Why a New Class of AI “Judgment Models” Could Have Big Business Implications — 文稿

M10:00

不是每天都能得到一个新模型来玩，也不是每天都能遇到一种全新的模型构建方法，而且这种方法对我们如何使用它有着相当不同的影响。

It's not every day that we get a new model to play around with, and it's certainly not every day that we get an entirely new approach to model building with some fairly different implications for how we even use it.

而今天，我们要讨论一种新的模型类别，你可以称之为AI判断模型。

Today though, we are talking about a new class of models which you might refer to as AI judgment models.

这些判断模型——比如我们今天要讨论的来自Typesafe的Jev——不是生成大段文本，而是围绕特定问题输出概率值。

Rather than producing long strings of text, these judgment models, like the one we're discussing today, Jev from Typesafe, produce probabilities around specific questions.

这个客户生气了吗？

Is this customer angry?

这封邮件里有没有新的依赖关系？

Is there a new dependency in this email?

我们需不需要因此改变运营计划？

Do we need to change the operational plan because of this?

今天我们要探索这些模型背后的理念、它们与众不同的训练方式、它们如何能更快更便宜地做出判断，以及最重要的是，它们将如何融入你的整体模型栈。

Today we're exploring the idea behind these models, how they're trained differently, how they can produce these judgments much more quickly and much less expensively, and most importantly, where they're going to fit in your overall model stack.

AI每日简报是一个关于AI领域最重要新闻和讨论的每日播客和视频。

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

好了朋友们，在开始之前先快速说几条通知。

Alright friends, quick announcements before we dive in.

M11:18

AI安全方面的讨论继续在科技行业和主流社会中发酵。

The AI safety discourse continues to trickle out through the tech industry as well as mainstream society.

不过目前，除非发生什么惊天动地的大事，我们会把这个话题放到新闻速报里，而不是做成主题节目。

But for now, unless something absolutely seismic happens, we're gonna move it into the headlines and away from the main episode.

基于此，在周末保持沉默之后，马克·扎克伯格已经就AI减速这个想法表明了自己的观点。

With that in mind, after staying quiet over the weekend, Mark Zuckerberg has made his thoughts known on this idea of an AI slowdown.

周二，扎克伯格在X上发帖写道，每个实验室都有责任和动力按照训练模型所需的安全速度前进，并且有能力采取自己的行动来确保这一点。

On Tuesday, Zuckerberg wrote in a post on X, every lab has the responsibility and incentive to move at the pace required to train its models safely and the ability to take its own actions to ensure that happens.

基本上，他的观点是，节奏是各个实验室的责任，而不是集体行动。

Basically, his view is that pacing is the responsibility of individual labs rather than a collective action.

这个观点基于两个核心想法。

And that view hinges on 2 core ideas.

第一，人们不想使用与自身不一致、不按他们要求行事的智能体。

First, that, quote, people don't want to use agents that are misaligned with them and that don't do what they ask.

所以实验室有很强的天然动力来让他们的模型更加对齐。

So labs have a strong natural incentive to make their models more aligned.

第二，如果模型造成伤害，实验室将面临重大责任。

And 2, labs face significant liability if their models cause harm.

所以他们也有很强的动力来防止这种情况。

So they have a strong incentive to prevent this as well.

为了强调这一点，扎克伯格表示，Meta已经将Muse的发布推迟了几个月，以进行安全方面的工作。

Emphasizing the point, Zuckerberg said that Meta had delayed the release of Muse by several months to work on safety.

他接着说，我们没有要求其他人在我们之前这样做。

He continued, we didn't call for everyone else to do this before we would.

我们只是把它作为日常工作的一部分，因为这显然是对人们正确的事情。基本上，扎克伯格的意思是，现有的个人激励和后果已经足以迫使AI实验室进行对齐工作，并正确把握前沿速度，而不需要某种外部政府强制实施的减速。

We just did it as part of our day-to-day work because it was clearly the right thing for people and Essentially, Zuckerberg is saying that the individual incentives and consequences that are already in place are enough to force AI labs to work on alignment and to pace the frontier correctly, rather than needing some exogenous government-enforced slowdown.

扎克伯格确实支持独立评估者和顾问的理念，认为这是最佳实践，而非监管要求。

Now, Zuckerberg did support the idea of independent evaluators and advisors as a matter of best practice rather than regulation.

他声称Meta已经聘请了外部评估人员，不是因为被要求，而是因为这有助于产出更好的成果。

He claimed that Meta has already engaged outside evaluators, not because it's required of them, but because it helps produce better work.

最后，他总结道，将绝大部分计算资源用于服务人类，而不是竞相追求递归自我改进，是确保我们安全开发这项技术的最佳方式之一。

Finally, he concluded, committing the significant majority of compute towards serving people, rather than racing towards recursive self-improvement, is one of the best ways to ensure that we develop this technology safely.

Meta已经做出了这一承诺，其他实验室也可以这样做。

Meta has made this commitment, and other labs can do this as well.

我相信为每个人建设一个积极的未来的关键，是维持正确的权力平衡。

I believe the key to building a positive future for everyone is maintaining the right balance of power.

这在我们能力范围内。

This is within our power to do.

措辞很谨慎，但基本上整段话的意思就是：得了伙计们，别整那些虚的了。

It's carefully worded, but basically this whole thing says, come on guys, let's please stop with the theatrics.

而且说实话，很多人觉得这像是一股清流。

And a lot of people frankly found this a breath of fresh air.

YouTuber Joseph Carlson 写道，慢着，你们是说公司可以放慢速度，确保安全，而不用告诉所有竞争对手也要放慢？

YouTuber Joseph Carlson wrote, hold up a minute, you are telling me companies can slow down, make sure things are safe without telling all their competitors to slow down?

我觉得大方向上，大家的反应可以分成两类。

And I would say that broadly speaking, reactions fell into one of two categories.

第一类就像刚才那个，还有 Matthew Berman 这条。

The first was like that one and like this from Matthew Berman.

喜欢这个。

Love this.

模型安全和对齐是一个特性，而且是有经济激励的。

Model safety and alignment is a feature and economically incentivized.

还有 Bill Ackman，他直接说这是开发 AI 的正确方式。

Or Bill Ackman, who simply called this the proper approach to AI development.

但另一方面，也有人认为 Zuckerberg 根本没有信誉或立场来做这个论证，无论论证本身有没有道理。

But then on the other side were those arguing effectively that Zuckerberg just does not have the trust or standing to make this argument, regardless of the merits of the argument itself.

Hard Fork 的 Kevin Roose 写道，不管你是民主党还是共和党，是 EA 还是 EACC，我想我们都同意，最适合保护我们免受强大新技术危害的人是 Mark Zuckerberg。

Hard Fork's Kevin Roose wrote, whether you're a Democrat or a Republican, an EA or an EACC, I think we can all agree that the person best suited to protect us against the harms of powerful new technology is Mark Zuckerberg.

与此同时，在 Strange Bedfellows Daily 那边，Bernie Sanders 和 Steve Bannon 联手呼吁以人为中心的 AI 监管，堪称这个政治周期里最奇怪的联盟。

Now meanwhile, over in Strange Bedfellows Daily, Bernie Sanders and Steve Bannon have joined forces to call for human-centric AI regulations in the strangest alliance of this political cycle.

这两位高度意识形态化的领袖周二在华盛顿的未来生命研究所举办的亲人类大会上，同台发表了讲话。

The 2 highly ideological leaders spoke from the same stage on Tuesday at the Future of Life Institute's Pro-Human Assembly in Washington.

Sanders 对人群说，如果每个男人、女人和孩子的生活都会被这项技术从根本上改变，那么这个国家的人民必须做出关于 AI 的决定，而不是仅仅几个寡头。

Sanders told the crowd, if the lives of every man, woman, and child are going to be fundamentally changed by this technology, then the people of this country must make decisions about AI and not just a handful of oligarchs.

Bannon 发表了非常类似的言论，他说美国公民再也不会向寡头们低头了。

Bannon had very similar remarks, stating, the American citizens are not going to be supplicants to the oligarchs anymore.

我们做不到。

We can't do it.

这是历史的转折点。

This is a hinge in history.

我们必须正确应对。

We have to handle this correctly.

要正确应对，第一，永远不能相信寡头说的话。

To handle it correctly, number one, we can never trust what an oligarch says.

我得说，虽然乍一看这两位高度意识形态对立的人同台有点奇怪，但他们所关联的更大运动在历史上是有不少共同背景的。

Now, I will note that while this seems strange at first, these 2 highly ideologically opposed people sharing the same stage, the broader movements they connect to have a fair bit of shared context in history.

2016 年总统大选——Bernie 以微弱劣势输给 Hillary 未能成为民主党候选人，而 Bannon 显然策划了第一届 Trump 政府——这两件事都植根于后全球金融危机金融格局下的民粹主义愤怒，以及对制造这场经济危机的机构和机构领导人缺乏问责。

The 2016 presidential election, in which Bernie was narrowly beaten by Hillary to miss out on becoming the Democrat nominee, and in which Bannon obviously architected the first Trump administration, both had their roots in populist anger at the post-GFC financial landscape And the lack of accountability for the institutions and institutional leaders who were involved in creating that particular economic crisis.

显然，左派和右派对那个特定背景的反应不同，但我想说，他们最终在某个点找到共同立场，可能比你想象的更不令人惊讶。

Obviously, the left and the right's reaction to that particular context were different, but it is, I would contend, perhaps less surprising than you think that at some point these two would find common ground.

需要说明的是，我并不是在否定他们基于这个问题本身的优点以及这个问题打乱现有政治联盟的力量所找到的共同立场。

And to be clear, I am not dismissing the common ground that they find just on the merits of this particular issue itself and the power of this particular issue to scramble existing political alliances.

我只是想说，他们的故事实际上比你第一眼看到的更加交织在一起。

I'm just making the point that their stories are actually more intertwined than you might think at first glance.

不管怎样，这场活动似乎不太关注本周充斥头条的 AI 存在风险，而更多关注技术所代表的阶级斗争。

In any case, the event seemed somewhat less about the existential risks of AI that have filled the headlines this week and more about the class struggle that technology has come to represent.

电视屏幕上播放了一段对虚构 AI CEO 的恶搞采访，描述为“一个爱人类但不怎么喜欢人类的人”。

TV screens played a parody interview from a fictional AI CEO described as someone who loves people but isn't crazy about humans.

整个活动中，似乎 AI 本身并不是需要解决的风险，而是寡头不受约束的权力。

Throughout the event, it seemed that AI itself wasn't the risk that needs addressing, but rather the unchecked power of oligarchs.

你也许还记得，八月份记者 Jasmine Sun 走遍全国，采访了那些真正反对数据中心的人。

Now, you might remember that back in August, journalist Jasmine Sun toured the country speaking to real people who were working in opposition to data centers.

她发现最常见的抱怨不是用电、用水或噪音污染，而是缺乏控制感，觉得未来是被强加给社区的，几乎没有发言权。

And found that the most common complaint was not about electricity, water use, or noise pollution, but a lack of control and the sense that the future was being forced upon their communities with little input.

这次活动也是这样。

The same was true about this event.

多位发言人都没有把 AI 的风险放在网络安全、生物恐怖主义或其他存在风险上。

Across multiple speakers, the risk of AI was not framed around cybersecurity, bioterrorism, or other x-risks.

而是关于缺乏决定未来形态的自主权。

It was about a lack of agency in determining the shape of the future.

由此引出了他们主要的担忧，简单来说就是失控。

From this followed their main concern, which was simply the idea of losing control of AI.

就像德克萨斯州民主党议员 Greg Casar——他正在支持伯尼·桑德斯的超级智能法案——说的那样，答案很简单。

As Texas Democrat Greg Casar, who is sponsoring Bernie Sanders' superintelligence bill, said, the answer is simple.

我们禁止那些强大到人类无法控制的人工智能系统。

We ban AI systems that are too powerful for humans to control.

尽管最近这些言论听起来很激烈，但你可能觉得奇怪，我仍然认为我们已经进入了一个新阶段，这个阶段的关键在于谈判关系——公民和政府都在其中有利益——而且既然现在基本上大家都认可这一点，那么下一阶段很可能会包含更多的具体性，我敢说还有更多的微妙之处。

And as strange as it might seem, despite all the intensity of this rhetoric recently, I still contend that we've moved into a new phase that's all about negotiating the relationship where citizens and governments have a stake in this, and where, given that that is now pretty much where everyone is, the next phase is likely to include a lot more specificity and dare I say nuance.

比如 Glenn Beck，他在节目中花了很长时间说，他既可以和他们一起签署支持人类的人工智能宣言，同时又可以在数据中心的具体问题上不同意他们的观点。

Glenn Beck, for example, had a long monologue on his show about how he can both have signed the Pro-Human AI Declaration alongside them, but also disagree with them on specifics of data centers.

说实话，虽然听起来很疯狂，但我认为，政治讨论越是让你脑袋发懵、觉得混乱又分散，那反而可能是一个信号——我们其实在取得进步。

And honestly, as crazy as it sounds, I think that the more that the political discourse kind of frags your brain for how confusing and all over the place it is, that might just be a sign that we're actually making progress.

这个话题也出现在了 Salesforce 的年度 Dreamforce 大会上。会上公司发布了一个新的人工智能模型，社交媒体上热议的焦点却是 Sam Altman、Jensen Huang 和 Dario Amodei 的出现，他们各自阐述了自己的安全观点。

The conversation also found its way into Salesforce's annual Dreamforce event, where the company unveiled a new AI model, but where a lot of the chatter on social media focused on appearances from Sam Altman, Jensen Huang, and Dario Amodei, who each iterated their own safety views.

Dario 说他只是试图提出一套行业可以共同遵循的标准，而 Altman 说他对自己公司和整个行业安全地做到这一点非常有信心。

Dario argued that he was just trying to put forward a set of standards that the industry could organize around, while Altman said that he was very confident in our company's ability and our industry's ability to do this safely.

与此同时，Jensen Huang 则重申了类似扎克伯格的看法：尽你所能地快速前进，但如果你在任何一个时间点觉得公司失控了，或者产品不安全了，就停下来，确保你做对了。

Jensen Huang, meanwhile, reiterated the Zuckerbergian view, saying, run as fast as you can, but if you feel at any given point in time the company's out of control or the product's not going to be safe, take a pause and make sure you get it right.

至于 Salesforce 的 CEO Marc Benioff，他认为每家公司都有责任维护道德标准。

And as for Salesforce CEO Marc Benioff, he believes that every company has a responsibility to uphold ethical standards.

不过，先不讨论安全争议，Salesforce 还公布了两项重大的实用人工智能消息。

Still, to give the safety debate a bit of a rest, Salesforce also had two big practical AI announcements.

首先，他们发布了一段时间以来的第一个自研模型。

First, they're releasing their first in-house model in quite some time.

这个模型叫做 Koa，是 NVIDIA 的 Nemotron 的一个微调版本，专门用于 CRM 系统内的销售管理。

Called Koa, the model is a fine-tune of NVIDIA's Nemotron and is designed to handle sales management within the CRM.

这一发布强调了开源在企业中所能发挥的作用——它让这种狭窄的垂直模型成为可能。

The announcement reinforces the role that open source has to play in the enterprise by enabling this sort of narrow vertical model.

其次，Salesforce 推出了一项新举措，叫 AI Force。

Secondly, Salesforce unveiled a new initiative called AI Force.

这将是一个总称，涵盖 Salesforce 的各类连接器，允许第三方智能体访问 Salesforce 的数据。

This will be the umbrella term for Salesforce's connectors that will allow third-party agents to access Salesforce data.

这次发布也再次表明了 Salesforce 的承诺：以平台无关的方式走向无头软件，让任何智能体都能成为界面。

The release reinforces Salesforce's commitment to moving towards headless software in a platform-agnostic way, allowing any agent to become the interface.

显然，这些都是重要且会持续的大讨论，但现在，我们的头条新闻就先到这里。

Now, obviously, these are big conversations that are important and are going to continue, but for now, that is where we will close the headlines.

接下来是本期的主节目。

Next up, the main episode.

M111:55

欢迎回到 AI Daily Brief！

Welcome back to the AI Daily Brief!

在今天的正片中，我们要聊一个非常不同、相当罕见的东西——简单说，是一种完全不依赖大语言模型的 AI 新方法。

On today's main episode, we are looking at something really different and quite rare, which is, in short, a totally different approach to AI that is not just another LLM.

在 ChatGPT 发布后的将近四年里，以及这档节目播出的三年半里，我们和整个行业关注的大部分内容，在某种程度上都与大语言模型相关。

Now, in the nearly 4 years since ChatGPT was released and the 3 and a half that this show has been around, the vast majority of the things that we and the rest of the industry have focused on have been in some ways related to large language models.

不过，正如一些人指出——尽管他们的声音常常没有得到多少关注，差不多是在对着虚空呐喊——大语言模型实际上并不是人工智能的全部。

And yet, as some have pointed out, although often in a way that didn't get much traction and was more or less screaming into the void, LLMs are not in fact the totality of artificial intelligence.

昨天，Diogo Almeida 在 X 上发文说：在共同发明了 ChatGPT 之后，我一直在问自己，为什么超人类聊天模型没有带来 AGI？

Yesterday, Diogo Almeida posted on X, after co-inventing ChatGPT, I kept asking myself, Why have superhuman chat models not led to AGI?

过去的两年里，我一直在秘密开发一种训练模型的新方法：RLCD，也就是用于校准决策的强化学习，以及我们今天发布的一种新型前沿 AI 模型：Jev。

I've spent the last 2 years in stealth building a new way to train models, RLCD, or Reinforcement Learning for Calibrated Decisions, and a new type of frontier AI model that we are releasing today: Jev.

Jev 的速度快 20 到 200 倍，成本低 40 到 400 倍，输出 token 完全免费，是一种面向决策优化的前沿可组合智能。

Jev is 20 to 200 times faster, 40 to 400 times cheaper with output tokens free, frontier composable intelligence optimized for decisions.

在我看来，这是通往基于 AI 的经济革命的最短路径。

As far as I can tell, the shortest path to AI-based economic revolution.

当然，看到像 20 到 200 倍更快、40 到 400 倍更便宜这样的数字，人们感到怀疑是完全情有可原的，至少可以说。

Now, one could absolutely be forgiven for seeing numbers like 20 to 200 times faster and 40 to 400 times cheaper and being a bit skeptical of the claims, to say the least.

如果这只是一个普通的大语言模型，那么这种怀疑完全合理。

And were this just another LLM, that skepticism would be entirely warranted.

不过，对于 Jev 和这种新策略来说，我们面对的是完全不同的东西。

However, with Jev and this new strategy, we're dealing with something that is quite different.

开发 Jev 的公司叫 Typesafe，在他们的官方公告博客中，他们进一步介绍了自己的方法有何不同。

The company behind Jev is called Typesafe, and in the announcement blog post, they talk a little bit more about what makes their approach different.

他们写道，我们构建了一个全新的、完全专注于自动化的技术栈，采用了新的模型架构、并行采样器以实现最高效率，以及我们称之为‘用于校准决策的强化学习’的训练方法。

They write, we built a new stack entirely focused on automation, with a new model architecture, parallel sampler for maximum efficiency, and training method we call reinforcement learning for calibrated decisions.

现有的LLM优化目标是人类偏好，也就是人类评分员更喜欢的写作和聊天回复，而新的System 1模型，第一个是Jev，则优化校准决策，即带有认知诚实概率的答案。

Whereas existing LLMs optimize for human preference, i.e., write-ups and chat responses that human raters prefer, the new System 1 models, the first of which is Jev, optimize for calibrated decisions, or answers with epistemically honest probabilities.

那么这到底意味着什么呢？

So what does that actually mean?

我们来看看Mike Taylor在Every上是怎么描述的。

Well, let's look at how Mike Taylor from Every describes it.

他写道，你可以把它想象成一个聪明的‘如果-那么’语句，当你在自动化工作流程时，它决定下一步该做什么。

He writes, Think of it as a smart if-then statement that determines what happens next when you're automating a workflow.

假设你在构建一个优先处理客服请求的软件，你写了代码去问模型：‘这个客户听起来生气吗？’

Say you're building software that prioritizes customer service requests, and you write code that asks the model, does this customer sound angry?

Jev可能会回答0.9，这意味着基于模型在训练中学到的内容，答案有90%的概率是‘是’。

Jev might answer 0.9, which means there's an estimated 90% probability that the answer is yes based on what the model learned in training.

你也可以提供你定义的类别，比如烦躁、恼怒、被冒犯、愤怒和暴怒，然后得知客户有60%的概率被归类为愤怒，只有10%的概率是暴怒。

You could also provide categories you define, like annoyed, irritated, offended, furious, and enraged, and learn that the customer was 60% likely to be classified as furious, with only a 10% probability of being enraged.

在解释这为什么重要以及它与LLM有何不同时，Mike继续说，如果答案是0.9，表示非常可能生气，软件可能会自动将客户的问题升级给经理。

Going on to explain why this matters and how it differs than LLMs, Mike continues, with an answer of 0.9, very likely to be angry, the software might automatically proceed to escalate the customer concern to a manager.

或者如果它回答0.1，表示不太可能生气，那么这个请求可能会被降低优先级。

Or if it answers 0.1, not likely to be angry, that request might be deprioritized.

然而，聊天机器人被训练成用华丽的文字回复，比如‘你说得完全正确，这个客户听起来确实很生气。’

However, chatbots are trained to respond with flowery text like, you're absolutely right, this customer does sound very angry.

‘要不要我用友好支持的语调帮你草拟一份邮件回复？’

Would you like me to compose a draft email response in a friendly, supportive tone?

这种文本回复会导致你正在构建的程序崩溃，因为它期望的是一个介于0和1之间的数字，而不是一篇文章。

This text response would cause the program you're building to crash because it was expecting a number between 0 and 1, not an essay.

这使得一类决策成为可能，这类决策既不适合笨拙无智能的代码，也不适合缓慢昂贵的LLM。

This allows a class of decision-making that was neither suited to dumb, unintelligent code nor to slow expensive LLMs.

正如Chubby总结的那样，JEV是一个为决策而非文本生成而构建的AI模型。

As Chubby sums up, JEV is an AI model built for decisions rather than text generation.

需要特别指出，这里的取舍是这个模型不生成文本。

And it is important to note that the trade-off here is that this model does not generate text.

换句话说，它并不是LLM的通用替代品。

It is not, in other words, a replacement for LLMs in general.

它替代的是LLM执行的某一类工作，这些工作就像是把方形的钉子硬塞进圆孔里来完成。

It's a replacement for a certain category of work that LLMs do, where they've been very square-peg-mashed into a round hole to do it.

Chubby接着说，这个想法是把快速且廉价的AI决策嵌入到软件中。

The idea, continues Chubby, is to embed fast, cheap AI decisions into software.

那么这个模型到底是用来做什么的呢？

So what is this model actually built for?

想想看，办公室工作中有多少是阅读某样东西，然后决定下一步该做什么。

Well, think of how much of office work consists of reading something and deciding what should happen next.

这条消息需要回复吗？

Does this message need a response?

应该由哪个部门处理？

Which department should handle it?

这份文档回答了问题吗？

Does this document answer the question?

这个客户描述的是bug还是请求新功能？

Is this customer describing a bug or asking for a feature?

这份草稿是否声称其来源并未支持的内容？

Does this draft make a claim its source doesn't support?

这个情况是常规到可以自动处理，还是应该有人来审核？

Is the situation routine enough to automate or should someone review it?

这些都是关于意义的判断，而且通常很难用固定的规则来表达。

These are judgments about meaning, and they're often difficult to express as fixed rules.

因此，Jev的设计目标是获取相关信息，并以概率、类别或分数的形式回答定义明确的问题。

Jev, then, is designed to take the relevant information and answer narrowly defined questions with probabilities, categories, or scores.

然后，周围的软件利用这些答案来路由、排序、标记或继续处理。

The surrounding software then uses those answers to route, rank, flag, or proceed.

在TypeSafe的文档中，他们明确建议将复杂的决策拆解成小问题，然后在代码中组合结果。

In the TypeSafe documentation, they explicitly recommend breaking complex decisions into small questions and then combining their results in code.

那么这在正常业务中会如何体现呢？

So where would this show up in normal business?

一个明显的领域是客户支持，模型能做的简单判断可能是，客户是否感到沮丧，而之前的回复是否未能解决这个问题。

One obvious area is customer support, where the small judgment the model could make would be something like, is the customer frustrated, and have previous replies failed to address it?

有了这些小判断，软件接下来就可以对工单进行路由、提高其优先级，或者请求人工审核。

With those small judgments in hand, the software could then next route the ticket, raise its priority, or request human review.

在销售领域，模型可以判断，这是否是购买咨询？

In the sales domain, the model could judge, is this a buying inquiry?

潜在客户是否符合产品？

Does the prospect fit the product?

他们是否在请求会议？

Are they requesting a meeting?

然后软件可以利用这些判断来分类进来的线索并分配后续跟进。

The software could then take those judgments to sort inbound leads and assign follow-up.

在市场和编辑领域，模型可能判断文案是否符合特定的风格规则，或者优惠是否清晰。

In marketing and editorial, the model might judge whether the copy meets specific style rules or whether the offer is clear.

围绕它的软件随后可以在发布前标记需要修改的段落。

The software that surrounds it could then flag passages for revision before publication.

重要的是，很多人关注的不仅是他们会在哪里使用这个而不是LLM，而是他们如何与LLM一起使用它。

And importantly, where a lot of people went was not just understanding where they would use this instead of LLMs, but how they might use it alongside LLMs.

YC的创始人Nathan Fleury写道，我可以想象很多工作流程看起来像LLM提出选项，Jev决定，代码执行。

YC founder Nathan Fleury wrote, I'd imagine a lot of workflows that look like LLM proposes options, Jev decides, code executes.

举一个明确的例子，想象一个客户写道：“这是我第三次联系你们。

To put a clear example on this, imagine that a customer writes, "This is the third time I've contacted you.

我们仍然无法导出报告，而且我们的续约在下周。”一个由类似Jev支持的支持工作流程可以同时询问几个问题：一，客户是否在描述一个产品问题？

We still can't export our reports, and our renewal is next week." A support workflow supported by something like Jev could ask several questions together: One, is the customer describing a product problem?

二，消息是否表明反复且未成功的技术支持？

Two, does the message indicate repeated unsuccessful support?

三，是否有商业上重要的截止日期即将到来？

Three, is a commercially significant deadline approaching?

四，哪个团队最有能力提供帮助？

Four, which team is best equipped to help?

周围的软件可以结合这些信号与实际的账户信息，比如续约日期。

The software that surrounds it could then combine those signals with actual account information, such as the renewal date.

并升级工单。

And escalate the ticket.

从那里，你仍然会有一个生成模型来起草回复，但是Jev可以再次根据狭窄的标准检查草稿：回复是否承认了反复的联系？

From there, you would still have a generative model draft the reply, but Jev once again could check the draft against narrow criteria: does the response acknowledge the repeated contacts?

是否解决了导出问题？

Does it address the export problem?

是否承诺了基于提供的信息无法支持的内容？

Does it promise something unsupported by the information provided?

人们对于这种新方法所开启的部分兴奋点在于，廉价的判断使得频繁检查变得更加实际。

And part of what people are excited about opening up with this new approach is that cheap judgment makes frequent checking more practical.

如果检查增加了明显的延迟或费用，团队可能只会在选定的案例或任务结束时运行它。

If a check adds a noticeable delay or expense, a team may run it only on selected cases or at the end of a task.

如果它变得足够快速和便宜，它可以在每个进来的请求上、每次草稿修订后、在代理采取关键步骤之前的许多候选文档上运行。

If it becomes sufficiently fast and inexpensive, it could run on every incoming request after each draft revision across many candidate documents before an agent takes a consequential step.

回到Mike Taylor在Every上的文章，他给了Jev所有27篇文章的文本以及10个故意采用AI风格的对立点，然后同时在所有文章上询问了相同的21个问题来检查AI的迹象。

Back in Mike Taylor's article from Every, he gave Jev the text from all 27 of his articles alongside 10 deliberately AI-styled counterpoints then asked the same 21 questions concurrently across all articles to check for AI tells.

基本上，他用Jev所做的检查是，这篇文章是否做了某些特定的事情，向人们表明这是由AI创作的？

Basically, the check that he was doing with Jev was, does this essay do specific things that indicate to people that it is AI composed?

比如，文本是否重复一个观点而没有增加证据？

Things like, does the text repeat an idea without adding evidence?

是否强行对称的双方论证？

Does it force a symmetrical both-sides argument?

是否对一个简单的观点过度解释？

Does it over-explain a straightforward point?

在不到0.7秒内，Mike说，Jev，所谓的，读完了全部37份文档，并且针对每份文档回答了全部21个问题。

In less than 0.7 seconds, Mike said, Jev quote-unquote read all 37 documents and answered all 21 questions for each.

返回了777条判断结果，花费估计只有四分之一美分。

Returning 777 judgments for an estimated quarter of a cent.

正如他指出的，这个速度快、成本低，足以用AI检查你公司所有人写过的所有内容，并且瞬间就能拿到结果。

As he points out, that's fast and cheap enough to AI-check everything everyone at your company has ever written and get the results back in an instant.

Mike做的类比是，这就像知识工作中的代码检查工具。

A comparison that Mike makes is a code linter for knowledge work.

他写道，在软件开发中，代码检查工具是一种分析你的工作、几乎立即标记出语法错误、捕捉bug、发现不良模式并强制风格一致性的工具。

He writes, in software development, a code linter is a tool that analyzes your work and almost instantly flags syntax errors, catches bugs, spots bad patterns, and enforces stylistic consistency.

Typesafe的模型在将模糊任务转化为清晰、结构化的答案方面非常快，以至于它可以作为知识工作中的一种代码检查工具。

Typesafe's model is so fast at turning fuzzy tasks into clear, structured answers that it could act as a kind of code linter for knowledge work.

让Codex或Claude接入Jev和一系列问题，它们就能快速检查自己的工作，看是否出现了你要求避免的问题。

Give Codex or Claude access to Jev and a list of questions, and it can quickly check its own work for problems you've told it to avoid.

大多数软件本质上就是一棵巨大的判断树：如果是这种情况，就做那个；如果是这种情况，就路由到这里；如果是这种情况，就升级处理；如果是这种情况，就拒绝；如果是这种情况，就询问人类。

Most software is ultimately a giant tree of if this, do that, if this, route here, if this, escalate, if this, reject, if this, ask a human.

Jev本质上在问：如果这些if语句能够理解混乱的人类上下文会怎样？

Jev is basically asking, what if those if statements could understand messy human context?

相比于再搞一个AI模型，这个框架要有趣得多。

That's a much more interesting framing than another AI model.

我认为这对于欺诈和风险评估、支持路由、内容审核、公关和QA自动化、潜在客户评分、合规、工作流编排以及代理路由等方面非常有用。

I can see this being very useful for fraud and risk, support routing, moderation, PR and QA automation, lead scoring, compliance, workflow orchestration, and agent routing.

显然，他说，这是早期技术，但这个类别本身很有道理。

Early tech, obviously, he says, but the category itself makes a lot of sense.

有趣的是，Matt Stockton指出，在某些方面，采用这种技术的公司相当于在用一种后LLM时代的AI技术，让前LLM时代的机器学习技术变得更容易使用。

Now, interestingly, Matt Stockton points out that in some ways, companies adopting this amounts to a post-LLM AI technology making pre-LLM machine learning techniques a little bit more accessible.

正如他所写，商业中的很多很多问题都是分类或回归问题。

As he writes, lots and lots of problems in business are classification or regression problems.

很多很多公司并不知道他们遇到的问题类型可以通过经典的机器学习方法解决。

Lots and lots of companies don't know that the types of problems they have are solvable by classic machine learning methods.

他们往往用人和流程来解决问题，而不是技术。

They often solve them with people and process instead of technology.

随着LLM的出现和普及，越来越多的公司开始想，也许我们可以用AI来做这件事，于是用LLM来解决分类和回归问题。

With the emergence and popularity of LLMs, more companies are thinking, maybe we can use AI for that, and are solving classification and regression problems with LLMs.

这在某些方面是好的，因为公司可能实现了一些手动工作的自动化，但在某些方面也是坏的，因为这往往是错误的工具，可能不如经典的ML方法好使。

This is good in some ways because companies are potentially automating some manual work, but also bad in some ways because it's often the wrong tool for the job, and possibly not as good as classic ML methods for what they are trying to do.

但是，他指出，经典技术需要你标注数据、训练模型，然后把模型托管在某个地方。

But, he points out, the classical techniques require you to label your data, train a model, and host that model somewhere.

它们不像调用LLM API那么容易使用。

They aren't as easy to use compared to calling an LLM API.

而且这需要你和你的组织了解这些技术并有能力投入其中。

And it requires you and your org to be aware of those techniques and capable of investing in them.

不用过度延伸这个类比，他基本是说，看待Jev的一种方式是，它提供了一种用户体验，让LLM风格的用户交互模式用于经典ML技术。

Without going too far on the analogy, he basically says one way to look at Jev is as a UX for using LLM-style user interaction patterns for classical ML techniques.

现在，一个有趣的问题是，这究竟是面向个人还是构建系统的团队。

Now, one interesting question that comes up is whether this is for individuals or teams building systems.

简短的答案是，虽然它当然两者都适用，但它也突出了我们最近一直在讨论的多玩家AI主题。

And the short answer is that while it is absolutely both, it also puts a fine point on the multiplayer AI themes that we've been talking about recently.

当然，个人可以使用这种能力来拥有个人工具，比如根据个人优先级对收件箱排序、根据编辑标准检查写作草稿、根据研究兴趣对收藏文章排序、或者在会议记录中标记承诺事项，诸如此类，能帮助个人更好地完成工作。

Certainly individuals could use this sort of capability to have personal tools that sort an inbox against their own priorities, check drafts of their writing against an editorial rubric, rank saved articles against research interests, or flag commitments in meeting transcripts, things like that that are going to personally help you do your work better.

但我认为，这类技术真正大放异彩的地方，是在团队协作领域——通过关于谁需要知情、谁应该行动、以及需要谁的批准这样的小判断来实现的。

But I think that where this sort of technique is going to really shine is in the domain of teamwork that happens through small judgments about who needs to know, who should act, and whose approval is required.

当代理服务于个人时，仅仅通过了解那个人的偏好就能取得很大进展。

When an agent is serving a single person, it gets pretty far simply by learning that person's preferences.

然而，跨团队运作的代理需要理解人们工作之间的关系。

An agent operating across a team, however, needs to understand the relationships between people's work.

这产生了一系列不同的问题。

That creates a different set of questions.

谁负责这个？

Who owns this?

这项工作影响到哪些人？

Whose work does this affect?

有人正在等待这个决定吗？

Is someone waiting on this decision?

这个承诺是否会给其他团队带来义务？

Does this promise create an obligation for another team?

是当前负责人能决定，还是需要更广泛的共识？

Can the current owner decide, or does this need broader agreement?

在某种程度上，有趣的工作单元变成了交接。

In some ways, the interesting unit of work becomes the handoff.

想想一个销售人员对客户说，我们应该能在你续约之前支持那个集成。

Consider a salesperson telling a customer, we should be able to support that integration before your renewal.

对于销售人员的个人代理来说，下一步可能很直接：更新客户记录，起草跟进邮件，并创建提醒。

For the salesperson's personal agent, the next steps might be straightforward: update the account record, draft a follow-up, and create a reminder.

但在组织内部，这句话涉及多个责任。

But inside the organization, that sentence implicates several responsibilities.

对销售团队来说，这句话意味着他们可能有机会确保续约，也就是支持那个集成。

For the sales folks, that sentence means for them a potential way to secure the renewal, i.e., supporting that integration.

对工程团队来说，这句话可能意味着涉及不确定工作的交付承诺。

For engineering, that sentence means a possible delivery commitment involving uncertain work.

对产品团队来说，这可能导致路线图优先级的改变。

For product, it means a potential change to roadmap priorities.

对客户成功团队来说，这是一个他们可能必须管理的期望。

And for customer success, it's an expectation they may have to manage.

一个多方代理需要识别这句话为可能的跨团队承诺，而一个判断模型可以评估具体问题：这条消息是否暗示了交付承诺？

A multiplayer agent would need to recognize the sentence as a possible cross-team commitment, and a judgment model could assess specific questions: Does this message imply a delivery promise?

这个承诺是否涉及说话者权限之外的工作？

Does the promise concern work outside the speaker's authority?

这是否与提供的路线图冲突？

Does it conflict with the supplied roadmap?

是否有证据表明负责团队已同意？

Is there evidence that the responsible team agreed?

更大的系统然后可以创建提议的承诺，确定必要的负责人，并请求缺失的决定。

The larger system could then create a proposed commitment, identify the necessary owners, and request the missing decisions.

现在再次强调成本与收益。

Now again, to reinforce costs and benefits.

JEV 和这类判断模型的重大成本在于，从某种意义上说，这是一个定义上不完整的类别，它本身就是一个不完整的类别。

The big cost of JEV and this type of judgment model in general, to the extent that this becomes a category, is that it is an incomplete category by definition.

它无法完成我们目前让生成式 AI 做的所有工作。

It cannot do all the work that we currently have generative AIs do.

判断模型必须成为更复杂模型架构的一部分，也就是我们过去几个月一直在讨论的那种模型堆栈。

Judgment models are going to have to be part of a more complex model architecture, the type of model stack that we've been discussing for the last several months.

当然，好处是它能以极低的成本做到这一点。

The benefit, of course, though, is that it can do this extraordinarily inexpensively.

而且因为这种判断智能可以如此廉价地应用，这意味着如果做得好，它可以极其深入地集成到我们都在构建的自动化系统中。

And because this sort of judgment intelligence can be applied so cheaply, it means that done well, it can be integrated incredibly deeply into the automated systems we're all building.

这显然只是一个新概念的起步阶段，而这个概念非常重要，以至于这个非常能干的团队已经花了两年时间进行研究。

This is obviously just the first day of a very new concept, and a concept which is significant enough that this very competent team has spent 2 years working on.

所以显然我们需要看看它在实践中会如何发展，但当你深入研究时，它确实给人一种既重要又明显的感觉，就像那种一旦出现，我们将来会对它存在了这么久感到惊讶的东西。

So obviously we're going to need to see how it all plays out in practice, but it does have the feel when you dig in of something both important and obvious, the type of thing that once it exists, we will be surprised in the future that we didn't have it for so long.

当然，我会持续关注这个，并且会继续寻找人们如何使用它的更多例子，以及公司在尝试构建这些新型系统时遇到的挑战。

Certainly I'm going to be keeping an eye on this and I will continue to look out for more examples of how people are using it, as well as where companies are running into challenges as they try to build these new types of systems.

不过现在，这真的是很酷的东西值得去关注，今天的 AI Daily Brief 就到这里。

For now though, very cool stuff to go check out, and that's going to do it for today's AI Daily Brief.

一如既往地感谢你的收听或观看。

Appreciate you listening or watching as always.

下次见，peace。

Until next time, peace.

Peace。

Peace.

已剔除 2 处广告（点击展开查看）

M10:54广告 · 已剔除

First of all, thank you to today's sponsors, KPMG, Blitzy, Section, and HyperAgent.

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

Subscriptions are just $3 a month for ad-free.

And if you want to learn about sponsoring the show, send us a note at [email protected], or just go to aidailybrief.ai where you can learn all about it.

M18:58广告 · 已剔除

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.

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.

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

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

Here's why most legacy modernization projects fail.

The AI doing the work can't understand codebases at scale.

It sees a small slice of context, examines syntax, and misses years of decisions distributed across the global application ecosystem.

Blitzy solves this the way it solves everything.

Grounded in your code before any migration begins, Blitzy's agents reverse engineer the entire legacy system into a persistent knowledge graph.

Every dependency, every constraint, every piece of tribal knowledge that used to live in one engineer's head.

From that understanding, Blitzy autonomously executes language migrations, framework upgrades, and monolith-to-microservices transformations, all validated end-to-end.

One Blitzy customer modernized a $10 million monolithic insurance stack in 16 weeks against a 137-week baseline with coding agents.

That's 9x compression.

Retire technical debt while accelerating your roadmap.

See how at blitzy.com.

That's blitzy.com.

Here's a harsh truth: your company is probably spending thousands or millions of dollars on AI tools that are being massively underutilized.

Half of companies have AI tools, but only 12% use them for business value.

Most employees are still using AI to summarize meeting notes.

If you're the one responsible for AI adoption at your company, you need Section.

Section is a platform that helps you manage AI transformation across your entire organization.

It coaches employees on real use cases, tracks who's using AI for business impact, and shows you exactly where AI is and isn't creating value.

The result?

You go from rolling out tools to driving measurable AI value.

Your employees move from meeting summaries to solving actual business problems.

And you can prove the ROI.

Stop guessing if your AI investment is working.

Check out Section at sectionai.com.

That's S-E-C-T-I-O-N-A-I dot com.

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

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

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

Marketing agents turn competitor moves into landing pages.

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

Ops agent chases the paperwork and tracks the budget.

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

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

Hire yours at HyperAgent.

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