Trump Lifts Claude Fable 5 Ban, Anthropic Preps Agent in Microsoft Teams, Elon Musk’s Inner Circle

欢迎各位收看《信息TITV》。

Welcome everyone to the Information TITV.

我是阿卡什·帕什里卡。

My name is Akash Pasricha.

今天是7月1日，星期三。

It is Wednesday, July 1.

本期节目中，白宫已解除对Anthropic Fable五的出口限制。

Today on the show, the White House has lifted export restrictions on Anthropic Fable five.

我们会深入了解政府是如何做出这一决定的。

We'll find out how the administration came to that decision.

我们还将探讨Anthropic如何悄然渗透办公应用领域，如今正逐步进入微软Teams。

We'll dive into Anthropic's quiet takeover of workplace apps, which is now expanding into Microsoft Teams.

接下来，我们会仔细分析埃隆·马斯克在特斯拉的核心团队，从中窥见特斯拉与SpaceX未来合并的可能性。

We're then taking a closer look at Elon Musk's inner circle at Tesla and what it tells us about a possible merger down the road with SpaceX.

节目最后，我们邀请到布拉德利·塔斯克，聊聊当前人工智能监管的现状。

And to close out the show, we've got Bradley Tusk coming on to talk about the current state of AI regulation.

这期节目会非常精彩，咱们马上开始。

It's going be a great show, so let's get right on into it.

Fable回来了。

Fable is back.

在最新模型短暂发布后被搁置了一段时间，白宫现在放宽了对Anthropic的出口限制。

The White House has eased its export restrictions on Anthropic after its latest model was effectively put in purgatory after a brief release.

关于这件事的更多细节，我们请来我们的AI与政治记者施瓦茨。

For more on that story, I want to bring on our AI and politics reporter, Schwartz.

莱奥，欢迎再次做客节目。

Leo, welcome back to the show.

很高兴你能来。

It's great to have you here.

很高兴能来。

Good to be here.

Fable回来了。

Fable's back.

Fable？它到底是怎么回事？

Fable's It was what was it?

两周？

Two weeks?

一周多

A week and

刚超过两周。

a Just over two weeks.

明白了。

Okay.

刚超过两周。

Just over two weeks.

是什么变了？

What changed?

政府为什么松口了？

Why did the government give?

有什么变化吗？

Did anything change?

我觉得这是个好问题。

I think that's a good question.

在过去这短短两周里，我们到底学到了多少，我也不太确定。

I don't know how much we actually learned over these very wild last two weeks.

昨晚经过大量报道后，Anthropic 终于发布了一篇长博文，解释了究竟发生了什么变化。

After a lot of reporting last night, Anthropic finally put out a lengthy blog, basically explaining what did change.

如果你还记得，这一切都源于亚马逊的一份报告，他们发现 Fable 存在一个漏洞或越狱方式，暴露了模型可能带来的各种安全风险。

And if you remember, this all stemmed because of an Amazon report where the company found a vulnerability or a jailbreak in Fable that basically surfaced different security risks the model could do.

当这个问题被白宫提出来时，Anthropic 的回应是，每个模型都有这个问题。

Anthropic, when this was surfaced by the White House, said, this is present in every model.

这其实算不上什么大问题。

It's not really an issue.

它并不会触发我们模型中那些被神化的功能。

It doesn't surface the mythos capabilities that are inherent in our model.

白宫显然不认同，直接出台了出口管制措施。

The White House obviously disagreed, put on these export controls.

从Anthropic昨天发布的博客来看，实际上没发生什么实质性变化。

What we can glean from the blog that Anthropic put out yesterday is really not that much changed.

Anthropic确实增加了一个分类器，用来阻止亚马逊报告中提到的那些请求类型。

Anthropic did put on what's called a classifier that would basically block those type of requests that were surfaced in the Amazon report.

但他们一再强调，自己仍不觉得这真是个大问题，不过加这一层防护也没关系，只是用户在做基本调试或其他操作时，可能会更频繁地被转到其他模型，稍微麻烦一点。

But as it said, over and over again, the report, we still don't really think this was a big big deal, but we're fine putting on this extra layer of protection, which all that might mean for users is a bit more annoyance when they try and do basic debugging or other tasks and they get rerouted to a different model.

只要安迪·贾西满意就行。

As long as Andy Jassy is happy.

你知道的，如果你还记得，我们之前报道过，正是亚马逊——就像你提到的——才引发了这一连串事情。

I mean, you know, he if you recall, we reported that it was it was really it was Amazon, as you said, that that kind of started this whole, you know, journey.

说说这个分类器吧。

Talk about this classifier thing.

Anthropic做出这个让步，难度有多大？

I mean, how hard a concession was that for Anthropic to give?

看起来这只是一个很小的调整。

It seems like it was a pretty minor change.

Anthropic在博客里说，这不过是给现有的防护系统再加一层保护。

What Anthropic said in the blog post is this will just it adds to the defense system that we already had.

基本上就是会拦截更多请求。

It will just basically block even more requests.

他们解释说，你可能会在做些基础操作时注意到这点，但总体来看，我们本来就不觉得这算个问题。

And the way they framed it is you'll probably notice this for more basic tasks that you're trying to do, but for the most part, we we don't really think it was an issue in the first place.

有意思的是，他们在博客里提到，早在之前就已经和商务部下属的AI安全机构Casey合作，共同设计了这套防护机制。

And what's interesting is they said in this blog post that they'd actually already worked with Casey, which is an AI safety institute under the commerce department before this to work out this type of defense system.

Casey早就批准了，现在又重新确认了一次。

Casey had already approved it, and they approved it now.

所以仔细琢磨一下，Anthropic的意思其实是：没太大变化，只是做点表面功夫来安抚政府。

So if you read between the lines, what Anthropic is basically saying is not much has changed, but we did this fix to appease the government.

抱歉，那这个是什么？

And sorry, what what is that?

Casey是商务部下属的一个小机构。

Casey is a small center within the commerce department.

它是人工智能标准与创新中心。

It's the the center for AI standards and innovation.

里面都是技术专家，负责评估不同模型的标准，确保它们安全可靠。

It's staffed with technical experts who can basically look at different standards within models and make sure that they're safe.

这个机构是在拜登政府时期成立的。

It was started under the Biden administration.

特朗普上台后，把它名字里的‘安全’二字删了，人员也大幅削减，但最近两周，它还是和Anthropic紧密合作，确保所有流程都合规。

It was renamed under Trump to take safety out of the name and significantly lowered in its amount of staff, but it worked closely with Anthropic in this last two weeks of negotiation to basically make sure that everything was above board.

眼下，背景里还有OpenAI，他们也在和政府合作，准备发布自己的模型。

Now, in the background here, we obviously have OpenAI who is working with the government as it seeks to release its own model.

你觉得OpenAI的举动，是不是给Anthropic施加了压力，逼他们赶紧推出这个分类系统？

How much of a fire do you think that lit under Anthropic to maybe develop this classifier system quickly?

你有没有觉得，他们之间有点竞争意味？要是OpenAI抢先发布模型，用户可能就转向他们了。

Do you think there was a little bit of a rivalry there in that, well, if OpenAI gets their model out there faster, adoption could suffer.

有趣的是，现在的角色几乎颠倒过来了。

What's interesting is that the roles are almost reversed now.

Fable 已经发布了，而 OpenAI 的 5.6 版本现在才进入有限发布阶段。

So Fable is out there, OpenAI's 5.6 is now in this limited release.

所以 Anthropic 在这方面确实领先了一点。

So Anthropic does have a bit of a lead there.

但我也不确定，从 Anthropic 的角度看，这是否真正改变了局面。

But I I don't know if it it necessarily changed the situation from Anthropic standpoint.

我认为，当政府实施这些出口管制时，Anthropic 就有很强的动力去开发解决方案。

I think when the government implemented these export controls, Anthropic was very incentivized to work on a solution.

他们很清楚，OpenAI 一直在谈论要发布功能相似的模型。

Obviously, they knew that OpenAI had been talking about releasing a model with similar capabilities.

不过看起来，政府意识到 OpenAI 马上就要推出产品，于是推动 Anthropic 加快进度，确保双方处在相对公平的竞争环境中。

It does seem like though that the government, understanding that OpenAI was right around the corner, did push it to move a little faster with Anthropic and make sure that there was a bit of a level playing field.

你昨晚写到，业界人士希望即将出台的行政令，能用一套可预期的评估机制，取代目前这种临时性的决策方式。

Now, you wrote last night that industry figures hope a forthcoming executive order will replace ad hoc decisions with a predictable evaluation process.

这有多大概率会发生？

How likely is that?

我觉得这就是每家公司都在盼着八月能出炉的万能解药。

I think this is the magic bullet that every company is hoping is going to come right around the corner in August.

你还记得吧，我们之前报道过，六月初的行政令要求在六十天内建立这个AI监管框架。

If you remember, we reported about the AI framework that's getting implemented by the early June executive order that gave sixty days for this framework to be created.

正如我们之前提到的，这方面其实没什么实质进展。

As we've reported, there hasn't really been that much progress on it.

政府只和业界开过一次会，讨论这个框架可能长什么样，不过我们知道，他们私下和一些公司有过不少沟通。

There's only been one session that the government has held with industry to talk about what that might look like, although we know that there's been some smaller communications and and back channeling with companies.

现在距离框架规定的生效日期只剩一个月了。

So we're now just a month away from when that's supposed to be implemented.

理论上，一旦这个框架落地，企业就能清楚知道发布模型需要满足哪些条件。

And theoretically, once this is there, this will be a framework to make sure that companies know what has to be done for a model to be released.

不过，你可能还记得，这在技术上其实是自愿性质的。

Although, if you remember, is technically voluntary.

我不确定这个‘自愿’到底有多自愿。

I don't know how voluntary it actually is.

这个框架能否真正起到兜底作用，防止类似情况再次发生，还有待观察，尤其是模型的发展速度似乎比政府的跟进还要快。

Whether this will be a catchall to make sure that this type of situation never happens again really remains to be seen, especially as models get more and more advanced in a way that I think is a little faster than the government can keep up with.

没错。

Right.

换句话说，关于未来Anthropic模型发布会怎样，答案其实取决于一个月后出台的这个框架具体内容。

In other words, the answer to the question, well, what happens to future Anthropic Model releases?

在没看到框架之前，根本没法判断模型发布后会不会陷入一种悬而未决的尴尬状态。

That really depends on what the framework that comes out probably in about a month actually says, because there's no way of knowing, without knowing that framework then, whether or not models will get stuck in this limbo or purgatory phase upon release.

是的。

Yeah.

这个框架会设定一些具体标准，只要模型达到这些标准，就必须向政府报备，经过一段时间后才能上线。

So the framework will have certain benchmarks that if models reach them, then they'll be covered and they have to submit the models to the government for some period of time before they launch.

但Anthropic那篇博客里还提到一个有意思的地方：他们正在开发另一个独立框架，专门用来和私营企业合作，评估模型的安全漏洞并制定应对策略。

But what's also interesting from this blog that Anthropic put out is they're developing a separate framework, and this is a framework for assessing with private industry how to look at security vulnerabilities and models and how to approach them.

所以我觉得，我们现在看到的是各种不同的解决方案纷纷出台，有些是政府提出的，有些则是私营行业可能自行推行的自我认证体系。

So I think what we're seeing is just a lot of different types of solutions being put forth, some by the government, some more of a a self certified system that private industry will probably do.

而所有这些方案背后，最大的疑问是：它们能否跟得上前沿模型飞速发展的步伐？

And and the looming question over all of these is will they be able to keep up with the very fast moving field of frontier models?

没错。

Right.

很好。

Great.

利奥，非常感谢你今天来做客。

Leo, well, I want to thank you for coming on.

这位是利奥·施瓦茨，我们《信息》杂志负责人工智能与政治报道的记者。

That is Leo Schwartz, our AI and politics reporter here at The Information.

我们会在节目后半段继续跟进关于Fable的报道。

We will have more coverage on the Fable story towards the end of the episode.

接下来我们请来的是布拉德利·塔斯克。

We've got Bradley Tusk coming on.

他在监管方面是专家，不过在这之前，我想先跟大家聊聊《信息报》的独家报道。

He is an expert on all things regulation, but in the meantime, the information has more exclusive reporting that I want to get through first.

Anthropic 正在为 Microsoft Teams 开发一个智能助手，这可能会和微软自家的 Copilot 竞争，我来请我们的微软记者霍姆斯谈谈。

Anthropic is prepping an agent for Microsoft Teams, which could be in competition with Microsoft's own Copilot AI tool, I want to bring on our Microsoft reporter, Holmes.

亚伦，欢迎再次做客节目。

Aaron, welcome back to the show.

很高兴你今天能来。

It's great to have you here.

一如既往，非常荣幸。

It's a pleasure as always.

说说这个即将在 Teams 上线的 Anthropic 智能助手吧。

Tell me about this new Anthropic agent that's going to come about in Teams.

是的。

Yeah.

简单说，Anthropic 上周刚在 Slack 上推出了这个产品，根据我的报道，他们最近已经通知微软，打算在 Teams 上推出一个类似的助手。

So, essentially, this is a product that Anthropic, you know, originally debuted for Slack, last week and, has recently told Microsoft, according to my reporting, that it's gonna release a similar agent for Teams.

基本上，现在你可以通过 Teams 和 Claude 对话，在 Teams 频道里@它，就像跟同事互动一样。

And, basically, the idea is that you can now use Teams to talk to Claude and tag it in your Microsoft Teams channels and interact with it in the way that you would interact with a coworker.

正如你提到的，这挺有意思，因为微软自己也有 Copilot 功能，直接和 Claude 竞争，他们当然希望客户更多地用自家的产品，而不是 Claude。

And, you know, with this, as you noted, is interesting because Microsoft has its own Copilot feature, which directly competes with Claude and which they would love more of their customers to use instead of Claude.

但他们采取了比较包容的态度，到目前为止，他们愿意让 Anthropic 这样的 AI 公司，把助手直接集成到 Teams 里，面向他们的客户。

But, you know, they're taking a conciliatory approach, and so far they are, you know, open to Anthropic or other AI agent makers having these plug ins directly in Teams in front of their customers.

那计算资源和 Anthropic 模型所需的开销，该由谁来承担呢？

And who would bear the cost then of the of the the compute and, you know, the what's needed under the Anthropic model?

总得有人为这些 token 买单吧。

Somebody has to be paying for those tokens.

对吧？

Right?

是的。

Yeah.

那费用就由客户自己承担了。

So that would just be the customer themselves.

简单说，如果你已经有Claude账号，只需把它连到你们团队的账号上，继续像往常一样付费，根据你使用Claude的量向Anthropic付费就行。

Basically, if you already have a Claude account, you would just connect it to your team's account and keep paying for both as you usually would and, you know, pay based on how much you use Claude to Anthropic.

明白了。

Got it.

那微软为什么允许这么做？

Why is Microsoft allowing this?

他们自己不是有Copilot吗？

I mean, it has Copilot.

这背后的策略到底是什么？

What's what's the strategy here?

是的。

Yeah.

这其实体现了微软所说的平台型公司理念。

So this is basically, you know, emblematic of this platform company approach as Microsoft calls it.

他们的想法是，理论上完全可以把产品封闭起来，只允许使用Copilot。

The idea is, you know, they could, in theory, kind of wall off their products and make it a place where you can only use Copilot.

但他们认为，如果客户本来就想用Claude这样的工具，那还不如直接在微软的应用里用，比如Teams。

But their bet is essentially that, you know, if customers already wanna use a tool like Claude, they might as well be using it inside a Microsoft app like Teams.

我觉得这里的策略核心，其实是掌控用户的行为模式。

And I think the strategy there is essentially about sort of owning user behavior.

如果Teams真成了你跟所有AI助手交流的唯一平台——这正是微软CEO萨提亚·纳德拉所描绘的愿景——那公司很可能更愿意续订团队许可证，毕竟员工已经离不开它了。

And, you know, if if Teams becomes the place that you're talking to all of your AI agents, which is kind of how Microsoft CEO Satya Nadella has pitched this vision for the product, that probably means that, you know, you're you're you as a company are more likely to keep renewing your team's license if if your employees are using it that way.

当然，这么做也有代价，毕竟Anthropic是竞争对手，一直在想办法把微软的客户拉走，但对微软来说，这就是他们的算盘。

Obviously, there are trade offs to that since, Anthropic is a competitor and in many ways is looking to shave away customers from Microsoft, but that's sort of the calculus on Microsoft's end.

昨天我们请了劳拉·布拉顿做客节目，她讲的几乎就是同一个故事：Slack和Anthropic推出的Claude Tag插件，当时在Slack里也发生了类似的情况。

Now we had Laura Bratton on the show yesterday, and we talked about this exact same story, almost to a tee, that was playing out with Slack and Claude Tag, which is the Anthropic agent that has been installed in Slack.

她提到，Salesforce的员工对自家公司为什么要推广这个工具感到困惑。

She was talking about Salesforce employees that were a little bit confused as to why Salesforce was promoting that tool.

你刚才说的，其实和这个思路一模一样。

I mean, you've sort of outlined the same approach here.

你觉得微软内部员工也有类似的困惑吗？还是说大家都完全认同这个理念：我们卖的是Teams，就要把它做到最好，不管付出什么代价都值得？

Do you get the sense that Microsoft employees are confused in the same way, or is everyone pretty bought into this vision of, you know, we're selling Teams, we want Teams to be the best product, we're on board with whatever it takes?

你知道吗，微软内部其实也一直在进行一场很健康的讨论。

You know, I think there's definitely a healthy debate going on inside Microsoft as well.

今年早些时候，Anthropic推出了可以直接接入Word、Excel和PowerPoint的新功能，让Claude能在这些工具里自动完成工作，那时候争议更大了，因为这直接跟Copilot向客户承诺的功能重叠了。

I think it was even more pronounced earlier this year when Anthropic came out with these new features that plug into Word and Excel and PowerPoint and let Claude automate work there because that's even more of a direct competitor to what, you know, Copilot has been pitched to customers as something that can do it can do.

但与此同时，微软内部也承认，尤其是在今年年初，Claude在企业级AI市场确实走在了前面。

But at the same time, I think that there's, you know, an acknowledgment inside Microsoft that right now and and especially earlier this year, you know, Claude was sort of the the front runner in the enterprise AI market.

微软并没有试图说服所有客户完全不用Claude，反而乐于借助Claude来进一步巩固自己软件的使用场景。

And rather than try to just convince all of their customers to not use Claude at all, I think Microsoft is happy to kind of use Claude to further entrench its software.

意思是，如果你要用Claude来自动发消息，或者自动生成Excel里的财务模型。

Meaning, you know, if you're gonna be using Claude to automate sending messages or automate, you know, creating financial models in Excel.

至少你还在用Teams和Excel，哪怕只是部分功能。

At least you're still using Teams and Excel in some form.

我觉得目前微软更愿意保持自己在客户工作环境中的这种地位。

And I think for now, is happy to to, you know, maintain that sort of stature within customers' environments.

我还在想另一件事：如果我是一名顶尖的AI研究员，或者说，我们聊聊那些仅次于顶尖的研究员，毕竟最顶尖的人才已经被数亿美元的薪酬抢光了。

The other thing that I'm thinking about is if I'm a top AI researcher, and maybe let's talk about the tier below the top AI researchers, because the best of the best are getting wooed with all these hundreds of million dollars pay packages.

但如果我是在一个研究团队里，而且我关心的是能真正影响最终用户的那类工作，那我就在想，微软越依赖Anthropic，而我又能在微软和Anthropic之间二选一，其他条件都一样的话，这反而更让我倾向于去Anthropic，对吧？

But if I'm on a research team, and I care about doing the work that is going to really impact the end customer, I sort of wonder if the more that Microsoft relies on Anthropic, and I have the choice of working in Anthropic or Microsoft, all things equal, this is kind of more of a reason to go work for Anthropic, right?

因为微软实际上等于在说：嘿，我们承认自己的模型可能没那么强。

Because Microsoft's basically saying that, Hey, we've sort of acknowledged that our models are not going to be as great.

我们觉得Anthropic照样能搞定这事。

We think Anthropic can do the job anyway.

我不太确定。

I I don't know.

我只是在想，这对人才吸引会不会有影响。

I just wonder if this has implications from a talent perspective.

如果微软的意思是：你哪怕去了Anthropic，我们最后也能把你拉回来。

If Microsoft is basically saying, well, if you work at Anthropic, we'll get to you anyway.

是的。

Yeah.

多年来，微软在吸引AI研究员方面根本竞争不过Anthropic，哪怕曾经有过机会也早就没了，一方面是因为Anthropic愿意给研究员开的薪酬太高，另一方面是因为它现在就是公认的顶尖AI实验室。

I don't think Microsoft has been able to compete in the talent market for AI researchers with Anthropic for years, if ever, both because of how much Anthropic is willing to pay researchers and just, you know, its stature is basically the leading AI lab right now.

但这确实挺有意思的，我觉得你说得对，微软一直在强调，他们不会去训练最前沿的模型。

But it it is definitely interesting, and I think you're right that Microsoft has made this sort of pitch that it's not going to train the the frontier cutting edge models.

他们会更专注于那些成本更低、但据微软说，对客户来说反而更有用的小模型。

It's going to focus more on these smaller models that could be cheaper, but, you know, according to Microsoft, in some ways, more useful to customers.

目前他们正全力推进这个定位。

And they're really run like, leaning into that pitch heavily at the moment.

所以我很想知道，客户会不会认同这一点，还有你提到的，训练这些模型所需的人才。

So I think it'll be interesting to see if customers agree with it, as well as to your point, the talent that they need to train these models.

对。

Right.

明白了。

Okay.

还有两个问题。

So, two more questions here.

OpenAI 有没有类似 Anthropic 集成到 Teams 的做法？毕竟 OpenAI 的模型已经在支撑微软很多自家产品了。

Does OpenAI have an equivalent here to Anthropic integrating into Teams given that OpenAI's models power a lot of Microsoft's own products.

对吧？

Right?

是的。

Yeah.

目前他们还没有把ChatGPT和Teams做集成的连接器。

They currently do not have any sort of connector for Teams and ChatGPT.

我不确定以后会不会有，但我觉得OpenAI基本上是想尽可能牢牢占据企业级聊天机器人市场。

I'm I'm not sure if that's something that is coming, but I think, you know, OpenAI, you know, is basically happy to sort of corner the chatbot market for enterprises as much as they can.

不过，目前还不清楚这会不会实现。

But, yeah, unsure if that's coming yet.

你之前深入报道过企业数据争夺战，比如公司对Slack、Teams这些平台里数据的访问权问题。

And then, you've reported extensively on the corporate data wars story and companies having access to data sitting within Slack Teams, all these types of interfaces.

这能说明企业数据争夺战有缓和的迹象吗？

Does this tell us anything about maybe the corporate data wars easing at all?

微软其实已经表态了，只要符合他们的政策，任何公司都可以为他们的应用开发连接器。

You know, it's I mean, Microsoft has basically said that they will let any company build a connector to their apps as long as they, you know, comport with Microsoft's policies.

与此同时，我们也看到他们屏蔽了一些功能。

At the same time, we have seen them block some features.

比如，Databricks 开发了一个功能，可以让用户把智能代理连接到微软的 Power BI 软件——这可是直接跟 Databricks 竞争的产品，但微软却把它给封了，我同事凯文报道过这事。

For example, Databricks built a feature that would have let people connect agents to Microsoft's, you know, Power BI software, which competes directly with Databricks, and Microsoft actually blocks that feature, as my colleague, Kevin, reported.

不过微软说，哦，这根本不是什么竞争手段。

But, you know, Microsoft said, oh, that that wasn't a competitive move.

他们只是担心这个功能的准确性和稳定性。

That was just due to, you know, we're our worries about the accuracy and reliability of this feature.

所以，到目前为止，这些做法还挺随意的，大概最终还是看微软希望用户在特定软件里做些什么。

So, you know, I I think it is so far a little bit ad hoc, and it'll probably depend, you know, exactly what Microsoft wants its customers to be doing in a given piece of software.

但要说数据大战快结束了？我觉得还远着呢。

But, no, I would say that the data wars are far from over.

很好。

Great.

非常感谢你今天来做客。

Well, I want to thank you for coming on.

这位是艾伦·霍尔姆斯，我们《信息》杂志负责微软报道的记者。

That is Aaron Holmes, our Microsoft reporter here at The Information.

特斯拉和SpaceX正越来越紧密地合作，共同推进埃隆·马斯克主导的这座价值550亿美元的巨型芯片制造项目Terafab。

Tesla and SpaceX are becoming increasingly intertwined as the two companies collaborate on Terafab, the giant $55,000,000,000 chip manufacturing project that Elon Musk has embarked on.

《信息》刚刚更新了特斯拉的组织架构图，展示了公司最高层的管理人员，这份架构图仅对我们的付费订阅用户开放。

The information just updated its teslaorg chart to reflect the senior most executives at the company, org chart that the information are accessible to our pro subscribers.

我想请我们的埃隆·马斯克特派记者格蕾丝·凯来聊聊，她更新了这张架构图，也写了这篇报道，让她带我们看看她目前最关注哪些方面。

I want to bring on Grace Kay, our Elon Musk reporter, who updated the org chart, wrote the story on it, and I want her to walk us through what she is paying attention to.

格蕾丝，欢迎回来，不对，是第一次上节目，很高兴你来。

Grace, welcome back welcome to the show, I should say, first time.

很高兴你能来。

It's great to have you here.

是的。

Yeah.

谢谢邀请我。

Thank you for having me.

好的。

Okay.

在我们看组织架构图之前，先帮我们回顾一下：Terafab这个项目的最终目标是什么？目前进展到哪一步了？

So remind us here before we get into the org chart, what is the ambition here with Terafab and where do things stand right now?

是的。

Yeah.

我认为这是埃隆·马斯克最雄心勃勃的项目之一。

I I would say it's one of Elon Musk's most ambitious projects.

这是一个横跨他三家核心公司的半导体制造设施。

So it's a semiconductor manufacturing facility spread across three of his most important companies.

涉及SpaceX、特斯拉和XAI三方的合作。

It's a collaboration between SpaceX, Tesla, and XAI.

他想通过这种垂直整合的方式，打造出能驱动全自动驾驶、擎天柱机器人和数据中心的芯片。

And basically, he is trying to bring this vertically integrated approach and create these chips that could power full self driving, optimists, and data centers.

简单说，就是贯穿了他旗下的所有主要公司。

So basically spanning across his companies.

但目前这还处于非常早期的阶段。

But right now it's very early in the process.

他正在从芯片行业招募人才，把来自特斯拉和太空探索技术公司的人员召集到一起，同时也在办理各种许可。

So he's still bringing on talent, across the chip industry, bringing them in from Tesla and SpaceX to work together and getting permits.

对。

Right.

他们还没真正开始制造，现在还处在规划阶段，什么都没建起来，真的还早。

So they haven't actually made any they're still sort of in the planning phase, so nothing's really been constructed Very

早。

early.

他三月份才刚宣布这个项目，但已经在推特上大力宣传，也在公司内部积极筹备。

He had just announced it in March, but it's something he's really hyped up on Twitter, something he's building up in the company right now.

好的。

Okay.

特斯拉和太空探索技术公司都为这个项目投入了大量资金吗？

Both Tesla and SpaceX have both contributed a significant amount of funding to this effort?

是的。

Yes.

他们最初共同同意出资550亿美元，之后几年内可能会增长到大约1010亿美元。

So they both jointly agreed to fund $55,000,000,000 initially, and that can grow over time, I think, to about $101,000,000,000 over the course of several years.

这是一笔巨大的投资。

So it's a significant investment.

好的。

Okay.

那我们现在来看看你发布的组织架构图。

So now let's get to the org chart that you published.

你提到特斯拉全公司大约有70人参与。

You documented the I think it was 70 people approximately at Tesla across the whole company.

你在文章里指出，这个架构图里一个明显的趋势是，不少人正在参与Terafab项目。

One of the themes that came of the org chart, which you point out in the story, is that a number of these people are working on the Terafab effort.

我们从中学到了哪些关于特斯拉哪些人直接参与这个项目的信息？

What did we learn about who a Tesla is directly involved in that program?

是的。

Yeah.

所以我们了解到，有很多来自SpaceX的人在特斯拉参与这个项目。

So we learned that there's a lot of SpaceX people in Tesla working on this endeavor.

约翰·费特施泰尔原本是SpaceX星链项目的高级产品工程师，现在担任这个项目的负责人，并直接向埃隆·马斯克汇报。

So John Fetterspiel, who was a senior level, product engineering person on Starlink at SpaceX, is the leader of the project and is reporting directly to Elon Musk.

他们从星舰、星链、龙飞船的工程和管理团队中调来了不少人参与这个项目，这些人在特斯拉内部也有正式记录。

They've brought in a number of people from Starship, Starlink, Dragon engineering and managers to work on the project, and they are also listed internally at Tesla.

他们还从英特尔招聘了不少人。

They've also been recruiting a bunch of people from Intel.

过去几个月我们发现了四位关键 hires，其中包括杨凯瑞，他在英特尔工作了整整十七年，对团队来说是个非常重要的引进。

We found four key hires over the past few months, including Gary Yang, who is a seventeen year veteran at Intel and is considered a pretty significant hire for the team.

当你提到从SpaceX引进人手时，我想确认一下我的理解对不对。

And when you say bringing them in from SpaceX, I just want make sure I understand this.

这是正式聘用这些人，还是他们只是暂时借调到特斯拉，类似那种安排？

Is hiring people from SpaceX or is this just people from SpaceX that are sort of on secondment over at Tesla or something like that?

是的。

Yeah.

这确实有点让人糊涂，因为这种安排很特别，也很有马斯克的风格。

It's pretty confusing because it's kind of a unique deal and it's a very Elon Musk thing.

他们同时在两家公司工作。

They're working at both companies.

在组织架构图上，他们被标记为SpaceX的合作伙伴。

They're brought in as SpaceX partners, is what they're identified in the org chart.

他们在SpaceX有各自的职位，但同时也出现在特斯拉的内部系统里，因为他们是在一起合作的。

They have their own roles at SpaceX, but then they're also in internal Tesla systems because they're working together.

他们依然受雇于SpaceX，但也在为特斯拉工作。

So they're still employed by SpaceX, but they're also working with Tesla.

换句话说，我觉得他们的领英资料上只能写‘我在马斯克公司工作’，因为实在找不到更准确的描述了。你刚才提到的那些人挺有意思的。

In other words, I think their LinkedIn literally has to say, I'm just working at the the Musk Corporation, because there's really So no other one to describe you talked about some of the people here, which are interesting.

我很好奇，当你整体看特斯拉的组织架构时，除了那些深度参与Terrafab项目的人之外，还有没有其他明显的趋势？或者你对公司的结构产生了哪些新的疑问？

I wondered, as you looked at the Tesla org chart as a whole, other than the number of people who are intimately involved in the Terrafab effort, were there other themes that you drew from the org chart or other questions that you came up with in terms of how the company is structured?

是的。

Yeah.

有几件挺有意思的事。

There was a few interesting things.

我注意到，今年只有埃隆·马斯克的一位直接下属离开了。

One thing I noticed is this year, only one of Elon Musk's direct reports has left.

这跟去年形成鲜明对比，去年他失去了好几位直接下属。

And that's quite a contrast to last year where he lost several direct reports.

他一向以高流动性著称。

Generally, he's known for having a lot of turnover.

他失去了拉吉，拉吉之前负责特斯拉的销售和人工智能业务。

So he lost Raj who was leading Tesla sales and AI efforts at the company.

拉吉的职责被分给了三位新的直接向埃隆·马斯克汇报的人。

And Raj's responsibilities were spread across three new direct reports to Elon Musk.

而且公司也在进一步强调人工智能这个方向。

And they're also kind of emphasizing this focus on AI in the company.

所以现在有很多新职位，专门负责把人工智能引入工厂，以及搭建公司的AI基础设施。

So there's multiple new roles around bringing AI to the factory, AI infrastructure at the company.

当然，还有康斯坦蒂诺斯、拉斯卡拉斯和阿卡什·帕里查的自动驾驶团队在推进人形机器人项目。

And then, of course, there's the humanoid robotics program under Constantinos, Lascaras, and Akash Pasricha's autopilot team.

现在有很多人直接向马斯克汇报，而且都专注于人工智能领域。

So there's a lot of different people reporting to Elon Musk working specifically on AI.

Grace，我最后再问你一个问题。

And last question for you, Grace.

我的意思是，这能说明特斯拉和SpaceX有可能合并吗？

I mean, what do you what did this tell us about a possible merger between Tesla and SpaceX?

从公司现在的组织结构来看，这会让合并的可能性更大吗？

Does it make it more likely based on the way the organization is structured?

这很难说。

It's hard to say.

公司内部确实有这种传闻，投资者们也经常在讨论这件事。

It's definitely a rumor inside the company and something that investors are talking a lot about.

这个话题我特别感兴趣。

It's something I'm very curious about.

我觉得这说明他们现在的合作比以往任何时候都更紧密，距离也越来越近。

I do think it shows us that they are collaborating more than ever and getting closer and closer.

至于会不会真的合并，说实话现在还太早下结论。

Whether that means there's gonna be a merger or not, I mean, it's too soon to say.

很好。

Great.

Grace，非常感谢你今天来做客。

Well, Grace, I want to thank you for coming on.

这位是Grace Kay，我们《The Information》的马斯克记者。

That is Grace Kay, our Elon Musk reporter, here at The Information.

接下来这位嘉宾是监管事务方面的专家，更重要的是，在科技政策比以往任何时候都更动荡的今天。

Our next guest is an expert on all things regulation, and more importantly, in an era where tech policy is just about as dynamic as it has ever been.

Bradley Tusk，Tusk Ventures的创始人兼首席执行官，是少数几个能帮我们理清这些头绪的人之一。

Bradley Tusk, founder and CEO of Tusk Ventures, is one of few people that can help us make sense of it.

布拉德利，欢迎来到节目。

Bradley, welcome to the show.

很高兴你再次来做客。

It's great to have you back.

是的。

Yeah.

谢谢你们再次邀请我。

Thanks for having me back.

好的。

Okay.

Fable 又回来了。

So Fable is back.

才过了两个多星期。

It was just over two weeks.

我们又开始运营了。

We're back in business.

那你对于这一整段经历有什么感想？

What, what were your reflections on this whole saga?

我觉得，从这一切中能得出的最大教训是，这其实是一片空白领域。

You know, I mean, I think that the broader lesson to take from all of this is this is all a white space.

对吧？

Right?

仔细想想，这其实也不算太罕见，毕竟优步当初也是这样一个空白领域。

And if you think about it, it's not totally unusual in the sense that Uber was a white space.

加密货币领域当时是一片空白。

Crypto was a white space.

自动驾驶汽车也是一片空白。

Autonomous vehicles were the white space.

按照定义，看这档节目的人通常都在探索并创造一些前所未有的新技术。

By definition, the people who are watching this show are involved in exploring and creating technology often that hasn't been done before.

不然的话，大家为什么要去做呢？

Otherwise, why would you do it?

说到监管，但凡是个新东西，往往根本就没有现成的框架可以套用。

And when it comes to regulation, frequently, if it's a new thing, there is no existing framework to work in.

要是政府官员自己能想出这些点子，然后去公司里干，他们早就去干了，何必还当官呢？

Because if the bureaucrats could have thought of the ideas and book them at the companies, they would do that instead of being bureaucrats.

所以不管是人工智能还是别的什么，关键问题就在于，我们该怎么应对这种新事物？

And so then whether it's AI or anything else, the question is how do we take this new thing?

人工智能的发展速度，简直是我见过最快的，我们得赶紧找到合适的监管方式。

And AI is obviously evolving faster than almost anything else I've ever seen and come up with the right regulatory framework for it.

我觉得在考虑人工智能监管时，不管是像Fable这样的产品，还是聊天机器人，都有几个关键因素需要理清楚。

And I think there are a few different competing factors that it's helpful to understand when you're thinking about AI regulation, whether it's something like Fable or a chatbot, right?

虽然像我们这样的人对AI充满热情，但普通大众并不买账。

Or anything else, which is one, as much as everyone like us likes AI and excited about it, the public does not.

公众对AI的反感非常强烈。

It is wildly unpopular.

不管是在红州、蓝州，还是在州级、联邦或地方层面，大家普遍都害怕自己会丢掉工作。

And that is true in red states, in blue states, at the state level, federal, local, people are very scared about losing their jobs.

他们非常担心前沿模型可能造成灾难性伤害。

They're very scared about catastrophic harm caused by frontier models.

他们害怕聊天机器人去干预心理健康，劝青少年去自杀。

They're scared about chatbots doing mental health, telling their teenagers to kill themselves.

这并不是说他们的所有担忧都有道理，但有没有道理其实并不重要。

And it's not to say that all of their fears are valid, but it doesn't matter if they're valid or not.

只要他们有这样的感受，代表他们的民选官员就会试图通过法律、政策和监管来回应这些焦虑。

If they feel that way, then the elected officials who represent them are going to want to reflect those anxieties in law, in policy, in regulation.

因为作为政客，你的目标就是赢得下一次选举，保住职位。

Because if you're a politician, your goal is to win the next election, stay in office.

而要做到这一点，就得让选民满意，对吧？

And the way you do that is making your voters happy, right?

所以第一件事是转变思维——要明白我们看待这个问题的方式，和他们看待的方式截然不同。

So the first thing is just a mind shift, which is to understand the way that we see this and the way that they see this are radically different.

据我观察，初创公司创始人在应对监管时最大的失误，就是觉得自己比他们聪明得多，懂得多得多。

And where founders in my experience fail the most around regulation is when they think, well, I'm so much smarter than they are and I know so much better than they do.

他们为什么不能听我的，照我说的做呢？

Why can't they just listen to me and do what I say?

答案就是他们根本不在乎你。

And the answer is because they don't care about you.

对吧？

Right?

他们在乎的是自己的政治利益、选举和职业生涯。

They care about their politics, their election, their career.

如果你没法用对他们政治利益有利的说法来解释你的需求，那你再聪明也没用。

And if you can't explain what you need in terms that work for them politically, it doesn't matter how smart you are.

不管你是斯坦福毕业的，还是YC或者AC十六号投了你的种子轮，这些都没关系。

It doesn't matter that you went to Stanford or you and Y Combinator or AC sixteen invested in your seed round.

这些对州长、州参议员或者其他官员来说，完全无关紧要。

Like that's all totally irrelevant to a governor, state senator or whatever else.

所以，就是这样。

So that's yeah.

你讲吧。

Go ahead.

不过，你认为这对我们未来的模型发布意味着什么？

What do you think this tells us about future model releases, though?

我是说，我们已经有了那个AI行政令。

I mean, we've got this we had the AI executive order.

大概有个六十天的等待期，或者从某个框架出台前三十天左右吧。

Have a sixty day period, I guess, waiting period, or thirty days out maybe from some kind

的

of a

框架。

framework.

你觉得以后模型是不是都会卡在这个中间状态了？

Do you think that models are going to get stuck in this purgatory phase from now on?

说实话，大概率会吧。

You know, yeah, probably.

但现实是，也许我说这话时对创新的支持还不够，但如果你推出的技术拥有如此强大的力量，能深刻影响社会，却指望完全不受任何监管审查，那简直太不现实了。

But the reality is, and maybe I'm not being pro innovation enough when I say this, but if you are releasing technology that has this level of power and this ability to to shape things, to expect there to be no sort of regulatory review is totally unreasonable.

对吧？

Right?

说真的，特朗普的那项命令可能太温和、太宽松了，你说是不是？

If anything, Trump's order was probably way too milquetoast and way too lax lax, right?

所以呢——

So like-

没错。

Right.

那只是自愿性质的。

It was voluntary.

就是说，你想干嘛就干嘛。

It was like, do what you want.

自愿的三十天审查期。

Voluntary thirty day review.

你想啊，要是有系统能黑进任何公司的基础设施，或者理论上能不经过任何人授权就发动无人机袭击，这种情况下，政府想介入、有必要介入，一点都不过分。

You know, if you're talking about systems that can hack through any corporation's, you know, infrastructure, if you're talking about systems that in theory could, you know, launch drone strikes without any sort of human, you know, authorization or whatever else, the idea that government would want and need to have a say in all of that is not unreasonable at all.

所以我觉得，不管你是不是Anthropic，只要在发布产品，就得把这一点考虑进去。

So I think that if you're doing a product release, whether you're Anthropic or anyone else, you have to build that into account.

再者，这可不是事后道歉就能解决的事儿。

And the other thing is, you know, this is not a beg for forgiveness situation, right?

这事儿要么提前请示，要么至少得主动一点。

This is some combination of ask for permission or at least be proactive.

还有，古希腊的伯里克利说过一句话：你不去关心政治，不代表政治就不会关注你。

And the more that you know, there's a quote from Pericles that just because you don't take an interest in politics doesn't mean politics won't take an interest in you.

你要是把头埋进沙子里，觉得他们不会注意到，或者等出了问题再说，又或者以为他们太蠢，等发现你有多厉害，就会乖乖听你的——

And the more that you put your head in the sand and either think, oh, they won't notice or we'll deal with that problem when it comes or they're so stupid that when they realize how smart I am, they'll just do what I want.

那你就输定了。

That's how you lose.

对吧？

Right?

你得主动去搞清楚，哪些政治力量对这件事有管辖权。

And you need to be proactive and thoughtful around who politically has jurisdiction over this.

他们会关心些什么？

What will their concerns be?

你该怎么打消他们的顾虑？

How do I alleviate them?

你拖得越久，到时候就越容易碰壁，因为他们一插手，就会拦下你的发布，或者你打算做的其他事。

And the more that you wait till the last minute, the more you're going be frustrated because then they are going to come in and stop you from your release or whatever else you want to do.

所以你得像对待工程、融资、市场推广那样，聪明又主动地处理这件事。

So you got to be smart and proactive about this, just like you are with your engineering, your fundraising, your go to market and everything else.

没错。

Right.

如果你要分析一下 Anthropic 和 OpenAI 在政治策略上的不同，就说 OpenAI 吧，我们刚请了他们的 AI 与政治记者来做节目，OpenAI 在 5.6 版本上就采取了更主动的做法。

So if you were to sort of take apart the differing political playbooks of Anthropic and OpenAI, I mean, OpenAI, we just had our AI and Politics Reporter on the show, OpenAI took a more proactive approach with 5.6, right?

他们主动去找政府，说：我们打算分阶段发布。

They went to the government, they said, Hey, we want to do a phased out release.

Anthropic 也做了类似的事，不过后来他们推出了 Fable，结果陷入了一场大麻烦。

Anthropic sort of did the same thing, although then they came out with Fable, and then they landed themselves in this whole debacle.

但归根结底，Fable 还是先发布了。

But at the end of the day, Fable is out sooner.

说白了，结果就是 Fable 发布了，而 5.6 还没出来。

Mean, the net result is Fable is out and 5.6 is not.

那你来分析一下这两家公司策略上的不同之处。

So unpack the differing playbooks here of these two companies.

你觉得哪家做得更好？

Which one do you think is doing it better?

你对双方各自的做法有什么批评意见？

What critiques do you have for both sides?

你怎么看？

What do you think?

总的来说，我觉得 Anthropic 比 OpenAI 做得更到位。

So overall, I would say Anthropic is doing it meaningfully better than OpenAI.

哦，有意思。

Oh, interesting.

我觉得，咱们先把这个具体问题放一放，回头再谈。

I think and, you know, we'll put this specific thing aside for a moment and come back to it.

但总的来说，这就像卢克·天行者和达斯·维达。

But overall, I think it's kind of Luke Skywalker and Darth Vader.

我觉得OpenAI在很多方面已经把自己扮成了达斯·维达。

And I think that OpenAI has in many ways embraced the role of Darth Vader.

而Anthropic可能并没有更道德、更讲伦理，但他们给人的感觉就是这样。

And I think that Anthropic, they may not be any more moral, any more ethical, anything else, but they come off that way.

对吧？

Right?

所以当他们拒绝做国防部的项目时，这个做法之所以高明，是因为第一，他们其实根本没法把国防部彻底排除在外。

So when they refused to do the DOD stuff, the reason that was so smart is one, they couldn't really actually take them out of the system.

不过说到底，那时候我们正处在一场真正的战争中。

Anyway, we're literally in the middle of a war when that happened.