and Nvidia are all down around 10% over the last five days.

SMCI今天下跌了10%。

SMCI is down 10% today.

顺便说一下，这是周四的事。

This, by the way, was from Thursday.

只有苹果上涨了，因为它最不涉及AI。

Only Apple's up, it's the least AI.

在我看来，这很奇怪。

Strange in my opinion.

我只能说，朋友们，系好安全带，因为我觉得我们将迎来一段有趣而令人困惑的时期。

All I can say is buckle up, friends, because I think we are in for an interesting and confusing period.

早在十月，OpenAI的Rune就写道：没有足够多的人在情感上做好准备，去面对如果这不是一个泡沫的情况。

Back in October, OpenAI's Rune wrote: Not enough people are emotionally prepared for if it's not a bubble.

我觉得这正是我们在这里看到的部分原因。

And I kind of think that's part of what we're seeing here.

正如Deirdre所说，过去几个月，AI已经进入了‘做’的阶段，而不是‘说’的阶段。

AI, as Deirdre put it, over the last couple months has entered the show, not tell phase.

它正在做事，而不是谈论它们。

It's doing things, not talking about them.

代理已经从一个非常酷的概念，转变为目前正在执行实际工作的工具。

Agents have turned the corner from a thing that would be really cool to a thing that is doing real work right now.

我们周围，工作方式发生改变的迹象非常明显。

And everywhere around us, the signals that the way that work is done has changed are profound.

以我们前几天分享的一个例子来说，OpenAI总裁格雷格·布罗克曼表示，到3月31日，公司内部任何技术性任务，人类首选的工具将是与代理交互，而不是使用编辑器或终端。

To take an example that we shared the other day, OpenAI President Greg Brockman says that by March 31, for any technical task that happens inside that company, the tool of first resort for humans is interacting with an agent rather than using an editor or a terminal.

到3月31日，实现代理优先的工作方式。

Agent first work by March 31.

正如我在引言中提到的，如果你也想跟上这个时间表，我决定像新年决心那样，再推出一个免费的自主学习项目，因为当然要这么做！

Now as I mentioned in the intro, if you want to get on that timeline as well, I decided to throw together another free self directed learning experience like the New Year's Resolution because heck yeah!

如果格雷格都要求他的团队实现这个目标，那我们其他人为什么不能也搞清楚呢？

If Greg is going to challenge his team to meet that goal, why shouldn't the rest of us figure it out, too?

无论如何，无论泰勒·科文是否正确——上周四Opus 4.6和5.3编码器发布的历史意义是否将成为某种转折点——但显然，一场转变已经发生。

In any case, whether Tyler Cowen is right and last Thursday when Opus four point six and five point three codecs were released goes down in history as some kind of turning point, what's clear is that a shift has happened.

事实上，这种转变已经持续了两个月，但现在它正全面渗透到整个系统中，每个人都在努力应对由此带来的影响。

It has in fact been happening for two months, but now it is fully working its way through system, and everyone is grappling with the implications.

衷心祝愿所有听众在这段时期一切顺利，我当然会继续尽我所能，帮助你们充分利用这段时光。

I wish all of you listeners nothing but the best navigating this period, and I will of course continue to do my best trying to help you make the most of it.

目前，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，感谢你们的收听或观看。

Appreciate you listening or watching, as always.

下次再见，平安！

Until next time, peace!