本集简介
双语字幕
仅展示文本字幕，不包含中文音频；想边听边看，请使用 Bayt 播客 App。
查莫普，你跟谁在偷笑呢？
Chamop, who are you giggling with?
你跟孩子在一起吗？
Are you with your kids?
这里发生什么事了？
What's going on here?
你摆出一副谷歌脸。
You're you're making Google face.
我收到了一条来自No的消息。
I got a I got a message from No.
她就像
She's like
是海姆斯吗？
Is it Hellmuth?
不是。
No.
不是。
No.
我收到了马特的消息。
I got a message from Matt.
她太搞笑了。
She's so funny.
所以，你知道的，我当时就说了一堆废话。
So, you know, we're I was like, blah blah blah.
她发了个关于‘天王星’的笑话，结果发到了聊天框里。
She did a Uranus joke into Enter.
我笑疯了。
By chat, and I lost it.
就像有人说的，天王星就在子宫边缘。
It is like a says, the Uranus is right there at the edge of the uterus.
不是。
No.
她问我：‘你饿了吗？’
She's like, are you hungry?
我都能吃下一个天王星。
I could eat a Uranus.
哦，天哪。
Oh, no.
我们可以去一家很棒的餐厅，天王星。
We could go to a great restaurant, Uranus.
那里提供的餐点非常棒，还有美味的巧克力熔岩蛋糕。
They had served a great meals there, and they have a great chocolate lava cake.
好了，各位。
Alright, everybody.
欢迎回到你们最爱的播客。
Welcome back to your favorite podcast of all time.
这是《全员参与》第169期。
It's episode one sixty nine of All In.
今天和我一起的有主席独裁者贾马夫·帕利哈皮蒂亚、大卫·弗赖伯格、我们的科学之王，还有雨人。
With me again, the chairman dictator, Jamaf Palihapitiya, David Freiberg, your sultan of science, and the rain man.
是的。
Yeah.
当然，大卫·萨克斯又戴上了他的Moncler帽子。
Definitely, David Saxe has his Moncler hat back again.
我想他们又补货了。
I guess they're back in stock.
兄弟们，现在怎么样？
How we doing, boys?
欢迎回到节目。
Welcome back to the show.
我们来谈谈今天的议程吧。
Let's get to the docket here.
第一个议题：埃隆已经起诉了OpenAI，事情已经开始了。
Issue one, Elon has sued OpenAI, begun.
meme战争在我们上周录完音后变得疯狂了。
The meme wars have After we've finished the recording last week the memes are incredible.
我们上周录完音后，埃隆起诉了萨姆·阿尔特曼、格雷格·布罗克曼和OpenAI组织。
After we finished recording last week, Elon sued Sam Waltman, Greg Brockman, and the OpenAI organization.
他指控他们违反合同、违背受托责任以及不正当竞争。
He's suing for breach of contract, fiduciary duty, and unfair competition.
他的论点基本上是：OpenAI最初是一个开源非营利组织。
And his argument is basically OpenAI started as an open source nonprofit.
如你所知，他给了他们大约五千万或七千五百万美元——这两个数字都被提及过。
As you know, he he gave them something like 50 or 75,000,000 of both those numbers quoted.
然后，当然，人人都知道OpenAI变成了ClosedAI。
And then, of course, everybody knows OpenAI turned into ClosedAI.
在技术部署后，他们转为闭源盈利型企业。
They became a closed source for profit venture after the tech was deployed.
这使他们能够在开发技术期间享受非营利组织的税收优惠，但如今他们已拥有两家公司实体。
This enabled them obviously to benefit from tax nonprofit tax breaks while building the tech, but they now have two corporate entities.
有一个名为OpenAI的非营利组织。
There's this nonprofit called OpenAI.
还有一个营利性实体。
There's a for profit.
它被称为OpenAI Global LLC，成立于2019年。
It's called OpenAI Global LLC that was created in 2019.
这两者之间有着复杂的关系。
And there's all this funky relationship between the two.
我们会深入探讨这一点，因为实际上挺有意思的。
We'll get into that because it's kind of interesting, actually.
埃隆表示，如果OpenAI可以这样做，那么这应该成为未来所有公司的标准。
Elon said that if OpenAI is allowed to do this, then it should be the standard for every company going forward.
这是一个有趣的观点。
That's an interesting point.
你可以先向非营利组织捐款，然后将其转变为营利性机构。
You can start by donating money to a nonprofit and then make it for a profit.
这看起来相当资本高效。
It seems pretty capital efficient.
那么根据诉讼，埃隆想从这件事中得到什么？
So what does Elon wanna get out of this according to the lawsuit?
他希望OpenAI开源他们的模型。
He wants OpenAI to open source their models.
顺便说一下，Facebook和Apple现在就是这样做的。
By the way, that's what Facebook and Apple are doing right now.
因此，既然公司当初就是这么成立的，这个要求似乎再合理不过了。
So that that seems more than reasonable since that's how the company was formed.
他还希望确保股东从OpenAI中得不到任何经济利益。
And he wants to make sure that shareholders receive no financial benefit from OpenAI.
我们可以进一步探讨OpenAI的非营利地位及其结构。
And we can get more into a bunch of the OpenAI nonprofit status and and their structure.
这实在是太复杂了。
It's super convoluted.
但我只想听听你的初步反应，萨克斯。
But I wanna just get your initial reaction, Sachs.
我没有看过这些法律文件，杰森。
I haven't read the legal filings, Jason.
我对这个案件的了解只来自于关注推特上的动态。
My knowledge of the case is just coming from following what's on Twitter.
上次我在这个播客里评论埃隆的一起诉讼时，结果被传唤去做了六小时的证词。
And the last time I commented on one of Elon's lawsuits on this pod, I ended up in six hours of deposition.
记得那次是推特诉讼案，对吧。
Remember that was the Yeah.
整个推特诉讼案。
The whole Twitter lawsuit.
无论如何，根据推特上来回转发的帖子，我理解埃隆主要提出了两点。
In any event, my understanding of the case from the tweets that are flying back and forth is I think Elon is making two points.
第一点是，OpenAI成立时，本意是作为一个非营利组织，推动人工智能成为一项开源技术，让每个人都能受益，而不是由任何一家大型科技公司控制。
One is that when OpenAI was set up, it was set up to be a nonprofit and to promote AI as an open source technology that everyone could benefit from and no one large tech company would control.
在这种情况下，埃隆主要担心的是谷歌，我想现在他更担心微软。
In that case, Elon was primarily concerned about Google, I think now he's more concerned about Microsoft.
然而，当初的设想是这将是一个开源项目。
Nonetheless, the idea was that this would be open source.
所以我认为第一点是，埃隆觉得在他捐赠了四千多万美元后，自己的信任被背叛了。
So I think point number one is Elon feels like the rug was pulled from under him after he donated 40 something million dollars to this.
他们完全改变了它的性质。
They completely changed what it was gonna be.
我想他以前用过‘欺诈’这个词。
And I think he's used the word swindled before.
他觉得自己被欺骗了。
He feels like he was cheated.
我认为，通常当出现这样的诉讼时，都是因为有人确实感到自己被根本性地欺骗了。
And I think usually when there's a lawsuit like this, it's usually because somebody does feel fundamentally cheated by what happened.
所以我认为这就是第一点。
So I think that's the first point.
第二点是，我认为埃隆的意思是：等等，如果你能用税前资金以非营利组织的形式创办一家公司，然后突然转为营利性公司，那为什么每个人不都这么做来避税呢？
The second is that I think that Elon's making is, well, wait a second, if you can start a company as a nonprofit using pretax dollars and then all of a sudden convert to for profit, then why wouldn't everybody do this in order to circumvent paying taxes?
我认为这是埃隆提出的另一个合理且有趣的观点。
And I think that's another, I'd say, valid and interesting point that that he's made.
现在，OpenAI 已通过发布埃隆发给他们的部分邮件作出回应，他们认为自己找到了确凿证据，因为据称埃隆曾告诉他们，必须筹集更多资金，才有可能与谷歌/深思抗衡。
Now, OpenAI has responded to this by publishing some of Elon's emails to them and they think that they have a smoking gun here because apparently, Elon told them they need to raise far more money in order to have a chance of taking on Google slash DeepMind.
我只是不确定这是否真的是他们声称的‘铁证’，但这就是他们目前所公布的内容。
I just don't know whether that's the smoking gun they claim it is, but that's that's basically what they're putting out there.
此外，风险投资人们也纷纷加入争论，就此事的不同方面展开辩论。
And then to add fuel to the fire, you've got VCs coming over the top debating different aspects of this.
有像维诺德这样的人在为 OpenAI 辩护。
You've got sort of Vinod defending OpenAI.
他是该公司的首位风险投资方。
He was the first VC investor in it.
因此他一直在公开批评埃隆，而我认为马克·安德森正在回应维诺德，但我不确定他是否真正是在为埃隆辩护。
And so he's been out there attacking Elon and then I think Marc Andreessen is responding to Vinod and may I don't know if he's exactly defending Elon.
总之，这件事在X平台上演变成了一场风暴，其他人也开始发布一堆疯狂的梗图。
Anyway, it's turned into a whole maelstrom on X and then everyone else is start starting to put out a whole bunch of crazy memes.
我不知道这场官司最终谁会赢，但这些梗图确实火了。
So I don't know who's gonna win this case in court, but the memes are definitely lit.
最搞笑的一个是，埃隆说，如果OpenAI只是把名字改成Closed.ai，他就撤诉。
And maybe the funniest one is that Elon has said that if OpenAI will simply change their name to Closed dot ai, he'll drop the lawsuit.
我觉得他是认真的。
I think he's serious about that.
我觉得他是认真的，但我觉得他想表达的是：你最终做了和你当初让我捐大笔款项、和我一起创办这个项目时所说的完全不同的事情。
I think he is serious about it, but I think he's making his point, which is you ended up doing something different than what you told me when you got me to write this big donation and when we co founded this thing together.
是的。
So Yeah.
我觉得这基本上概括了埃隆对此事的立场。
I think that sort of sums up Elon's position on this.
查马斯，一年后，你认为我们会处在什么状况？
Chamath, a year from now, where do you think we'll be with this?
我认为还有一件事，就是他们曾大力使用他的名字来筹集更多资金，而他在招募一些早期关键人才，尤其是伊利亚方面起到了关键作用，这一点自他离开谷歌以来已有大量报道。
Well, I think I think there's one more thing, which is that they used his name pretty aggressively to get more money, and he was very instrumental in getting some of the early critical hires, particularly Ilya, which has been, I think, well documented since Geva to leave Google.
对吧？
Right?
所以，这可能也是萨克斯感到被欺骗的部分原因。
So that's probably Saxe where also some of this feeling that he was bamboozled comes from.
但我认为，所有这些情绪都比不上法治，也就是他的第二个观点，这才是关键点。
But I think that all of those emotions matter less than the rule of law, which is his second point, which is the important point.
无论某封邮件怎么说，另一封邮件又怎么说，如果突然出现一个巨大的漏洞并被利用，这对美国的税收体系来说都不是好事。
Irrespective of whatever one email says or another email, it's not great for The US tax system if all of a sudden a big gaping loophole is identified and taken advantage of.
我认为，对于加州以及所有这些OpenAI员工曾经居住、获得股权并现已获得报酬的其他州来说，都存在巨大的经济动机。
I think that there's a huge economic incentive for the state of California, every other state where these OpenAI employees lived and have gotten equity and now have gotten paid.
财政部也有动机。
There's an incentive for treasury.
国税局也有动机。
There's an incentive for the IRS.
这涉及到许多复杂的税法方面。
It touches a lot of aspects of very complicated tax law.
所以我认为这个问题最终会得到解决，因为这个答案非常重要。
And so I think that's why there will be a resolution to this case because I think that that answer is very important.
而且我认为，正如他正确指出的那样，这将推动许多人未来的行为。
And I think it will, as he has correctly identified, motivate a lot of people's actions going forward.
因此，不管这场戏中所有参与者的情感状况如何，现实是他发现了一个漏洞，而这个漏洞必须被修补。
So independent of what the emotional situation is amongst all the actors in this play, the reality is he's identified a loophole, and that loophole needs to get fixed.
我给你一个更温和的类似例子，但这个案例持续了八九年之久。
I'll give you a different example of this that's much more benign, but it dragged on for eight or nine years.
在Facebook，曾经有一段时间——这是公开的信息，所以我可以告诉你——
At Facebook, there was a period where and this was public, so I can tell you.
我们曾一度将大量知识产权转移到爱尔兰，当时有一笔转让付款之类的操作。
We transferred a lot of our IP to Ireland at one moment, and there was a transfer payment and whatnot.
但几年后，当所有人都意识到——包括我们自己——Facebook的价值时，才发现这笔转让付款并不合理。
And then a few years afterwards when when everybody realized, including us, and we had miscalculated, but when everybody realized the value of Facebook, that transfer payment did not seem correct.
而且这是我们促成的一个巨大的税务套利行为。
And it was a huge tax arb that we had facilitated.
对吧？
Right?
因为所有这些位于爱尔兰的知识产权所适用的税率，与这些知识产权在美国适用的税率完全不同，其他公司也效仿了这种做法。
Because all of that IP sitting inside of Ireland gets taxed at, you know, a very different rate than that IP would have gotten taxed in The United States, and other companies had copied this.
简而言之，国税局提起了诉讼。
Long story short, the IRS sued.
就像你，大卫一样，我也经历了多年的取证和面谈等各种事情。
You know, similar to you, David, I was in years of depositions and interviews and all of this stuff.
所以重点是，政府非常重视这类问题，因为涉及的资金数额巨大。
So the point is that the government really cares about these kinds of things because so much money is on the line.
如果OpenAI最终成长为一家价值数千亿美元的巨头，这件事一定会在法庭上被厘清，因为牵涉的利益实在太大了。
And if OpenAI turns out to be this multi $100,000,000,000 behemoth, this will get figured out in court because there's just too much money at stake.
弗赖贝格，你有什么看法？如果可以的话，斯蒂尔曼，你对OpenAI的立场有什么看法？
Freiberg, any thoughts and maybe Steelman, if you if you feel like it, OpenAI's position?
我
I
我认为我们所有人都忽略了OpenAI昨天发布的文件，这些文件展示了OpenAI与马斯克之间的邮件和互动，马斯克承认并认可设立一个营利性子公司以通过筹集数十亿美元外部资本来实现基金会目标的必要性。
think we're all ignoring the documents that OpenAI put out yesterday showing emails and interactions with Elon where Elon acknowledged and recognized the necessity of having a for profit subsidiary that could pursue the interests of the foundation by raising billions of dollars of outside capital.
我认为这是一组非常有趣的事实，为这个故事提供了不同的视角，而且OpenAI发布这些文件非常重要。
I think it's a really interesting set of facts that provides a different light on the story and it was really important that OpenAI released it.
我打算分享一下
I'm gonna share a
等等，弗里伯格。
Well, hold on, Fieber.
我提到过这一点。
I did mention that.
弗里伯格，我也提到过。
Friedberg, I mentioned that too.
我的观点是，这并不能让你违法。
My point is it doesn't allow you to break the law.
是的，让我跟你们讲讲另一个例子。
Yeah, so let me just tell you guys a little bit about another example.
我妹妹在一家名为囊性纤维化基金会的非营利组织工作。
My sister works at a nonprofit called the Cystic Fibrosis Foundation.
囊性纤维化是一种遗传性疾病，影响肺部和消化道。
Cystic fibrosis is an inherited disorder, affects the lungs and digestive tract.
这种病令人衰弱，对儿童影响尤其严重。
It's debilitating, really affects children.
1955年成立了一家名为囊性纤维化基金会的非营利组织，多年来他们一直在药物研发方面做了大量工作，直到他们意识到，必须建立一个市场机制，才能提供必要的激励，吸引所需资本来研发治疗囊性纤维化的药物。
There was a nonprofit called the Cystic Fibrosis Foundation established in 1955, they did tons of work on drug development research and it was a nonprofit for years until they realized that there needed to be a market based system to create the necessary incentives to drive the capital needed to find a drug that could cure cystic fibrosis.
2000年，他们从基金会的捐赠基金中投资了4000万美元给一家名为Aurora Biosciences的公司。
And in the year 2000 they invested out of their foundation endowment $40,000,000 in a company called Aurora Biosciences.
后来这家公司被Vertex收购，他们又追加投资了6000万美元，总计1亿美元投入到这种疾病治疗药物的研发过程中。
Subsequently it was acquired by Vertex and they continued to invest another $60,000,000 so a total of $100,000,000 in the lifetime of the development of a drug that could cure this disease.
2005年，一种药物被发现，并于2012年获得美国食品药品监督管理局批准。
In 2005, a drug was discovered and in 2012 the FDA approved it.
该非营利组织随后将其对该营利性实体的权利和利益以33亿美元的价格出售。
The nonprofit then sold their rights, their interest in this for profit entity for $3,300,000,000.
这是一笔惊人的回报。
It was an incredible return.
哇哦。
Was Wow.
这是历史上非营利组织获得的最大投资回报。
The largest investment return by a nonprofit in history.
他们卖给了谁？
Who do they sell it to?
卖回给公司了吗？
Back to the company?
实际上，我认为他们把权利卖给了Royalty Pharma，这个公司你我都熟悉。
I actually think they sold the rights to Royalty Pharma, who you and I know well.
这33亿美元继续为他们开辟了持续投资的能力。
And that $3,300,000,000 continued opened up this ability for them to continue to make investments.
他们现在资助生物技术领域的风险投资公司，并进行直接投资以及其他活动。
They now fund biotech VCs and they make direct investments and other things.
但这真正确立了所谓‘风险慈善’的标杆——非营利性母公司可以投资于营利性实体，从而筹集所需资金，以实现非营利组织广泛而艰巨的使命。
But it really set the benchmark for this concept of what's called venture philanthropy where a nonprofit parent company can make investments in for profits that can raise additional capital that's needed to pursue the broad difficult interests of the nonprofit.
我认为这个观点实际上与OpenAI发生的事情紧密相关，你可以在埃隆的邮件往来中看到这一点，他极具远见，我必须给予埃隆极大的赞誉，因为他早就意识到，要实现OpenAI的目标，每年需要数十亿美元的投资。
And I think that this argument really kind of ties into what happened with OpenAI and you can see it in the email exchanges with Elon where he was so prescient that I I give Elon extraordinary credit for this, that he saw that this is gonna take billions of dollars a year of investment to realize the pursuit that OpenAI was going after.
而埃隆自己不可能独自筹集到这么多资金，雷德或其他人也不可能单独拿出这笔钱。
And there was no way that Elon was gonna be able to generate that cash himself or that Reed or others were gonna just be able to pony up that money.
他们需要一个营利性实体，让市场发挥作用，让资本家能够将资金投入这个组织，支持这项投资计划。
They needed to have some sort of for profit vehicle that would allow the market to work and allow capitalists to find their capital into this organization to make this this investment interest.
因此，关于OpenAI作为基金会是否陷入困境的真正问题在于：非营利组织是否拥有营利性实体的实质性股权？
So I think the real question with respect to is OpenAI in trouble as a foundation is does the nonprofit own a meaningful piece of the for profit entity?
我不知道答案。
And I don't know the answer to that.
我不知道你们是否知道。
I don't know if you guys do.
听起来是这样。
Sounds like it.
我的意思是，确实应该是这样，对吧？
I mean, yeah, it must, right?
所以如果它们拥有股份，那就意味着有投资利益。
So if they do, then they have an investment interest.
第二个问题是，非营利性母公司是否仍在从事慈善活动？
The second question is does the nonprofit parent still do charitable stuff?
因为如果它不再从事，那么非营利地位就真的成问题了。
Because if it doesn't, then there's a real question on the nonprofit status.
我认为，为了符合501(c)(3)资格，你每年必须将一定比例的资产用于慈善用途。
I think you have to have a certain amount of your assets deployed every year in order to qualify for five zero one c three.
所以我认为，法院的判断很可能会是：设立营利性实体、投资营利性实体完全合理，其他非营利组织也这么做过，特别是当你需要吸引数十亿美元的外部资本才能实现目标时。
So I think the test in the courts will likely end up being, look, it's totally reasonable to have a for profit entity, to fund a for profit entity, other nonprofits have done it, particularly when you need to attract billions of dollars of outside capital to make it work.
真正的问题是，非营利组织是否仍在做非营利的事情？
The real question is does the nonprofit still do nonprofit stuff?
这是个有趣的问题，Freebird，你觉得他们把非营利组织专家从董事会中踢出去时，是不是已经触及了某个界限？
That's an interesting question because, Freebird, do you think that when they booted off the the nonprofit experts from the board that they Right.
他们可能在那里踩了雷区？
They may have crossed a tripwire there?
是的。
Yes.
对。
Yeah.
我是律师。
I'm the lawyer.
我是律师。
I'm the lawyer.
这是个好问题。
That's a great question.
我做过一些新闻调查，和一些参与或了解另外两个类似情况的人聊过。
I did a little journalism and I talked to a couple of people who are in and around two other of these scenarios.
这里的一个关键问题是，在你的例子中，弗里德伯格，这个囊性纤维化组织并没有进行药物发现。
One of the key issues here is that the IP, in your example, Friedberg, it wasn't like that cystic fibrosis organization did the drug discovery.
对吧？
Right?
他们并没有做这件事。
They didn't do it.
对吗？
Correct?
嗯，他们资助了，我认为他们当时移交了一些公开的、公开的研究成果，然后他们获得了它。
Well, they they funded I think they they handed over some some research that was open and public at the time and they they Got it.
但他们资助了一家独立的营利性机构。
But they funded an independent for profit.
所以这里发生的情况是，知识产权和非营利组织的员工被给予了、赠予了。
And so what happened here was you had the IP and the employees of the nonprofit given, gifted.
他们被转移给了这家营利性组织。
They were transferred over to the the for profit organization.
是的。
Right.
当Mozilla基金会和另一家公司Samasource这样做时，它们都采用了非营利组织和营利性公司两种结构。
And they when the Mozilla Foundation did this and another company called Samasource, and these both have two structures, a nonprofit and a for profit.
但在它们的案例中，都设立了独立的董事会。
But in both of their cases, they set up separate boards.
这两个不同的董事会分别对营利性事务和非营利性事务负责。
And the two different boards, you know, answer to the for profit concerns and the nonprofit concerns.
它们也没有将大量股权赠予所有员工，然后让员工在二级市场出售。
They also didn't give huge chunks of equity to all of the employees, which then sold them in secondary.
因此，这里发生的情况是，所有知识产权很可能都已转移，财富也流向了这些员工。
So what's happened here is all of the IP, in all likelihood, has been and the wealth transfer has gone to all these employees.
他们现在已经完成了二级市场交易。
They've now done the secondary.
这实际上引发了大量美国国税局的调查。
And this actually triggers a lot of IRS investigations.
Mozilla基金会曾开发Netscape浏览器，你们很多人可能都用过，他们每年从与谷歌的广告合作中赚取数亿美元，大卫。
The Mozilla Foundation, which was making the Netscape browser, which many of you probably have used over the years, they were making hundreds of millions of dollars a year, David, from advertising with Google.
国税局说，嘿。
The IRS said, hey.
等等。
Wait a second.
在我们看来，你们看起来就像一家营利性公司。
You look like a for profit to us.
于是他们成立了一个营利性实体。
So they made a for profit entity.
米切尔·贝克和她的团队设立这套机制，让所有资金都回流到Mozilla基金会及其非营利事业中，他们只是支付高于市场水平的薪资。
All of the and Mitchell Baker and the team over there set this up so all the money went back into the Mozilla Foundation and their nonprofit efforts, and they just paid above market salaries.
但至关重要的是，这些实体必须分开，员工不能持有股权，所有资金都必须向上回流。
But it was very important that these things be separated and that the employees not have equity and that all that money flowed back up.
我认为，这就是OpenAI即将陷入困境的地方。
And I think this is where OpenAI is gonna get super tripped up.
我认为国税局在这里会大有可为。
And I think the IRS is gonna have a field day here.
尼克，你能把图片放出来吗？
Here's the Nick, can you please throw the image up?
这是来自OpenAI官网对其结构的说明。
This is from the OpenAI website that explains their structure.
我们来稍微推测一下，因为我们只是在猜测。
So let's just speculate for a second because we're we're just guessing.
但有报道称微软拥有49%的股份。
But it has been reported that Microsoft owns 49%.
对吧？
Right?
这个数字对吗？
That's the number?
指的是这家营利性实体的股份。
Of the for profit.
在这个例子中，微软在‘少数经济利益’这一栏对应的数字是49，这代表他们拥有OpenAI Global LLC的这部分股权。
In this example, Microsoft, where it says minority economic interest, that number there would be 49, and that's what they own of this OpenAI Global LLC.
而多数股权由这家控股公司持有。
And the majority owner is this holding company.
好的。
Okay.
这意味着51%的股权在这里。
So that means 51% is there.
但这里写的是，这51%由非营利组织、员工和投资者共同持有。
But it says here that that 51% is owned by a combination of the nonprofit, employees, and investors.
所以问题来了，回到萨克斯的问题，我们可以通过剔除投资者和员工的份额，精确地算出非营利组织持有多少。
So the question is, back to Sax's question, we can know very precisely how much the the nonprofit owns by just x ing out the investors and employees.
我猜至少有30%到40%。
And I'm gonna guess it's at least 30 or 40%.
我觉得这个推测应该相当合理。
I think that that's probably a pretty reasonable guess.
这意味着OpenAI基金会可能拥有5%到20%之间的份额。
So it means that the OpenAI Foundation probably owns somewhere between five to 20%.
这个最大估值的推测合理吗？
Is that fairly reasonable guess at the maximum?
这看起来是对的。
That seems right.
是的。
Yeah.
所以我认为他们必须展示出这种所有权结构，拆解所有这些实体，并明确实际的持股比例。
And so I think that they're gonna have to show that ownership structure and and decompose all these entities and show the the actual percentages.
如果实际比例远低于此，比如只有1%左右，那就会显得有点可疑，我觉得。
If it's a lot less than that, if it's, say, like a 1% thing, then it's a little bit more sketch, I think.
但如果真是15%或20%，他们仍然必须证明这一切都是合规完成的，因为这些人所做的关键事情是，通过创建这种LPGP结构和有限利润机制，很明显他们是在试图规避某些东西。
But if it's really, like, 15 or 20%, they're still gonna have to prove that all of this was done in a clean way because the big thing that these guys did was because by creating this LPGP structure and these capped profits, it's clear that they were trying to avoid something.
问题是，在调查过程中，大家会不会了解到他们究竟想规避什么？
And the question is in the discovery, will everybody learn what it is that they were trying to avoid?
因为我经历过这些。
Because I've been through this.
我认为，萨克斯，你也经历过。
I think, Sachs, you've been through it.
弗里德伯格，你也经历过。
Friedberg, you were through it.
当你设立这些大型基金并汇集资金时，各种复杂的结构就会冒出来，而纪律始终是告诉律师和会计师‘不行’。
When you set up these large vehicles and pooled amounts of money, folks come out of the woodwork with all kinds of convoluted structures, and the discipline is always to tell these lawyers and accountants no.
他们会说：我们设立一个主基金- feeder 结构，通过百慕大，再走其他路径。
And they'll be like, let's set up a master feeder, you'll go through Bermuda and this and that.
说‘好’很容易，因为他们推销的东西非常诱人，但本质上总是为了规避某些东西。
And it's so easy to say yes because it's very seductive what they're selling you, but it's always about trying to avoid something.
所以真正的问题是，为什么要有这么复杂的结构？
So the real question is why this convoluted structure?
他们到底想规避什么？
What was it trying to avoid?
这个知识产权是从哪里来的？
And where did this IP come from?
那些知识产权都是来自一家非营利组织。
Like, all that IP sacks came from a nonprofit.
所有那些员工都曾为这家非营利组织工作。
All those employees worked for the nonprofit.
然后他们决定从开源模式转向闭源。
Then they decided to go from an open source model to closed.
通过这样做，他们攫取了所有价值，并将这些价值赠予了这些员工和投资者群体。
And then by doing that, they capture all the value, and then it's gifted to these employees and this investor base.
我认为这一点是
That's where I think
但请记住，很重要的一点是：如今OpenAI所拥有的大部分知识产权，都是在营利性实体成立之后产生的。
But remember, something important Well, it's important to remember the majority of the IP that exists at OpenAI today was generated after the for profit was set up.
真正的问题是，在成立这家营利性实体时，是否公平地将价值转移给了它？
The real question was there a fair transfer at the time when it was set up of value into this for profit entity?
我猜你刚才看的那个GP结构，非营利组织通过它获得的报酬，大概是为了公平地反映当时所作的贡献。
And my guess is that GP box that you just looked at, the compensation that they're earning through that GP box that the nonprofit earns is likely meant to be kind of a fair amount relative to what was contributed at the time.
对我来说，真正的问题仍然是：OpenAI的501(c)(3)非营利组织究竟还从事哪些其他慈善或公益工作？
The real question for me still remains, what other nonprofit and charitable work does the OpenAI five zero one c three actually do?
如果答案是什么都没有，它只是一个空壳非营利组织，而它名下唯一拥有的就是一家营利性实体的权益，那么我就要质疑：它的实际活动到底是什么？
And if there's the answer is nothing and it's just a shell nonprofit and under it, the only thing that it owns is a for profit interest, then I think there's a real question on what's the activity.
我并不是想对这些事情妄加评论。
And I I don't mean to speak out of turn on all this stuff.
我对这些了解得不够多，但我觉得，这可能是我看待这些问题的唯一角度。
I don't know enough, but I think that's the only sort of lens I would kinda look at this stuff through.
萨克斯，你刚才想说什么吗？
Sachs, you were gonna say something?
嗯，有几点要说。
Well, a couple of points.
首先，我一直主张：绝不要在结构上做创新。
One is, one of my contentions has been never innovate on structure.
你所做的只是制造法律问题。
All you do is create legal problems.
创建初创公司的传统且可靠的方式是成立C型公司。
There's a tried and true way of creating a startup, which is a a c corp.
当你试图在法律结构上创新，而不是在产品上创新时，几乎总是会适得其反。
When you try to innovate on legal instead of product, it almost always backfires.
第二个观点是，我认为OpenAI正处在一个进退两难的境地。
The second point here is I think OpenAI is in a little bit of a damned if you do, damned if you don't situation.
回到Freebird提到的点，关于非营利组织位于顶层的问题，回想一下当萨姆短暂被解雇又复职时，问题是如何产生的：当时非营利组织派来的董事们似乎完全不了解初创公司是如何运作的，因此他们要么毫无理由地解雇了萨姆，这显得非常无能；要么他们确实有充分理由，却没有妥善沟通，这同样显得无能。
To go back to Freebird's point about the nonprofit sitting at the top there, remember what created all the problems when Sam briefly got fired as CEO and then came back is they had these directors from the nonprofit world who didn't really seem to understand how startups worked and so either they fired Sam for no reason, which was kind of incompetent, or they had a really good reason but didn't communicate it properly, which was also incompetent.
无论如何，整个事件都是一场闹剧，暴露了营利性公司与硅谷标准的、以最大化成果为目标的创业模式，与 atop 的非营利董事会之间存在的文化冲突。
Either way, that whole thing was a it was a spectacle and it just kind of showed that there was this like culture clash going on between the for profit company and like the standard way of running a Silicon Valley startup to maximize the outcome and then this nonprofit board that was sitting on top of the whole thing.
所以我认为，现在他们已经更换了非营利董事会，清除了非营利背景的人士，这可能对营利实体来说是正确的决定，但如今这反而可能让他们陷入麻烦，因为这会为埃隆的诉讼提供依据——即他们彻底改变了这个组织的原始使命。
So I think that now they're in the situation where they've changed that nonprofit board, they've booted off the nonprofit people, and that may have been the right thing for the for profit entity, but now it might get them in trouble because it lends credence to Elon's lawsuit that they've completely changed the original mission of of this organization.
对吧？
Right?
本应如此
Was supposed to
是使命
be mission
它本应是开源的，是的。
it was supposed to be open source Yes.
它本应是非营利的，而现在却变成了营利性且闭源的。
And it was supposed to be nonprofit and now it's for profit and closed source.
想想看，你知道，我们这里只是在推测。
Think about this, you know, and again, we are speculating here.
但如果初衷是开源，而你后来发现你拥有了极其宝贵的东西，于是转为闭源，成立营利性公司，把所有员工和所有知识产权都转移到营利性实体中，彻底封闭，背离了最初的使命——
But if it if the mission was to be open source and then you realize you've got something super valuable and you close source it, create a for profit, and then take all the employees and all the IP and put it into the for profit, lock it down, break the original mission that, hey.
这原本是为了帮助人类，因为我们打算让它开源。
This was gonna help humanity because we're gonna make it open source.
所以如果你在这里捐款，我们获得免税资格，人类就能受益，因为每个人都能同时查看代码、参与开发和迭代。
So if you donate money here and we get the tax exemption, humanity benefits because everybody gets to look at that code at the same time and work on it and iterate.
现在他们基本上窃取了这一切，只为让自己致富。
Now they they've just basically stolen all that in order to enrich themselves.
我认为，再加上萨姆可能进行了一些交易，根据报道，这似乎是其中一个触发因素，比如OpenAI风险基金等等。
And I think then you add to it, Sam, maybe doing some deal making, which seems to be one of the triggers according to reports and this OpenAI venture fund, etcetera.
也许这些交易让非营利组织的人说：嘿。
Maybe that deal making made the nonprofit people say, hey.
听好了。
Listen.
你们用OpenAI的名义在做更多营利性的事情。
You're doing even more for profit stuff with the OpenAI name.
我们本应是非营利组织。
We're supposed to be a nonprofit here.
那个风险基金是怎么回事？
What's the venture fund thing?
他成立了一个名为OpenAI风险基金的风险基金来投资公司，并且他是该基金的唯一所有者，现在他们说这只是一个行政错误之类的问题。
He started a venture fund to invest in companies called the OpenAI Ventures Fund, and he was the sole owner of it, which they're saying now it's a clerical mistake or something.
但他投资了一大批初创公司，这些公司能独特地接触到OpenAI的基础设施，我觉得是这样。
But he's invested in a bunch of startups that have unique access to the OpenAI, you know, I think So I mean infrastructure.
他是代表OpenAI的普通合伙人。
He's the he's the GP on behalf of OpenAI.
基金。
Fund.
是的。
Yes.
那笔钱是谁的？
Whose money is that?
萨姆筹集了资金，创建了OpenAI基金，用于投资使用OpenAI软件的公司。
Sam raised the money to create an OpenAI Fund to o to invest in companies using OpenAI software.
好的。
Okay.
但这是另一个层面的问题。
So but that's a different kind of issue.
这被称为公司机会问题，即OpenAI的基金在做这些事情。
That's called a corporate opportunity issue where OpenAI's Stop fund doing.
停止做。
Stop doing.
这本质上是一种利益冲突，OpenAI基金的经济机会本应属于OpenAI。
It's basically a conflict of interest where the economic opportunity of the OpenAI start fund should belong to OpenAI.
如果萨姆设立了一个由独立有限合伙人出资、他担任普通合伙人并从中获利的独立基金，那就构成了对公司机会的潜在侵占。
If Sam created a separate fund with separate LPs that he's the GP of and gets economics in that, then that's a potential usurpation of a corporate opportunity.
我我不清楚这件事的真相，我只是说这是一个稍微不同的问题。
I I don't I I don't know the truth of that matter, I'm just saying that's a slightly different issue.
回到之前的话题
To go back
而这一点是
to And the point this is
OpenAI的名称。
the OpenAI name.
是的。
Yeah.
那就是企业机会问题变得非常明确的地方。
That that's where it, like, the corporate opportunity becomes really explicit.
回到你刚才提到的关于员工的观点，我只是觉得，你用了‘窃取’这个词。
To go back to the point you were making a minute ago about the employees, I just think I mean, you used the word stealing.
我觉得这个词太强烈了。
I think it's like a really strong word.
我的意思是，我们必须肯定这家公司的员工——如果它是一家公司的话——包括萨姆、格雷格和伊利亚，他们创造了一个了不起的产品。
I mean, I think we have to complement the employees of this company, if it is a company, including Sam, including Greg, including Ilya for creating an amazing product.
对吧？
Right?
不可思议的东西。
Something incredible.
他们创造了
They've created
一些了不起的东西。
something incredible.
当然。
Absolutely.
他们还创建了一个了不起的生态系统，许多开发者都在此基础上进行开发，我认为没有人是在攻击他们的成果。
And they've created, I think, an amazing ecosystem and there's a lot of developers building on top of this and I think no one is attacking their work.
我认为说有人在窃取什么太过苛刻了，因为他们创造了所有的价值。
And I and I think it's a little too harsh to say that anyone's stealing anything because they have created all the value.
问题在于，我认为你必须把萨姆也包含在内，承认他作为CEO做得非常出色，因为他们打造了如此了不起的产品。
The question and and I think you have to include Sam in that and say say that he's done a great job, okay, as CEO because they've built something amazing.
还有，萨克斯，说到你的观点
And, Sax, to your point
对。
Right.
这不算窃取，因为非常聪明的投资者是带着清醒的眼睛投入资金的。
It's not stealing because very smart investors are putting in money eyes wide open.
是的。
Yeah.
我认为这里的批评与结构有关。
I think that the criticism here is related to structure.
从我的角度来看，这也与结构有关。
From my standpoint, it's related to structure.
无论埃隆最初是否被告知了不同的情况，后来内容被更改了，而且显然更改的方式与他最初投入这笔资金时的条款不一致。
And whether Elon was told something at the beginning that's different, it got changed, and clearly it got changed in a way that was not consistent with the terms under which he initially contributed all this money.
是的。
Yeah.
我的观点是，我正在陈述对这里发生事情的最愤世嫉俗的解读。
Mine is I'm stating what is the most cynical interpretation of what happened here.
当然，你也可以给出最善意、最仁慈的解释，而真相或许介于两者之间。
There is the most, you know, benign and benevolent interpretation you could have as well, and and maybe the truth is in between the two.
但这一系列事件，当你肩负着将这些知识产权奉献给世界的使命时，看起来并不光彩。
But the series of events that occur does not look good when you have a mission to give this intellectual property to the world.
所以没有人从中获益。
So no one person benefits from it.
所有人都应该从中获益。
All of humanity is supposed to benefit from it.
这就是开源的初衷。
That was the point of being open source.
对。
Right.
然后你把它关闭了。
Then you close it.
如果它被关闭了，谁会从中获益？
Now who gets the benefit if it's closed?
对。
Right.
员工们。
The employees.
是的。
Right.
这是个好观点。
That's a good point.
而这将是这个问题的关键。
And that's gonna be the crux of this.
而且如果你真的
The and if you really
想要投资者。
The want investors.
投资者。
The investors.
投资者。
The investors.
但不行。
But no.
但正是这样，他们才能吸引到数十亿美元的投资，因为你不可能让别人为一个开源项目投入数十亿美元，而且
But that's how they were able to attract billions of dollars because you can't get billions of dollars to invest in something that's gonna be open sourced and
然后员工们通过二级市场套现了20亿美元。
And then the employees sell 2,000,000,000 in secondary.
所以，如果你真的想以最愤世嫉俗的角度来看待或解读这件事——再次强调，这仅仅是一种解读方式。
So now you if you really wanna take the most cynical approach to this and or interpretation, again, this is just one interpretation.
他们拿了一个开源项目，把它闭源了，融了资，然后在接下来的两年里，靠着这项惊人的创新，套现了20亿美元，装进了自己的口袋。
They took an open source project, they closed it, they raised money, and then within the next two years on this incredible innovation, they sold $2,000,000,000 and put that in their pockets.
现在，如果这些员工是在一家营利性公司，那当然没什么问题。
Now those employees, if it was a for profit, obviously, no problem with that.
但他们却宣称自己的使命是将这项技术无偿奉献给全人类，然后从中攫取了3%到4%的收益，装进了自己的口袋。
But they took a they stated mission of giving this to all of humanity and then took three or 4% of that gain and put it in their pockets.
我认为，这件事接下来会变得非常疯狂。
That's where I think this whole thing is gonna get really crazy.
而这正是国税局关注的焦点。
And this is where the IRS point.
展开剩余字幕（还有 480 条）
是的。
Yeah.
国税局还对Mozilla进行了认定，Mozilla只值1亿美元，但没有任何员工拿走一分钱。
And the IRS also interpreted Mozilla, which was just a $100,000,000, and none of the employees took any of it.
来自Mozilla的那1亿美元，他们只是说，我认为你应该为这笔钱缴税。
The 100,000,000 that came out of Mozilla, they were just saying, like, I think you should pay tax on that.
你应该为这笔钱缴税吗？
Should you pay tax on that?
然后他们被调查了多年。
And then they were investigated for years.
所以我认为，基于Mozilla发生的事情，国税局会对此事非常关注，
So I think the IRS is gonna be on this, like, crazy based upon what happened to Mozilla, which was
杰森，你认为个别员工卖家会因为这件事被国税局审计吗？
Jason, do you think do you think individual employee sellers are gonna get audited by the IRS because of this?
我不确定，我的意思是，我们正处在未知领域。
I don't I mean, I'm no we're in uncharted territory here.
当我发了一些推文，说：嘿。
When I did some tweets and said, hey.
所有了解这些结构的人能联系我吗？
Can everybody contact me who has information on these structures?
我找到了一些Samasource的例子，他们的做法非常干净，没有知识产权转移，也没有员工中饱私囊，更没有像Sam那样的‘上帝之王’做各种交易来谋取私利。
I was able to find the some Samasource example, and they did it very clean, and there was no, like, IP transfer or employees enriching themselves or, you know, a god king like Sam doing all kinds of deals and enriching himself.
这是一个非常干净的结构。
It was a very clean structure.
然后，创立Mozilla的米切尔·贝克也做了同样的事情。
And then Mitchell Baker who set up Mozilla also did the same thing.
嘿。
Hey.
我们只是给员工的工资额外增加五万美元，他们在硅谷算是略高薪，但并没有拿到数十亿美元的股权，而我对营利性公司的人获得数十亿美元股权本身并没有意见。
We'll just pay the employees an extra 50 k on their salaries and they'll be a little overpaid in Silicon Valley, but they're not getting billions of dollars in equity, which, again, I have no problem with people getting billion dollar billions of dollars in equity in a for profit.
但这个本应是为了造福人类。
But this was supposed to benefit humanity.
它本应是开放的。
It was supposed to be open.
如果这是开源的，他们拿了20亿美元，我其实也不会有什么意见，因为我们可以所有人都看到这个使命，然后萨克斯、我和弗里德伯格，以及任何想使用这段代码的人，都可以进去进行调整。
If this was open source and they took the 2,000,000,000, I would actually not have much of a problem with it because we could all be looking at that mission and then Sachs and myself and Friedberg and anybody who wanted to use that code could go in there and adapt it.
你知道吗？
And you know what?
有一些营利性公司，比如苹果和Meta，正在开发开源软件。
There are for profits, Apple and Meta, which are producing open source software.
所以，苹果在这方面相当出色，Meta在这方面更是顶尖。
So if they're producing Apple's pretty competent and Meta's super competent in this respect.
他们能够做开源，但SAM或OpenAI却声称开源代码太危险，不能让我们看到？
They're able to do open source, but SAM is claiming or OpenAI is claiming it's too dangerous to show us the open source code?
我对此表示怀疑。
I call bullshit on that.
他们的代码并没有危险到不能让我们看到。
Their code is not too dangerous for us to see.
这完全是胡说八道
That's complete utter bull
所以你觉得这主要是出于盈利动机吗？
So you you think that there is a motivation for profit, basically?
你的意思实际上是，这件事主要是由利润驱动，而不是像那些邮件所描绘的那样，双方都是勤奋的人，试图解决一个非常困难的问题，应对一个围绕资本的难题？
What you're saying is essentially that this was highly motivated by profit versus versus, like, eve like like, some of the emails, the emails try to paint a picture of very industrious people on both sides who are trying to solve a very hard problem and dealing with a conundrum that's really around capital.
对吧？
Right?
是的
Yeah.
我们需要进行所有这些训练。
We need to we need to do all this training.
这会非常昂贵。
It's gonna be so expensive.
但我觉得你真正想说的是，这其实偏离了那个观点——得了吧，各位。
But I think what you're saying, which is a deviation from that, it's actually, come on, guys.
这些都是非常聪明的人，他们看到了赚钱的机会，于是就抓住了。
These are all very smart people, and those folks saw an opportunity to make a ton of money, they took it.
我的意思是，对我来说，这看起来就是这里发生的事情。
I mean, that seems like what happened here to me.
我的意思是，这简直就是奥卡姆剃刀的情况。
I mean, it's Hockam's razor kind of situation.
我认为，如果你仔细看的话，他们可能后悔把这件事做成非营利组织，然后试图找到办法逆转它。
I think they if you look at it, I think they probably regretted making this a nonprofit and then tried to figure out a way to reverse it.
这实际上就是我认为正在发生的事情。
That's actually what I think is going on here.
而且我认为其中确实有一部分，沙莫特。
And I do think there is part of it, Shammoth.
你说得对。
You're right.
他们需要服务器，也需要算力。
That they needed servers and they needed capacity.
但他们本可以不给员工大量股权、不卖出数十亿美元、不保持开源，然后告诉微软：嘿。
But they could have done that without giving the employees tons of equity, without selling billions of dollars, keeping it open source, and then telling Microsoft, hey.
如果你想要访问这个开源项目或其他什么，他们本来可以只提供
If you want access to this open source or whatever and, you know, they could give
你提出了一个非常非常好的观点。
you're making a very, very good point.
原因是，如果它是开源的，微软完全可以直接拿去，自己做训练，然后付钱就行。
And and and the reason is that if it was open source, Microsoft could have just taken it and done its own training and just paid for it.
是的。
Yeah.
对吧？
Right?
或者向基金会捐款。
Or make a donation to the foundation.
这很有趣，因为
This is interesting because
这是个非常好的观点。
It's a very good point.
因为OpenAI，是的。
Because OpenAI yeah.
OpenAI公布了他们与埃隆的一些邮件。
OpenAI dropped a few of the emails that they had with Elon.
显然，这些邮件是经过挑选的，以最有利于他们的立场。
Obviously, are cherry picked to help their case the most.
但我们不知道其他邮件显示了什么？
What we don't know is what do the other emails show?
比如，他们对这种结构还有什么其他选择？
Like what were their alternatives to this particular structure?
他们是否能以更符合公司原始使命的方式筹集所需资金？
Could they have raised the necessary funds in a way that was more consistent with the original mission of the company?
例如，他们是否可以保持开源？
Could they have remained open source, for example?
另外，有没有员工谈到私有化对他们个人可能带来的好处的邮件？
Also, are there any emails where employees talk about the potential benefit to them of Private.
追求利润，对吧？
Going going for profit, right?
那些邮件已经被泄露了。
Those emails have already been leaked.
在泄露的OpenAI邮件中，他们提到可以将OpenAI的虚拟股权置换为YC的股权。
There was this thing where in the OpenAI emails that were leaked, they talked about being able to swap this Phantom equity in OpenAI into either YC equity.
然后埃隆说，也许还有SpaceX，但我得再想想。
And then Elon said, oh, maybe SpaceX too, but I'll have to figure that out.
还有，微软进来收购了49%的股份，这也是我想在证据开示中看到的另一部分。
Well, then also, like, Microsoft coming in and buying 49%, and then it's this is the other piece that I would wanna see in Discovery as well.
而且，我本人对这些人没有任何意见。
And, again, I don't have any problem with any of the people.
我很喜欢萨姆·阿尔特曼。
I love Sam Allman.
我觉得他很棒。
Think he's great.
我觉得那里所有人都很棒。
I think all the people there are great.
我觉得他们为人类所做的贡献非常伟大。
I think what they've done for humanity is great.
我只是觉得他们本该坚持开源，这才是最初的使命。
I just think they should have kept this thing open source, which was the mission.
但他们后来闭源了，查马斯，还把49%的权重、全部源代码都给了微软。
But then they closed source it, Chamath, and then give 49% of it all the weights, all the source code to Microsoft.
所以对我来说，这简直太离谱了——你想说，把这家非营利组织的知识产权拿走，其中一部分价值被转化为数十亿美元分给员工，而微软则拿走了这家非营利组织全部努力成果的49%去商业化。
So that to me was like a really like, you you wanna talk about taking this nonprofit's IP and then some amount of that bag gets given to the employees for billions of dollars, and then Microsoft gets 49% of all that nonprofit's effort to then go commercialize.
自从这件事公布以来，微软的市值增加了5000亿美元。
And Microsoft has added, what, $500,000,000,000 in market cap since this whole thing has been announced.
所以现在，对。
So now Right.
所有这些利润都集中到了微软的股票中。
All that profit has been aggregated into Microsoft stock.
我认为这就是你必须理解埃隆为何感到被欺骗的原因：因为我们不仅从非营利组织变成了营利性企业，从开源变成了闭源，他还特别担心人工智能的所有好处都会集中到一家强大的科技巨头手中。
And I I think this is where you gotta understand why Elon feels swindled is because not only are we going from nonprofit to for profit and open source to closed source, he was specifically concerned about all the benefits of AI accruing to one powerful big tech company.
当时他认为是谷歌，现在则是微软。
Now at that time, he thought it was Google, now it's Microsoft.
其实并没有太大区别。
There's not really that much of a difference.
他从不希望人工智能的所有好处都掌握在一家极其强大的科技公司手中。
He never wanted all the benefits of AI in the hands of one really powerful tech company.
对。
Right.
而微软现在是最大的什么？
And Microsoft is the most it's now the what?
全球市值最高的公司。
Biggest company in the world by market cap.
这完全违背了他最初的意图。
So this is like the opposite of what he intended.
他们可能白白送了微软一万亿美元。
They gifted a trillion dollars to Microsoft probably.
有个简单的解决方案。
There's an easy solution.
如果他们确实是出于善意，并且是为了正确的原因，那就把代码开源吧。
If they are people of good faith and they're doing this for the right reasons, open source it.
回到原点吧。
Just go back
但正如你之前所说，这或许能解决诉讼，让埃隆撤诉，但却打开了一个关于税务和结构安排的巨大潘多拉魔盒，其影响远不止于OpenAI。
and open that may Jason, but to your earlier point, that may solve the lawsuit and Elon may drop the lawsuit, but it's opened a can of worms with respect to tax and structuring that's much bigger than just OpenAI.
是的。
Yeah.
这关系到整个非营利行业。
It's entire nonprofit industry.
这对每一家新公司来说都是一个税务问题。
It's a tax problem for every new company.
每一个研究这种模式并试图为自己谋取私利的创业者都会效仿，即使这里并非本意。
It's every other entrepreneur that studies this model and tries to replicate it for their own personal gain even if that wasn't intentional here.
所以我们现在真的把蛋打碎了，引发了一整套问题。
So there's a whole set of issues that we've we've really cracked the egg here.
我们必须想办法把蛋重新拼回去。
We gotta we gotta figure out how to unscramble it.
这里有国税局的问题，有道德上正确的问题，还有埃隆的不满。
There there's the IRS issues, then there's what's morally right, and then there's Elon's beef.
如果这是一家营利性公司，而埃隆在种子轮投入了五千万，他应该拥有多少股份，萨克斯？
And if it is a for profit company and Elon put in 50,000,000 when it was a seed round, what would he own, Sachs?
他在这家营利性公司中应该拥有多少股权？
What would his ownership in this for profit be?
天啊。
Oh my god.
我的意思是，大概是
I mean, probably
98%
98%.
好的。
Okay.
所以
And so
是的。
Yeah.
我们需要知道这一轮的总规模。
We we need know the total size of that round.
他投入了最初的4000万美元。
He put in the first 40,000,000.
我们知道其他人投了多少吗？
Do we know what other people put in?
那是大部分资金。
It was most of the money.
对吧？
Right?
是的。
Yeah.
那是大部分资金。
Was most of the money.
所以，我的意思是，如果我们给它一个疯狂的估值，比如5亿。
So I mean, if you put let's just put a crazy valuation on it, 500,000,000.
好的。
Okay.
他拥有大约10%。
He owns whatever, 10%.
不。
No.
在那个时候，AI还没有像今天这样。
That point at that point, it it would AI was not what it is today.
当然。
Of course.
你会以5000万的估值融资1.5亿。
Raised that, like you would have raised a 150 on 50 pre.
好的。
Okay.
这是一笔资本支出非常高的交易，发生在人们不认为能带来巨大回报的领域，这真不幸。
That's that's the way that that that a very CapEx intensive deal in a space that wasn't thought to yield big outcomes, that's unfortunate.
埃隆至少会拥有这家公司。
Elon would have the company at least.
至少。
At least.
所以给他公司20%的股份。
So give him 20% of the company.
这是200亿美元。
It's $20,000,000,000.
这甚至是你所说的扎克的一半。
That's even half of what you're saying, Zach.
对吧？
Right?
我们在这档节目中讨论过，当那些非营利组织董事们发生那场风波时，我们说应该回去重新结构整个组织，让它回归本来应有的样子——一个纯粹的营利性实体。
We talked about on the show when that whole fracas with these nonprofit directors went down, is we said go back and restructure the whole thing to make it what it always should have been, which is just a clean for profit entity.
给埃隆他的股权，给萨姆他的股权，因为原来的那种方式从来就不合理。
Give Elon his equity, Give Sam his equity because it never made sense of another the way.
萨姆有
Sam has
没有股权，但他有一家风险投资公司。
no equity, but he's got a venture firm.
这一切都太复杂了。
It's it's all so convoluted.
这是一种奇怪的补偿形式，他们通过赋予他企业机会来变相补偿他，而实际上他应该直接拥有公司的股权，而公司则应拥有其所有的机会。
It's a weird form of compensation in which they're giving him corporate opportunities in effect as like a type of compensation when really he should just have compensation in the corporation and then the corporation should own all of its opportunities.
哦，这样啊。
Oh, that's okay.
你刚才说的这一点真的很有趣。
That's a really interesting thing you just said.
所以基本上，是的，他以没有股权而闻名，但他却保留了从整个生态系统中获利的选项。
So basically, yeah, it's like he famously has no equity, but then he has this retained optionality to monetize the ecosystem.
因此，尽管他没有直接从这个平台获利，他却能从他周围的一切中获利。
So even though he's not monetizing the thing, he's he gets to monetize everything around him.
而且看起来董事会——至少是新董事会——对此是认可的，因为他在核心事务上的报酬确实偏低。
And it seems like the the board or at least the new board's okay with that because he is being under compensated with respect to the main thing.
所以他得到了一些附加利益，但这种方式其实也不应该是正确的。
So then he gets the side things, but that's not really the way it should work either.
你知道，比尔·盖茨曾经在我们去Facebook的时候对我们说过，一个生态系统的价值在于，它所产生的经济价值超过了平台本身的价值。
You know, Bill Gates said to us famously when we were at when I was at Facebook, the value of an ecosystem is when the economic value generated by the ecosystem exceeds that of the platform.
在这种情况下，如果真是这样，你实际上更愿意拥有生态系统0.5%的份额，而不是OpenAI的5%。
Now in this case, you'd actually rather have 50 basis points of the ecosystem than 5% of OpenAI if that's true.
因为如果这件事真的如此具有革命性，那涉及的金额可能是10万亿、20万亿、30万亿美元。
Because if this thing could be so revolutionary, you're talking $10.20, $30,000,000,000,000.
真是一团糟。
What a mess.
哎哟哟哟哟。
Oyoyoyoy.
萨姆就应该
Sam should just
获得OpenAI的大额期权授予，但与此同时，OpenAI应该拥有自己的风险投资基金。
be given, a a huge option grant in OpenAI, but then OpenAI should own its own venture fund.
是的。
Yeah.
美国证券交易委员会正在调查所有这些事情。
And the SEC is looking into all this stuff.
你知道，他们确实会关注很多事，公平地说，他们确实在调查这件事。
You know, they look into a lot of things in fairness, but they're looking into it.
真是一团糟。
What a mess.
我们会持续关注这件事。
And, we'll keep track of it.
还有另一件疯狂的事，我不知道你们知不知道，尼克，就在我们结束之前，OpenAI有限合伙人投资协议中最荒谬的部分——你们可以在网上搜索‘OpenAI LP协议’，就是这一部分。
The other thing that's crazy, I don't know if you guys know this, but, Nick, just as we close here, the most insane part of OpenAI's LP investment agreement, which is on their website, you can just search for OpenAI LP agreement, is this part.
该合伙企业的存在是为了推进OpenAI公司的使命，确保安全的通用人工智能得到部署，并造福全人类。
The partnership exists to advance OpenAI Inc's mission of ensuring that safe artificial general intentions is deployed and benefits all of humanity.
普通合伙人对这一使命和原则的职责优先于对OpenAI公司的义务。
The general partner's duty to this mission and principles advance in OpenAI Inc.
图表之类的条款优先于盈利义务。
Chart, yada yada yada, take precedent over the obligation to generate a profit.
该合伙企业可能永远不会盈利，普通合伙人也没有义务使其盈利。
The partnership may never make a profit and the general partner is under no obligation to do so.
普通合伙人可以自由地将运营实体的全部现金流再投资于研发活动或相关支出，而无需对有限合伙人承担任何义务。
The general partner is free to reinvest any and all of the operating entities cash flow into research and development activities and or related expenses without any obligation to the limited partners.
所以他们基本上告诉所有人，无论是文德还是员工之类的，我们可以直接让你的股权归零，然后我们可以
And so they basically told everybody, Vinod and employees or whatever, we can just basically make wipe your equity out and we can
随意处置这些
do whatever we want with the
利润，而你很可能就会亏掉你的钱。
profits and you're probably gonna lose your money.
所以这个结构很奇怪。
So this this structure is weird.
你们觉得，当投资者进入时，尤其是在最近这一轮860亿美元的融资中，他们是否评估了这种结构固有的法律风险，还是只是轻描淡写地忽略了它，说
Do you guys think that when the investors came in, especially in this latest round, the $86,000,000,000 round, do you think they underwrote the legal risk inherent in the structure or they just sort of hand waved over it and said
没有。
No.
通常在这种交易中，你们会聘请像毕马威、德勤或安永这样的公司来做完整的财务尽职调查文件。
Typically what happens in these deals is you hire somebody like KPMG or Deloitte or Ernst and Young to do, like, the full financial diligence packet.
对吧？
Right?
通常情况下，许多后期阶段的组织都会通过这种方式来记录他们为有限合伙人履行受托责任的过程。
And that typically tends to be how a lot of late stage organizations document that they've done and manage their fiduciary responsibilities on behalf of their limited partners.
所以，当你完成一笔交易时，
So what happens is you will do a deal.
你会签署一份条款清单。
You'll sign a term sheet.
然后把它交给德勤。
You'll turn it over to Deloitte.
再交给普华永道，让他们去仔细核查一下。
You'll turn it over to KPMG and say, please go and run this down.
然后他们会提供一份报告，确认：
And then what they do is they will furnish a report that says, yes.
这符合所有常规预期。
This meets all the customary expectations.
我猜，如果这些人能好好管理后期资金，他们很可能就把这些工作交给了那些机构。
I suspect that if these folks were doing a decent job of of running late stage money, they probably sent it to those folks.
而那些机构很可能出具了一份报告，说看起来没问题。
And those folks probably produced something that said, this looks fine.
如果你今年读过巴菲特的信，他会有一段非常精彩的评论，谈到德勤、毕马威以及这类报告，用最委婉的说法，那并不是特别支持的态度。
Now if you read Buffett's letter this year, he has a really great commentary on Deloitte and KPMG and these sorts of letters, which is not exactly the most supportive is the best way to say it.
尼克，你可以找到这段提及。
Nick, you can find that mention.
在他那个案例中，是德勤与通达。
In in his case, it's Deloitte and Touche.
但通常来说，萨克斯，这就是他们所做的。
But typically, Sacks, that's what they do.
他们会去找毕马威、德勤或者安永，顺便说一句，99%的情况下，他们的工作其实做得很好，但这是个标准流程。
They go to a KPMG or a Deloitte or a Nurse and Young and say and and by and by the way, 99% of the time, they're I mean, they do really good work, but it's a standard structure.
他们只是想确保没有不正当的事情发生。
And they just wanna make sure that nothing nefarious is amok.
在这种情况下，我怀疑根本没什么问题。
In this case, I doubt anything was amok anyways.
但我怀疑他们是否仔细审视过这个结构，并且还提升了这项交易被拆解所带来的诉讼风险。
I doubt, though, that they looked at the structure and then also elevated the litigation risk of this getting unwound.
我只是看不到那份尽职调查报告——我看过不少这类报告——通常里面会有这一部分。
I just don't see that that diligence report, and I've seen enough of them, typically has that section in it.
另一件事是
The other thing
这里完全虚伪的地方在于，当他们声称达到AGI、即将出现一个有感知的人工智能时，弗里德伯格，他们说会停止运营，不再作为非营利组织等。
that's completely hypocritical here is they said it when they hit AGI and they're gonna be, like, a sentient artificial intelligence going on here, Friedberg, that they would wrap up shop, and they're gonna no longer be a nonprofit, etcetera.
但他们并没有这么做，却声称自己还没达到那个阶段，却又关闭了软件。
But they haven't and but they're claiming they haven't hit that, but they closed the software.
所以，如果他们没达到AGI，而你也不认为他们达到了通用智能，那代码就应该是开源的，对吧，弗里德伯格？
So it should be open source if they haven't hit AGI and you don't think they've hit general intelligence, right, Freeberg?
或者接近通用智能的任何东西？
Or anything close to it?
也许你可以向观众解释一下那是什么，以及他们所声称的要成为营利性组织的事
Maybe you could educate the audience on what that is and and that claim that they have to
闭嘴
shut up
营利性。
the for profit.
是的。
Yeah.
他们并没有。
They haven't.
但我觉得我们一直在重复这个概念：模型应该是开源的还是闭源的。
But I think the we keep repeating this concept of the models should be open source versus closed source.
将人工智能用于造福人类，可以有多种解读方式。
Making AI for the benefit of humanity can be interpreted in a lot of ways.
也许曾经有人，比如埃隆或其他人，提到过我们要让模型开源，但后来这个决定被改变了，原因是为了吸引资金，而这些资金是私人投资，需要有资本回报。
There may have been some anecdotal conversation at some point with Elon or others about we're gonna make the models open source, but there was a reason that that change was made along the way, which was to attract dollars and those dollars need to have some return of capital available to them because they're private investor dollars.
所以我认为这并不一定是他们的使命，如果我错了请纠正我，我认为OpenAI的软件模型并不一定是开源的。
And so I don't think that that was necessarily, and correct me if I'm wrong, I don't think that's in the mission, that the open AI software models will be open source.
为人类利益而开发人工智能可能有多种解读方式，我们拭目以待。
Making AI for the benefit of humanity could probably be interpreted in a lot of different ways and we'll see.
但不，我认为没有人已经实现了通用人工智能这个终极目标。
But no, I I don't think that anyone has this achieved this holy grail of general intelligence.
是的。
Yeah.
我认为OpenAI其中一个有趣又古怪的方面是，他们的使命明确是创造AGI，而大多数人会将AGI与某种科幻式的反乌托邦结局联系起来。
I think one of the the one of the more interesting and kind of wacky things about OpenAI is that their mission is explicitly to create AGI, which most people would associate with some sort of sci fi dystopian outcome.
对。
And Yeah.
我认为这加剧了人们对人工智能的恐惧，因为他们明确试图创造一种能够取代人类的意识。
I think this has like raised the fear factor around AI because they're explicitly trying to create the sentience that's gonna replace humanity.
现在我认为他们对AGI的定义是不同的。
Now I think they define AGI in a different way.
他们说这能取代80%的工作，但我想我们都清楚它真正的本质是什么。
They say it's something that can replace 80% of the jobs, but I think we all kind of know what it really is.
所以我只是想知道
So I just wonder
我不知道。
I don't know.
我觉得
I think
我认为这假设了世界处于一种稳定状态。
I think that that's like, that assumes a steady state in the world.
所以如果你拥有一个系统，它具备了大量高素质知识工作者的所有能力，你可以坐在电脑前说：让我们设计一项火星任务。
So if you end up with a system and the system has all the capabilities of a bunch of really highly qualified knowledge workers and I can sit in front of a computer terminal and I can say, let's design a mission to Mars.
一项火星任务可能需要二十年的工程努力，涉及数百人来设计建筑、规划飞行路径、计算燃料需求，以及研究如何改造火星环境。
A mission to Mars could be a twenty year engineering project with hundreds of people involved to design the buildings, to design the flight path, to figure out the fuel needs, to figure out how you would then be able to terraform Mars.
但如果一个人就能通过电脑交互，设计出一个前往并定居火星的计划呢？
And what if one person could interact with a computer and design a plan to go and inhabit Mars?
所有的技术细节文档都可以生成。
All of the technical detail docs could be produced.
所有的工程规范都可以自动生成。
All of the engineering specifications could be generated.
所有的运营计划、所有的时间节点、所需的人力、所需的产量、所需的资金都可以计算出来。
All of the operating plans, all of the dates, the amount of labor needed, the amount of production needed, the amount of capital needed.
原本需要NASA或某些知名国际公司或资金雄厚的私营企业花费数十年才能完成的工作，一段软件可以在很短的时间内完成。
What would otherwise take NASA or some, you know, international or well funded private company many, many decades to do, a piece of software could do in a very short order.
我认为，这对我而言是一个非常深刻的例子，说明了当这些工具被广泛使用时，人类的潜力将变得无比广阔。
I think that's like a really, like, for me, poignant example of the potential of having these tools broadly available, that the potential of humanity starts to become much broader.
我们可以提出：我想在海底建造一座城市，因为我希望探索地球的更多未知领域。
We could say, I wanna develop a city underneath the ocean because I wanna explore more of the earth.
我认为人类需要去攻克癌症，根据每位患者的基因型，研发出能够治愈癌症的生物药物及其组合方案。
I I think humans need to go go solve cancer, figure out the biologic drugs and the combination of biologic drugs that would be needed to solve cancer based on this this patient's genotype.
高度知识化或他人所称的通用智能型工具的可扩展性是惊人的，单个个体将拥有整个知识工作者团队可供调遣，去完成我们今天甚至无法想象的事情。
The extensibility of highly knowledgeable or what other people might call general intelligence type tooling is extraordinary, that one individual starts to have an entire cohort of knowledge workers available at their disposal to do things that we can't even imagine today.
所以我认为这并不是恶意的。
So I don't think that it is nefarious.
之所以说它是恶意的，是因为我们假设世界如今处于一种静止状态，什么都不会改变，因此一款软件就能取代我们所有人。
It's nefarious because we assume a steady state of the world today that nothing changes, therefore a piece of software replaces all of us.
但人类的潜力正迈向一个我们尚未适应的新时代，因为我们还不真正了解或理解它。
But the potential of humanity starts to stretch into a new era that we're not really comfortable with because we don't really know know it or understand it yet.
我并不是说开发人工智能是恶意的，因为我同意你关于它非凡潜力的所有观点。
I'm not saying it's nefarious to wanna develop AI, because I agree with you about all the extraordinary potential of it.
我想说的是，专门致力于通用人工智能（AGI）这件事有点像狂热崇拜，显得怪异，而在我看来，普通人所说的AGI就是天网。
I'm saying there's something a little bit cultish and weird about explicitly devoting yourself to AGI, which I think in common parlance means Skynet.
是的。
Yeah.
它意味着
It means
你的意思是
mean, you talk about
这种鼓舞人心的观念可能是需要被正视的，即AGI实际上能够实现与人类知识工作者同等的能力，而这可能会释放出一系列全新的机遇。
the inspirational think that's maybe that that parlance is what needs to be addressed, which is AGI effectively enables equivalence to a a human knowledge worker and that, you know, that can kind of unleash a new kind of set of opportunities.
所以你觉得这就是它的本质吗？
So you think that's what it is?
我对AGI的定义是：比有史以来最聪明的人还要聪明。
My definition for AGI is smarter than the smartest human being who ever lived.
是的。
Yeah.
这周我跟一位身处要职的人聊过，他说AGI的定义非常模糊，根本没有清晰的界定，因此每个人都可以根据自己的理解来锚定这个术语，从而合理化自己的立场。
I was talking to somebody this week who's in a position who said the definition of AGI is very fuzzy, that there isn't a clear definition, and therefore, it allows every side to kind of anchor on their interpretation of what that term means and therefore kind of justifies their position.
所以，你知道，我不太认同简单地断言‘我们是否已经达到AGI’这种说法。
So, you know, I don't really feel great about, like, just saying, are we at AGI?
它到底是什么？
What is it?
我们对它的真正含义并没有清晰的认识。
We don't have a clear sense of what it means.
我认为，如果你看看Anthropic本周发布的Cloud 3模型所做的部分工作，你们有人看过这个模型的演示吗？
I do think if you look at some of the work that was done by Anthropic and published in the Cloud three model this week did any of you guys see the demos that were done of the output of that model?
有个人写了一篇关于量子物理中一个非常深奥复杂问题集的论文，他让Claude 3来解决这个问题集，结果模型生成了他完整的论文。
There was a guy who wrote a thesis in quantum physics on a very esoteric complicated problem set, and he asked Claude three to solve this problem set, and it came up with his thesis.
这真的非常了不起。
It was really, like, extraordinary.
而这是他所说的，世界上没人懂这些东西。
And this is something he's like, no one in the world knows this stuff.
他说，我简直不敢相信这个模型居然能写出我的论文。
And he's like, I can't believe this model, like, came up with my thesis.
这种东西，地球上几乎没人读得懂，更别提理解了。
That's the sort of thing that very few people on Earth even read or understand.
而Cloud 3模型能够重现其基础、推导过程，最终生成了他论文的完整输出。
And the Cloud three model was able to kind of recreate the basis, the buildup, and then the the the output of his thesis was Could
可能是
be like
有人写了一个剧本，只给出前两幕，然后让别人猜第三幕。
somebody writing a screenplay and then giving it the first two acts and say, guess the third act.
就像，是的。
It's like, yeah.
这就是第三幕。
This is the third act.
接下来会发生这些事。
Here's what happens.
这相当令人印象深刻。
Like, it's pretty impressive.
这些像，
These like,
是的。
yeah.
很复杂。
Complicated.
你内心深处，当这些人说他们要创造通用人工智能时，你认为他们真正想表达的是什么？
You, like, deep down, when these guys say they're gonna create AGI, what do you think they really mean in their heart of hearts?
哦，他们的意思是
Oh, they mean
《终结者》。
the Terminator.
是的。
Yeah.
他们指的是那个有意识的上帝般的存在。
They mean the sentient guys god.
你们是不是在传播一些错误的观点，你们不应该这样
Are you guys are propagating some bad You guys shouldn't be
我觉得这就是他们所想的。
saying I think that's what they think.
还记得拉里·佩奇说过的话吗。
Remember what Larry Page said said it.
还记得拉里·佩吉对埃隆说了什么吗？
Remember what Larry Page said to Elon?
不要有种族主义。
Don't be speciesist.
是的。
Yeah.
我认为科技界确实有不少人有意推动超级智能的诞生。
I think there's like I think there's a meaningful number of people in the tech community who deliberately wanna give rise to the super intelligence.
关于超级智能，还有另一种观点，认为超级智能意味着软件的智能水平超过了所有人类。
There's another point of view of super intelligence where super intelligence means that the software is now more intelligent than all humans.
因此，软件可能会产生自己的动机，试图在地球上超越人类。
And as a result, the software may have its own motivations to figure out how to supersede humans on Earth.
至于拉里·佩吉的那番话，我并没有亲耳听过，只是和你读了同一篇文章，有些人可能会说，这是进化的必然。
Now the the Larry Page statement, which I I don't know firsthand, I read the same article you did, is one that a group of people might say evolution is evolution.
对。
Yeah.
会这么说。
Would say that.
没错。
That's right.
我知道埃隆持这种观点，认为我们需要保持人类的主导地位，但是
And I know Elon's taken this point of view that, like, you know, we need to maintain human human supremacy, but
我最喜欢的测试是，你知道的，有图灵测试，就是你无法分辨对方是人类还是机器人。
My my favorite test is, you know, there was there's the Turing test, which, like, you can't tell if it's a human or if it's a robot.
但穆斯塔法提出了现代图灵测试，即一个AI模型获得10万美元，必须将其变成100万美元。
But Mustafa came up with the with the modern Turing test, which is an AI model is given a $100,000 and has to obtain 1,000,000.
去吧。
Go.
这挺有意思的。
It's kind of interesting.
另一个是
Then the other It's
好主意。
good idea.
另一个有趣的是加里·马库斯。
The other interesting one was Gary Marcus.
他提出了宜家测试，也就是平板家具测试。
He said the IKEA test, the flat pack furniture test.
人工智能查看宜家平板家具的零件和说明书，控制机器人正确组装家具。
An AI views the parts and instructions of an IKEA flat pack product that controls a robot to assemble the furniture correctly.
哦，看。
Oh, look.
嗨，Sunny，怎么了？
What's up, Sunny?
现在，我们要向我们亲爱的朋友桑迪普·曼德拉表示一下简短的祝贺。
It's now time to do just a quick little congratulations to our dear friend, Sandeep Mandra.
我们叫他Sunny。
We call him Sunny.
那是他的绰号。
That's his nickname.
他也是我们的扑克牌友之一。
And he's one of our poker buddies.
他一直在创办优秀的企业。
He keeps building great companies.
而这次，Sandeep是第一个集齐全部四个Vessis的人。
And in this case, Sandeep is the first person to collect all four Vessis.
我们都投资了他的公司Definitive。
We've all invested in his company, Definitive.
Definitive原本在做AI领域，但本周我们得到了一个好消息，所以想给他一些肯定。
And Definitive was working in AI, but we got some great news this week, and we thought we would give him his flowers.
Sunny，你能跟我们说说这周在你的公司Definitive Intelligence（我记得你的域名是definitive.io）发生了什么吗？
Sunny, you wanna tell us what happened this week with our investment in your company, Definitive Intelligence at definitive.io, I believe, is your domain.
是的。
Yeah.
嗯，有你们的支持，我们公司一直在成长，也看到了一个绝佳的机会与Grok合作，我们已经和他们合作了几个月。
Well, you know, with your guys' support and, you know, we've been growing our company, and we saw a really great opportunity to work together with Grok, and we've been working with them for a couple of months.
你们所看到的所有热潮，都源于我们合作打造的云服务和API产品。
And all the hype that you've seen has been built on the collaboration that we've done, building the cloud offering, the API offerings.
因此，我们决定与他们合并，对此我们非常兴奋。
And so, you know, we've decided to merge with them, and we're super excited.
以前所有的好友们都是Definitive的股东，现在也都成了Grok的股东。
And all the besties are now not only shareholders in Definitive previously, but now shareholders in Grok.
我想这应该是我们第一次所有人都出现在资本结构表上。
This is this is the first investment, I think, where we're all on the cap table.
对吧？
Right?
所以我们现在都朝着同一个方向努力。
So we can we're all rooting in the same direction.
开发者这边发生了什么？
What's happening in the developer side?
势头怎么样？
How's the momentum going?
嗯，势头非常强劲。
Well, the the momentum's incredible.
你知道，我们现在有超过一万六千名开发者在我们的自助实验平台上。
You know, we have now 16,000 plus developers in our, you know, self serve playground.
已经有上千个应用是人们使用API开发出来的，还包含了各种新功能。
There's well over, like, a thousand apps that people have developed using the API and all kinds of new functionality.
你知道，API让开发者能够获得更高的令牌吞吐量和更低的延迟，因此出现了各种新应用，从语音到网页实时翻译都有。
You know, the API allows people to get a higher rate of throughput on tokens and low latency, so there's all kinds of new applications from voice to real time translation of web pages.
我们正在我们的Discord上收集这些应用。
You know, we're collecting them on our Discord.
我们的Discord里也有三千多人。
We have, you know, 3,000 people in the Discord as well.
这是一个围绕Grok真正凝聚起来的社区，我要说的是，开发者们现在经历的这种飞跃，就像当年从拨号上网转向宽带网络时一样。
It is a real community that's come together building around Grok, and what I will say is, you know, sort of the same jump that developers saw when we went from dial up Internet to broadband.
他们现在正从使用传统的LLM API转向使用我们提供的API。
They're seeing that now and from using, like, traditional APIs for LLMs to using, you know, the ones that we offer.
你们支持他们刚刚发布的最新和最强大的模型吗？这些模型看起来相当厉害？
You guys support the latest and profit models that they just launched that seem pretty kick ass?
不。
No.
我们还没有这些模型，但我们正在与外界各方进行讨论，希望支持他们的模型。
We don't have those yet, but we're we're in, you know, we're having discussions with everyone out there and we wanna support their models.
目前，由于需求巨大，我们暂时只限于Llama 2 70B和Mixtral。
Right now, what we've done given all the demand is we've kinda limited it to Lama two seventy b and Mixtrol.
实际上，我们还为用户提供了其他一些模型，但仅限私有模式，我们对此非常兴奋。
And we actually have a bunch of other models that we make available in private mode for folks, we're pretty excited.
但如果有人希望我们为他们提供支持，请随时联系我们，我们会帮助您接入我们的系统。
But if there's anyone out there that wants to have us, you know, give us a call, and we'll we'll get you going on our systems.
那么，Sunny，恭喜你。
Well, Sunny, congrats to you.
你是一位非凡的创业者，和你一起经历这段旅程总是很有趣。
You're an incredible entrepreneur, and it's always fun to kind of be on the journey with you.
这是我与桑尼·马德拉合作的第四家公司。
This is my fourth business that I've done with Sunny Madra.
太棒了。
That's amazing.
ExtremeLens。
ExtremeLens.
是的。
Yeah.
然后中间还有一家公司。
Then there was the company in between.
没有。
No.
没有。
No.
没有。
No.
中间还有一家公司，你应该知道。
There was a company you should
你得理解一下。
you have to understand.
Sunny和我相识，是因为Sunny去了我长大的地方上学。
Sunny and I met because Sunny went to school where I grew up.
他上了渥太华大学。
He went to the University of Ottawa.
我是在那里长大的。
I grew up there.
哦。
Oh.
我们通过一个共同的朋友认识，他第一家公司的名字叫SpongeFish，我想我是在2006年左右投资了它。
And we met through a mutual friend, and his first company was called SpongeFish, which I backed, I think, in 2006 maybe.
我的意思是，我那时候一无所有。
I mean, I had no money.
我可能写了一张一万美元的支票，大概吧。
I may have written a $10,000 10 k maybe.
可能更少。
Maybe less.
一万或五千。
10 k, 5 k.
是的。
Yeah.
我的意思是，我有什么就给什么。
I mean, whatever I have.
我没多少钱。
I didn't have much.
我喜欢这个。
I like it.
凑齐的
Scraped together.
这是我们第四次创业了。
And it's it's our fourth business.
Sunny是个变量兔脚。
Sunny's a variable rabbit's foot.
你就是个兔脚，Sunny。
You're a rabbit's foot, Sunny.
他是。
He is.
他是个幸运盒。
He's a luck box.
Lex更喜欢有准备的。
Lex prefers the prepared.
嗯，我参与过Sandeep的两家公司，我感谢他。
Well, I mean, I've been in two of Sandeep's companies, and, I thank him for that.
然后，Extreme Labs，再次感谢你们。
And then Extreme Labs, actually, to to thank you again.
十年前我刚起步的时候，你们赞助了我很多活动，我非常感激。
You sponsored a lot of my events ten years ago when I was coming up, so I appreciate that.
所以看到一个好人赢，而不是我在这档播客里一起工作的那三个倒霉蛋一直赢，真是太好了。
So it's just great to see a nice guy win as opposed to the three miserable people who I work with on this podcast keep winning.
听起来不错。
Sounds good.
终于，一个好人赢了，而不是我们这四个人。
Now finally, a good guy wins as opposed to the four of us.
非常感谢。
Well, appreciate
你们的支持对我们意义重大。
all your guys' support.
这对我们来说非常重要。
It means a lot to us.
你们这些开发者持续推出产品，也让我们保持诚实。
You guys getting the developers and keep keep putting it out there and and keep us honest as well.
如果我们有什么做得不对的地方，一定要告诉我们。
So if we're not doing something right, let us know.
我们非常感谢你们所有人。
We really appreciate you guys all.
萨克斯，你有什么话想对桑尼说吗？
Sax, you wanna say something motivational to Sunny here?
我知道你总能给朋友一些鼓舞人心的话，那种让人感受到投身其中的激情。
I know you always have a great motivational word for your friends, something you say that just gives people that thrill of being in the game.
说吧，扎克。
Go ahead, Zach.
你总是有话要说。
You always have something
要表达出来。
coming to say.
我们要把两千万元打入你的安全账户。
We're gonna we're gonna ship 20,000,000 into your safe note.
我知道你已经关闭了，但我们还是要把它打开，哦。
I know you closed it, but we're gonna pry it open and Oh.
拿一些配额吧。
Get get some get some allocation.
等一下。
Wait a second.
你是核心用户，但是
You're prime, but
这很安全。
that's safe.
安全账户的
The safe's
如果有个楔子的话，
if there's a wedge,
我认识一个
I know a
有个拥有80万粉丝的播客的人。
guy with a podcast who could it's got 800,000 followers.
也许我可以悄悄塞进去500。
Maybe I could slip in a quick 500.
这可能吗？
Is that possible?
你觉得呢，桑迪普？
What do you think, Sandeep?
告诉我们一下。
Let us know
对，我们这周和沙特人做了个大公告，所以那个价格可能上涨了。
right We we did a really big announcement with the Saudis this week, so that price might be up.
所以让我们往那个安全笔记里注入一些东西，这样我们才能真正有切身利益。
So let us ship something into that safe note so we actually have some real skin in the game.
大卫，给我发条短信。
David, text me.
好的。
Alright.
明白。
Will do.
嗯，不行。
Well, no.
但说真的，嘿，桑德普。
But seriously, hey, Sandeep.
等等，等等。
But wait a second.
如果他——我的意思是，如果你派那个人去，得再确认一次。
If he's I mean, if you're send that guy, gotta get a second check-in.
朱莉想投入500。
July wants in for 500.
他的意思是500美元。
By that, he means $500.
所以我们让你
So we'll let you
你送了我一张礼品卡。
I got a gift card from you
来自星巴克。
from Starbucks.
我收到了
I got
我给你和团队准备了一张500美元的礼品卡。
a $500 gift card for you and the team.
我觉得我拿到了大约100张票。
I think I got, like, a 100 tickets.
我要去兑换那张星巴克礼品卡。
I'm gonna redeem that that Starbucks gift card.
你会以每美元20美分的价格卖掉它
You'll sell it
然后把它运进来
for 20¢ on the dollar and then ship it in
把它运进来。
the it in.
也许是我
Maybe me
我可以把它放大100倍。
I can 100 x it.
不行。
No.
认真的，你那个朋友Jay Cal给了你50万美元。
Seriously, 500 k from from your boy, Jay Cal.
好了，各位。
Alright, everybody.
感谢Sunny加入。
Thanks to Sunny for jumping on.
祝贺你们完成合并。
Congratulations on the merger.
谢谢大家。
Thanks, guys.
好的。
Alright.
第二个问题。
Issue two.
第二个问题。
Issue two.
苹果正在与两位主要的iOS开发者对抗，而监管机构站在了开发者一边。
Apple is battling two major iOS developers and regulators are siding with the devs.
你可能听说过苹果与Epic Games的纠纷。
You may know about the Apple versus Epic Games saga.
我们之前在这里讨论过。
We've talked about it here.
Epic Games 计划在 iOS 上创建一个自定义应用商店，因为欧洲的《数字市场法案》（DMA）规定，苹果现在必须允许第三方应用商店在欧盟运营。
Epic Games is planning to create a custom app store on iOS because Europe's DMA, the the Digital Markets Act, has said that Apple now has to allow third party app stores in The EU.
因此，Epic 在瑞典注册了一个开发者账户，苹果两周前实际上批准了这个账户。
So Epic created a developer account based in Sweden, and Apple actually approved the account two weeks ago.
但到了周三，苹果态度逆转，终止了 Epic 的欧盟开发者账户。
Then on Wednesday, Apple flipped, terminated Epic's EU developer account.
苹果表示，终止该账户的原因之一是 Epic 首席执行官公开批评了他们的 DMA 合规计划。
And Apple said one of the reasons they terminated the account was because Epic CEO publicly criticized their DMA compliance plan.
此外，周一，苹果被欧盟反垄断监管机构处以 20 亿欧元罚款，并被迫取消音乐应用（如 Spotify）中的反引导规则。
Additionally, on Monday, Apple was fined 2,000,000,000 by the EU's antitrust regulators and was forced to remove its anti steering rules, from music apps like Spotify.
基本上，苹果一直限制音乐应用告知用户价格和折扣信息，而欧盟委员会认为，由于苹果自己运营 Apple Music，这种做法具有反竞争性。
Basically, Apple has been restricting music apps from informing users about pricing and discounts, and the European Commission consider this anticompetitive since Apple runs Apple Music.
他们要求 Spotify 支付 30% 的费用。
And they want Spotify to pay 30%.
丹尼尔·埃克，这位播客的朋友，你知道的，他专门做了一个视频，讲他们不能多收30%的费用之类的。
Daniel Ek, a friend of the pod, you know, basically did a whole video on this about how they can't charge 30% more, yada yada.
萨克斯，你以前谈过苹果的垄断问题。
Sax, you've spoken about Apple's monopoly before.
你对欧盟正在发生的事情有什么看法？之后我们再聊聊苹果更广泛的问题。
Your thoughts on what's happening in the EU, and then we'll get into Apple's wider problems.
你是说苹果因为不喜欢Epic对他们的评价，就把Epic从应用商店里踢出去了？
I mean, did you just say that Apple booted Epic from their App Store because they didn't like what Epic said about them?
我觉得他们这么做是为了让Epic吃点苦头。
I think they took their killings to hurt.
他们这是在侵犯Epic的言论自由，因为不喜欢Epic说的话。
Well, they're violating Epic's free speech because they don't like what Epic is saying.
我的意思是
I mean
根据Epic的说法，是的。
According to Epic, yes.
苹果此举太过强势了。
This is crazily heavy handed by Apple.
苹果对开发者很强势吗？
Apple heavy handed with developers?
是的。
Yeah.
他们疯了吗？
Have they lost their minds?
我的意思是，这正是‘权力导致腐败，绝对权力导致绝对腐败’的典型表现。
I mean, this is right out of power corrupts and absolute power corrupts, absolutely.
是的。
Yeah.
我的意思是，无论你和Epic有什么纠纷，都不该因为不喜欢他们的批评就将他们踢出应用商店。
I mean, whatever dispute you have with Epic, you don't boot them out of your App Store because you don't like their criticism of you.
我的意思是，这恰恰证明了大家一直以来对苹果的指控完全正确。
I mean, this is basically proving exactly what everyone's been saying about Apple Totally.
这正是。
Which is Totally.
他们太强大且专横了。
They're too powerful and heavy handed.
苹果却跑来说：让我通过对Epic采取专制行为来给你们确认一下。
And Apple's coming along and saying, let me confirm it for you guys by acting tyrannically against Epic.
没错。
Exactly.
我的天，这简直是搬起石头砸自己的脚。
I mean, talk about a backfire.
这在我看来太疯狂了。
This seems insane to me.
是的。
Yeah.
这简直太荒谬了。
It's super nuts.
沙马茨，我们在讨论苹果的其他问题之前，你对此有什么看法吗？
Shamatz, any thoughts on this before we get into Apple's other problems?
我认为这是苹果衰落的开始。
I think it's the beginning of of the decay of Apple.
苹果的巅峰。
Peak Apple.
好吧，我们来深入谈谈这个。
Well, let's let's get into that.
苹果正面临诸多逆风。
There's tons of headwinds facing Apple.
你还可以加上
You can add
杰森，清单上还可以加上很多其他因素。
a bunch of other things to the list as well, Jason.
是的。
Yeah.
问题是，过去几年苹果本质上是一家所谓的GDP加增长公司，也就是说，取GDP增长的两到四个百分点。
The thing is Apple for the last couple of years has been what is effectively what we call a GDP plus growth company, which means that take GDP two, three, four percent.
也许它们的增长能比这个再高出几个百分点，但它们实际上与GDP紧密挂钩。
Maybe they can grow by a couple of percentage points more than that, but they are effectively levered to GDP.
当你看Facebook或英伟达时，它们的增长率是50%、200%、2000%，不管是多少。
Meaning, when you look at a Facebook or an Nvidia, they're growing at 50%, 200%, 2000%, whatever it is.
这些增长与GDP无关。
That's not tied to GDP.
它们只是在抢占市场份额。
They're just taking share.
但苹果现在是一家随着全球经济扩张而增长的公司。
But Apple is a company now that grows as the economy of the world grows.
因此，除非它能拓展其业务覆盖的范围，否则这对它的未来前景并不乐观。
So that's not super great for its future prospects unless it can expand the surface area of where they operate.
在这方面，有几个万亿美元规模的市场，它们真的可以渗透进去。
And then on that dimension, the there are a few trillion dollar markets that they can really penetrate.
他们刚刚宣布终止了在其中一个领域的项目，那就是汽车。
And they just announced that they've killed a project in one of those areas, which is autos.
对吧？
Right?
泰坦项目，耗资一百亿美元，最终以失败告终。
Project Titan, which was $10,000,000,000 turned out to be a failure.
所以，我认为所有这些迹象都表明，它实际上正逐渐变成一只周期性、对利率敏感的股票。
So all of these things, I think, mean to me that it is effectively becoming a cyclical rate sensitive stock.
而最后一击来自沃伦·巴菲特。
And then the coup de grace is Warren Buffett.
这周我和尼克聊到了这件事。
And Nick and I were talking about it this week.
巴菲特年报中一个有趣的现象是，你可以通过他提及某家公司次数的多少，判断他是否已经对该企业失去兴趣。
And what was interesting about Buffett's letter is that you can tell when Buffett has gotten disengaged with a company based on the number of times he mentions it in his annual letter.
在这个例子中，这是苹果被提及的次数。
So in this example, this is the number of times Apple was mentioned.
为了明确一下，我所说的‘提及’并不是指它出现在图表中或作为披露内容的一部分。
And just to be clear, what I mean by mentions is not when it's included in the chart or part of a disclosure.
我的意思是，沃伦是否明确地以正面或负面方式提到它，或者根本没提，而后者在我看来意义非常重大。
What I mean is when Warren actually explicitly mentions it in a positive or even negative way or he doesn't mention it at all, which I think rings very loudly.
他曾经几乎把苹果视为一切的终极目标，但现在你可以看到这种提及在逐渐减少，从原本频繁出现变成了几乎完全没有。
He went from basically saying Apple was the absolute end all and be all, and now what you can start to see is this shrinking, and it's gone from basically a bunch of times to almost none.
他确实提了一次，但那是在正面评价可口可乐和美国运通的语境下提到的。
He did mention it once, but he mentioned it in the context of talking positively about Coca Cola and Amex.
他赞扬了这两项投资，只是顺便提到它们相比苹果的规模要小一些。
And he was lauding these two positions, but just mentioning that they were not as large in comparison to Apple.
这是今年年报中唯一一次提及苹果。
That's the only mention in this year's annual letter.
有趣的是，上一次出现这种情况是在富国银行身上，鼓点响起——
What's interesting about that is the last time that that happened was with, drum roll, Wells Fargo.
在十五到二十年间，巴菲特建立了庞大的持股。
Over fifteen and twenty years, Buffett built up a huge position.
我认为他能够应对市场的起伏变化。
I think he was able to weather the vicissitudes of the market.
所以即使市场收缩时，他也知道何时该坚持持有这家公司，直到意识到它并非真正属于他的永久性持股，然后才退出。
So even when the markets would contract, he knew when to hold on to that company until he realized that that company was not really one of his forever stocks, and he got out of it.
而在他的信中，提到它的次数几乎降为零。
And the number of times it basically was mentioned in his letter went to zero.
因此，有趣的是，我认为这个巴菲特指标对苹果来说非常重要：它从一个他声称会永远持有的股票，变成了几乎不再被提及。
So, interestingly, I think this Buffett index is a really important one for Apple, which is it went from a forever holding that he said he would own forever to barely getting mentioned.
在他之上、排在苹果前面的，都是巴菲特持有的日本贸易公司、美国运通和可口可乐。
And above him above Apple on that list were all the Japanese trading companies that Buffett owns, American Express, Coca Cola.
所以，我认为这个人对经济的理解，比世界上任何人都要深刻。
So that is a person that understands the economy, I think, better than anybody else in the world.
因此，如果你以押注经济作为持有苹果的理由，而最懂经济的这个人已经开始转向——他在第四季度就开始减持，接着你又看到这些现象：反垄断规则正在扼杀万亿级市场规模的项目，
And so if you're basically taking a levered bet to the economy as a reason to own Apple and the person that understands the economy the most has now started to pivot away, he started to sell in quarter four, and then you see all of these things, antitrust rules, killing projects in in trillion dollar TAMs.
不幸的是，这预示着除非苹果能找到新的出路，否则未来五到十年前景堪忧。
Unfortunately, it speaks for a very bad next five to ten years for this company unless they figure something out.
关于 Bayt 播客
Bayt 提供中文+原文双语音频和字幕，帮助你打破语言障碍，轻松听懂全球优质播客。