如何使用最新的AI工具

这周新发布的AI和模型产品太多了，我不得不做两期节目才能把所有的内容都覆盖到。

This week had so many new AI and model product releases that I've had to do not one, but two episodes just to capture everything that's been going on.

有意思的是，这第二批产品中有很多都展现了一些非常明显且重要的趋势，预示着我们未来的发展方向。

What's interesting about this second set is that many of them show some pretty distinct and important trends about where we're headed.

比如，很明显我们将越来越多地通过语音来管理我们与电脑的互动。

For example, it is very clear that we are going to be managing more and more of our interactions with computers via our voice.

这不会一下子发生，但现在我们有了像ChatGPT Live这样的东西，开发者可以基于它进行开发，而且它的能力集显著增强，能更好地分辨谁在说话、处理背景噪音，你会看到越来越多的应用会涉及语音。

It won't happen all at once, but now that we have things like ChatGPT Live available to developers build around with a significantly increased capability set, better ability to distinguish who's talking, background noise, you're just going to see more and more applications that involve voice.

这些公告中展示的另一个趋势是，人们继续朝着复杂的模型架构发展，可以在成本和足够好的能力之间进行优化，而不是一味追求最高的能力。

Another example of a trend shown off in these announcements is the continued push towards complex model architectures where people can optimize for cost and good enough capability as opposed to just always seeking the highest capability.

关键是，有时候新产品发布的意义在于你能用它们做什么，有时候则在于它们预示着我们大家很快都会做什么。

The point is that sometimes new product releases are about what you can do with them, and sometimes they're about what they say about what we're all going to be doing soon.

《AI每日简报》是一个每日播客和视频节目，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.

好了，朋友们。

Alright, friends.

在开始之前，先快速说几个通知。

Quick announcements before we dive in.

首先，感谢今天的赞助商：毕马威、Blitzy、Robots and Pencils 和 Hyperagent。

First of all, thank you to today's sponsors, KPMG, Blitzy, Robots and Pencils, and Hyperagent.

如果你想听无广告版本，可以去 patreon.com/aideallybrief，或者也可以在 Apple Podcasts 上订阅。

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

如果你想了解如何赞助这个节目，请发送邮件到 sponsorsaidailybrief dot ai。

And if you wanna learn more about sponsoring the show, send us a note at sponsorsaidailybrief dot ai.

欢迎回到《AI 每日简报》头条版，每天五分钟，带你了解所有重要的 AI 新闻。

Welcome back to the AI Daily Brief headlines edition all the daily AI news you need in around five minutes.

今天我们先从 Anthropic 的一份新报告说起，这份报告详细介绍了他们检测和应对 AI 滥用的努力。

We kick off today with a new report from Anthropic that details their efforts to detect and counter AI misuse.

这份报告里有很多有意思的细节，要不是这周有更大的 AI 安全讨论，它可能会成为更热门的话题。

There were a lot of juicy nuggets in this thing, and were it not for the larger AI safety conversation happening this week, it probably would have been a bigger topic of conversation.

我们来看看几个重点：Anthropic 表示，他们成功阻止了来自阿里巴巴、DeepSeek 和小米的多起重大蒸馏攻击。

Going through some of the highlights, Anthropic said that they had disrupted major distillation attacks from Alibaba, DeepSeek, and Xiaomi.

每家公司都利用大量虚假账户，从Anthropic的模型中提取推理痕迹，用作训练数据。

Each company used networks of fraudulent accounts to extract reasoning traces from Anthropic models for use as training data.

不过，Anthropic还声称，他们发现DeepSeq和Kimi K3 Maker Moonshot都在把请求转发给Claude，然后把响应提供给自己的用户。

However, Anthropic also claimed that they had detected both DeepSeq and Kimi K3 Maker Moonshot routing requests to Claude and serving the responses to their users.

Anthropic写道：在一个案例中，十天之内，Moonshot将近千个客户请求转发给了Anthropic，其中绝大多数都路由到了Opus。

Anthropic wrote: In one instance, over a ten day period, Moonshot relayed almost thousand customer requests to Anthropic, the vast majority of which were routed to Opus.

根据这份报告，Moonshot似乎并不是用这种方法来伪装自己的模型，用Anthropic的后端来让它们看起来更强大。

Based on the report, doesn't seem like Moonshot was using this method to spoof their models, using an Anthropic backend to make them appear more capable.

相反，这看起来像是为了收集真实的用户查询，用于他们的蒸馏流程。

Instead, it seems like this was a way to gather realistic user queries for their distillation pipeline.

不过，Anthropic指出，这种方法多次暴露了中国政府和企业用户的敏感信息。

Still, Anthropic notes that this method exposed sensitive information from Chinese government and corporate users on multiple occasions.

在生物学部分，Anthropic提到了五月份的一起事件，当时他们的分类器阻止了一份科研经费申请的工作。

In the Biology section, Anthropic discussed an incident in May when their classifier blocked work on a grant application for scientific funding.

那份申请讨论的是对一种叫基孔肯雅热的蚊媒病毒进行功能获得研究。

The application discussed gain of function research on a mosquito borne virus called chikungunya.

Anthropic写道：因为基孔肯雅病毒在自然界中自然传播，所以如果作为生物武器故意释放，很难与自然爆发区分开来。

Anthropic wrote: Because chikungunya circulates naturally, a deliberate release as part of a bioweapon would be difficult to distinguish from a natural outbreak.

该资助项目旨在识别基孔肯雅病毒的增强突变，将其工程化到感染性克隆中，并在体内筛选毒力。

The grant sought to identify enhancing mutations of the chikungunya virus, engineer them into infectious clones, and select for virulence in vivo.

换句话说，病毒在反复感染活体动物时会逐渐变得更危险，研究人员每轮都会保留致病性最强的变体。

In other words, the virus would become progressively more harmful as it repeatedly infected live animals, with researchers keeping the most disease causing variants in each round.

Anthropic指出，这项研究可能在疫苗开发中有正当用途，但他们担心它可能是恶意的，因为研究计划在军事研究机构进行。

Anthropic noted that this research could have innocent applications in vaccine development, but they were concerned it might have been malicious because research was intended to be conducted at a military research institute.

报告没有指明是哪个国家资助了这项研究，但指出在研究人员被禁止后，AI提示词继续通过灰色市场经销商传播。

The report did not identify which nation was sponsoring the research, but did note the AI prompts continued through gray market resellers after the researchers were banned.

报告还包含了几十个其他被阻止的滥用案例，包括俄罗斯行为者利用Claude进行间谍活动和宣传，中国、俄罗斯和也门行为者试图编写常规武器软件，以及一系列网络攻击和金融诈骗企图。

The report also contained dozens of other examples of blocked misuse, including Russian actors using Claude for espionage and propaganda, Chinese, Russian, and Yemeni actors attempting to write software for conventional weaponry, and a host of attempted cyberattacks and financial scams.

Anthropic指出，这份报告是精心编写的，写道：我们在这里分享的案例不是典型的滥用，而是我们迄今为止识别出的最显著和新颖的威胁活动的例子。

Anthropic noted the report was put together with intention, writing: The cases we share here aren't typical misuse, but rather examples of the most notable and novel threat activity we've identified to date.

我们发布这项工作，是因为我们认为我们有责任披露对我们服务的恶意滥用。

We're publishing this work because we believe we have a responsibility to disclose malicious misuse of our service.

随着模型能力越来越强，风险也会增加，除非AI开发者和社会的守护者采取行动让它们更安全。

As models become increasingly capable, their risks will increase unless AI developers and society's defenders act to make them safer.

值得注意的是，这份报告完全聚焦于对Haiku、Sonnet和Opus模型的滥用，只有一例与Fable和Mythos相关的蒸馏案例。

Notably, the report was entirely focused on misuse of haiku, sonnet, and opus models, with only a single case of distillation related to fable and mythos.

说到AI风险方面的变化，萨姆·奥尔特曼告诉员工，他对AI放缓持开放态度。

Now speaking of the changing wins around AI risk, Sam Altman has told staff he's open to an AI slowdown.

在本周的一次全员会议上，奥尔特曼告诉团队，OpenAI可以放慢新AI模型的开发节奏。

During an all hands meeting this week, Altman told his team that OpenAI could pace the development of new AI models.

不过，他说这可能会与其他AI实验室协调进行，并表达了对并非所有竞争对手都会同意的担忧。

However, he said this would likely happen in coordination with other AI labs, and expressed concerns that not all rivals would agree.

这些评论是在关于放缓AI开发速度的讨论浪潮之后发表的，这场讨论甚至在雅各布·科克森本周早些时候的病毒式推文之前就已经开始了。

These comments come after a wave of discussion about slowing the pace of AI development that started even before Jacob Coxons viral tweet earlier this week.

许多OpenAI员工在七月份签署了《Pacing the Frontier》公开信，首席科学家雅各布·帕查基最近警告说，递归自我改进加剧了放缓的理由。

Many OpenAI staffers signed on to the Pacing the Frontier open letter in July, and chief scientist Jacob Pachacchi recently warned that recursive self improvement amplified the case for a slowdown.

在上周末发布的一篇博客文章中，帕查基写道，他希望行业协调和“自愿放缓”能成为常态，直到建立共享的安全标准。

In a blog post published last weekend, Pachacchi wrote that he hopes for industry coordination and quote, voluntary slowdowns to become commonplace until shared safety bars are established.

这似乎是奥特曼第一次表示愿意放慢速度。

This appears to be the first time that Altman has indicated a willingness to slow down.

与他在Anthropic的同行达里奥·阿莫代不同，他没有签署《前沿放缓》公开信，而且直到最近在公开场合都一直不置可否。

Unlike his counterpart Dario Ahmaday at Anthropic, he did not sign the Pacing the Frontier open letter and had been non committal in public statements until very recently.

周二，在谈到OpenAI解决了一个千禧年大奖难题时，奥特曼写道：世界现在拥有极其强大的模型。

On Tuesday, remarking on OpenAI finding a solution to a Millennium Prize problem Altman wrote: The world has extremely capable models now.

我没想到这么重大的成果会这么快出现。

I did not expect a result of this magnitude to happen so soon.

我们一直在讨论需要放缓进展以确保安全。

We've been talking a lot about the need to pace progress to ensure safety.

对我来说，这是迄今为止最有力的紧迫性证据。

For me, this is the strongest evidence yet of the urgency.

再加上本周早些时候AI安全研究员保罗·克里斯蒂亚诺加入OpenAI董事会，奥特曼的言论可能表明，行业协调放缓正变得越来越有可能。

Alongside the addition of AI safety researcher Paul Cristiano to the board of OpenAI earlier this week, Altman s comments could indicate a coordinated industry slowdown is becoming more likely.

另外在OpenAI的新闻中，该公司似乎遇到了计算瓶颈，因此将暂停高端订阅服务。

Now also in OpenAI news, the company appears to have hit a compute wall, and will pause top end subscriptions as a result.

我们之前讨论过这个问题，当时它遇到麻烦的时候，我们就说过这会是每个人都会面临的问题，现在看来确实已经来了。

We talked about this back when was having all of its trouble, that this was going to be a problem coming for everyone, and it is apparently now here.

现在，在GPT-6 Astra发布后不久，OpenAI的产品负责人Thibault Sotiao发帖说：对Astra的需求真是前所未有。

Now, shortly after the release of GPT six Astra, OpenAI product leader Thibault Sotiao posted: Demand for Astra is really unprecedented.

我们正在尽一切可能来满足需求，但我之前从没见过这种情况，而且我们之前也经历过非常快速的增长。

We're pulling all the levers possible to sustain the demand, but I've not seen anything like it until now, and we went through very steep growth before.

优先事项始终是为现有用户提供优质服务，但如果这种情况持续下去，我们可能不得不暂停新的Pro订阅。

Priority will always be to keep excellent service for existing users, but we might have to pause new pro subscriptions for a bit if this continues.

Sam Altman转发了这条消息，并说：这确实很糟糕，但我们会优先为客户提供优质服务，直到我们能够重新掌控局面。

Sam Altman reposted that saying, This would suck, but we will prioritize great service for customers until we can get back on top of things.

看来他们现在已经无计可施了，因为Tebow在周四发帖说。

Well, it seems like they have now run out of levers to pull with Tebow posting on Thursday.

为了确保我们现有用户能获得出色的体验并继续使用Astra，我们将暂停200美元Pro计划的订阅。

To make sure our current users have an incredible experience and continued access to Astra, we are going to pause subscriptions to our $200 Pro plan.

这些订阅对我们的系统压力最大，我们想采取最小的措施，以便继续提供尽可能广泛的访问。

These put the most strain on our systems and we wanted to take the smallest step that allows us to continue giving the broadest access possible.

API里的其他套餐都还能用。

All other plans in the API remain available.

现有账户不受影响，我们正在尽快增加容量。

There is no impact to existing accounts, and we are working on adding more capacity as fast as we can.

马特·舒默直截了当地说，补贴代币的时代要结束了。

Stating it bluntly, Matt Schumer wrote, The era of subsidized tokens is ending.

大家做好准备吧。

Prepare accordingly.

亚历克斯·巴里什则认为，200美元的套餐不会再回来了。

And Alex Barish thinks, The $200 plan will not be returning.

之后他们会发一篇文章讲怎么选合适的套餐，然后告诉我们，200美元套餐的用户里，大部分其实用100美元的套餐就够了，只有一小部分人老是顶到上限。

We will get a post talking about right size fit, and they will tell us there was a large split between people on the $200 plan, where most would have been better off on the $100 plan, and a small percentage of users that were constantly maxing out limits.

说到算力，微软在跟自己的算力短缺问题斗了几年之后，计划把数据中心容量翻三倍。

Now, speaking of compute, Microsoft plans to triple their data center capacity after battling their own compute shortages over recent years.

知情人士告诉彭博社，微软在新计划下已经加速推进了AI建设战略。

Sources told Bloomberg that Microsoft has turbocharged their AI build out strategy under a new plan.

他们现在计划到2032年实现38吉瓦的全球容量，而目前是12吉瓦。

They now intend to have 38 GW of global capacity by 2032, up from their current 12 GW.

目前他们的容量中只有2吉瓦专门用于AI计算，但现在他们打算将这一比例提高到总容量的三分之一。

Only two GW of their current fleet is dedicated to AI compute, but they will now aim to grow the share to a third of total capacity.

此外，该计划还将增加CPU，以支持智能体工作流。

In addition, the plan will see them add CPUs to power agentic workflows.

重要的是，在整个AI建设过程中，微软一直是超大规模云服务商中最保守的。

Importantly, throughout the AI build out, Microsoft has been the most conservative of the hyperscalers.

众所周知，他们在2025年初缩减了计划，取消了一些大规模租赁，这导致了市场的重大回调。

They famously scaled back plans in early twenty twenty five, canceling some large scale leases, which contributed to a major market pullback.

但在过去一年里，微软多次表示，计算能力限制正在制约增长，导致他们不得不拒绝客户。

But over the past year, Microsoft has repeatedly said that compute constraints are limiting growth and causing them to turn away customers.

微软的新计划表明，AI基础设施远未过度建设，推动建设的公司预计至少还需要五年的大规模投入才能满足需求。

Microsoft s new plan reinforces that AI infrastructure is nowhere near overbuilt, and the companies driving the build out expect at least five more years of elevated spend before they can meet demand.

说到建设持续推进，NVIDIA首席执行官黄仁勋在高盛会议上重申了70%增长的预测。

Speaking of the build out continuing apace, NVIDIA CEO Jensen Huang has doubled down on forecasts of 70% growth at a Goldman Sachs conference.

黄仁勋在周四的发言中说，增长只受供应链限制，他评论道：即使我们的需求远超过70%，我们的供应也能让我们有信心实现70%的增长。

Speaking on Thursday, Jensen said that growth is limited only by the supply chain, commenting: even though our demand is much greater than 70%, our supply allows us to confidently deliver 70%.

黄仁勋向来不吝于自我宣传，但这是他过去一年来首次讨论前瞻性预测，这表明他更加自信了。

Huang is never shy about self promotion, but this is the first time he's discussed forward estimates over the past year, suggesting an added layer of confidence.

在投资者面前，黄仁勋解释说，很多人没有跟上英伟达产品近年来的演变。

Before the crowd of investors, Huang explained that many of them haven't kept up with how NVIDIA's product has evolved over recent years.

他说，大多数人以为英伟达只是造芯片的。

He said, Most people think NVIDIA builds a chip.

我的意思是，我们造的东西得用飞机来运。

I mean, you need airplanes to ship what we build.

现在一块GPU不是399美元，而是850万美元，那是一块GPU，全部通过NVLink连接，有200万个部件，对吧？

One GPU now is not $399 it's $8,500,000 That's one GPU, all connected with NVLink, 2,000,000 parts, right?

25万千瓦，那才是一块GPU，而我们出货的是成千上万块。

250,000 kilowatts, that's a GPU and we ship thousands of them.

黄仁勋还补充说，他们的机架级系统订单目前正以每月27%的惊人速度增长。

Huang added that orders of their Rackscale systems are currently growing an astonishing 27% month over month.

总的来说，黄仁勋想传达的核心观点是，尽管每隔一周就有关于新竞争对手的头条新闻，但英伟达仍然处于人工智能建设浪潮的中心。

And overall, the core idea that Huang tried to communicate was that NVIDIA is still at the center of the AI build out despite headlines about new competition every other week.

不过，他们现在面临的一个新情况是，司法部正在调查他们200亿美元的Grok交易，以确定这笔交易是否是为了规避反垄断法而设计的。

Still, one new thing they have to contend with is the fact that the DOJ is investigating their $20,000,000,000 Grok deal to determine whether it was structured to circumvent antitrust laws.

这笔交易在去年12月宣布时，芯片初创公司Grok将其描述为非排他性许可协议，允许英伟达使用他们的技术，以便整合到未来的产品中。

When the deal was announced in December, chipmaking startup Grok framed it as a non exclusive licensing agreement that gave NVIDIA access to their technology for integration into future products.

作为交易的一部分，首席执行官约翰·罗斯和首席运营官桑尼·马德拉也加入了英伟达。

As part of the deal, CEO John Ross and COO Sunny Madra also joined NVIDIA.

据《纽约时报》报道，司法部在交易宣布后不久就开始调查此事。

The New York Times reports that the DOJ has been investigating the deal since shortly after it was announced.

这笔交易是近年来人工智能行业众多“非收购”交易之一，其中包括谷歌与Character AI和Windsurf的交易，以及Meta与Scale AI的交易。

The deal was one of many not acquisitions in the AI industry over recent years, including Google's deals with Character AI and Windsurf, as well as Meta's deal with Scale AI.

通过将交易设计为非排他性许可协议，它们不会触发联邦贸易委员会的自动审查，也不会像传统并购那样被阻止。

By structuring the deals as non exclusive licensing agreements, they don trigger automatic review by the FTC, and can t be blocked in the same way as traditional mergers and acquisitions.

然而，这种结构在二月份受到了批评，多位参议员呼吁联邦贸易委员会和司法部调查这一特定问题。

Still, that structure came under fire in February, with a number of Senators calling on the FTC and the DOJ to investigate this particular issue.

在信中，参议员们写道，这些交易实际上等同于事实上的合并，使得公司能够整合人才、信息和资源，同时显然试图绕过通常适用于并购的审查。

In a letter, the senators wrote that the deals function as de facto mergers, allowing the companies to consolidate talent, information, and resources, all while apparently attempting to bypass the scrutiny typically applied to mergers and acquisitions.

目前这个阶段，这只是一项调查，还不清楚会有什么结果。

Now at this stage, it is just an investigation, and it's unclear whether anything will come of it.

但是，当这类交易开始出现时，这种调查迟早会发生，这是不可避免的。

But when they still started happening, it was always inevitable that this sort of investigation was going to happen.

最后，Meta 的新个人智能体 Muse 似乎在消费者中取得了一些不错的进展。

Lastly, Meta's new personal agent, Muse, seems to be finding some decent traction with consumers.

除此之外，它在华尔街也确实大受欢迎。

And beyond that, it is certainly a big hit on Wall Street.

Sensor Tower 的新数据显示，该智能体在美国 iOS 应用商店上线首日被下载了 8.3 万次。

New data from Sensor Tower shows the agent was downloaded 83,000 times on launch day in the iOS app store for The US.

对于 Meta 的新产品来说，这个数字并不特别令人印象深刻。

Now that's not particularly impressive for a new Meta product.

Threads 上线首日下载量达到 430 万次，而 Meta AI 应用则达到了 10.8 万次下载。

Threads achieved 4,300,000 downloads on launch day, while the Meta AI app reached 108,000 downloads.

不过，这个首发成绩还是不错的，足以让Muse在App Store里冲到第二名。

Still, the debut was good enough to push Muse to second place in the App Store.

但在华尔街看来，Muse被视为Meta在AI领域仍有竞争力的一个证明。

However, on Wall Street, Muse is being viewed as evidence that Meta can still compete in AI.

发布当天股价上涨了6%，而且基本保住了这些涨幅。

The stock jumped 6% on release day and is largely holding onto the gains.

周四，摩根大通的分析师将Meta的股票评级上调为“买入”，他们认为Muse预示着未来会有更多动作。

On Thursday, JPMorgan analysts upgraded the stock to a buy, viewing Muse as a sign of more to come.

在一份研究报告中，他们写道：Meta在向大约40亿用户提供面向消费者的AI产品方面处于有利位置，这种规模化的分发能力是一个显著的竞争优势。

In a research note, they wrote: Meta is well positioned to deliver consumer driven AI products to its base of around 4,000,000,000 users and that scale distribution is a significant competitive advantage.

随着Meta在发布前沿模型和广告之外的AI驱动产品方面还处于早期阶段，仍有很大的上行空间。

There's still meaningful upside potential as Meta is in the early stages of releasing frontier models and AI driven products beyond advertising.

对于一个重要的新产品类别来说，这是不错的进展，但今天就先报道到这里。

Good progress for what is an important new product category, but for now, that is going do it for the headlines.

接下来，是本期的主节目。

Next up, the main episode.

毕马威和德克萨斯大学奥斯汀分校的一项新研究发现，当人们与人工智能合作时，相似的技能并不能保证相似的结果。

A new study from KPMG and the University of Texas at Austin found that when people work with AI, similar skills don't guarantee similar outcomes.

研究人员研究了500多名早期职业专业人士，发现表现最好的人通过指导、评估和完善人工智能的输出，持续放大了人工智能的价值。

Researchers studied more than 500 early career professionals and found that the best performers consistently amplified the value of AI by guiding, evaluating, and refining its outputs.

这些顶尖表现者被称为“人工智能放大器”，他们的定义不仅仅在于他们知道什么，而在于他们如何与人工智能合作。

These top performers, called AI amplifiers, weren't defined by what they knew alone, but by how they worked with AI.

想了解更多关于人工智能放大器与其他人的区别，请访问kpmg.com/us/aiamplifiers。

Learn more about what separates AI amplifiers from everyone else at kpmg.com/us/aiamplifiers.

Blitzy对庞大代码库的理解，使其能够实现自主安全修复、现代化改造和新功能开发。

Blitzy's understanding of massive codebases unlocks autonomous security fixes, modernization, and new features.

那么，当完全没有遗留代码时，会发生什么呢？

So what happens when there's no legacy code at all?

绿地项目本该是容易的部分——从零开始，没有技术债务。

Greenfield is supposed to be the easy part clean slate, no technical debt.

但即使是绿地项目，也只能以人类的速度推进，一次一个冲刺。

But even Greenfield moves at human speed, one sprint at a time.

Blitzy把工作单元从开发者变成了整个项目，自主规划、构建、测试和验证整个应用程序，从零开始生成数十万行可直接投入生产的代码。

Blitzy changes the unit of work from the developer to the project, autonomously planning, building, testing, and validating entire applications from scratch hundreds of thousands of lines of production ready code.

Blitzy的一个客户搭建了一个全新的应用程序，共53.4万行代码，把原本65周的路线图压缩到了两周。

One Blitzy customer stood up a brand new application, 534,000 lines of code, compressing a sixty five week roadmap into two weeks.

另一个客户在没有前端工程师的情况下，交付了整个应用程序。

Another shipped an entire application with no front end engineer.

无论是遗留系统还是全新项目，答案都一样：软件以计算的速度运行。

Legacy or greenfield, the answer is the same: Software at the speed of compute.

在blitzy.com上构建你的下一个项目。

Build what's next at blitzy.com.

网址是blitzy.com。

That's blitzy.com.

到了这个阶段，问题已经不再是企业是否在积极使用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 and Pencils 这家公司，我觉得就是为这个时代而生的。

Robots and 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，解决对业务真正重要的问题，而不是那些只停留在幻灯片里的实验。

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.

每次合作，他们都是从客户真正需要的结果出发，倒推着来。

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工程和那些噪音，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.

本期《AI每日简报》由 Hyperagent 赞助播出，你可以在这里运行你的团队可以共同管理的智能体集群。

This episode of the AI Daily Brief is brought to you by Hyperagent, where you run fleets of agents your team can manage together.

别再依赖本地智能体和聊天工作流，等着你的笔记本电脑被提示了。

Forget local agents and chat workflows waiting on your laptop to be prompted.

Hyperagent 在云端部署了始终在线的智能体，在你团队已经使用的工具中完成实际工作。

Hyperagent deploys always on agents in the cloud, doing real work across the tools your team already uses.

营销代理能把竞争对手的动作变成落地页。

Marketing agents turn competitor moves into landing pages.

销售代理会丰富潜在客户信息、起草邮件，并更新客户关系管理系统。

Sales agents enrich leads, draft emails, and updates the CRM.

运营代理负责跟进文书工作，并跟踪预算。

Ops Agent chases the paperwork and tracks the budget.

每个代理都能访问共享上下文，并遵循你关于范围和审批的规则。

Every agent has access to shared context and follows your rules about scope and approvals.

是时候让代理像团队成员一样工作了。

It's time you add agents that feel like teammates.

去 Hyperagent 雇佣你的代理吧。

Hire yours at Hyperagent.

在 hyperagent.com/aidailybrief 获取 100 美元积分。

Get $100 in credits at hyperagent.com/aidailybrief.

欢迎回到 AI 每日简报。

Welcome back to the AI Daily Brief.

今天我们回到实用话题，因为即使最新一轮的AI风险讨论已经席卷主流，那些现在对你就有用、而且基本上没人认为有实际危险的新产品，还是在不断推出，创造新的机会。

Today, we are back to the practical because even as the latest round of the AI risk conversation has exploded into the mainstream, New products that are useful for you right now and which basically no one is arguing are actively dangerous have continued to come out and create new opportunities.

所以今天，我们来聊聊最新AI工具可以用来做什么。

And so today, we're going to talk through what to use the latest AI tools for.

我会介绍过去几天发布的八九个产品，简单说说每个产品是什么、谁应该考虑使用，以及应该用它来做什么。

I'll be going through eight or nine product launches from the last couple of days with quick notes on what the launch is, who should be considering using it, and what they should be thinking about using them to do.

首先，OpenAI在API中发布了GPT Live One。

First up, OpenAI has released GPT Live One in the API.

这是OpenAI在7月份发布的语音模型，就是那个让人印象深刻的奶奶视频里的那个。

This is the voice model that OpenAI released back in July with that memorable grandma's video.

这是一个实时语音模型，也就是说它可以同时说话和听，不需要轮流来。

This is a live voice model, meaning it can talk and listen at the same time without needing to take turns.

它声称能更好地处理嘈杂环境和打断，而且架构上可以把任务交给后端的推理模型。

It boasts better handling of noisy environments and interruptions, as well as architecture that hands off tasks to a back end reasoning model.

这样模型就可以在后台完成任务，同时继续对话。

This allows the model to complete tasks in the background while still carrying on the conversation.

这次API发布还附带了一个很酷的演示视频，展示了这个模型对开发者来说为什么有用。

The API release came with another nifty launch video demonstrating why this model could be useful for developers.

OpenAI展示了模型在用户几英尺外拍篮球的情况下，依然能识别语音的能力。

OpenAI showed off the model's ability to detect voice even with a basketball being dribbled a few feet away from the user.

视频还展示了模型在对话的同时，能控制机器人和显示屏，演示了同时使用多个工具的能力。

It also shows the model can control a robot and a display while carrying on the conversation demonstrating multiple simultaneous tool use.

这个模型现在音频部分每分钟收费5美分，后端模型按标准API价格另外计费。

The model is now available for 5¢ per minute for the audio model and standard API pricing for the backend model on top.

OpenAI的Tibo写道，你现在可以在我们为ChatGPT十亿用户提供的同一语音系统上进行开发。

Wrote OpenAI's Tibo, you can now build on top of the same voice system that we shipped to a billion users in ChatGPT.

全双工系统加上可靠的工具调用，有很多有趣的应用程序。

So many fun applications of a full duplex system with solid tool calling.

用过这个之后，再回到纯文本体验，感觉比我想象的要难。

Going back to text only experiences after this feels harder than I had expected.

AI实验者Alex Finn写道，这其实是我期待已久的发布。

AI experimenter Alex Finn writes, This is actually the release I've been waiting for for a long time.

黑客的梦想成真了。

A hacker's dream come true.

你现在可以在他们发布之前就打造一个乔尼·艾维风格的OpenAI设备。

You can now build a Johnny Ive OpenAI device before they even release it.

我打算在亚马逊上买一个小开发屏幕和麦克风。

Going to be buying a small dev screen slash mic on Amazon.

插到电脑上，告诉Astra把它改装成Live One API设备。

Plug it into my computer, tell Astra to hack it to put Live One API on it.

现在无论你走到哪里，都可以直接跟设备说话，让它帮你干活。

Now anywhere you go, you can just talk to the device and have it do work for you.

基本上，ChatGPT的语音功能无处不在。

Basically, ChatGPT's voice everywhere you go.

给它下命令，它就开始在你家里的电脑上干活。

Give it commands, it starts doing work on your home computer.

你现在基本上可以打造自己的个人助理了。

You can basically now build your own personal assistant.

在亚马逊上有很多便宜的方法，用小型开发设备就能搞定，搜一下就能找到。

Tons of cheap ways to do this on Amazon with small dev devices, just search for them.

像这样的东西让AI变得有趣，我觉得我们太不注重玩得开心了。

Things like this just make AI fun, and I don't think we focus enough on having fun.

所以说到谁该用这个产品，很多方面来说，这其实就是对现有许多语音体验的一次直接升级。

So when it comes to who should be using this product, in a lot of ways, this is just going to be a strict upgrade to many existing voice based experiences.

显然，这对客服主管、呼叫中心、以及那些因为忙于工作而错过很多电话的小型服务企业来说，会非常有价值。

Obviously, this is going to be extremely valuable for customer support leaders, contact centers, small service businesses that have a lot of missed calls because they can't be talking on the phone while they're doing their work.

但我也觉得，大家应该更多地尝试把语音引入到新的体验中去。

But I also think that people should be experimenting a lot more with bringing voice into new types of experiences.

我总的看法是，随着语音技术越来越好，人们会越来越多地从打字转向跟设备说话。

My general sense is that as voice technology gets better, people are going to shift more and more from typing to their devices to talking to their devices.

这种转变其实已经在发生了，这是代际性的，而且说实话，某种程度上，之所以还没大规模普及，就是因为像Siri这样的原生语音识别技术一直都很差。

This is a shift that's already been happening, it's generational, and honestly, to some extent, the only reason it hasn't happened more is that the native voice recognition technology in places like Siri have been so bad historically.

随着更多不同类型的用户开始习惯语音，这会带来很多新的可能性。

As more different types of users get comfortable with voice, it opens up a lot of possibilities.

比如说，对于B2B销售和入站营销团队来说。

For example, for a b to b sales and inbound marketing teams.

很多这样的体验都是从强迫人们预约演示开始的，但也许你可以让感兴趣的潜在客户先聊聊他们的需求，在进入演示阶段之前就得到相关的产品介绍。

So many of those experiences start with forcing people to schedule a demo, but maybe instead you could just let interested prospects talk through their needs and receive a relevant product explanation before they even get to the demo stage.

显然，这里在语言学习和更广泛的辅导方面有巨大的机会。

It obviously feels like there's a ton of opportunity here for language learning and for tutoring more broadly.

这为现实练习和模拟场景创造了一些非常有趣的机会，特别是当你在构建让人们练习工作中会遇到的特定互动类型的体验时。

It creates some really interesting opportunities for realistic practice and simulation type of scenarios to the extent you're building experiences where people are practicing a particular type of interaction that's going to come up in their jobs.

而这些只是冰山一角。

And these are just the tip of the iceberg.

已经有一家公司在GPT Live上构建了产品，那就是Cognition，他们刚刚发布了Devin Voice，宣传语非常简单。

One company that has already built on GPT Live is Cognition, who have just announced Devin Voice with a very simple pitch.

你最喜欢的AI软件工程师现在有了座机。

Your favorite AI software engineer just got a landline.

你说，Devin就交付。

You say it, Devin ships it.

这个功能在语音层由OpenAI的GPT Live驱动，后端则使用Cognition新推出的SWE 2模型。

The feature is powered by OpenAI's GPT Live on the voice layer and Cognition's new SWE two model on the back end.

你可能已经猜到了，我们马上就会讲到SWE 2。

And as you might have guessed, we will get to SWE two in just a minute.

这些用语音进行编程的更新，相比传统的回合制语音模式，真是大大提升了使用体验。

Now these sort of updates while using voice for coding are huge quality of life upgrades over turn based voice modes.

能够打断、随意说、自我纠正，这让整个过程自然多了，也更贴近真实交流。

The ability to interrupt, ramble, and correct yourself just makes the experience a lot more natural and native.

Cognition团队的Nader Davitt说，我日常工作中90%的任务现在都是靠语音完成的。

Nader Davitt from the Cognition team says 90% of my day to day work is now done with my voice.

所以把它作为Devin的一等公民功能，完全说得通。

It only makes sense to make it a first class citizen in Devon.

OpenAI的Kova写道，语音是与AI互动最神奇的方式。

Writes OpenAI's Kova, Voice is the most magical way to interact with AI.

未来，把一个想法变成现实，从把它说出口开始，这真的很美好。

There's something beautiful about a future where bringing an idea to life starts with saying it out loud.

所以在这种情况下，谁应该用这个以及他们应该用它来做什么，就稍微简单一些了。

So in this case, the who should be using this and what they should be using it for is a little bit more simple.

应该使用这个的人，当然就是Devon的用户。

The people who should be using this are, of course, Devon users.

他们应该做的，就是把他们的行为从打字提示和交互，转变为用语音来提示和交互。

And what they should be doing with it is shifting their behavior from typing to prompt and interact to talking to prompt and interact.

即使你不用Devon，我也会借此机会再次恳求你，如果你还没开始用语音，那就开始用吧。

And even if you are not using Devon, I'll use this chance once again to implore you to start using voice if you haven't yet.

一个很好的方式就是，如果你是ChatGPT用户，可以用原生的ChatGPT语音功能。

A really great way to do that is the native ChatGPT voice feature if you are a ChatGPT user.

但即使你用的是其他工具和其他大语言模型，我也强烈建议你在电脑上设置类似WhisperFlow这样的工具，然后开始有意识地尝试把行为转向语音。

But even if you're using other tools and other LLMs, I highly recommend setting up something like WhisperFlow on your computer and starting to very intentionally try to shift your behavior over to voice.

一旦你这么做，我猜你不仅会发现速度提升了，还会发现你可以一边做其他事情，比如散步，一边完成更多工作。

Once you do, my guess is that you'll not only find your speed improved, but you'll also find you can do more of your work while doing other things like taking a walk.

从最初这几项公告可以很清楚地看出，这个行业反正也在推动你往那个方向走。

It's pretty clear from these first couple announcements that the industry is pushing you there anyways.

你不如就去抓住这个新机会吧。

You might as well go surf this new opportunity.

接下来要讨论的新发布是之前提到的SWE 2。

Next new launch to discuss is the previously mentioned SWE two.

这是Cognition最新的内部编程模型，也是他们取代SWE 1.7的第一个主要迭代版本。

It is Cognition's latest in house coding model, and their first major iteration of their model which replaces SWE 1.7.

这个版本是基于Kimi K3进行后训练优化的，专门用于编程，其他方面不做优化。

This version is a post trained version of Kymi K3 optimized for coding and nothing else.

过去一年里，我们已经看到Cognition和Cursor推出了好几代这类模型。

We ve seen a few iterations of these types of models from Cognition and Cursor over the past year.

通常，这些模型的目标是以低廉的价格提供足够好的性能，高效处理日常任务。

Generally, they ve been aimed at delivering a good enough model at a cheap price point to efficiently handle everyday tasks.

实际上，Cursor和Cognition都稍微领先一步，他们意识到我们在能力上已经达到了某个拐点，同时AI承担的任务量也在大幅增加，因此不仅优化纯能力，还要优化效率，这开始变得非常重要。

In fact, Cursor and Cognition were both a little ahead of the curve in understanding that we had reached a certain inflection point in capability, as well as a magnitude of tasks that are being undertaken by AI, where optimizing not just for pure capability but also for efficiency was going to start to really matter.

有趣的是，随着SWE 2的推出，很明显“足够好”的定义在最近几个月已经大幅提高了。

What's interesting is that with SWEET two, it's clear that the definition of good enough has dramatically increased in recent months.

既然这个模型是基于Kimi K3构建的，毫不意外，它的性能接近前沿水平。

Given that the model is built on Kimi K3, it unsurprisingly offers near frontier performance.

比如，这个模型在Frontier Code 1.1上得了50分，略高于GPT56 solo，略低于Fable 5.1，但成本降低了64%。

For example, the model scored 50% on Frontier Code 1.1, putting it slightly ahead of GPT56 sole and just behind Fable 5.1 but at a 64% reduction in cost.

这个模型甚至比基于更小的Kimi K2.7构建的SWE 1.7还要便宜。

The model is even cheaper than SWE 1.7 which was built on the much smaller Kimi K2.7.

Cognition还增加了努力级别，这样你可以降低强度，用起来更便宜。

Cognition has also added effort levels so you can turn down the juice for even cheaper usage.

现在，关于这个时机，有趣的一点是，即使Cursor越来越融入SpaceX AI，导致像OpenAI切断Cursor内部对其模型的访问这样的后果，Cognition却表示他们想保持独立，刚刚以480亿美元的估值筹集了20亿美元。

Now, one of the interesting things about the timing for this is that even as Cursor gets further integrated into SpaceX AI, leading to consequences like OpenAI cutting off access to its models inside Cursor, Cognition has signaled that they want to stay independent, just raising $2,000,000,000 at a $48,000,000,000 valuation.

我猜这意味着，一些重视独立性和能够不受约束地选择他们想要模型的开发者，可能会去看看Cognition的产品，比如Devon。

My guess is that that means some number of developers who prize independence and the ability to select the models they want without constraint might go check out Cognition's products like Devon.

当然，这次新模型的发布也提供了一个好机会，去看看这个更便宜的原生模型在开发环境、桌面和CLI中能做什么。

And this new model release, of course, creates a good context to go see what's possible with this cheaper native model that lives inside the dev and desktop and CLI.

基本上，问题是谁应该试试这个，以及他们应该用它来做什么。任何已经在尝试构建一个完整的模型栈，以更好地将能力与需求对齐并控制成本的人，可能会对看看这个感兴趣。

Basically, it comes to who should try this and what they should use it for, anyone who is already in the midst of trying to build a complete model stack that can better align capability with need to keep costs down might be interested in checking this one out.

说到省钱模型，我们下一个发布来自DeepSeek，他们推出了新版本的flash模型，就是v 4.1 flash。

And speaking of models to keep costs down, our next release comes from DeepSeek who have dropped a new version of their flash model with the release of v 4.1 flash.

这个模型有5520亿参数，专为高延迟任务设计。

The model has five fifty two billion parameters and is designed for high latency work.

基准测试成绩看起来相当不错，在DeepSUI上得了74.2分，和GPT56 SOUL以及OPUS five差不多。

The benchmarks look pretty solid with a score of 74.2 on DeepSUI, which puts it right in line with GPT56 SOUL and OPUS five.

在Terminal Bench 4.0上，这个测试目前还很难刷分，v 4.1 flash得了31.2%，而SOL是39.9%，OPIS是74%。

On Terminal Bench 4.0, which is so far much less bench maxable, v 4.1 flash scored 31.2% compared to 39.9 for SOL and 74% for OPIS.

不过，这个模型定价在市场上是最低的，输入每百万token只要30美分，输出每百万token是1.2美元。

Still, the model is priced at the bottom of the market, charging 30¢ per million input tokens and a buck 20 per million output tokens.

有意思的是，Artificial Analysis发现，这个flash模型在智能指数上其实比DeepSeek全尺寸的Pro模型表现还好，但成本只有四分之一。

Interestingly, artificial analysis found that the flash model actually outperforms DeepSeek's full size Pro model on the intelligence index at a quarter of the cost.

这是Artificial Analysis新出的4.3版指数，最高分是Fable 5.1和GPT six Astra，都是53分，而DeepSeek 4.1离前沿还有距离，总分只有40。

Now this is artificial analysis's new version 4.3 index, where the top score is Fable 5.1 and GPT six Astra at 53, and DeepSeek 4.1 is still away off the frontier with an overall score of 40.

这大概和GPT-five 0.6 Luna以及Gemini 3.8 Flash差不多，再次说明这可能是适合那些想构建更全面架构的团队的模型。

That puts it roughly in line with GPT-five 0.6 Luna and Gemini 3.8 Flash, and again means that this is likely a model for teams that are looking to build a more holistic architecture.

特别是对于那些基础但高频、计算密集型的任务，这个模型很值得一试。

Especially for teams that have basic but highly recurring, compute intensive types of tasks, this is one that's likely going to be worth a test.

接下来，我们再次回到OpenAI的世界，他们推出了小企业插件合集。

Next, we move once again back over into OpenAI world with the introduction of the Small Business Plugin Collection.

这个其实并不复杂。

Now, this is not particularly complex.

OpenAI把他们一些有用的插件打包成了一个小企业套装。

OpenAI has collected some of their useful plugins as a pack for small business.

里面包括Dropbox、HubSpot、Canva、Figma、Shopify、DocuSign、PayPal、QuickBooks、Stripe、Gusto、Slack、Wix、Mercury等等。

It includes things like Dropbox, HubSpot, Canva, Figma, Shopify, DocuSign, PayPal, QuickBooks, Stripe, Gusto, Slack, Wix, Mercury, etcetera.

这其实就是为了让使用ChatGPT的小企业主们能更方便，把所有东西都放在一个容易找到的地方。

Now this is really just trying to make things a little bit simpler for small business owners who are using ChatGPT by putting everything all in one easily and accessible place.

不过有意思的是，有些人开始意识到，在智能体时代，插件的整个概念已经变得不太一样了。

What's interesting though is that some people are recognizing that the entire idea of plugins gets a little bit different when it comes to the agentic era.

针对OpenAI总裁Greg Brockman发的关于插件合集的帖子，Abdul Wasi写道：插件1.0之所以失败，是因为它们只是聊天框后面的API包装器。

Responding to OpenAI President Greg Brockman posting about the plugin collection, Abdul Wasi writes, Plugins v one died because they were API wrappers behind a chat box.

这次，智能体可以运行整个工作流程。

This time, the agent can run the whole workflow.

不过，阿卜杜勒也指出，尚未解决的仍然是分发问题。

However, Abdul also points out, the unsolved part is still distribution.

中小企业主不会去浏览插件目录。

SMB owners don't browse plugin directories.

不过，既然这些插件现在被集中在一起，那正是这类用户应该去了解一下的。

Although given that they are now collected, that is exactly the type of user who should check this out.

如果你是小企业主或者团队很小，而且已经在用这些工具，看看ChatGPT里的原生集成能做什么，可能会让你的体验大大提升。

If you are a small business owner or have a very small team and you are using these already, seeing what the sort of native integrations in ChatGPT can do could be a significant upgrade to your experience.

OpenAI还发布了ChatGPT for Finance，并且对GPT-6 Astra进行了重大更新。

OpenAI has also released ChatGPT for Finance, with a big update for GPT six Astra.

这个产品的新版本仍然是面向金融专业人士的数据连接器和技能包，但已经针对OpenAI的新产品线进行了更新。

The new version of this product is still a bundle of data connectors and skills built for financial professionals, but it's been updated for OpenAI's new product lineup.

该产品现在位于ChatGPT工作区内，并内置了高级数据源访问权限，包括Dilupa、PitchBook、LSEG News和Crunchbase。

The product now lives within ChatGPT work and includes built in access to premium data feeds including Dilupa, PitchBook, LSEG News, and Crunchbase.

这意味着用户不需要自己去搞懂MCP连接，几乎不用怎么配置就能直接上手。

This means that users don't need to figure out MCP connections and can get started right away with minimal configuration.

这个产品还针对ChatGPT的企业安全和管理控制进行了设计，让组织层面可以控制数据访问。

The product is also designed for ChatGPT's enterprise security and governance controls to allow for organizational level control of data access.

OpenAI与摩根士丹利和Evercore合作进行了重新设计，力求更贴近金融专业人士的需求。

OpenAI partnered with Morgan Stanley and Evercore for the redesign, trying to closely align it to the needs of financial professionals.

它主要用于构建金融模型、开展研究和制作客户材料。在新闻发布会上，产品副总裁Nick Turley表示，我们实际上是在教ChatGPT像分析师一样做研究，并且像分析师一样为结论提供依据。

It's designed for tasks including building financial models, developing research, and building client materials, and during a press briefing, VP of Product Nick Turley said, We're effectively teaching ChatGPT to research like an analyst and back up its conclusions like an analyst as well.

它原生使用GPT六Astra，OpenAI计划随着新模型的推出不断更新它。

It uses GPT six Astra natively, and OpenAI intends to keep it updated for new models as they become available.

OpenAI的Ryan Brewer写道，通过ChatGPT金融服务版，我们已经完成了索引所需数据的繁重工作，并让模型可以轻松访问这些数据以进行金融分析。

OpenAI's Ryan Brewer wrote, with ChatGPT for financial services, we've done the hard work of indexing the data that you need and making it easily accessible to the model for financial analysis.

我们打造了定制化的图表体验、更好的引用、SEC文件查看器等等。

We've created a custom charting experience, better citations, an SEC filing viewer, and much more.

Sundeep Srivastava写道：这可不是一个披着金融外衣的通用聊天机器人。

Sundeep Srivastava writes: It s not a general chatbot with a finance skin.

它瞄准的正是初级投资银行家们每周花时间做的事情：公司研究和股票分析、LBO建模和买方筛选、以及按照公司自己的模板格式化的推介材料和客户演示文稿。

It s aimed squarely at what junior investment bankers spend their weeks doing: company research and equity analysis, LBO modeling and buyer screening, pitch books and client decks formatted to the firm s own templates.

现在，正如你可能预料的那样，这引发了一些关于那些初级员工是否会被取代的疑问。

Now, as you might expect, this is opening up some questions about whether those junior associates are going to be replaced.

但我认为重要的是要指出，认为高级银行家会因为智能体在这些核心工作上变得更擅长，就自己去做模型和幻灯片，这种想法并不现实。

But I think it's important to note that it's not particularly realistic to think that high level bankers are going to be making their own models and slide decks, just because the agents have gotten much better at that core work.

仍然存在问责、迭代等问题，而这些是更高级别的银行家不愿意去做的。

There's still questions of accountability, iteration, and all these things which more senior level bankers are not gonna wanna do.

相反，我们在金融行业看到的是，初级银行家开始被要求展示出AI熟练度。

Instead, what we're seeing in the finance industry is that junior bankers are starting to be required to demonstrate AI proficiency.

例如，瑞银最近开始要求应聘的初级银行家展示AI熟练度，作为招聘条件。

UBS, for example, recently started demanding that prospective junior bankers show proficiency in AI as a hiring requirement.

换句话说，我不认为这个产品是初级银行家的替代品。

In other words, I don't think this product is a replacement for junior bankers.

我认为它是初级银行家的一个强力工具，能让他们在工作中更高效、更出色。

I think it is a power tool for junior bankers, which is going to allow them to be much more efficient, and much better at their jobs.

OpenAI 最后发布的是他们为 ChatGPT Work 推出的新数据代理。

The last one from OpenAI is their new Data Agent for ChatGPT Work.

这个代理专为处理组织内部的专有数据而设计，支持连接 Amazon Redshift、Datadog、Google BigQuery、ClickHouse、Databricks、MongoDB 和 Snowflake 等数据源。

The agent is designed for handling proprietary data within an organization, with connections to data providers Amazon Redshift, Datadog, Google BigQuery, ClickHouse, Databricks, MongoDB, and Snowflake.

OpenAI 表示，这些功能可以用来导入销售数据，并围绕销售转化率和留存率等核心指标生成洞察。

OpenAI says the features can be used to ingest sales data and generate insights around core metrics like sales conversions and retention.

目标就是实现那种能直接跟数据对话、做真实分析，而不用再去碰其他工具的愿景。

The goal is to realize the promise of being able to talk to your data and perform real analysis without needing to touch additional tools.

Tableau 的首席产品官 Southern Jones 说：把 Tableau 和 ChatGPT Work 连接起来，就把可信的业务语义带到了员工已经工作的环境里。

Southern Jones, the Chief Product Officer at Tableau, said: Connecting Tableau with ChatGPT work brings trusted business semantics into a place employees already work.

这样他们得到的答案，是基于团队所依赖的同一个数据模型的。

So the answers they get are grounded in the same data model their teams rely on.

用户可以通过提问、探索证据，以及利用内置的设计和分析最佳实践，直接把新的视图发布到 Tableau，轻松把洞察转化为行动。

Users can easily transform insights into action by asking questions, exploring evidence, and publishing new views directly to Tableau using built in design and analytics best practices.

退一步看，OpenAI 这些功能更新汇总起来，都是在针对特定类型的业务用户，更好地连接他们工作所需的工具和上下文，通过更好的工具接入和上下文连接，让更多工作流转移到 ChatGPT 变得更容易，也更有吸引力。

Taking a step back then, if you look at the aggregate of these feature updates from OpenAI, they are all about honing in on specific types of business users, better connecting both the tools and context they need to do their work, and through that better tool access and better context connections make it vastly easier and frankly more inviting to move more of their workflows into ChatGPT.

这不是一个新趋势，但我认为我们以后会看到更多这样的情况。

This is not a new trend, but it's certainly one that I think we're going to see a lot more of.

换句话说，就是垂直化，真正理解特定类型的人是如何工作的。

In other words, verticalization that really understands how very specific types of people work.

这个趋势的另一个小例子是Grockbot，他们这周刚宣布，因为原生集成了Salesforce、HubSpot、Gong、Clay、Granola和其他市场推广工具，现在对销售团队来说更加强大。

Another small example of that trend is Grockbot, who just announced this week that they are now more powerful for sales teams because of native integrations with Salesforce, HubSpot, Gong, Clay, Granola, and other go to market tools.

在这里，我们看到垂直化、工具接入和上下文管理这些趋势并不是OpenAI独有的，当然，整个行业都在共享这些趋势。

And in this, we see the verticalization, tool access, and context management trends are not unique to OpenAI, but are, of course, shared across the industry.

在我们结束之前，还有两个快速的内容。

Two more quick ones before we get out of here.

第一个是Cursor中的项目功能。

The first is projects in Cursor.

这个有点意思，因为虽然我们都习惯了在ChatGPT和Claude中使用项目功能，但这个功能把它带到了稍微新的领域。

Now, this is a little interesting because while we're all used to using projects in ChatGPT and Claude, this feature takes it to slightly new places.

它不仅仅是一个围绕特定项目积累上下文的地方，这个功能还设计用来容纳定时任务、长时间运行的代理和编排。

Rather than just being a place to accumulate context around a particular project, this feature is also designed to house scheduled tasks, long running agents, and orchestration.

换句话说，项目不再只是一个文件夹，而是一个真正的项目架构。

In other words, projects aren't just a folder, they are an actual project architecture.

Cursor 的“项目”标签页采用了与 GrokBot 类似的方法，让用户可以启动一个持久运行的代理，按需执行工作。

The Projects tab in Cursor uses a similar approach to GrokBot, allowing users to spin up persistent agent that performs work as required.

在发布视频中，Cursor 表示：它已经从回合制聊天中毕业了。

In their launch video, Cursor said: It's graduated from turn based chat.

你不再需要事无巨细地管理每个代理，而是与一个更加自主的同事合作。

Instead of you micromanaging every agent, you're working with a much more autonomous colleague.

比如，如果你在打开 PR 后有一个标准工作流程，项目可以在你下次触发该流程时自动弄清楚如何执行并完成任务。

For example, if you have a standard workflow after opening a PR, the project can figure out how to do and carry out the tasks automatically the next time you trigger the workflow.

Cursor 表示，测试人员使用项目功能后，合并的 PR 数量是原来的六倍，合并率提升了 30%。

Cursor said that testers have merged six times as many PRs using projects and merge rates have increased by 30%.

这个功能最初显然主要面向开发者，但非开发者的知识工作者也值得关注其中的模式。

Now this is one where the obvious users initially are going to be developers, but the patterns are worth paying attention to for non developer knowledge workers as well.

Prasanjit Sarkar 写道：从按任务聊天到持久协调者的转变，在我看来，是 Cursor 项目中比原始子代理数量更重要的变化。

Prasanjit Sarkar writes, The move from chat per task to persistent coordinator is, in my read, the more important shift in Cursor projects than the raw sub agent count.

我之前看过的每个编程代理都是这么工作的。

Every coding agent I've watched before this worked the same way.

打开一个会话，描述任务，代理执行，会话结束。

Open a session, describe a task, agent executes, session ends.

Cursor 的项目功能把这个模式颠倒过来了。

Cursor projects inverts the model.

一个协调线程在整个项目生命周期内保持开启。

One coordinator thread stays open for the life of a project.

协调者不写代码，它负责规划、分配子代理，并把结果带回来检查。

The coordinator doesn't write code, it plans, delegates the sub agents, and brings results back to check.

正如 Press 和 g 指出的，这并非独特的架构。

Now as Press and g points out, this isn't a unique architecture.

有 Cognition 的 Devon、OpenAI Codecs 等，但他认为持久线程和主动触发系统是不同的。

You've got Devon by Cognition, OpenAI Codecs, and others, but argues that the persistent thread and the proactive trigger system are different.

我觉得持久线程这个模式很有意思。

Now the persistent thread pattern is one that I think is interesting.

我们很多构建项目都得在不同实例之间手动传递上下文，比如 Cloud Code，因为一个实例的上下文窗口满了，就得重新开始。

A lot of our build projects involved having to hand off context between different instances of something like Cloud Code because the context window in one was filled and you needed to start fresh.

到了2026年，情况发生了很大变化。

That has changed a lot in 2026.

Kansar 描述了具体的变化。

Kansar described what changed.

他写道：2025年第四季度，我主导了一项重要工作，教 Block 的工程师们高级上下文工程，以便充分利用 Claude Code 和 Co。

He wrote: In Q4 of twenty twenty five, I led a big push to teach block engineers advanced context engineering to get the most out of Claude Code and Co.

三月和四月，我们开始大规模转向使用 codecs。

In March and April, we started switching to codecs in large numbers.

到了5.4，尤其是5.5版本，它能解决其他工具无法解决的问题，构建出其他工具无法构建的东西。

With 5.4 and especially 5.5, it could solve problems and build things that other tools could not.

但我最喜欢的是它在上下文压缩方面的出色表现。

But my favorite was how good it was at compaction.

不需要那些花哨的子代理技巧或主动压缩，你只要一直继续下去就行。

Instead of all the fancy sub agent tricks and intentional compaction, you could just keep going and going.

我有一个线程，它构建了一整套同步协议、服务器和客户端，跨越多个仓库和语言，还把它们都部署了。

I had one thread that built a whole sync protocol, server, and client across multiple repos and languages, and also deployed them.

从那以后，压缩功能越来越好。

Since then, compaction continues to get better and better.

现在单线程是个热门话题，但我还是惊讶于在一个线程里能完成这么多事情，而且没有任何问题。

Mono threads are a common topic now, but I'm still surprised by how much you can do in one thread without any issues.

我发现自己在Claude和Codex里都更常使用这种模式。

This is a pattern I found myself using a lot more in both Claude and Codex.

比如，我把AI每日简报网站相关的所有内容都放在一个很长的线程里，这个线程做了大量工作，运行了很长时间，而且从来没有遇到过上下文问题。

I have, for example, everything related to the AI Daily Brief website pretty much in one long thread that has done a ton of work and has been going for a very long time and hasn't really ever run into any context issues.

所以提醒大家，有时候新产品或功能的发布，与其说你是否会使用那个特定功能，不如说它表明了整体发展方向。

So the reminder is that sometimes a new product or feature launch is less about whether you are going to use that particular feature, and more about what it says about where things are headed in general.

好了，朋友们，这就是对最近几天发布的产品和功能的快速回顾，我觉得它们适合谁，有什么用，希望你们在周末前有一些新想法。

Anyways, folks, that is a quick rip through the products and features that have launched in the last couple of days who I think they're for, what I think they're useful to do, and hopefully, you have some new ideas heading into the weekend.

今天就到这里，这就是今天的AI每日简报。

For now, that's gonna do it for today's AI daily brief.

感谢你的收听或观看，一如既往。下次再见，祝一切顺利。

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