Did you know you can't steal a charity? Don't worry. Elon Musk will remind you.

大家好，欢迎回到《Equity》，这是TechCrunch关于初创企业商业动态的播客。

Hello, and welcome back to Equity, TechCrunch's podcast about the business of startups.

今天是5月1日，星期五。

Today is Friday, May 1.

我是TechCrunch的交通领域编辑柯尔斯滕·科罗塞克。

And I'm Kirsten Korosec, transportation editor here at TechCrunch.

今天和我一起的是资深记者肖恩·奥凯恩。

And today is just senior reporter Sean O'Kane.

安东尼今天不在。

Anthony isn't here today.

我们会想他的。

We will miss him.

通常我们会从科技领域轻松有趣的话题开始这档节目，比如一些好玩的故事。

Normally, we kick off the show usually on the lighter side of tech, you know, sometimes a fun story.

但今天有点不一样。

This is a little bit different.

肖恩，你和多米尼克·米多里·戴维斯做的这个调查，能给我们捋一捋吗？我觉得这对那些正准备或刚刚庆祝公司被收购的创业者来说，是个特别有价值的教训。

Sean, catch us up on this investigation that you and Dominic Midori Davis did on a startup, I think it has a valuable lesson for founders who maybe are celebrating or about to celebrate their startup getting acquired.

这正是我们想说的核心。

That's really the thrust of this.

这周我们写了一家叫 Sallie 的公司，创始人叫克里斯托弗·格雷，2013 年就创办了。

We wrote this week about this company called Sallie that was founded by this guy called named Christopher Gray back in 2013.

那会儿正是 iPhone 和智能手机兴起的早期，他本身出身普通，没多少机会上大学，但发现了一个机会——做一款应用，帮学生更精准地找到那些他们完全够资格申的奖学金。

A really like early iPhone and like smartphone revolution kind of company where he, you know, as as someone who came from like a relatively poor background and didn't have a lot of access to college, realized that there was an opportunity to build something, you know, in particular at the time an app to better match not only find, but match students with scholarships that they were well qualified for.

结果这东西真火了。

And, you know, it it really caught on.

在应用商店刚起步那会儿，它一度是热门的 99 美分应用之一，后来还上了《鲨鱼坦克》。

It was like one of those like really popular 99¢ apps in some of the earlier days in the app stores, and that led them to Shark Tank.

2015 年他上了《鲨鱼坦克》，根据我们的调查，确实拿到了几位鲨鱼的投资。

And he was on Shark Tank in 2015, got some money from some of the sharks in our reporting.

我回头翻了翻，发现节目主持人之一发过一个 Instagram 短视频，说那是他们录节目以来最激烈的争抢战之一，谁都不想错过投资这家公司，我觉得特别有意思。

I went back and found an Instagram reel from one of the hosts of the show who sort of referred to it as like one of the all caps worst fights that they ever had of who got to invest in the company, which I thought was pretty funny.

从那以后，业务就稳步发展，累计收入达到了大约三千万美元。

And then, you know, it sort of cruised along from there and grew to I believe it was something along the lines of like $30,000,000 in cumulative revenues.

算不上特别庞大，但在这个大学资助的小领域里，已经是个非常稳健的生意了。

So nothing incredibly huge, but a really solid business on this one particular piece of the college puzzle.

到了2023年，这家公司被Sallie Mae收购了。

And then in 2023, they got acquired by Sallie Mae.

Sallie Mae，就是那个很多人熟悉的长期学生贷款机构。

Sallie Mae, the sort of like long time student lender that a lot of people are familiar with.

对Gray来说，这不仅是对他和小团队打造的产品的极大肯定，更意味着他们终于能实现Sallie的真正目标——让服务完全免费。

And for Gray, this was, you know, not just supreme validation of the product that he and his small team built, but also it was the way that he felt they were gonna be able to really true reach the true goal of Scally, which was to make the service free.

而Sallie Mae当时承诺的是，收购后整合进平台，就彻底免费。

And that was Sallie Mae's big promise was this is gonna be free after we acquire it and and sort of integrate it into our platform.

但实际情况并没有完全如预期那样发展。

But it didn't quite turn out exactly as expected there.

我的理解是，收购刚完成的那段时间，一切还算顺利。

So my understanding is that in the beginning of, right after the acquisition, it was okay.

但后来他开始对Sallie Mae如何打包这些服务以及数据收集的方式产生了严重疑虑。

But then he started having some serious questions about how Sallie Mae was sort of packaging these and the the data collection piece.

而且他担心，参与这项服务的用户，也就是那些客户，是否真的清楚发生了什么。

And and whether the people who are participating in this, meaning the customers, actually knew what was going on.

是的。

Yeah.

我觉得这里有两点重要的教训。

Think there's two big lessons here.

第一，公司被收购后，很难继续掌控自己亲手打造的产品，哪怕你自以为在协议里已经加了再多的法律保护。

One is that it is very hard when your company gets acquired to, you know, keep control over the thing that you've built, regardless of sort of how many protections you think you are able to sort of lawyer into the agreement.

第二，天下没有免费的午餐。

The other is that nothing is free.

这道理其实不新鲜，但我们每个人都应该时不时提醒自己这一点。

And and that's not necessarily surprising lesson, but it's one that, you know, we should all kind of come back to as much as we can.

在Gray的位置上，他非常保护使用Sallie的学生们的隐私数据，不只是因为他想服务像他一样出身于少数群体或贫困家庭的人，更因为他和这些人有共鸣，觉得有责任保护他们。

You know, in in Gray's position, he was very protective of the data of the students who were using Sallie, not just because he was really going after people who were like him and were coming from minority communities or impoverished communities, and so therefore he felt like a kinship with them and and a need to protect them.

而且学生大多是未成年人，就算文件上有父母签字，他们也可能根本没真正明白自己在做什么，不管是背上了巨额债务，还是只是申请奖学金时交出了个人信息。

But also because students tend to be minors and they might not you know, even if they have a parental signature on a document, they might not really fully understand the trade offs of whatever they're getting into, whether that's a ton of debt or even just searching for scholarships and handing over some data.

这周我们报道的是，格雷已经对萨利梅公司提起了诉讼，指控其不当解雇，核心论点是萨利梅违背了合同——他声称公司高管曾承诺不会出售这些数据，结果却偷偷卖了。

And, you know, what we wrote about this week was that Gray has filed a lawsuit against Sallie Mae, citing wrongful termination, basically arguing that the contract was breached by Sallie Mae going out and actually selling this data when he says that he was promised by executives at Sallie Mae not to sell it.

我们还深入探讨了一下，萨利梅公司具体是怎么操作的。

And, you know, we really get into a little bit of the mechanics of how we think Sallie Mae is doing this.

他们在拓展业务、不再只做贷款之后，另外成立了一家公司。

They essentially started this other company as they expand beyond just lending.

这家公司就叫‘Sallie’，但你要是把两家公司的网站放在一起看，简直让人分不清——用的还是同一个标志。

Just called Sallie is frankly very confusing when you look at the two websites side by side, they use the same logo.

这确实容易让人混淆。

And that can be kind of tricky.

萨利梅是一家上市公司。

Sallie Mae is a public company.

我在他们的SEC文件里没看到多少关于这件事的提及。

I didn't see a lot of mentions of this in their SEC filings.

这正是格雷所质疑的问题之一。

That's one of the things that Gray was contending was wrong here.

但最终结果是，现在实际在运营Scally的这家公司，确实在隐私政策中说明，它会收集这些用户的数据，并出售给广告网络、品牌，甚至其他正在寻找生源的大学。

And but the, you know, the end the end result here is that Sallie, the company that's actually using Scally now, is you know, does appear to and then discloses in its privacy policy that it is taking data from these users and selling it to ad networks and brands and even other universities that are looking for leads on enrollment.

是的。

Yeah.

我们在发布前就多次讨论过这个问题。

We've talked a lot about this prior to publication.

我很想知道他这场诉讼会不会成功，也可以猜一猜他们能不能赢。

And to me, it will be interesting to see if he has any success in the lawsuit, and and we can speculate if they will.

但这里其实传递了一个公开的信息，不只是对创业者，也对用户——我们都觉得自己的数据迟早会被用掉，但你真得仔细看那些小字条款。

But there is sort of a public message here, not just to founders, but also on the customer side, which is I think we all think of, well, our data is probably gonna be used, but you really need to read the fine print.

关键在于，即使他们写出来了，这信息够明显吗？

And it really comes down to this whole question of even if they disclose it, is it obvious?

写得够清楚吗？

Is it clearly stated?

是不是写得明明白白，让所有参与这些公司活动的人——比如萨利梅——都能一眼看懂？

Is it boldly stated so that anyone who is participating in whatever these companies are doing, and in this case, Sallie Mae, that it's very clear.

不过这周我们要聊的内容可真不少。

But we have a lot to get into this week.

我们有宝马iVentures的三亿美元基金，还有国防科技公司Scout AI最近的一轮融资。

We have deals from BMW I Ventures and their 300,000,000 fund, a recent fundraise from a defense tech firm called Scout AI.

接下来我们会聊OpenAI对埃隆·马斯克的诉讼，再深入分析‘七巨头’的财报。

We're gonna get into the OpenAI Elon Musk trial and then dig into the earnings from some of the magnificent seven.

咱们先说说宝马iVentures。

Let's get into BMW I Ventures.

其实这篇报道是我写的。

I actually wrote this story.

所以，肖恩，你有什么问题想问我吗？

So, you know, Sean, do you have any questions for me?

三十亿美元的基金。

300,000,300 million dollar fund.

你想知道些什么？

What do you wanna know?

我一直对企业的风险投资部门很感兴趣。

I've long been fascinated with corporate venture arms.

我记得刚入行那几年，很久以前曾采访过捷蓝航空的投资部门。

I remember one of the first few years that I started doing this job a a long time ago talking with like JetBlue's venture arm.

当时我对他们居然会做这种事感到特别有意思。

And I was just very tickled at the idea that they would even do something like that.

那时候，企业做投资的动机差别很大，有些纯粹是跟风，为了能参加硅谷那些时髦的派对；另一些则是真心想通过投资，最终为自己的主营业务带来好处，比如发现潜在的收购目标，或者找到能用得上的技术。

And, you know, I think back then, there was a a bigger divide between companies that were clearly doing it just because it was like the hot thing to do and get them into some, like, cool Silicon Valley parties and others that were really using it in a way that could eventually benefit, you know, their core business in some way, whether that was, you know, finding acquisition targets or finding technologies that make sense to them.

我第一个问题是，对你来说，这感觉怎么样？

My first question to you is sort of like, you know, what what does this feel like to you?

你觉得宝马这个基金，只是单纯为了做投资而做投资吗？

Does this feel like one of those ones that's a venture fund at BMW just for the sake of doing it?

还是你觉得他们真的在认真寻找那些未来能帮到他们的项目——不只是带来财务回报，也许还有技术上的突破？

Or do you feel like they're actually putting some effort behind this to go out and find stuff that that could actually benefit them down the road, not just financially, but maybe even technologically?

是的。

Yeah.

我觉得是后者。

I would say the latter.

首先，这是他们的第三支基金了。

I mean, they've had, first of all, this is their third fund.

之前的基金，每次间隔大约五年。

So their previous fund and each time, it's about five years in between.

目前，他们管理的资产总额已达11亿美元。

So they have, $1,100,000,000 of assets under management at this point.

这一轮基金规模是3亿美元。

This fund is 300,000,000.

上一轮也是3亿美元。

The previous one was 300,000,000.

第一轮稍微大一点。

The first one was a little bit larger than that.

大概在5亿多一点。

It was more around, a little bit more than 500,000,000.

每次他们都会聚焦某个特定领域，这次押注人工智能，其实也不意外。

And each time they do kind of a specific focus, this one, not surprising on AI.

不过我觉得，他们早就认真布局很久了。

But they've been, I think, pretty serious for a while now.

他们有两个办公室。

I mean, they have two offices.

他们的管理合伙人之一常驻硅谷的办公室。

They have one of their managing partners is located in their office in Silicon Valley.

另一个在德国。

The other is Germany.

他们投资的这些公司，对母公司宝马集团——也就是唯一的有限合伙人——都有很明确的助益。

The these are companies that they're, investing in that may have a very clear benefit to BMW AG, the parent company, which is the only LP.

我觉得这点特别有意思。

I think that's really interesting too.

有时候，这些战略性的风投部门也会引入其他资金。

Sometimes with these strategic venture arms, they're taking in other money.

虽然这里只有一个有限合伙人，但他们依然保持独立运作。

There is one one LP in this, but they're still considered independent.

所以简短回答就是，他们确实是认真的。

So I do think that they're serious is the the short answer.

有意思的是，从采访一开始他们就说‘我们不追风口’，可他们偏偏聚焦在人工智能上——而这恰恰是个大热领域。

You know, the question to me that was really interesting is that from the very beginning of the interview, they're like, we don't chase trends, but they are focused on AI, which, you know, is pretty trendy.

对他们来说，人工智能不是一时的潮流，而是一项基础性技术，会催生出一系列其他技术，最终重塑汽车行业。

For them, they think that this is AI is really more foundational and that will create a number of other technologies that in their view will reshape automotive.

有趣的是，我不知道我之前有没有提到，他们还说这项技术也能用在其他行业。

Interestingly, I don't know if you caught this in my story, but they say also though that this could be applied to other industries.

他们也在寻找那些业务超出汽车领域的初创公司，尽管汽车仍是他们的主要方向。

They they are looking for for startups that go outside of automotive as well even though that's the primary focus.

这让我觉得特别有意思，原因大概有两个。

This is interesting to me for probably two big reasons.

第一，去年我写过一个叫Leitmotif的新基金，它跟你刚才说的那种技术上独立的基金很像，但它的唯一有限合伙人是大众汽车。

One, last year I wrote about this new fund called Leitmotif, which is kind of the other version of of what you were just just describing of a technically independent fund, but the single LP in that same $300,000,000 fund is is Volkswagen.

所以它们和这家基金相比，关系上稍微远一点，但目标类似，都是想切入这些新兴且快速发展的领域。

And so they're they're a bit more at an arm's length than it seems like this one is, but it was similar remit of, like, getting into these emerging and fast moving industries.

我觉得有意思的是，现在全球两大德国汽车巨头都在加大投入，很可能是为了投资彼此竞争的同类技术，这让我觉得挺有意思。

And and I think it's interesting that now we have two of the biggest German automakers and automakers on the planet really stepping up their efforts to, you know, I would imagine invest in like competing versions of some of the same technologies, which I find that sort of tickles me a bit.

但另一点是，现在有大量初创公司正试图在汽车产业链的各个环节抢占位置，尤其是从人工智能的角度。

But the the other thing is there are a ton of startups that are trying to carve out space in the entire automotive chain, especially from an AI perspective.

我甚至写过一些从消费者端切入的公司。

I've written about a few even from the customer side.

有些公司正在打造基于AI语音助手的业务，帮助客户安排经销商预约和车辆保养。

There are companies that are trying to build businesses around, you know, sort of like AI voice agents helping out with dealership scheduling and maintenance.

还有些公司利用AI扫描技术检测车辆划痕或缺陷，合作对象包括赫兹这类租车公司。

There's ones that are working on detecting defects or scratches, you know, using sort of AI assisted scanners that are working with company rental companies like Hertz and things like that.

这里头有太多值得评估的机会了，希望能推动发展，也许未来还能帮上宝马。

So there's a ton of stuff to be evaluating here and and hopefully help grow and maybe find a way to to help BMW down the road.

但我认为，总的来说，过去一两年刚起步的这些AI原生初创公司，我跟他们聊的时候，最常问的问题就是：你们有足够的底气能独立撑好几年吗？

But I also think that, you know, broadly speaking, a lot of these AI first startups that are just getting off the ground over the last year or so, The big question when I talk to them is always like, do you have enough to stand alone as a company for for years?

还是说，你们只是暂时领先，最终会被收购？

Or is this something that like you are getting ahead of others and you'll eventually be acquired?

我不禁想，宝马在观察这些公司频繁更替的时候，会不会觉得：有些可能出乎意料地对我们有用，但也有不少很快就会被并购、整合进别的公司里。

And I wonder how much BMW is looking at some of the churn here and just saying, like, some of these might pan out for us and and in surprising ways even, but also some of these are probably gonna get exited fairly quickly by getting rolled up into something else.

这算是一种不错的副业收入，比单靠汽车主业多赚点钱也挺好。

And like, it's a, you know, an okay way to make a couple extra bucks on the side compared to just the automotive business.

对。

Yeah.

我觉得，从战略投资的角度来看，很多基金都是这么想的：他们当然期待其中一些能带来大额退出，但也有不少是抱着广撒网的心态在布局。

I mean, I think that's always I think generally speaking, a lot of strategic funds look at look at it in that view, which is, you know, there there are ones that they're absolutely expecting to have, like, pretty, hopefully, large exits, but there are ones that they're sort of spreading their butts out.

有意思的是，这些初创公司不少，而且整个行业都在出现这种现象。

You know, I think what's interesting here is a number of these startups and we're seeing this across the industry.

这不只是宝马i在投的公司，而是越来越多AI原生的公司，目标都是想重塑工作流程。

This isn't just, you know, the companies that BMW I is is investing in, but the the idea of companies that are AI first but that are seeking to change workflows.

Scenera 就是他们投资的公司之一，最初是一家做 SaaS 的公司，主要在自动化一些工作流程。

So Scenera is one of these companies that they have invested in, and they started out as a SaaS company that were sort of automating some workflows.

但现在他们在平台基础上添加了不少 AI 代理。

But now what they've done on top of their platform is they've added a bunch of AI agents.

所以他们的产品完全聚焦在设计和工程流程上。

So it is completely focused on the design and engineering process.

这样一来，你就能明白为什么这对汽车行业特别有颠覆性，同时也能应用到任何涉及工程和设计的行业。

So suddenly, you can see how that could be really transformative for automotive, but also apply to really any kind of industry that uses engineering and design.

这个产品让他们能快速迭代。

And this basically, this product allows them to, like, iterate really fast.

我觉得未来会有更多类似的初创公司冒出来。

And we're I think we're gonna see a lot more of these types of startups pop up.

这是宝马投资的那一家。

This is the one that BMW invested in.

你知道，现在只要是跟 AI 有关的东西，对很多公司来说都特别有吸引力。

You know, I think anything AI related is very, very sexy to a lot of companies.

我得提一下，跟宝马没关系，但软银最近投资了一家机器人公司，据说他们在建数据中心。

I have to mention, separate from BMW, but, you know, SoftBank just had this, investment in a robotics company that is, supposedly building data centers.

两个热门话题：人工智能和数据中心。

Two hot topics, AI and data centers.

而且，他们已经瞄准了千亿级别的IPO，简直疯狂。

And, apparently, they're already eyeing a $100,000,000,000 IPO, which is absolutely insane.

我们直接跳过了所有中间步骤。

Like, we're just skipping all the steps.

我都怀疑我们是不是已经过了AI的巅峰期了。

I I don't even know if we're peak AI anymore.

我觉得我们现在根本不在地球上，而是在外太空。

I think we are, like, in somewhere in space.

我们现在完全在轨道上运行了。

We're in orbit right now.

我不确定什么时候能回落到现实，但这说明了这些基金对任何沾点AI边的东西都极度感兴趣。

I mean, I don't know when it's gonna come back down to Earth, but it just shows how interested these funds are in literally anything that is touching AI.

是的。

Yeah.

我觉得软银这件事挺有意思，因为我也在想，宝马接下来会在实体AI领域怎么布局。

I mean, the the SoftBank thing is interesting to me in the sense of I also wonder about where v BMW will look as far as the physical AI stuff that's happening.

比如，我马上想到Mind Robotics，就是瑞拜最近刚拆分出来的那家公司。

Like, we you know, my mind goes to Mind Robotics, the company that Ribian just sort of recently spun out.

未来几年，肯定会有大量非炫酷人形机器人、但更偏向工业场景的AI与机器人应用出现，目的就是升级和优化我们的制造方式。

Like, there's so much that people are gonna try over the next couple years that may not be flashy humanoid robotics or anything like that, but is like an industrial application of AI and robotics meant to modernize and make more efficient the way that we build things.

所以我觉得宝马也可能通过这个基金进入这个领域，这其实也正是软银投资Grose AI这家公司的意图所在。

And so, like, I could see BMW playing in that space here too with this fund, and that's clearly an element of what SoftBank is aiming at with this this company called Grose AI.

是啊，他说他们已经打算搞IPO了，听上去是这么回事。

That, yes, he said they're already, you know, supposedly eyeing an IPO, which, you know, listen.

马赛的野心可不是一般的强，用‘有野心’来形容他都算克制了。

It Masa is nothing if not ambitious, which is maybe the most restrained way of describing Masayoshi's own.

这恐怕都不是他第一次提出百亿美金级别的上市计划了。

That's probably not even the first one that he's floated at a $100,000,000,000

哦，我肯定还有其他的

Oh, I'm sure there are

有的。

others.

今年就有。

This year.

比如说

Like

是的。

Yeah.

我肯定还有其他的。

I'm sure there are others.

你知道吗，现在另一个热门领域，这个趋势已经持续了十八个月，甚至两年了，就是国防科技。

You know, what another hot area right now and and this has been going on for for about eighteen months, even two years, is defense tech.

我们之前有个故事，是蒂姆·弗罗恩霍尔茨写的。

And we had a story Tim Frohneholz wrote about.

他亲自去实地考察了他们的工作。

He went and actually, you know, visited what they're doing.

这家公司叫Scout AI，正在开发用于操作军用车辆的AI模型。

And this is a company called Scout AI, and they're building AI models to operate military vehicles.

他们在5000万美元的种子轮之后，又融了1亿美元。

It's they've they've raised a 100,000,000 after a 50,000,000 seed round.

这里面有很多值得深入探讨的地方。

There is a lot to unpack here.

在我看来，有意思的是，三年前你可能很难看到哪家公司公开说自己在搞武器或军事相关的项目。

To me, what's really interesting is I don't know if three years ago, you would see companies flouting the fact that they are working on anything to around weapons or the military.

但现在情况真的变了。

And that has really changed.

我也不太确定。

And I don't know.

你对Scout AI，还有其他类似公司，有什么看法？

What are your thoughts about not just Scout AI, but, you know, other companies?

你现在是不是经常看到，他们公开谈论并宣传自己如何把AI用在武器上？

Are you seeing this a lot more often where they're really talking and promoting how they're using AI in weaponry.

几年前确实有公司这么做，但那时候它们绝对是少数派。

There were companies that were doing it a few years ago that they were certainly the outliers.

对吧？

Right?

那时候像Palantir、Anduril，可能再加一两家，算是比较出名的。

You had Palantir or Anderil and, and maybe one or two others that were really high profile enough.

但特朗普第二次当选后，一切都变了——政府跟这些人走得特别近，甚至直接请了他们以前的同事。

But everything changed when Trump was elected for a second time, and and the way that the administration has aligned itself so closely with, you know, quite literally some of these folks, like former coworkers.

对吧？

Right?

比如特拉维斯·卡兰尼克在Uber时期的得力助手埃米尔·迈克尔，现在就在政府里任职。

Like, we have one of the top sort of lieutenants to Travis Kalanick in his tenure at Uber, Emil Michael, working in the administration.

像这样的例子还有很多，遍布整个特朗普政府。

That's just one of many who are, like, spread across the Trump administration.

所以我觉得这给了他们很大的掩护，再加上当时大规模放松监管的浪潮，以及国防领域大量资金涌入，让更多人敢站出来，更公开地宣称自己在做这些事。

And so I think that bought a lot of cover, you know, that plus this idea that, you know, there was also, like, just a crazy deregulation push happening and also a lot of funding in the defense space, you know, gave a lot of cover for more people to come out and more boldly proclaim that this is what they're doing.

而且确实，这类事情多得不得了。

And and, yeah, there's just a ton of this stuff.

比如赫尔墨斯这家公司，他们正在研发高超音速飞机，我几周前还写过他们，最近刚融了一大笔钱——虽然他们的名气还没大到像安德罗尔那种程度。

I mean, like, Hermes is a company that's building, like, hypersonic plane that I wrote about a couple weeks ago that just raised a bunch of money, and and is not maybe a company that people have heard of on the scale of, like, an Andoril.

总之这类公司越来越多，以后还会更多。

So there's just a lot of it, and there's gonna be a lot more of it.

我觉得这篇报道挺有意思，蒂姆采访了科比·阿戈克，他是这家公司的负责人，同时也是布雷特·阿戈克的弟弟，而布雷特掌管着人形机器人公司Figure AI。

I thought this piece was interesting in the way that Tim sort of talked to Colby Agcock, who runs this company, and and is the brother of Brett Agcock, who runs Figure AI, the the humanoid robotics company.

报道里谈到他们如何利用这种公开表态来争取融资和抢订单，跟谷歌这类大科技公司形成鲜明对比——尽管谷歌也跟特朗普政府走得很近，但内部仍有不少人反对自家技术被用于战场和武器系统。

You know, about that piece of this and how they are sort of, you know, trying to press their advantage as far as fundraising and snapping up contracts goes by being proud of it versus being part of the larger, you know, conglomerate tech companies like a Google, where there is still even though they have aligned themselves with the Trump administration in many ways, there is still resistance inside those companies to see their technology be used on the battlefield and to direct weapons.

关于Scout AI，有两点特别值得注意：他们的目标是打造用于军事领域的大型语言模型，最终实现军事级通用人工智能。

A couple of things I think worth noting specifically about Scout AI is that their goal here is that they're building large language models to create military AGI.

这确实是个相当宏大的目标，他们正在为此努力。

So that's that's a lofty goal that they're working on.

他们的说法是，他们使用的是视觉-语言-动作模型，也就是VLA模型，我最近听到越来越多的初创公司都在采用这种技术，目的是减少附带损伤。

And their argument is that they're using vision language action models or VLA models, which I am hearing so much more from startups that this is this is the approach that they're taking, to reduce collateral damage.

他们整个宣传点就是，通过这些AI工具，能实现更高的精准度，从而减少附带伤亡和人员死亡。

Their whole pitch is that by using these AI tools, it will be far more precise, and it will reduce collateral damage, the loss of human life.

这正是他们在推介时的核心论点。

That is their argument in their pitch.

至于他们能不能真正实现这一点，现在还不好说。

Whether they're, you know, successful in that, we have we have yet to see.

但国防部显然对这类初创公司很感兴趣，我相信未来还会有更多类似企业，其中一些可能连我们都没听说过。

But there's clearly interest, from the DOD in all sorts of startups like this, and I expect so many more and probably a a number of them that we don't we haven't heard of yet.

我还得提一下，他们正在和DARPA以及美国陆军合作。

I should also mention they're working with DARPA and the US Army.

所以他们已经有合同在手，看起来也已经取得了一些进展。

So they they have contracts and, you know, seem to be making progress already.

不过也得打个补丁，确实如此。

Caveat that with, yeah.

我不是说这不难，但很多公司都能拿到国防部和政府的合同。

I I don't wanna say it's not hard, but, like, a lot of companies get contracts with the DOD and and with the government in general.

他们一直在尝试各种新技术。

Like, they're always trying out new technology.

我不是想贬低他们，但这种助力未必像很多人想象的那么大。

I don't say that to dismiss these guys, but it is it is not necessarily the, like, tailwind that I think a lot of people make it out to be sometimes.

另外，词是有含义的，有时候我们真该记着这一点。

The other thing is, like, words have meanings and it's it's good to try and remember that sometimes.

好吧，也许他们真想减少战场上的附带损伤。

Like, okay, maybe they wanna help reduce collateral damage on the battlefield.

这话听起来不错，但当公司公开宣称这些时，我们该问问他们到底什么意思。

That sounds great in a vacuum, but I think it's worth when companies come out and and do say this stuff proudly, asking them what they mean when they say these things.

他们有没有一条明确的红线？比如，一旦发现政府——不管是现在的、未来的 administrations——在滥用技术，哪怕表面理由说得冠冕堂皇，他们会不会真的退出合作？

And, you know, do do they have like a bright red line that where they consider something to be misused and that they would actually walk away from a deal in some way because they see, you know, this administration or the next or the next, you know, using it in a different way but justifying it, you know, on the surface in order to get away with how they're doing it.

所以，每当我听到这些公司一边吹自己多安全、多理性，一边谈这些事，我就忍不住想这些。

So, that's just what comes to mind when I hear a lot of these companies talk about this this way and and and try to profess that they're also being safe and reasonable and and all these other things.

是的。

Yeah.

我的意思是，我们确实得继续往下聊，毕竟还有几件特别有意思的事要讨论。

I mean, we we do have to move on because we have a couple of other really interesting things to talk about.

但我要说，从技术角度讲，我们造出了一种极其精准、能减少附带损伤的武器，这和它实际被怎么用是两回事。

But I will say that there is a difference between saying technologically speaking, we have created something so precise that it will reduce collateral damage.

因为任何武器都可能被滥用，美国和其他国家在武器使用上的滥用历史已经很长了。

It is a very different thing of how it is used because it could certainly any weapon can be misused, and we have a long, long history, within The US and other countries of misuse of weaponry.

而且一旦签了合同，想抽身就非常困难了。

And and there's once you're under contract, it's very difficult to, you know, pull back from that.

现在Anthropic和美国政府在AI领域的互动，就正在上演这种情况。

And we're seeing that play out on the AI stage with Anthropic right now and the US government.

我觉得未来几期节目里，这方面还有更多内容可以聊。

So there's gonna be a lot more to talk about that, I think, in future episodes.

不过我更期待谈的是OpenAI和埃隆·马斯克的这场官司。

But we have something that I've been looking forward to talking about, which is the OpenAI Elon Musk trial.

我笔记里记了一句话，是马斯克最近反复提到的——‘你偷不了一个慈善机构’，今天他又说了这句话，还发了推特。

And I wrote down in my notes a a quote that Elon Musk seems to be, continuing to to come back to, which is you can't steal a charity, which was something he said today and that was tweeted out.

首先，咱们得把背景说清楚。

First of all, let's set the stage here.

这事儿到底怎么回事？

What is this about?

这是吃不到葡萄说葡萄酸，还是真有法律上的依据？

Is this sour grapes, or is there a legitimate reason behind the lawsuit?

是的。

Yes.

我觉得这更多是酸葡萄心理，目的就是尽可能把OpenAI，尤其是阿尔特曼，拖进泥潭，而不是真的在乎案子本身的是非曲直。

Like, I think it is more sour grapes and an attempt to, you know, drag OpenAI and and in particular, Altman, through the mud as much as possible, than it is about the merits.

不过我也觉得，马斯克确实够偏执，他真觉得这两者从根本上就是一回事。

Although, I do think Musk is fanatical enough that he really thinks the two are basically intertwined.

简单来说，事情是这样的。

Really short version of this.

这是马斯克对OpenAI提起的诉讼，大意是：我参与创办了你们。

This is the lawsuit that Elon Musk filed against OpenAI, basically saying, I helped found you.

我当初参与创办并投资时，以为你们会是一家非营利机构。

My understanding when I helped found you and fund you originally was that this was going to be a nonprofit.

但过去几年，我们眼看着这场闹剧一步步上演——萨姆·阿尔特曼决定，嘿，其实我们该转成营利性公司，否则没法持续融资、回报投资者，也撑不起我们想成为超大规模企业的野心。

We've seen over the last couple years this big drama sort of play out as Sam Altman has decided, hey, actually, we should be a for profit company and we need to be in order to keep raising all the money that we wanna raise and and pay back our investors and all this stuff as we become a hyperscaler.

这事一路闹到了法庭，现在就摆在我们面前。

And and it made it all the way to trial and and here we are.

当然，也有很多人觉得马斯克的立场其实站得住脚。

And, you know, I think, you know, there are certainly plenty of people who actually think that Musk has a good case here.

但更多人，尤其是法律专家，认为他的主张根本站不住脚。

I think there are more people who, especially legal experts, who's believed that he doesn't.

说白了，就是他钱多到能随便打官司，换别人早就不折腾了。

And that it's just the fact that he has more money than anybody else in the world that he can bring this to trial and that most other people wouldn't bother.

所以现在，我们就在这儿了。

And and so here we are.

不过话说回来，这一切已经为一场混乱的法庭对决铺好了路。

And, you know, all that said, that has set the stage for just a, you know, a pretty messy courtroom fight.

我觉得我们现在还只是刚开始，就在我们开始录这段的时候，马斯克才刚结束他的证词。

And we're only in the early stages of it, I think, you know, as as of the time that we started recording of it, Musk was maybe only just finally coming off the stand.

他是这个案子中第一个被传唤出庭的重要证人。

He was the first real big witness in this case called to testify.

他差不多连续在庭上待了三天之久。

And, you know, he's been on the stand for basically the better part of three days.

从现在开始，事情只会越来越离谱。

And and it's only gonna get weirder from here.

我的意思是，早在开庭前，就已经有不少料了。

I mean, there's already been a ton in the lead up to this.

已经曝光了大量短信、邮件和笔记、日记。

We've seen a ton of text messages released, emails, notes, diaries.

真的，太多了。

I mean, like, so much stuff.

就算庭审前筛掉了一大堆证据，剩下的东西还是乱得不行。

And even with all the stuff that gets excluded before we get to trial, there's still a lot of a lot of mess.

所以那些推崇近藤麻理惠整理法的人，估计挺开心的。

So Marie Kondo fans out there are probably pretty happy.

哦，这个冷知识引用得真到位。

Oh, that's a really good deep cut reference.

我喜欢这个说法。

I like that.

马斯克提出的一个有意思的观点是，人们坚持认为，收益有上限的投资者和收益无上限的投资者之间存在巨大差别。

I think that the interesting argument that Musk is making is that there's this insistence that there is a big difference between investors whose profits are capped and those whose profits are unlimited.

这是他核心论点之一，而且这个观点可以从好几个角度来理解。

And that's one of the key arguments he's making, and and that can be viewed a couple of different ways.

一种是酸葡萄心态，就是说——

The sour grapes way, which is, hey.

你其实自己也动过这个念头。

You actually thought about doing this yourself.

然后呢，把这些全都分开，这确实引出了一个关于像OpenAI这样的公司的整体性问题。

And and and then there's separating all of that, that is an interesting sort of overall question about a company like OpenAI.

抛开那些个人恩怨不谈——我觉得这让这场庭审变得有点像马戏团——真正让我觉得有意思的是，埃隆·马斯克的推文正在狠狠打他的脸。

What's really fascinating to me aside taking out all the personal stuff, which I think is makes this trial a little bit more circus like.

这里面涉及很多人物关系，而马斯克的那些推文确实成了他的把柄。

There's a lot of personalities in here, is that Elon Musk's tweets are absolutely coming back to bite him in the butt.

这其实也是我们报道这篇新闻时的一个重点。

And that was actually one of one of the focuses of of one of our articles and covering this.

之前提到的Tim Fernholz现在就在法庭里，一直在报道这场庭审。

Tim Fernholz, who we mentioned earlier, is actually in the courtroom and has been covering this.

说实话，这种情况我以前也见过。

And and, really, that's been something that I have watched happen before.

现在我们就站在这里。

And here we are.

你知道的，他是在宣誓后作证，有人拿出了一些推文来反驳他的观点，我觉得这特别有意思，也让我们对特斯拉在做什么有了更清晰的了解。因为其中一个问题就是，他之前发推说特斯拉正在研发通用人工智能（AGI）。

You know, he's under oath, and he's being presented with tweets that absolutely have contradicted some of his arguments, which I find really fascinating and also been kind of illuminating about what Tesla is working on because one of the questions was he had tweeted out that Tesla was working on AGI.

他在证人席上明确回答：没有。

And on the stand, he confirmed, no.

他们根本没有在做。

They're not.

正如蒂姆提到的，股东们要注意了。

Shareholders take note as as Tim mentioned.

我真希望这样的庭审能多一些，这样我们就能更清楚地了解特斯拉和SpaceX真正正在做什么，而不是只看他发在X（前推特）上的那些内容。

I wish we had more of these trials because I think that we would find out a lot more about what Tesla's and SpaceX are actually working on as opposed to what he, presents out on x, formerly Twitter.

我们每天看到的，往往是那些有权势的人试图自己操控舆论，而不必面对这种公开质询。

So much of what we see every day is, like, people in power who are trying to, you know, shape the narrative in their own way without having to do this.

这些人最终不得不在法庭上，或在审判前的证词环节中，面对宣誓后的直接提问。

Those people wind up in these situations like in a courtroom or in a deposition ahead of a trial where they're under oath and they actually have to answer these questions.

当然他们事先会做准备，但不可能完全预演到位——比如回到那些个人细节，马斯克第一天就被问到：西奥班·齐勒是谁。

Certainly, they're prepped, but they can't prep I mean, like, to get back into some of the personal stuff, Musk was asked on the first day to say who Siobhan Zilles was.

作为旁观者，我们知道西奥班·齐勒是他四个孩子的母亲。

And we we, as observers, know Siobhan Zilles is the mother of four of his children.

我们知道她曾经或现在在多家他的公司任职。

We know that she works and or has worked at many of his companies.

在我们这场庭审所讨论的那段争议时期，她曾担任过OpenAI的董事会成员。

She was on OpenAI's boards during a lot of the the sort of back and forth period that we're talking about in this trial.

她是个广为人知的人物，但马斯克当时却支支吾吾，估计是没准备好回答这种私人问题。

She's a very well known quantity and Musk just sort of stammered because he, I guess, wasn't prepared or wasn't ready to sort of answer that kind of question.

这种事说白了，就是这种案子常会挖出来的八卦和私人恩怨。

You know, that's more like the kind of looky loo personal drama stuff that comes out in a case like this.

但更重要的是，这种场合能真正逼他们给出实话，而他们最不想面对的就是这个。就连特斯拉的财报电话会，过去五六年也一直被夸奖，说他们带头重视散户，邀请散户参与电话会，回答他们通过Say Technologies等平台提出的问题，现在好多公司都学着这么做了。

But it's also this other stuff where you can really finally pin them down on answers in a way that, you know, they don't like, even Tesla's earnings calls, like Tesla's gotten a lot of credit for kind of leading the charge over the last five or six years of focusing on retail investors and bringing people bringing retail investors into the earnings calls and answering their questions from the, you know, the Say Technologies platform and now so many other companies do this too.

这确实不错。

And that's great.

让散户在决策中也有发言权，是好事。

Like, it's good to give retail investors a voice at the table.

不过那些问题都是提前提交的，他们有时间准备回应，这跟真正被逼问的情况不一样。

Those questions are also pre submitted and they get to prepare the remarks to them and everything and it's not

而且他们往往是粉丝。

And they're oftentimes fans.

对吧？

Right?

他们是有切身利益的。

There is a stake there.

他们可不是普通观众。

They're not DBS.

他们确实是股东，因为只有股东才能提交问题。

They definitely are because they're literal shareholders.

你得先持有股份，才有资格提交这些问题。

Like, you have to own shares to be able to to even submit those questions.

尤其是特斯拉，有些股东早就烦了，他们的提问经常被顶到最前面。

I will say, like, there are especially with Tesla, there are some fed up shareholders who often get their questions voted to the top of the queue.

但公司和马斯克本人，其实都准备好回答这些问题了。

But like, you know, the company and Musk himself are prepared to answer those questions.

这种情况在这批里可不一样。

That is not the case in this set.

而且他也没发誓作证。

Well, and he's not under oath.

我的意思是，他并没有发誓。

I mean, it's he's not

对，没错。

on Yeah.

没错。

Exactly.

他没发誓作证。

He's not under oath.

你懂的。

You know?

我觉得这才是关键区别。

And I think that that's, like, the big distinction.

就是说，咱们得承认，他在财报电话会议上确实挺开放的，大家也都在关注。

It's like, yeah, let's give him credit for, like, you know, having these fairly open, conversations on earnings calls, and people are paying attention.

这确实能直接影响股价，但他毕竟不是在法庭上宣誓作证。

And it can have a direct effect on stock price, but you're still not under oath in a court of law.

这两者的标准还是不太一样的。

There's a little bit of a different bar.

我们知道他对此很不满，因为本周他的证词里有个细节特别引人注目：他和OpenAI的一位律师有过一段交锋。那位律师一直在某个问题上紧追不舍，而马斯克的回应是，你的问题就是故意设套想忽悠我、骗我。

And we know he doesn't like it because one of the sort of things that really popped this week from his testimony is an exchange with one of OpenAI's lawyers where, you know, the lawyer was really pestering Musk on a particular issue and Musk's response was, like, your questions are designed to to fool me or deceive me.

而且就像你说的，我们可以看到，之前也有报道提到，法官在很多情况下不得不让双方都退一步。

And it's just, like, you know, you can see like, we've also seen reporting about a number of instances where the judge has really kinda had to back both sides really off.

但尤其是马斯克，法官主要是想让他回答这个问题。

But but especially Musk to just sort of answer the question.

法官的意思是，你今天是来配合调查、把信息记录在案的，而不是来争论说，比如，OpenAI律师的某些做法是不是越界了之类的。

Like, you're here to do a thing and and and get the information out into the record, not to argue whether or not, like, you know, the tactics of OpenAI's lawyer are really stepping over a line or something like that.

所以我们也知道，他其实不太喜欢这种局面，尽管这局面是他自己一手造成的。

So we know that he doesn't like being in these situations as much as this is, a situation of his own creation.

是的

Yeah.

他完全是自找的。

I mean, he put himself in this position.

我觉得我们接下来还会听到更多内容，而且不只是关于埃隆·马斯克的。

I think we're gonna hear a lot more, and not just about Elon Musk, obviously.

还有很多我们经常关注的人物即将出庭，比如萨姆·奥尔特曼和其他人，还有格雷格·格罗克曼。

There's so many more characters who we cover a lot that will be on who will be on the stand, Sam Altman and and others, Greg Grokman.

我真的很期待这些庭审。

So really looking forward to that.

我想在节目结束前聊聊所谓的‘七巨头’。

I want to close the show by talking about some of the magnificent seven.

这些上市公司占据了整个股市总市值的很大一部分。

These are the publicly traded companies that, really make up a huge amount of the total of the public exchanges value.

但并不是所有公司都已公布了财报。

Not all of them have, reported earnings.

苹果公司可能就在我们录这期节目的时候刚发布财报，不过已经有好几家公司公布了。

Apple is actually probably, as we're recording this, just about to, report earnings, but we have a number who have.

我说的是谷歌、微软、亚马逊和Meta。

And I'm talking about Alphabet, Microsoft, Amazon, Meta.

至少我看到的趋势是，如果你在做AI基础设施的‘卖铲人’——比如有云业务，那你就赚到钱了。

And at least the theme I saw was that if you are in the picks and shovels business of AI infrastructure so if you have a cloud business, you're making some money.

赚得不少。

You're making lots of money.

但同时花的钱也特别多。

You're also spending lots of money.

那这样算下来是打平了，还是有人真正赢了，肖恩？

So is it equalized out or is there a winner here, Sean?

我觉得这周最值得关注的一件事，是Meta又上调了本年度的资本支出预算，而他们原本的预期就已经很高了。

You know what I think is one of the most interesting things to come out of this this week is Meta announced another bump to already really high CapEx expenditures, like expectations for this year.

具体数字我一时想不起来，但年初的时候他们预计大概是1200亿美元左右。

Something like I forget the exact figure off the top of my head, but it's like previously coming into the year, they had promoted something around say, a 120,000,000,000.

现在他们预计要达到一千四百亿美元左右。

Now they're expecting like a 140,000,000,000.

如果不是精确这个数字，也差不多在这个范围内。

If it's not those exact numbers, it's like in that range.

当初他们公布第一个数字时，市场普遍认为这是个强烈的乐观信号。

And when they really led this year with that first number, that was, like, seen as a big sign of optimism.

看看他们打算建多少数据中心、扩充多少算力、做多少推理运算，这些都得砸钱。

Look at all this build out they're gonna do with with data centers and and capacity and and how much inference they're gonna be doing, like, all this stuff.

就在三个月前，这个数字还被看作是巨大的信心体现，以至于上周我们分析特斯拉财报时，感觉特斯拉上调资本支出预期，简直就像个小数目——毕竟其他科技公司都计划今年花上一千亿甚至更多，这都是从上季度财报里透出来的趋势。

Like, it was such a sign of where we were even just, like, three months ago to the point that like when you and I were covering Tesla earnings last week, like, it very much felt like the reason Tesla bumped up its CapEx expectations for this year was because like, it looked like a a paltry little number compared to the all all the other tech companies, which were all planning to spend like a 100 or more billion dollars this year, coming out of the the sort of fourth quarter results.

而这一周，Meta又把预算推高到一千四百亿美元左右，结果股价大跌。

And now this week, Meta, you know, bumped that up again to like around 140,000,000,000 and their stock price went way down.

昨晚到今天开盘，他们的股价遭遇了多年未见的大跳水。

Like they had a big hit last night and into today's trading, one of the biggest drops they've seen in a long time.

我觉得这其实是个很明显的信号，说明有些投资者已经开始踩刹车了：他们可能意识到，对AI、数据中心和相关基建的投资，也不是无底洞，投资人心里其实有条线。

And I feel like that's just a really good signal of like how some people have been tapping the brakes on some of this stuff where like, oh, there actually might be a limit on the investor side of how deep they want these companies to go into AI and AI build out and data centers and all of this stuff.

但就在三四个月前，根本感觉不到有什么上限。

Whereas just three or four months ago, didn't feel like there was a limit.

那时候大家都觉得数字只会往上走，而且永远停不下来。

It felt like number was gonna go up and it was never gonna stop going up.

所以这对我来说是最让人印象深刻的转变。

So that was the thing that stuck out the most to me.

是的。

Yeah.

关键是这么回事。

I mean, here's the thing.

公司特别喜欢那个黄金点，只要被当成增长型公司，股东就会给奖励。

Like, companies love that sweet spot, and they're rewarded by shareholders when they are not they are a growth company.

当然，也有已经成熟的股票。

And, you know, there's the mature stocks.

你懂的。

You know?

很长一段时间里，微软一直被视为成熟型股票，但现在他们努力把自己重塑为成长型公司，毕竟他们正在投资前沿技术。

For for a long time, Microsoft was considered a mature stock, but they've tried to pivot themselves as kind of a growth stock now because here they are investing in cutting edge technology.

看看我们的支出，其中一个关键指标就是资本开支。

Look at what we're spending, and one of the indicators is capital spending.

投资者通常很认可这种做法，也会用股价来奖励这样的公司。

And oftentimes, investors are happy with that, and they reward companies for that.

但我觉得，我们所有人——说实话，我根本不投资这些公司。

But I think all of us I mean, I don't invest in any of these companies.

我们根本不投资这些公司。

We don't invest in any of these companies.

但这些投资者现在都在说，就像你之前提到的，开始踩刹车了。

But any of these investors are now saying to your point earlier, like, plumbing the brakes.

意思是，明白了。

Like, okay.

我们懂了。

We get it.

那你现在把钱投到哪儿了？

Like, now you're what are you putting your money into?

现在感觉，或者看起来，这些额外的支出似乎没什么意义。

Now it feels, or looks as if there isn't a point to this extra spending.

我觉得这正是很多散户和机构投资者担心的地方，因为他们已经花得够多了。

And I think that that's what a lot of retail and institutional investors are worried about because they've already they're already spending a ton.

我的意思是，谷歌的资本支出大概有两千亿美元，或者可能是亚马逊，差不多这个数。

I mean, believe Google is something around 200 or maybe it's Amazon, 200,000,000,000 in capital spending.

那到底什么时候该停下来呢？

I mean, at what point, do we stop doing that?

而且，这些投入到底带来了什么成果？

And, also, what is the result of it?

我觉得另一个问题是，虽然他们承诺要大笔投入，但还没看到明显的回报。

And I think that that's the other thing is that there is all these commitments to spending a lot, but they haven't quite seen the upside of it yet.

我前面提到过一个正面效果，就是云业务表现相当不错。

I mean, the one upside I mentioned at the beginning of this segment, which is the cloud businesses are doing quite well.

谷歌的云业务和AWS表现得非常出色。

Google's cloud business and AWS are doing incredible.

安迪和贾斯西在亚马逊上季度财报电话会上提到，云业务的增长速度是过去十五个季度以来最快的。

Andy, Jassy, yesterday during the earnings call for Amazon, noted that the growth rate for its cloud business is the fastest growth rate it's had in 15 quarters.

这确实值得关注，但值得投入这么多资本吗？

That's notable, but does it warrant this amount of capital spending?

这种增长能持续下去吗？

And is it sustainable?

未来几个季度才会见分晓，但我觉得，这确实是个大问号。

And, you know, we won't know until future quarters, but, you know, I think that that's the big question mark.

对。

Yeah.

就像你说的，如今一些最大项目的前景变得特别模糊，这在一年前可没有过。

I mean, like, you know, to your point, there's just been so much haze around some of the largest projects in a way that that, you know, that didn't exist a year ago.

一年前，我们刚听说‘天际’项目，那是个五千亿的计划，OpenAI、甲骨文和软银都参与其中，还得到了特朗普政府的支持。

You know, a year ago, we were coming off of the announcement of Stargate, this $500,000,000,000 project that, like, OpenAI and Oracle and SoftBank were all gonna do, and it was backed by the Trump administration.