Closing time / 闭幕时刻

What did we learn throughout Musk v. Altman? A lot of gossip. 在马斯克诉奥特曼案中我们学到了什么？一大堆八卦。

Today was closing arguments in the Musk v. Altman trial, and I almost feel bad writing about the unbelievable demolition derby I just witnessed. Steven Molo, Musk’s lawyer, stumbled over his words. He at one point called Greg Brockman — a co-defendant — Greg Altman. He erroneously claimed that Musk wasn’t asking for money and had to be corrected by the judge. He made it clear we’ve heard from many liars over the past few weeks, but offered little evidence for Musk’s actual legal claims.

今天是马斯克诉奥特曼案的结案陈词，写下我刚刚目睹的那场令人难以置信的“拆迁式”庭审，我甚至感到有些愧疚。马斯克的律师史蒂文·莫洛（Steven Molo）在陈述时语无伦次。他一度将共同被告格雷格·布罗克曼（Greg Brockman）称为“格雷格·奥特曼”。他错误地声称马斯克并未索要赔偿，结果不得不由法官予以纠正。他清楚地表明，过去几周我们听到了许多谎言，但他却几乎没有为马斯克实际的法律诉求提供任何证据。

OpenAI’s lawyer, Sarah Eddy, countered this by simply arranging the mountain of evidence that the company introduced in chronological order. She didn’t spend time trying to pretend anyone in this trial is especially reliable. She did, however, get the zinger of the day, about Musk: “Even the mother of his children can’t back his story.” William Savitt, who took the defendant baton after her presentation, demonstrated the number of times Musk “didn’t recall” some critical detail — and wondered how a sophisticated businessman couldn’t understand or read a four-page term sheet OpenAI had sent to him.

OpenAI 的律师莎拉·埃迪（Sarah Eddy）对此进行了反击，她只是将公司提交的如山证据按时间顺序进行了梳理。她没有浪费时间去假装这场审判中的任何人特别可靠。不过，她还是贡献了当天的金句，关于马斯克：“就连他孩子的母亲都无法支持他的说法。”在她的陈述之后接过接力棒的威廉·萨维特（William Savitt）列举了马斯克多次“不记得”关键细节的情况，并质疑一位老练的商人怎么会看不懂或读不完 OpenAI 发给他的那份四页纸的条款清单。

I found myself wondering, again, why we were all wasting our time here. So let’s discuss the gossip, which is the real point of this trial. How good was it? Here are my favorite nuggets.

我再次自问，为什么我们都要在这里浪费时间。所以，让我们来聊聊八卦吧，这才是这场审判的真正意义所在。这些八卦有多精彩？以下是我最喜欢的几个片段。

• Musk used OpenAI to improve xAI. I’m afraid this trial was front-loaded; revealed the first week, this was the best bit of gossip we got. As I suggested in 2024, Grok was developed very fast — so fast I was skeptical it was entirely independent. Turns out, it wasn’t. Musk admitted that xAI distilled other models, including OpenAI’s. What was all that investor money for, again? 马斯克利用 OpenAI 来改进 xAI。 恐怕这场审判是“虎头蛇尾”；第一周披露的内容就是我们得到的最好的八卦。正如我在 2024 年所暗示的那样，Grok 的开发速度非常快——快到我怀疑它是否完全独立。事实证明，并非如此。马斯克承认 xAI 提炼了其他模型，包括 OpenAI 的模型。话说回来，那些投资者的钱到底花哪儿去了？

• Tesla AI failed to run at AGI. Musk in fact failed at this twice, once by unsuccessfully trying to acquire OpenAI and again by unsuccessfully trying to kneecap OpenAI by recruiting its employees, including Sam Altman himself, for a “world-class AI lab.” 特斯拉 AI 在实现通用人工智能（AGI）方面失败了。 事实上，马斯克在这方面失败了两次：一次是试图收购 OpenAI 未果，另一次是试图通过招募包括萨姆·奥特曼本人在内的员工来建立一个“世界级 AI 实验室”，从而削弱 OpenAI，但也失败了。

• Sam Altman confirmed decade-old reports that he thought about running for governor of California. 萨姆·奥特曼证实了十年前的报道，即他曾考虑竞选加州州长。

• Everyone, but especially Musk, is obsessed with Demis Hassabis. 每个人，尤其是马斯克，都对德米斯·哈萨比斯（Demis Hassabis）着迷。

• Musk claimed he doesn’t lose his temper. During the cross examination, he lost his temper at OpenAI’s Savitt. 马斯克声称他不会发脾气。 但在交叉盘问期间，他却对 OpenAI 的律师萨维特发了火。

• Mira Murati played both sides of the Altman ouster — providing some chats that got him removed, then feeding Altman details about what the board was up to, publicly complaining about the event, and refusing to tell employees about her role in the ouster. I would call this “directionally very bad” for Murati. Everyone hammered on Altman as a snake, but Murati is no better. 米拉·穆拉蒂（Mira Murati）在奥特曼被罢免事件中两头下注——她提供了导致奥特曼被罢免的聊天记录，随后又向奥特曼透露董事会的动向，公开抱怨此事，并拒绝向员工说明她在罢免事件中的角色。我认为这对穆拉蒂来说是“方向性的大错”。大家都抨击奥特曼是条毒蛇，但穆拉蒂也好不到哪儿去。

• Musk wanted his kids to inherit OpenAI. (At least, according to Altman!) 马斯克希望他的孩子们继承 OpenAI。（至少奥特曼是这么说的！）

• Altman hired Savitt, who kicked Musk’s ass on Twitter, and may very well kick his ass again. I was amused by Altman’s commitment to knifework under oath, managing to suggest both that Musk had little to do with OpenAI, say he had “front-runner-itis,” and imply Musk was no longer respected, but managing to annoy someone just procedurally… From one petty bitch to another: beautifully done. 奥特曼聘请了萨维特，他曾在 Twitter 上把马斯克批得体无完肤，而且很可能还会再来一次。 奥特曼在宣誓后那种“捅刀子”的执着让我觉得很有趣：他既暗示马斯克与 OpenAI 关系不大，又说他患有“领跑者综合症”，还暗示马斯克已不再受人尊重，同时在程序上又成功地激怒了对方……从一个小心眼的人对另一个小心眼的人来说：干得漂亮。

• Brockman and other OpenAI engineers worked at Tesla, while they were theoretically employed by OpenAI, on self-driving software. 布罗克曼和其他 OpenAI 工程师在理论上受雇于 OpenAI 时，却在特斯拉从事自动驾驶软件的工作。

• Shivon Zillis didn’t tell her fellow board members that Musk was the father of her children until Business Insider got ahold of some court documents. She didn’t even tell her own father! 希冯·齐利斯（Shivon Zillis）没有告诉其他董事会成员马斯克是她孩子的父亲，直到《商业内幕》拿到了一些法庭文件。她甚至连她自己的父亲都没告诉！

• “Rumor has it that, on top of the folks that secretly converse on Twitter DM because they don’t trust Demis not to spy on their email and gchat…” “有传言称，除了那些因为不信任德米斯会监视他们的电子邮件和 Gchat 而在 Twitter 私信中秘密交流的人之外……”

• “I don’t want to be IBM and OpenAI to be Microsoft.” “我不想成为 IBM，也不想让 OpenAI 成为微软。”

• The jackass trophy. I am so sorry that the jurors didn’t get to see it. It was like a little league trophy but just the back half of a donkey. Even Judge Yvonne Gonzalez Rogers thought this was pretty funny. “蠢驴”奖杯。 很遗憾陪审员们没能看到它。它看起来像个小联盟奖杯，但上面只有驴的后半身。连伊冯·冈萨雷斯·罗杰斯（Yvonne Gonzalez Rogers）法官都觉得这很有趣。

While this trial was meant to punish Altman and arguably already has, I’d like to focus on my actual takeaway here: Elon Musk sucks at AI. 虽然这场审判旨在惩罚奥特曼，而且可以说已经达到了目的，但我还是想谈谈我真正的收获：埃隆·马斯克在 AI 领域真的很烂。

Look, Musk said multiple times that OpenAI wouldn’t succeed. He’s tried, repeatedly, to kneecap it and steal its researchers and in one case — that of Andrej Karpathy, a founding team member Musk lured to Tesla — succeeded. But how’s xAI doing? Well, it’s a black hole for money that’s been acquired by SpaceX. It’s hemorrhaging researchers. One of its huge data centers isn’t going to be occupied by xAI — there’s a deal with Anthropic instead. It might buy Cursor, in an attempt to match the programming-focused products put forward by Anthropic and OpenAI. xAI’s enterprise users, whether that’s the US government or private companies, have been strong-armed into using it. To the degree that its bespoke CSAM machine Grok, aka MechaHitler, works, it apparently works because Musk distilled other people’s models. 听着，马斯克多次说过 OpenAI 不会成功。他曾多次试图削弱它并窃取其研究人员，其中有一次——即马斯克诱导特斯拉创始团队成员安德烈·卡帕西（Andrej Karpathy）加入——他确实成功了。但 xAI 现在怎么样了？它是一个被 SpaceX 收购的“烧钱黑洞”。它正在流失研究人员。它的一座大型数据中心最终不会由 xAI 使用，而是与 Anthropic 达成了协议。它可能会收购 Cursor，试图与 Anthropic 和 OpenAI 推出的编程产品相抗衡。xAI 的企业用户，无论是美国政府还是私营公司，都是被强迫使用它的。至于它那台定制的 CSAM（儿童性虐待材料）机器 Grok（又名“机械希特勒”）之所以能运作，显然是因为马斯克提炼了别人的模型。

Zilis wrote in 2018 that Brockman and Sutskever thought that Musk “really hasn’t done his homework [on] AI / AGI and that concerns them about working with him.” I am leaving this trial thinking all these fucking liars deserve each other, but in fairness to Brockman and Sutskever, they were absolutely right about this. The question now is if anyone who’s thinking of investing in the upcoming SpaceX IPO has noticed, or cares. 齐利斯在 2018 年写道，布罗克曼和苏茨克维（Sutskever）认为马斯克“在 AI/AGI 方面真的没有做功课，这让他们担心与他共事”。离开法庭时，我心想这些该死的骗子真是绝配，但平心而论，布罗克曼和苏茨克维在这一点上绝对是正确的。现在的问题是，那些考虑投资即将到来的 SpaceX IPO 的人是否注意到了这一点，或者他们是否在乎。