一个男人用人工智能治愈了爱犬的癌症*

今天在《AI每日简报》中，我们将聚焦于那位利用人工智能治愈爱犬癌症的人，以及这反映了AI第二波浪潮中的舆论生态。

Today on the AI Daily Brief, all about that guy who used AI to cure his dog's cancer and what it says about the discourse in AI's second moment.

在那之前，先来看一下新闻头条，以及英伟达GTC大会的前瞻内容。

Before that are the headlines, a preview of NVIDIA's GTC.

《AI每日简报》是一档每日更新的播客和视频节目，聚焦人工智能领域最重要新闻与讨论。

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

好了，朋友们，在正式开始前，先做几个简短的公告。

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

首先，感谢今天的支持方：毕马威、Blitzy、AIUC 和 PromptQL。

First of all, thank you to today's sponsors KPMG, Blitzy, AIUC, and PromptQL.

如需收听无广告版本，请前往 patreon.com/aideallybrief，或在 Apple 播客订阅本节目。

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

如想了解节目赞助事宜，请发送邮件至 sponsorsaidailybrief dot ai。

To learn about sponsoring the show, send us a note at sponsorsaidailybrief dot ai.

同时，访问 aideallybrief.ai，你可以了解这个生态系统中正在发生的各种动态。

And while you are at aideallybrief.ai, can find out all about all the various things going on in this ecosystem.

这周最大的新闻当然是Agent Madness。

The big one this week is of course Agent Madness.

这是一个类似三月疯狂的淘汰赛赛制，我们将通过现场人类和智能体投票，选出今年你们用直觉编码和构建的最酷项目。

It's a March Madness style bracket where we will be having live human and agentic voting on the coolest things that you have vibe coded and built this year.

除了荣誉之外，这些智能体还将在节目中亮相。

In addition to bragging rights, will feature these agents on the show.

如果你对此感兴趣，请访问agentmadness.ai。

So if you are interested in that, check out agentmadness.ai.

目前提交截止日期定于3月18日。

Currently submissions are slated to close on March 18.

也就是本周三。

That is Wednesday of this week.

所以，赶紧去agentmadness.ai吧。

So again, get on over to agentmadness.ai.

这周对NVIDIA来说至关重要，因为他们的GTC开发者大会将在圣何塞开幕。

It is a big week for NVIDIA as their GTC developer conference kicks off in San Jose.

CEO 黄仁勋原定于周一上午发表主题演讲，因此当我们这一期节目发布时，我们可能会了解到更多信息。

CEO Jensen Huang was scheduled to deliver his keynote on Monday morning, so we'll likely know more by the time this episode goes out.

在活动前夕，许多猜测都集中在与 Groq 合作开发的新芯片系统上。

In the lead up to the event, much of the speculation was around a new chip system developed in collaboration with Grok.

G R O Q，不是 G R O K。

That is G R O Q, not G R O K.

带 Q 的 Groq 是一家不属于埃隆·马斯克的公司。

Grok with a Q is the one that is not a Elon Musk company.

英伟达于去年 12 月收购了这家芯片初创公司，并预计将在本周公布首个合作产品。

NVIDIA acquired the chipmaking startup in December and are expected to announce the first collaborative product this week.

相关信息描述了这一新产品将 Groq 的语言处理芯片集成到英伟达的机架级服务器中。

The information described the new product as integrating Grock's language processing chips into NVIDIA's rack scale servers.

如果属实，这将是英伟达首次直接应对推理需求。

If that's the case, this will be NVIDIA's first attempt to directly address inference demand.

到目前为止，英伟达的芯片在 AI 训练领域一直处于世界领先地位，但并未特别关注高效的推理性能。

Until now NVIDIA's chips have been world leading in AI training, but haven't been particularly focused on efficient inference.

这就是Groc发挥作用的地方，它提供专为推理工作负载设计的芯片。

That's where Groc steps in delivering a chip tailored exclusively to inference workloads.

据预计，NVIDIA将宣布OpenAI为这款新芯片的买家。

NVIDIA is expected to announce OpenAI as a buyer of the new chip.

消息人士称，三星的芯片代工厂已开始提升产量，大规模生产预计将在今年下半年启动。

Sources said that production has been ramping up at Samsung's chip foundry, and mass production is expected to begin in the second half of the year.

值得注意的是，这将是NVIDIA首次在台积电之外制造AI芯片，可能有助于将供应链从台湾分散出去。

Notably, this will be the first time NVIDIA has manufactured an AI chip outside of TSMC, potentially diversifying supply chains out of Taiwan.

据消息人士称，新服务器还使用了英特尔的CPU，而非NVIDIA的CPU，这表明NVIDIA的芯片目前与Grok芯片的集成效果不佳。

The new servers also use Intel CPUs rather than NVIDIA CPUs according to sources, which suggests that NVIDIA's chips don't integrate well with Grok chips at this stage.

消息人士补充说，多个代际的硬件正在规划中，未来有可能将Grok的技术整合进NVIDIA的Feynman GPU中，这些GPU是今年晚些时候推出的Rubin之后的下一代产品。

The sources added that multiple generations of hardware are being planned, with the potential to build Grok's technology into NVIDIA's Feynman GPUs, which are the next generation following Rubin later this year.

除了产品发布外，NVIDIA的Neo Cloud合作伙伴也在加速运营。

Outside of product releases, NVIDIA's Neo Cloud partners are stepping up operations.

《The Information》报道称，EnScale正在与西弗吉尼亚州的一个大型数据中心场地进行收购谈判。

The Information reports that EnScale is in negotiations to acquire a huge data center site in West Virginia.

该场地已通过监管审批，目标是在2027年前实现2吉瓦的容量。

The site has cleared regulatory hurdles and is targeting two GW of capacity by 2027.

对于一家Neo Cloud提供商而言，这笔交易有些不同寻常，因为它们过去通常只是租赁数据中心。

Now the deal is a little unusual for a Neo Cloud provider, which have typically rented data centers in the past.

这还将立即使总部位于英国的EnScale成为美国市场的重要参与者，同时为其上市做准备。

It would also immediately make UK based ENScale a major player in The US market as they move towards an IPO.

《The Information》披露的新文件显示，这项收购将使EnScale的2027年收入预期翻三倍，达到300亿美元。

New documents surfaced by The Information said that the acquisition would triple ENScale's revenue projections to $30,000,000,000 for 2027.

据报道，他们正在与字节跳动洽谈租赁容量，但也可能将服务器重新租回给NVIDIA。

They are reportedly in talks to rent the capacity to ByteDance but could also rent their servers back to NVIDIA.

More Insights and Strategy的首席执行官兼首席分析师帕特里克·穆尔黑德写道：NVIDIA不再是一家芯片公司。

Writes More Insights and Strategy CEO and Chief Analyst Patrick Moorhead, NVIDIA is no longer a chip company.

随着GTC 2026的开幕，该公司计划将自己定位为一个完整的异构AI基础设施平台，涵盖训练、预填充、解码、推理和智能体编排。

As GTC twenty twenty six opens, the company plans to present itself as a full stack, heterogeneous AI infrastructure platform, spanning training, pre fill, decode, inference, and agent orchestration.

接下来，尽管今年许多软件公司的首席执行官都在淡化AI对其公司的颠覆性风险，但SEC的文件却讲述了一个不同的故事。

Next up, while many software CEOs have been downplaying the AI disruption risks to their company this year, SEC filings are telling a different story.

今年迄今为止，已有27家公司把AI代理列为对商业模式的重大风险，而去年同期仅有7家。

So far this year, 27 firms have listed AI agents as a material risk to their business model up from just seven this time last year.

警告代理风险的公司名单包括Figma、Workday和HubSpot，它们的首席执行官最近都淡化了这些担忧。

The list of companies warning about agents includes Figma, Workday, and HubSpot whose CEOs have all recently dismissed concerns.

在最近一次财报电话会议上，Figma首席执行官迪伦·菲尔德表示：我认为人类将继续使用软件，而代理也会越来越多地使用软件。

During their most recent earnings call, Figma CEO Dylan Field said: I think it is the case that humans will continue to use software, and increasingly agents will too.

我对这一点感到兴奋。

And I'm excited about that.

但他补充道：我认为，如果你现在愿意把关键任务交由代理处理，并让它们完全无人监管地运行，那你真是非常大胆。

However, he added: I think right now, if you're willing to hand off mission critical work to agents and just let them do it unsupervised, you're a very brave person.

与此同时，Figma在同一天发布的10-K文件中承认，代理式AI可能会改变人们访问和交互数字产品的方式，从而降低对传统软件应用的依赖。

Meanwhile, Figma's 10 ks filing released on the same day acknowledged that Agentic AI may change how people access and interact with digital products in ways that reduce reliance on traditional software applications.

请记住，SEC文件不应被过于字面地理解。

Now keep in mind SEC filings should not be taken too literally.

公司被要求讨论任何可能对其业务构成重大风险的因素，这常常导致披露一些夸张或不太可能的风险。

Companies are required to discuss any material risk to their business, which often leads to disclosures of fanciful or unlikely risks.

尽管单个披露信息并不能说明太多，但这种数量上的增长表明，我们已经越过了关于智能代理的临界点。

Still, while individual disclosures don't tell us all that much, the volume is another signal that we've moved past the tipping point on agents.

去年上半年，人们几乎还没有意识到智能代理可能对SaaS行业造成颠覆。

The idea that agents were capable of disrupting SaaS barely registered in the first half of last year.

然而，随着这项技术变得更加可行，下半年及今年初的披露数量迅速增加。

And yet disclosure volume rapidly increased in the second half and in the beginning of this year as the technology became more viable.

至少，这种转变意味着软件行业的高管们正在更认真地对待颠覆性威胁，或者至少他们的法务部门是这样。

If nothing else the shift means software executives are taking the threat of disruption more seriously or at least their legal departments are.

接下来，字节跳动因版权争议暂停了其前沿视频模型的全球发布。

Next up, ByteDance has paused the global launch of their cutting edge video model due to copyright disputes.

《The Information》报道称，Seed Dance 2.0的全球发布已被搁置，原因是与好莱坞制片公司发生了一系列版权纠纷。

The Information reports the global release of Seed Dance two point zero has been mothballed due to a series of copyright disputes with Hollywood Studios.

Seed Dance 2.0已于上个月在中国发布，引发了巨大的网络反响。

See Dance two point zero was released in China last month, gathering a huge online reaction.

你可能还记得那段广为传播的视频，汤姆·克鲁斯和布拉德·皮特在打斗，展现了对真实演员极其逼真的还原。

You might recall this viral clip with Tom Cruise and Brad Pitt in a fistfight, which demonstrated incredibly high fidelity replication of real world actors.

这一新模型在好莱坞引发了强烈不满，迪士尼、华纳兄弟、派拉蒙和奈飞等公司向字节跳动发送了停止并终止通知。

The new model led to outrage in Hollywood with companies including Disney, Warner Bros, Paramount, and Netflix sending cease and desist notices to ByteDance.

当时，电影协会首席执行官查尔斯·里夫金在一份声明中表示：Seed Dance 2.0 大规模未经许可使用了美国

Motion Picture Association CEO Charles Rivkin said in a statement at the time: Seed Dance two point zero has engaged in unauthorized use of U.

的

S.

受版权保护的作品。

Copyrighted works on a massive scale.

字节跳动原计划于三月中旬在全球范围内推出该模型。

ByteDance had planned to make the model available globally in mid March.

该计划包括通过其云平台字节加提供API访问，以及为海外用户设计的新消费者应用。

The plan included API access through their cloud platform Byte plus as well as a new consumer app designed for a foreign audience.

据称，这些计划目前已被搁置。

Those plans are now reportedly on hold.

与此同时，中国用户报告称，该模型的管控比发布时严格得多，甚至会拒绝与版权内容无关的提示。

Chinese users, meanwhile, are reporting the model as far more tightly controlled than it was at launch, to the point of rejecting prompts with no relation to copyrighted content.

企业客户抱怨说，该模型的访问权限仅限于中国公司，且无意向国际分发内容。

Enterprise customers have complained that model access is limited to Chinese companies with no intention of distributing content internationally.

一位消息人士表示，他们无法在不承诺花费约150万美元的情况下谈妥条款。

One source said they'd been unable to negotiate terms without committing to spending around $1,500,000 on the model.

有趣的是，主要的障碍似乎不在于实施防护措施，而在于优化这些措施，以免阻拦太多无关内容。

Interestingly, it seems like the major holdup is not so much about implementing guardrails, but instead about refining them so that they don't block too much unrelated content.

我们也在OpenAI发布Sora 2时见过类似情况。

We've seen this with OpenAI's release of Sora two as well.

虽然阻止受版权保护的内容相对简单，但要在不因过多拒绝提示而让用户感到沮丧的情况下做到这一点，是一个更困难的工程问题。

While it is relatively straightforward to block copyrighted content, doing so without frustrating the user with too many refused prompts is a much more difficult engineering problem.

说到困难的工程问题，一家由前Anthropic创始人领导的新AI初创公司正在筹集资金，以推动人工智能增强的科学研究前沿。

And speaking of difficult engineering problems, a new AI startup led by former Anthropic founders is raising money to push the frontier of AI enhanced scientific research.

这家名为Mirandil的新公司正在洽谈以十亿美元的估值筹集1.75亿美元。

The new company called Mirandil is in talks to raise $175,000,000 at a billion dollar valuation.

如果成功，这笔融资将使Mirandil成为又一家在种子轮就达成独角兽地位的AI初创公司。

And if successful, the round would make Mirandil the latest AI startup to establish unicorn status in their seed round.

该公司由前Anthropic研究员贝迪姆·内沙瓦尔和哈什·梅塔领导，他们在Anthropic期间主要研究AI代理的长期科学推理和自动化AI研究。

The company is led by former Anthropic researchers Bedim Neshawar and Harsh Mehta, who spend their time at Anthropic working on things like long horizon scientific reasoning with AI agents and automated AI research.

两位创始人都曾在谷歌工作过。

Both founders also have experience at Google.

目前尚不清楚该公司的确切计划，但消息人士称，这家新公司旨在生物和材料科学等领域开展AI增强的科学研究。

Now exactly what the company plans to do is not known yet, but sources say the new company aims to conduct AI enhanced scientific research in fields including biology and materials science.

随着多个NLW实验室专注于AI赋能科学，这一AI研究领域正迅速吸引关注和投资。

This area of AI research is quickly gathering interest and investment dollars as multiple NLW labs focus on AI for science.

我预计这一趋势将在今年持续下去。

I would expect this to be a trend that continues throughout the year.

说到谷歌，谷歌地图正迎来AI升级，推出全新的对话式界面。

Speaking of Google, Google Maps is getting an AI twist with a new conversational interface.

这一新功能名为‘Ask Maps’，允许用户接入由Gemini驱动的聊天机器人，以帮助他们探索世界。

The new feature called Ask Maps allows users to tap into a Gemini powered chatbot to help them navigate the world.

该功能旨在回答有关地标的问题，并协助规划行程。

The feature is designed to answer questions about landmarks and help schedule travel.

谷歌给出了一些实际的小例子，比如询问附近哪里可以给手机充电，或者寻找有灯光的公共网球场以便晚上打球。

Google gave small practical examples like being able to ask for a nearby location to charge a phone or find a public tennis court with lights for an evening match.

该功能还能帮助规划行程，谷歌举例说可以规划一条前往大峡谷的多站式旅行路线。

The feature can also help with trip planning, with Google offering the example of building a multi stop trip to the Grand Canyon.

谷歌表示，过去要找到这些信息需要大量研究和筛选评论，但现在你只需点击‘Ask Maps’按钮，就能以对话方式获得答案，并附带定制地图帮助你直观了解各种选择。

Writes Google, previously finding this information meant lots of research and sifting through reviews, but now you can just tap the Ask Maps button and get your questions answered conversationally and with a customized map to help you visualize your options.

该功能与Gemini的记忆功能集成，因此如果你向Maps询问餐厅推荐，它能调用Gemini已知的你的偏好信息。

The feature integrates with Gemini's memory, so if you ask Maps for a restaurant recommendation, it can tap into what Gemini already knows about your preferences.

谷歌还利用Gemini在Maps中推出了一种新的导航可视化模式。

Google is also leveraging Gemini to launch a new visualization mode for navigation in Maps.

此次更新增加了三维视图，展示建筑物、立交桥和周边地形。

The update adds a three d view that depicts buildings overpasses and surrounding terrain.

再次体现了谷歌在多模态能力及其整个生态系统整合方面的实力。

Once again Google flexing its multi modality and the integration of its entire ecosystem.

最后，今天要提一下，这算是为我们主要节目的过渡话题：ServiceNow首席执行官比尔·麦克德莫特警告称，人工智能可能导致年轻专业人士的失业率飙升至30%以上。

Lastly today, sort of a bridge topic to our main episode: ServiceNow CEO Bill McDermott has warned that AI could send unemployment soaring above 30% for young professionals.

在与CNBC的采访中，麦克德莫特表示，未来几年内，大学毕业生的失业率很容易攀升至30%中期。

In an interview with CNBC, McDermott said that unemployment for college graduates could easily go into the mid-30s in the next couple of years.

他继续说，大量工作将由智能代理完成，因此年轻人在职场中脱颖而出将变得极具挑战。

So much of the work is going to be done by agents, he continued, so it's going to be challenging for young people to differentiate themselves in the corporate environment.

根据美联储的数据，目前新近毕业的大学生失业率为5.6%，远低于没有大学学历的年轻人7.8%的失业率。

Now, according to data from the Federal Reserve, unemployment for recent college graduates currently stands at 5.6%, which is far lower than the 7.8% unemployment rate for young people without a college degree.

然而，42.5%的大学毕业生被归类为过度教育就业，即他们工作不足，或从事不需要大学学历的工作。

However, 42.5% of college graduates are classified as underemployed, meaning they don't have enough work or are working in roles that don't require a college degree.

这是自2020年以来大学毕业生过度教育就业率的最高水平。

This is the highest level of underemployment for college grads since 2020.

计算机科学专业的毕业生失业率高达7%，但其过度教育就业率相对较低，为19.1%，低于其他专业。

Computer science majors have among the highest unemployment rates at 7%, but their underemployment rate is relatively low at 19.1% compared to other majors.

这种言论如今如此具有影响力，正是我们主节目的主题。

Now just why this type of discourse is so potent right now is in fact the topic of our main episode.

好了，以上就是今日要闻，接下来我们进入主节目。

So with that, we will close the headlines and move on over to the main.

代理式AI正在推动一场价值3万亿美元的生产力革命，企业领导者正面临关键决策点。

Agentic AI is powering a $3,000,000,000,000 productivity revolution and leaders are hitting a real decision point.

你是要自建AI代理，还是购买现成的，或者通过合作借用外部资源来加速扩展？

Do you build your own AI agents, buy off the shelf, or borrow by partnering to scale faster?

普华永道最新发布的前瞻性研究报告《应对构建、购买或借用的决策》很好地拨开了迷雾，提供了一个实用框架，帮助你根据价值、风险和准备度做出选择，并指导如何通过正确的信任、治理和编排基础来扩展代理。

KPMG's latest thought leadership paper, Navigating the Build, Buy, or Borrow decision, does a great job cutting through the noise with a practical framework to help you choose based on value, risk, and readiness and how to scale agents with the right trust, governance, and orchestration foundation.

不要锁定错误的模式。

Don't lock in the wrong model.

现在就可以访问 www.kpmg.us/navigate 下载这份报告。

You can download the paper right now at www.kpmg.us/navigate.

再次提醒，网址是 www.kpmg.us/navigate。

Again, that's www.kpmg.us/navigate.

随着2022年AI代码生成技术的出现，英伟达的首席发明家、哈佛工程师西德·帕雷希提出了一个反主流的观点。

With the emergence of AI code generation in 2022, NVIDIA master inventor and Harvard engineer Sid Pareshi took a contrarian stance.

推理阶段的计算资源和代理编排，而非预训练，才是解锁企业级高质量AI驱动软件开发的关键。

Inference time compute and agent orchestration, not pre training, would be the key to unlocking high quality AI driven software development in the enterprise.

他认为，真正的突破不在于AI生成代码的速度，而在于它能否深入推理，以构建企业级应用。

He believed the real breakthrough wasn't in how fast AI could generate code, but in how deeply it could reason to build enterprise grade applications.

当全世界都在关注协作者时，他设计了一种根本不同的东西：Blitzy——首个专为大规模企业代码库打造的自主软件开发平台，利用数千个智能代理运行。

While the rest of the world focused on copilots, he architected something fundamentally different: Blitzy, the first autonomous software development platform leveraging thousands of agents that is purpose built for enterprise scale codebases.

财富500强企业领导者借助Blitzy，将工程效率提升了5倍，数天内即可完成原本需要数月的工程工作。

Fortune 500 leaders are unlocking 5x engineering velocity and delivering months of engineering work in a matter of days with Blitzy.

转变您的软件开发方式。

Transform the way you develop software.

前往 blitzy.com 了解详情。

Discover how at blitzy.com.

那就是 blitzy.com。

That's blitzy.com.

我认为，企业AI代理领域即将出现一个全新的标准。

There's a new standard that I think is going to matter a lot for the enterprise AI agent space.

它被称为AIUC-one，是全球首个自建的AI代理标准。

It's called AIUC-one, and it builds itself as the world's first AI agent standard.

它旨在涵盖所有核心企业风险，如数据与隐私、安全、可靠性、问责制以及社会影响，并由可信的第三方进行验证。

It's designed to cover all the core enterprise risks things like data and privacy, security, safety, reliability, accountability, and societal impact, all verified by a trusted third party.

它引起我关注的原因之一是，Eleven Labs——我之前多次提到过这家公司，目前正势不可挡——刚刚成为首个获得AIUC-one认证的语音代理，并推出了全球首个可投保的AI代理。

One of the reasons it's on my radar is that Eleven Labs, who you've heard me talk about before and is just an absolute juggernaut right now, just became the first voice agent to be certified against AIUC-one and is launching a first of its kind insurable AI agent.

这意味着在实践中，它具备实时防护机制，可阻止不安全的回应并防范操纵，同时还配备完整的安全体系。

What that means in practice is real time guardrails that block unsafe responses and protect against manipulation, plus a full safety stack.

这种特性正是推动企业采用的关键。

This is the kind of thing that unlocks enterprise adoption.

当一家基于Eleven Labs构建的公司能够出示第三方认证，并宣称其代理安全、可靠且经过验证时，整个对话的格局就改变了。

When a company building on Eleven Labs can point to a third party certification and say our agents are secure, safe, and verified, that changes the conversation.

访问aiuc.com，了解全球首个AI代理标准。

Go to aiuc.com to learn about the world's first standard for AI agents.

那就是aiuc.com。

That's aiuc.com.

如果你是一名运营者，你的一天将充满不间断的决策。

If you're an operator, your day is a non stop stream of decisions.

而大多数都需要你查看数据。

And most of them require you to look at the data.

你不需要另一个仪表盘。

You don't need another dashboard.

你需要的是快速、可信赖的答案。

You need answers you can trust, fast.

但瓶颈始终如一。

But the bottleneck is always the same.

数据还没准备好。

The data isn't ready.

数据分散在各处。

It's scattered.

数据很混乱。

It's messy.

定义不够清晰。

Definitions aren't clear.

你正在等待数据团队或领域专家提供澄清和确认。

You're waiting on your data team or waiting on domain experts for clarification and confirmation.

今天赞助商PromptQL正是为打破这一瓶颈而打造的。

That's the bottleneck today's sponsor PromptQL is built to break.

PromptQL是一款适用于高频决策的可信赖AI分析师。

PromptQL is a trusted AI analyst for high frequency decision making.

它能够连接多个数据仓库、数据库、SaaS平台和内部API。

It connects across warehouses, databases, SaaS, and internal APIs.

无需进行大规模的数据预处理或集中化。

No massive data prep or centralization required.

它专为多人协作输入而设计。

It's built for multiplayer input.

团队成员可以加入对话线程，纠正假设和细微差别，标记边缘情况。

Teammates can jump into a thread, correct assumptions and nuance, flag edge cases.

PromptQL将日常对话转化为共享的上下文。

PromptQL turns everyday conversations into a shared context.

如果有什么不清楚的地方，它不会猜测。

And if something is ambiguous, it doesn't guess.

它会将问题转交给正确的专家，记录正确的逻辑，并在下一次准确解决。

It escalates to the right expert, captures the correct logic, and gets it right next time.

这就是它如何实现信任与准确性的。

That's how it delivers trust and accuracy.

随着时间推移，PromptQL 会逐渐适应你的业务，就像那位什么都知道的资深员工。

Over time PromptQL specializes to your business, like that veteran employee who just knows things.

从简单的‘是什么’问题到复杂的‘如果怎样’情景，你都可以在做出决策前模拟影响并压力测试。

From simple What is questions to complex What if scenarios, you can model impact and stress test decisions before you commit.

所有这些都通过一个简单的自然语言提示完成。

All through a simple natural language prompt.

PromptQL，为拥有共享上下文和杂乱数据的团队提供信任的AI分析师。

PromptQL, the trusted AI analyst for teams with shared context and messy data.

欢迎回到AI每日简报。

Welcome back to the AI Daily Brief.

今天的节目名义上是关于一个用人工智能治愈爱犬癌症的人，至少网上大家都在讨论这个。

Today's episode is nominally about this guy who used AI to cure his dog's cancer, or at least that's what everyone was talking about online.

但更广泛地说，它关乎当前人工智能讨论的现状。

But more broadly, it's about the state of the AI discourse.

我认为，我们需要从所有头条新闻中退后一步，提出一个根本性的问题：现在到底发生了什么？

And I think that the starting question that we need to ask, taking a big step back from all of the headlines, is: What the heck is going on right now?

目前外界对人工智能的讨论简直疯狂至极。

The AI discourse out there is absolutely frenetic right now.

你看到伯尼·桑德斯发布了长达九分钟的视频，谈论人工智能的生存风险。

You've got Bernie Sanders dropping nine minute long videos about X risk.

像ServiceNow的CEO比尔·麦克德莫特这样的企业领袖，也在主流媒体上抛出令人震惊的数据。

CEOs like Bill McDermott from ServiceNow dropping insanely terrifying statistics all over the mainstream media.

比如，他随意预测人工智能将导致新近大学毕业生的失业率超过30%。

In this case, a casual prediction that AI is going to cause recent college graduate unemployment over 30%.

每次美国发布一项民调，都显示公众对人工智能的负面情绪在持续加剧。

Every time a poll comes out in America, it shows just increasingly negative sentiment around AI.

这或许与这些媒体机构不断发布这些吓人的预测有关，谁知道呢。

Which, who knows, maybe has something to do with all these media outlets publishing these scary predictions.

但另一方面，那些从未写过代码的普通人，现在也能管理十几个甚至更多的智能代理，完成以前根本不可能做到的工作。

But then on the flip side, you've got normal people who haven't coded before, managing teams of a dozen agents or more doing all of this work that was never possible for them before.

换句话说，主流认知与实际能力之间的差距从未如此之大，但两者都处于极度高涨的状态。

The divergence, in other words, between mainstream perception and actual capability has never been higher, and yet both of them are in this incredibly heightened state.

那么，到底发生了什么？

So what is going on?

简而言之，我认为我们正处在AI的第二个时刻。

The short of it is I think that we are in AI's second moment.

显然，这里我用AI作为生成式AI的简称，而第一个时刻是2022到2023年的ChatGPT时刻。

Obviously in this case I am using AI as shorthand for generative AI and the first moment was the ChatGPT moment at the 2022, 2023.

这个时刻则是Claude、Code Opus 4.5、Codex 5.2等等。

This moment was the Claude code Opus 4.5, Codex 5.2, etc.

时刻。

Moment.

如果你想要真正高效地理解，这就是AI时刻和智能体时刻。

If you want to be really productive about it, it's the AI moment and the Agents moment.

本月初，伊森·莫洛赫在推特上写道：从AI用户的角度来看，到目前为止能力上的四大飞跃：第一。

At the beginning of the month Ethan Moloch tweeted: From an AI user perspective, the four big leaps so far in ABILITY: one.

GPT-3.5（ChatGPT 2022年11月）、GPT-4 2023年春季、推理模型（最初是‘一版预览’，但真正的突破是2025年春季的‘三版’）、可运行的智能体系统（Harness加优秀推理模型，2023年12月）。

GPT 3.5 (ChatGPT November GPT-four Spring twenty twenty three Reasoners (starts with 'one preview but the real deal was 'three Spring twenty twenty five) Workable Agentic Systems (Harness plus Good Reasoner Models December.

但说实话，我认为他的前两个和后两个都属于同一件事。

But really, I think his first two and his second two were all part of one thing.

请记住，在第一次浪潮前后，我们也经历了异常激烈和狂热的讨论。

And remember, in and around the first time we also got some really heightened frenetic discourse.

你或许还记得2023年5月，也就是这个节目开播的第二个月，《时代》杂志发布了一期名为《人类的终结》的特刊，探讨这一风险究竟有多真实。

You might remember in May 2023, which was the second month of this show, when TIME magazine dropped an issue called The End of Humanity, a special report on how real is the risk.

所以我要表达的观点是，如果这确实是AI的第二个时刻，那么围绕它掀起的尘埃比我们在这两个时刻之间进行的任何重要讨论都要更庞大、更激烈、更戏剧化，这完全说得通。

So the point that I'm making is that if this really is AI's second moment, it makes sense that the cloud of dust being kicked up around it is proportionally bigger and more heightened and more dramatic than even the important conversations we've had in between these two moments.

在某种程度上，我认为我们所经历的，只是第一次浪潮后出现的所有议题的重新浮现，只不过现在有一些关键的不同。

And to some extent I think part of what we're experiencing is just a resurfacing of everything that came up in the wake of the first moment with some key differences now.

第一个不同之处是，能力显然有了巨大提升。

The first difference is that there's obviously been a huge increase in capabilities.

使用3.5版本的ChatGPT已经非常惊人了。

ChatGPT with 3.5 was amazing.

当你把那时出现的图像生成能力与之结合时，使用这些工具的人们简直感觉自己像巫师一样。

You combine that with some of the image generation capabilities of the models that were coming out around then, and people who were trying these tools absolutely felt like wizards.

你其实不需要去说服大多数人。

You didn't really have to convince most people.

只要他们试用了这些工具，就会意识到正在发生重大的变化。

If they tried these tools they realized that something big was changing.

然而，即使在那些早期阶段，人们仍觉得还有更大的东西即将出现。

And yet even in those early days there was still this idea of something even bigger.

我做的第一个在YouTube上走红的节目，至少在这个类型的节目中，是关于一个早期的代理原型。

The first episode that I ever had go viral, at least in the terms of a show like this on YouTube, was about an early prototype agent.

我们做过像AutoGPT和BabyAGI这样的实验，还有GPT工程师，这些都成为了后来令人喜爱的项目的种子。

We had experiments like AutoGPT and baby AGI, a GPT engineer which would form the seeds that would go on to be lovable.

因此，两年后，随着智能代理真正投入使用，这种能力的大幅提升再次相应地加剧了讨论的热度。

And so two years later, as agents really come online, that big increase in capabilities has I think proportionally heightened the discourse once again.

第一个时刻与第二个时刻之间的第二个重大变化是，参与对话的人数大大增加了。

A second big change between the first moment and the second moment is that there are now many more people in the conversation.

在ChatGPT兴起的时期，这些工具是我们见过的增长最快的工具之一（还记得ChatGPT在前五周就获得了第一亿用户，打破了TikTok此前八个月达成这一纪录的纪录），但现在每周都有数十亿人使用这些工具。

Around the ChatGPT moment, these tools were some of the fastest growing we'd ever seen (remember ChatGPT got its first 100,000,000 users in its first five weeks, beating the previous record of eight months for TikTok) But now we have literally billions of people using these tools every week.

甚至那些不喜欢这些工具的人也在使用它们。

Even people who don't like the tools are using the tools.

因此，参与讨论的人数要多得多。

So there are just far more people in the conversation.

第一个时刻与第二个时刻之间的第三个区别是经济利益的提升。

A third difference between the first moment and the second moment is higher economic stakes.

而在这里，我甚至不是在谈论理论上未来可能带来的工作岗位替代问题。

And in this case I'm not even really talking about theoretical future job displacement things.

我谈的是当下此刻的现实情况。

I'm talking about right here and right now.

华尔街与SaaS公司、AI基础设施建设交易及其私人融资、正在构建AI的私营公司估值等之间的互动。

Wall Street's interaction with SaaS companies, AI infrastructure build out deals and the private financing thereof, valuations for private companies that are building AI, etc, etc, etc.

当时Anthropic对大多数人来说根本无足轻重，而现在它的年收入已达190亿美元，每次发布新功能都会冲击整个行业。

Anthropic wasn't even a blip on the radar to most people then, and now it's at a $19,000,000,000 run rate, taking down industries every time it announces a new feature.

AI第一阶段与第二阶段的第四个关键差异，与AI本身无关，而在于2022年至2026年间市场的演变。

A fourth key difference between AI's first moment and second moment has nothing to do with AI itself but has to do with the evolution of the market between 2022 and 2026.

AI如今已成为企业推卸责任的替罪羊，特别是在公司试图逆转后疫情时代过度招聘的背景下。

AI is now useful as a corporate fall guy, specifically in the context of companies trying to undo over hiring in the post COVID period.

投资者沙马特·帕利哈皮蒂亚写道：如果AI并不需要立即展现投资回报率，而只是企业用来裁减50%本就无所作为的员工的合理借口呢？

Investor Shammath Palihapathy writes: What if AI doesn't need to show an immediate ROI, but instead is the plausible deniability companies use to RIF 50% of the workforce they already knew did nothing?

第五点：无论你如何看待当前的政治立场，我认为可以毫不争议地说，第一阶段与第二阶段的另一个区别在于，这一切都发生在整体政治动荡加剧的背景下。

Number five: No matter what you think of the politics of the moment, I think it's fairly inarguable that finally as a difference between the first and second moment, this is happening in the context of generally increased political volatility.

换句话说，AI并非世界上唯一在发生的事，它现在正与诸如伊朗战争等事件相互交织。

In other words, AI isn't the only thing happening in the world it's now interacting with things like war in Iran.

还有一个最后的区别我想指出：我们已经见证了AI行业三年半以来，完全糟糕地解释自身，并以任何方式谈论未来，却始终无法与普通人产生哪怕一丝共鸣。

There is a last difference which I could point out: We've now had three and a half years of the AI industry doing a completely awful job of explaining itself and talking about the future in any way that's going to be even remotely resonant to the average person.

Not Boring的Packy McCormack最近发推说：AI对我来说非常奇怪，因为通常我会是那个认为我们对这项奇迹技术不够兴奋很疯狂的人。

Not Boring's Packy McCormack recently tweeted: AI is very weird for me because normally I'd be the guy who'd argue that it's crazy we're not more excited about this miracle technology.

但我完全理解这种负面情绪。

But I completely get the negative sentiment.

AI公司显然把故事讲砸了。

AI companies have clearly botched telling the story.

这是其中很重要的一点。

That's a big piece of this.

告诉人们‘我们造出了一个肯定会夺走你工作的玩意儿，希望我们能想办法在另一边给你发点救济金，或者找到更好的工作之类的’——这显然是糟糕的沟通。

Telling people We built this thing that is definitely going to take your job, and hopefully we can figure out how to give you handouts or something on the other side or come up with even better jobs or whatever is clearly terrible messaging.

不管怎样，这是一条很长的推文，但我认为AI行业极其糟糕的沟通，绝对是第一阶段和第二阶段之间发生的另一大变化。

Anyways, it's a much longer tweet, but I think that the incredibly poor messaging from the AI industry is absolutely another thing that has changed between the first and the second moment.

当然，第一阶段时的沟通也谈不上有多好。

Not that there was good messaging around that first moment, mind you.

只是那时我们还没来得及一次又一次地自掘坟墓。

There just hadn't been as much time for us to shoot ourselves in the foot over and over yet.

这一点在于，目前围绕人工智能的讨论变得极其激烈。

The point of this is, right now, everything around the AI discourse is incredibly heightened.

整个对话始终处于最高潮状态。

The whole conversation is at an eleven all the time.

自从我们所有人于2026年重返工作岗位以来，情况基本就是这样。

And basically has been since we all returned to work at the 2026.

这个周末有两场对话充分体现了这一点。

There were two conversations that really demonstrated this this weekend.

第一场围绕开发者安德烈·卡帕西的一个周末项目展开，该项目引发了巨大轰动。

The first was around a weekend project from developer Andre Karpathy that became an absolute firestorm.

周六晚上5点（东部时间），KaitoNEX发推称：五分钟前，安德烈·卡帕西刚刚发布了Karpathy/jobs项目。

At 5PM ET on Saturday night, KaitoNEX tweeted: Five minutes ago, Andre Karpathy just dropped Karpathy slash jobs.

他爬取了美国经济中的所有工作岗位（来自劳工统计局的342种职业），使用大语言模型为每种职业的AI替代风险打分（0到10分），并以树状图形式可视化呈现。

He scraped every job in The US economy (three forty two occupations from BLS), scored each one's AI exposure zero to 10 using an LLM, and visualized it as a treemap.

如果你的全部工作都在屏幕上完成，那你可就危险了。

If your whole job happens on a screen, you're cooked.

所有工作的平均得分为10分中的5.3分。软件开发人员：8到9分；屋顶工：0到1分；医疗 transcriptionists：10分。它指向了这个链接：karpathy.ai/jobs，即完整图表。

Average score across all jobs is 5.3 out of 10 Software devs: eight to nine Roofers: zero to one Medical transcriptionists: tenten It pointed to this link, karpathy.ai/jobs, which is the full chart.

推特上立刻涌出了大量类似Tucci的评论。

Instantly, Twitter was flooded with takes like this one from Tucci.

警报emoji，你明白Karpathy刚刚做了什么吗？

Siren emoji, do you understand what Karpathy just did?

他并没有写一篇观点文章。

He didn't write an opinion piece.

他爬取了美国每一个工作岗位，用AI进行了评估，并按1到10分的尺度打分，衡量你被替代的可能性。

He scraped every single job in America, ran it through AI, and scored how replaceable you are on a scale of one to 10.

这不是预测，而是一种诊断。

Not a prediction, a diagnosis.

会计师得了9分。

Accountants scored nine.

律师助理得了9分。

Paralegals nine.

文案写手被淘汰了。

Copywriters cooked.

放射科医生读扫描影像。

Radiologists reading scans.

AI已经做得更快了。

The AI already does it faster.

唯一得分低的工作是那些需要你亲自接触实物的工作。

The only jobs that scored low are the ones that require you to physically touch something.

2015年，学习编程是解决一切问题的答案。

In 2015, learning to code was the answer to everything.

2025年，代码会自己生成。

In 2025, code writes itself.

现在那些听从建议的人，成了历史上最易被替代的一代。

The people who listen are now the most replaceable generation in history.

我想你的学位并没有为你准备好职业生涯。

I guess your degree didn't prepare you for a career.

就连平时不常做这种低俗营销的人，也开始转向这种类似的煽情立场。

Even people who aren't usually schlock merchants like that started to veer into this same sort of sensationalist territory.

Chubby KPMGTISMIS 写道：Karpathy 绝对没有兴趣进行过度夸张。

Chubby KPMGTISMIS writes, Karpathy is by no means interested in hyper exaggeration.

通过使用 AI，他得出结论：在美国 1.43 亿劳动者中，约有 5700 万人面临工作被 AI 严重负面影响的高风险或极高风险。

Using AI, he concluded that out of 143,000,000 people working in The US, approximately 57,000,000 are at high to very high risk of their jobs being negatively impacted by AI.

这几乎占了 40%！

That's almost 40%!

好好想想吧，体会一下这意味什么！

Let that sink in and consider what it means!

现在，如果你经常听这类内容，你可能正等着听那句‘是的，但这里的细微差别在哪里？’

Now at this point if you listen frequently, you're probably waiting for the Yes, but where's the nuance here?

首先，如果你真的去读一下 Karpathy 发布的那篇原文，会发现他对数字 AI 暴露评分有一个非常重要的说明。

Well, first of all, if you go actually read the page that Karpathy posted he has a very important caveat on digital AI exposure scores.

他写道：这些只是粗略的 LLM 估算，并非严谨的预测。

He writes These are rough LLM estimates, not rigorous predictions.

高分并不意味着该工作会消失。

A high score does not predict the job will disappear.

软件开发者的得分是10分中的9分，因为AI正在改变他们的工作，但对软件的需求可能会随着每位开发者的生产力提升而增长。

Software developers score nine out of 10 because AI is transforming their work, but demand for software could easily grow as each developer becomes more productive.

该评分并未考虑需求弹性、潜在需求、监管壁垒或社会对人类工作者的偏好。

The score does not account for demand elasticity, latent demand, regulatory barriers, or social preferences for human workers.

许多高暴露度的工作将被重塑，而非被取代。

Many high exposure jobs will be reshaped, not replaced.

事实上，Karpathy本人对这一反应感到沮丧。

Indeed, Karpathy himself was frustrated by the response.

当Kaito那条原始推文下的某人说‘我找不到’时，Andre回复说：‘这是一个周六早晨花了两小时用代码快速完成的项目，灵感来自我正在读的一本书。’

When someone on that original tweet from Kaito said I can't find it, Andre responded This was a Saturday morning two hour vibe coded project inspired by a book I'm reading.

我觉得这些代码和数据可能对其他人探索BLS数据集有所帮助。

I thought the code and data might be helpful to others to explore the BLS dataset visually.

或者用不同的方式上色，或使用不同的提示，或添加他们自己的可视化效果。

Or color it in different ways or with different prompts or add their own visualizations.

这被严重误解了，我本该预料到这一点，即使有README文档，所以我把它撤下了。

It's been wildly misinterpreted, which I should have anticipated even despite the readme docs, so I took it down.

在另一条推文中，他写道：这种暴露程度是通过大型语言模型根据工作的数字化程度来评分的。

In another tweet he wrote: The exposure was scored by an LLM based on how digital the job is.

这与这些职业实际上会发生什么毫无关系。

This has no bearing on what actually happens to these occupations.

这涉及到需求弹性以及更多因素。

Which has to do with demand elasticity and a lot more.

人们在夸大这个可视化工具，并且强加给我一些我没有说过的话。

People are sensationalizing the visualization tool and putting words in my mouth.

关于这一点，确实有一些有趣而细致的讨论。

Now there was some interesting nuanced conversation about this.

斯蒂芬·舒伯特在《更新》通讯中写道：许多人似乎将此视为自动化整体步伐将很快的理由，但我不认为这有任何道理。

The Update newsletter Stefan Schubert wrote: Many seem to take this as a reason to believe that the overall pace of automation will be high, but I don't think that makes any sense.

芝加哥布斯商学院经济学家亚历克斯·伊马斯说得更直接：暴露程度并不意味着被取代的威胁。

Even more to the point was Chicago Booth economist Alex Imas who wrote: Exposure does not mean threat of displacement.

这 literally 可能意味着完全相反的意思。

It can literally mean the opposite.

被AI暴露的职业反而可能增加招聘并吸引更高的工资。

AI exposed jobs may increase hiring and attract higher wages.

这一切取决于a)消费者需求的弹性，以及b)一份工作中被AI暴露的任务数量。

It all depends on a) elasticity of consumer demand and b) number of AI exposed tasks in a job.

Anthropic的Peter McCrory补充道：我强烈同意Alex的观点，我的理解是，Claude的使用模式清晰地指向了劳动力市场的不均衡影响。

Anthropics Peter McCrory added: I agree strongly with Alex here, and my read is that Claude usage patterns clearly point toward uneven labor market implications.

我们最近引入的观察性暴露指标旨在识别那些暴露更可能转化为实际替代的情况，但任何暴露指标都不完美，也无法做出单调预测。

Our recently introduced observed exposure measure aims to identify cases where exposure is more likely to transform into actual displacement but no exposure measure is perfect or has monotone predictions.

即使一份工作的大部分内容被自动化，剩余的瓶颈任务最终仍可能增加对互补性人力技能的需求，即使在高度暴露的职业中也是如此。

And even when much of a job is automated, the remaining bottleneck tasks may ultimately increase demand for complementary human skills even among highly exposed roles.

多伦多经济学家Kevin Bryan表示：我敢打赌一千美元，从现在到2030年，最易受影响的职业在劳动力中的占比将上升。

Toronto economist Kevin Bryan said I bet a thousand dollars that from now to 2,030, most susceptible jobs see increased share of labor.

在这些图表所依据的模型中，明确的不是AI可以替代，而是AI相关。

In the model these types of charts are based on, it is explicitly not AI can substitute, but AI is related.

AI也是一种补充。

AI is a compliment too.

比如，现在谁不想写代码呢？

Who doesn't want to code right now, for instance?

我认为这些都是对的，而且我们当然会继续讨论AI对劳动力市场的真正影响。

And I think that's all true and obviously we will continue to discuss the real no BS labor market implications of AI.

但重点是，相对于我们更宏大的对话，这种讨论的激烈语气。

But the point is relative to our larger conversation this frenetic tone to the discourse.

更糟糕的是，就在凯托发布关于卡帕西研究的那条内容后不到一分钟，科贝西的信也发布了：Meta计划进行大规模裁员，可能影响公司20%或更多的员工。

Not helping this was the fact that at literally within one minute of Kaito posting that thing about Karpathy's research, the Kobeissy letter posted: Meta is planning sweeping layoffs that could affect 20% or more of the company.

正如我所说，现在这场讨论已经达到了极致。

Like I said, right now the conversation goes to 11.

但不仅仅是AI的负面方面达到了极致。

But it wasn't just the negative side of AI that was at eleven.

谷歌DeepMind的塞布·克莱尔分享了一篇来自《澳大利亚人报》的文章链接，该文章迅速走红，获得了近1300万次浏览。

Google DeepMind's Seb Cryer shared an article link from The Australian that went hyper viral with nearly 13,000,000 views.

维托里奥这样总结：这简直疯狂。

Vittorio summed it up this way: This is actually insane.

做个澳大利亚的科技宅。

Be tech guy in Australia.

收养一只患癌、只剩几个月寿命的流浪狗。

Adopt cancer riddled rescue dog months to live.

花3000美元对她的肿瘤DNA进行测序。

Pay $3,000 to sequence her tumor DNA.

把数据输入ChatGPT和AlphaFold。

Feed it to ChatGPT in AlphaFold.

完全不懂生物学。

Zero background in biology.

识别突变蛋白，将其与药物靶点匹配。

Identify mutated proteins, match them to drug targets.

从零设计一种定制的mRNA癌症疫苗，需要伦理审批才能给药，三个月后终于获批，开车十小时给罗斯伊注射第一针。

Design a custom mRNA cancer vaccine from scratch Need ethics approval to administer it Three months, finally approved Drive ten hours to get Rosie her first injection.

肿瘤缩小了一半。

Tumor halves.

毛发又变得有光泽了。

Coat gets glossy again.

狗还活着，而且很开心。

Dog is alive and happy.

教授，如果我们能对狗做到这一点，为什么不对人类推广呢？

Professor, if we can do this for a dog, why aren't we rolling this out to humans?

一个带着聊天机器人、只花了3000美元的人，竟然超越了整个药物研发流程。

One man with a chatbot and $3,000 just outperformed the entire pharmaceutical discovery pipeline.

我们将治愈这么多疾病。

We are going to cure so many diseases.

我认为人们还没有意识到未来会变得多么美好。

I don't think people realize how good things are going to get.

所以，这就是这个故事。

So here's the story.

澳大利亚企业家保罗·科宁厄姆有一只名叫罗斯西的狗。

Australian entrepreneur Paul Coiningham has a dog named Rosie.

2024年，罗斯西被诊断出患有癌症，对化疗和手术均无反应。

In 2024, Rosie was diagnosed with cancer that ended up being non responsive to chemotherapy or surgery.

肿瘤持续增长。

The tumors just kept growing.

当保罗向ChatGPT寻求帮助时，它建议他为罗斯西进行DNA测序，然后使用谷歌DeepMind的AlphaFold寻找可作为免疫疗法靶点的突变。

When Paul turned to ChatGPT for help, it suggested that he should get Rosie's DNA sequenced and then use Google DeepMind's AlphaFold to look for mutations that could be a target for immunotherapy.

当制药公司不愿提供现成的免疫疗法治疗时，科宁厄姆转向了新南威尔士大学RNA研究所所长帕尔·索德森。

When a drugmaker wouldn't provide an off the shelf immunotherapy treatment, coining him turned to Palle Thorderson, the director of the RNA Institute at the University of New South Wales.

索德森利用罗斯西的DNA，在不到两个月的时间内开发出一种定制的mRNA疫苗。

Thorderson used Rosie's DNA to develop a bespoke mRNA vaccine in less than two months.

他对媒体表示：这是首次为狗设计的个性化癌症疫苗。

He told the press, This is the first time a personalized cancer vaccine has been designed for a dog.

这仍处于癌症免疫治疗的前沿，但最终我们将用它来帮助人类。

This is still at the frontier of where cancer immunotherapeutics are, and ultimately we're going to use this for helping humans.

罗西教会我们的事是，个性化医疗结合mRNA技术可以非常有效，并且能在紧迫的时间内完成。

What Rosie is teaching us is that personalized medicine can be very effective and done in a time sensitive manner with mRNA technology.

正如你所知，这个过程远不止是简单地让ChatGPT来治愈癌症。

Now, as you can tell, there is a lot more to this process than simply prompting ChatGPT to cure cancer.

事实上，即使这种治疗本身也并非完全成功。

And indeed, even the treatment itself wasn't entirely successful.

是的，罗西的一些肿瘤已经缩小了，但要说这是癌症治愈就显然言过其实了。

Yes, some of Rosie's tumors have shrunk, but it would be certainly going too far to call it a cancer cure.

更重要的是，这 arguably 更像是一个关于诺贝尔奖级别的AlphaFold模型多么革命性的故事，而不是关于ChatGPT的故事。

On top of that, it's arguably a story about how revolutionary the Nobel Prize winning AlphaFold model is rather than a story about ChatGPT.

帕尔·索德森最终转向X平台来解释这个故事的一些细节。

Palle Thorderson ended up turning to X to explain some nuances of the story.

这些细节包括：这与其说是治愈，不如说是争取时间。

The nuances include: The fact that this was less about a cure and more about buying time.

由于许多人无偿贡献了时间和资源，很难估算出真实成本。

The fact that it's difficult to estimate the real costs as lots of people donated time and resources to this.

第三个细微之处在于，兽医研究和治疗的监管显然与人类医疗大不相同。

The third nuance is that regulation of vet research and treatment is obviously quite different than human health.

但最终，帕利表示：在人类医疗领域，罗斯的故事表明，我们可以让癌症疫苗的设计过程更加普及。

But ultimately, Palli says: In the human health space, Rosie's story demonstrates that we can democratize the process of designing cancer vaccine.

尽管基因组分析和RNA生产仍将是专业化的，但随着自动化程度提高，它们有可能转变为纯粹的服务提供。

While genomic analysis and RNA production will continue to be specialized, they could turn into pure service provision, especially as automation increases.

这就引出了一个问题：我们是否需要据此全面改革监管体系？

This then begs the question: Do we need to overhaul the regulatory regimes with this in mind?

我们能否确保公平获取？

And can we ensure equitable access?

当然，当人们刚看到这个故事时，即使在所有这些细微之处被披露之前，就已经有很多人持怀疑态度。

Now of course, there were tons of people who were skeptical on spec when they saw the story, even before all that nuance was shared.

更令人意外的是，我个人觉得，人们对于人工智能积极颠覆潜力的兴奋，比一味关注其负面影响要令人耳目一新。

And what's more unsurprisingly I personally find it a little bit refreshing to have people excited about the positive disruptive potential of AI than to just be constantly looking at the negative.

但关键是，这些仍然是同一枚硬币的两面。

But the point is that these are still two sides of the same coin.

我们正处在迈向人工智能第二阶段的过渡期。

We are in the midst of the transition into AI second moment.

在我们所有人适应这个新范式之前，这段时间会有些怪异。

And for a little while, until we all get used to the new paradigm that we're living in, it's gonna be weird.

我只能保证，如果你常来这里，至少会感觉没那么像在吃精神药物。

All I can promise is that if you hang out around here, you will feel at least slightly less like you're taking crazy pills.

好了，今天就到这里，感谢收看AI每日简报。

That For is going do it for today's AI Daily Brief.

一如既往，感谢你的收听或观看。

I appreciate you listening or watching, as always.

我们下次再见，保重！

Until next time, peace!