上海的月亮: CN — LARRY ROMANOFF: 揭穿埃隆·马斯克 – 部分 11 — 开放人工智能 — Debunking Elon Musk – Part 11

揭穿埃隆·马斯克 – 部分 11 — Debunking Elon Musk – Part 11

译者：珍珠

I’m deeply offended. Source

我感到非常被冒犯。来源

The OpenAI founding story is murky, with conflicting accounts from most parties. There is a Douyin video (on YouTube as well, I assume) where Elon Musk claims he “founded” OpenAI, gave it the name, and that “OpenAI would not exist without me”. [1] The facts do not bear out these assertions. This essay contains dramatic evidence of the clear patterns of Elon Musk’s character leading to an obsessive demand for control, his natural self-serving tendency for deceit, conniving, and manipulation, and his almost obsessive vindictiveness when thwarted.

OpenAI的创始故事很模糊，大多数方面的说法都相互矛盾。有一个抖音视频（我想YouTube上也有）显示，埃隆·马斯克声称他“创立”了OpenAI，并给它起了这个名字，“没有我，OpenAI就不会存在”。[1]事实并不支持这些断言。本文包含戏剧性的证据，证明了埃隆·马斯克性格的明显模式，导致了他对控制的痴迷需求，他天生的自利倾向，包括欺骗、纵容和操纵，以及他在受挫时近乎痴迷的报复心理。

Elon Musk was indeed present in 2015 when a large group of people more or less formally gathered to discuss the production of an AI model. There were many individuals at the table. The founding team comprised Greg Brockman, Trevor Blackwell, Vicki Cheung, Andrej Karpathy, Durk Kingma, Jessica Livingston, John Schulman, Pamela Vagata, and Wojciech Zaremba. Also among them were the primary funders of this initially vague enterprise: Reid Hoffman, Peter Thiel, and Jessica Livingston. Elon Musk also contributed to the initial funding. SEC filings and OpenAI’s internal records show Musk contributed $44.4 million in various tranches between 2016–2020, with the others contributing $90 million. To this point, they relied on donated funds.

2015年，当一大群人或多或少正式聚集在一起讨论人工智能模型的生产时，埃隆·马斯克确实在场。桌子上有许多人。创始团队包括格雷格·布罗克曼、特雷弗·布莱克威尔、张维琪、安德烈·卡尔帕西、杜克·金马、杰西卡·利文斯顿、约翰·舒尔曼、帕梅拉·瓦加塔和沃伊切赫·扎伦巴。其中还有这个最初模糊企业的主要资助者：里德·霍夫曼、彼得·蒂尔和杰西卡·利文斯顿。埃隆·马斯克也为初始资金做出了贡献。美国证券交易委员会的文件和OpenAI的内部记录显示，马斯克在2016年至2020年期间分批捐款4440万美元，其他人捐款9000万美元。到目前为止，他们依靠的是捐赠资金。

Sam Altman and Elon Musk being co-chairmen of Open AI.

Sam Altman 和 Elon Musk 共同担任 Open AI 的联席董事长。

At the beginning, OpenAI was just a research discussion group that included the names listed above, and others. They eventually began serious research, some development, and began hiring engineers and other experts in the field. On December 8, 2015, they incorporated a “non-profit” entity to organise and direct their efforts, with Sam Altman and Elon Musk being co-chairmen of this venture. However, immediately upon its formation, the company found itself embroiled in control disputes entirely related to Elon Musk.

起初，OpenAI只是一个研究讨论小组，包括上述人员和其他人。他们最终开始进行认真的研究，一些开发，并开始招聘该领域的工程师和其他专家。2015年12月8日，他们成立了一个“非营利”实体来组织和指导他们的努力，Sam Altman和Elon Musk是该企业的联合主席。然而，公司成立后不久，就发现自己陷入了与Elon Musk完全相关的控制权纠纷。

As they researched and discussed, it became obvious that the capital requirements to create an AGI were much beyond their initial estimates. From the OpenAI website archive: By early 2017, “Our research progress led us to realize we would need billions of dollars for the compute to build AGI.” [2] [By the fall of 2017] “It also became obvious that a non-profit entity lacked the ability to attract capital in the amounts necessary”, and that they would need a for-profit enterprise to attract not donors but investors. “We and Elon agreed that a for-profit was the next step for OpenAI to advance the mission.” It is important to note that it was Elon Musk who first stated that necessary research funds would be in the billions and that a for-profit company was the only way to attract that volume of funding.

随着他们的研究和讨论，很明显，创建AGI所需的资金远远超出了他们最初的估计。从OpenAI网站档案中可以看出：到2017年初，“我们的研究进展使我们意识到，我们需要数十亿美元的计算能力来构建AGI。”[2] [到2017年秋季]“很明显，非营利实体缺乏吸引所需资金的能力”，他们需要一个营利性企业来吸引投资者，而不是捐赠者。“我们和埃隆都认为，营利性企业是OpenAI推进使命的下一步。”值得注意的是，是埃隆·马斯克首先表示，必要的研究资金将达到数十亿美元，而营利性公司是吸引如此大量资金的唯一途径。

Almost immediately (September, 2017), Elon Musk independently created a Delaware-based public benefit corporation called “Open Artificial Intelligence Technologies, Inc.”, ** with himself as the head. OpenAI was to be folded into Musk’s new company, and [he] “demanded majority equity, absolute control, and to be CEO of the for-profit. We rejected Elon’s terms because giving him unilateral control of OpenAI and its technology would be contrary to the mission”. The details of all this, including full email records, are available on the OpenAI website. [3] When Musk’s demands for control were refused, he withheld all further funding in a fit of pique, leaving Reid Hoffman to provide the cash to support the continuing enterprise.

几乎就在同时（2017年9月），埃隆·马斯克（Elon Musk）独立创建了一家总部位于特拉华州的公益公司，名为“开放人工智能技术股份有限公司”（Open Artificial Intelligence Technologies, Inc.），**并由他本人担任负责人。OpenAI将被并入马斯克的新公司，并且[他]“要求获得多数股权、绝对控制权，并担任这家营利性公司的首席执行官。我们拒绝了埃隆的条件，因为让他单方面控制OpenAI及其技术将违背我们的使命”。所有这些细节，包括完整的电子邮件记录，都可以在OpenAI的网站上找到。[3]当马斯克的控制权要求被拒绝后，他一时赌气停止了所有后续资金投入，让里德·霍夫曼（Reid Hoffman）提供资金以支持企业继续运营。

** Note that this is the official name under which Musk formed this company. He did not name it “OpenAI” as he later claimed, the final name being selected by the group after much debate.

** 请注意，这是马斯克创立这家公司的官方名称。他并没有像后来声称的那样将其命名为“OpenAI”，最终名称是由该团队经过多次辩论后选定的。

“Musk doesn’t invest in companies — he invests in choke points.”

“马斯克并不直接投资于公司，而是投资于那些关键的瓶颈环节”

At about the same time, the group contemplated buying Cerebras, a startup company with a new and successful methodology for manufacturing chips – of which OpenAI would require a great many. The valid reasoning was that manufacturing their own chips would substantially reduce their computing costs, but Musk was planning to route the purchase through Tesla, meaning that he, and not OpenAI would own (and control) the chip manufacturer. The pitfalls of this approach were immediately apparent to the others, and they refused to pursue the purchase. As a former OpenAI engineer testified in court in 2024, “Musk doesn’t invest in companies—he invests in choke points.”

大约在同一时间，该集团考虑收购Cerebras，这是一家初创公司，拥有一种新的、成功的芯片制造方法——OpenAI需要大量这种芯片。合理的推理是，制造自己的芯片将大大降低其计算成本，但马斯克计划通过特斯拉进行购买，这意味着他，而不是OpenAI，将拥有（并控制）芯片制造商。其他人都立即看到了这种方法的陷阱，并拒绝继续购买。正如一位前OpenAI工程师在2024年的法庭上作证时所说，“马斯克不投资公司，他投资瓶颈。”

Musk’s proposal was cleverly designed as a form of permanent entrapment. By tying OpenAI’s core infrastructure to Tesla, Musk would have gained veto power over OpenAI’s operations. Switching chip suppliers later (like Nvidia) would have been prohibitively expensive, locking OpenAI into Musk’s ecosystem. This fits Musk’s established pattern of exploiting transitional moments: just as he tried later to absorb OpenAI into Tesla during its nonprofit-to-profit shift, here he attempted to insert Tesla as an indispensable hardware supplier during OpenAI’s GPU sourcing crisis.

马斯克的提议巧妙地设计成了一种永久性的陷阱。通过将OpenAI的核心基础设施与特斯拉绑定，马斯克将获得对OpenAI运营的否决权。之后更换芯片供应商（如英伟达）的成本将高得令人望而却步，这将使OpenAI陷入马斯克的生态系统。这符合马斯克利用过渡时刻的既定模式：正如他后来试图在特斯拉从非营利转向盈利的过程中将OpenAI纳入特斯拉一样，在这里他试图在OpenAI的GPU采购危机中插入特斯拉作为不可或缺的硬件供应商。

Having failed with that ploy, Musk’s next move was a prolonged attempt to convince the founding group that their billions in financing needs were unattainable and that their only salvation lay in merging OpenAI into Tesla, suggesting they could use Tesla as “a cash cow” to finance all their needs. The proposition included a (more or less imaginary) threat from Google, whom Musk presented as a dangerous enemy whom OpenAI had to beat and outsmart at all costs. It was already clear that Musk was intent on obtaining absolute control of OpenAI, by whatever means.

这一策略失败后，马斯克的下一步行动是长期试图说服创始团队，他们数十亿美元的融资需求是无法实现的，他们唯一的救赎在于将OpenAI合并到特斯拉，暗示他们可以利用特斯拉作为“摇钱树”来满足他们所有的需求。这一提议包括谷歌（或多或少是虚构的）的威胁，马斯克将其描述为OpenAI必须不惜一切代价击败和超越的危险敌人。很明显，马斯克决心通过任何手段获得对OpenAI的绝对控制权。

As quoted from an email by some of the founders to Elon Musk: “The current [proposed] structure provides you with a path where you end up with unilateral absolute control over the AGI. You stated that you don’t want to control the final AGI, but during this negotiation, you’ve shown to us that absolute control is extremely important to you. The goal of OpenAI is to … avoid an AGI dictatorship. Thus, we are concerned that … you will choose to retain your absolute control of the company despite current intent to the contrary. So, it is a bad idea to create a structure where you could become a dictator if you chose to, especially given that we can create some other structure that avoids this possibility.” [4] Within hours, Musk wrote back that “this was the last straw”.

正如一些创始人给埃隆·马斯克的电子邮件中所引用的那样：“目前的[拟议的]结构为你提供了一条最终实现对AGI单方面绝对控制的途径。你表示你不想控制最终的AGI，但在这次谈判中，你已经向我们表明，绝对控制对你来说极其重要。OpenAI的目标是……避免AGI独裁。因此，我们担心……你会选择保留你对公司的绝对控制权，尽管目前的意图与此相反。所以，创建一个如果你选择成为独裁者就可以成为独裁者的结构是一个坏主意，特别是考虑到我们可以创建一些其他结构来避免这种可能性。”[4]几小时内，马斯克回信说“这是最后一根稻草”。

The story now becomes a bit bizarre. Musk, having pushed so hard for the creation of a for-profit subsidiary, was now firmly against the idea, demanding that OpenAI remain a non-profit entity. The reason was obvious to everyone: a non-profit company – by definition – has no value. You cannot “sell” a non-profit because it has no residual revenue and the purchaser could never recover his investment. What this means, is that Elon Musk would absorb OpenAI into Tesla at no cost, and would have absolute control. But upon that absorption, Tesla’s value would automatically increase by the value of OpenAI. Musk claimed OpenAI was meant to be “open-source” and non-profit – yet his own emails (cited in the lawsuit discussed below) show he planned to commercialize it via Tesla. As one ex-OpenAI Researcher said in an interview with Fortune magazine: “Elon’s play was obvious: Own the IP via Tesla, license it back to OpenAI. A win-win … for him.”

现在故事变得有点奇怪了。马斯克曾大力推动创建一家营利性子公司，现在却坚决反对这个想法，要求OpenAI保持非营利实体。原因对每个人来说都很明显：非营利公司——顾名思义——没有价值。你不能“出售”一家非营利公司，因为它没有剩余收入，购买者永远无法收回投资。这意味着，埃隆·马斯克将无偿将OpenAI并入特斯拉，并拥有绝对控制权。但一旦合并，特斯拉的价值将自动增加OpenAI的价值。马斯克声称OpenAI是“开源”和非营利的，但他自己的电子邮件（在下文讨论的诉讼中引用）显示，他计划通过特斯拉将其商业化。正如一位前OpenAI研究员在接受《财富》杂志采访时所说：“埃隆的策略很明显：通过特斯拉拥有知识产权，再将其授权给OpenAI。对他来说，这是双赢。”

Elon Musk opposed commercialisation he didn’t control.

埃隆·马斯克反对不受他控制的商业化。

This speaks volumes about Musk’s opportunism, and also his hypocrisy. Elon Musk didn’t object to the commercialisation of AI; he opposed commercialisation he didn’t control. His subsequent efforts to kill the creation of the for-profit entity reflected only his obsession with control. The contradiction is apparent: Elon Musk wanted OpenAI’s research to be “open” and non-commercial unless he controlled it – which he would have done had his Tesla ploy succeeded. OpenAI would have become a closed-source de facto subsidiary of Tesla. Musk would have captured for pennies what later became a ~$90B valuation (2023). Even more, Musk owned ~13% of Tesla in 2018. Absorbing OpenAI into Tesla would have effectively gifted him personally ~$10–12B+ in value by 2023 — with potential for $200B+ long-term.

这充分说明了马斯克的机会主义，以及他的虚伪。埃隆·马斯克并不反对人工智能的商业化；他反对的是他无法控制的商业化。他随后试图扼杀营利实体的创建，这只能反映出他对控制的痴迷。矛盾显而易见：埃隆·马斯克希望OpenAI的研究是“开放”的，非商业化的，除非他控制它——如果他的特斯拉策略成功，他就会这么做。OpenAI将成为特斯拉事实上的闭源子公司。马斯克本可以以几美分的价格获得后来估值约900亿美元（2023年）的东西。更重要的是，马斯克在2018年拥有特斯拉约13%的股份。将OpenAI并入特斯拉，到2023年将有效地为他个人带来约100亿至120亿美元以上的价值，长期来看可能达到2000亿美元以上。

The management of OpenAI continued with the creation of a for-profit subsidiary which they expected (correctly, as future developments would prove) would attract the investors they needed. Elon Musk countered with the Tesla merger, and committed to a $1 billion investment in OpenAI if the others would agree to giving him full control of the for-profit arm. They refused, and Musk reneged on his $1 billion commitment. Thiel and Hoffman grasped this subterfuge, as would any rational person. Ceding control would have turned OpenAI into a Tesla subsidiary enriching Musk personally while betraying its mission. Musk’s offer of $1 billion in future funding was conditional on them giving him full control and agreeing to the merger into Tesla. However, it should be obvious that if events were to proceed in that sequence, the $1 billion funding would be moot because Musk would have effectively been giving the money to himself.

OpenAI的管理层继续创建了一个营利性子公司，他们预计（正如未来的发展所证明的那样，这是正确的）这将吸引他们所需的投资者。埃隆·马斯克（Elon Musk）以特斯拉合并案作为反击，并承诺如果其他人同意让他完全控制营利性子公司，他将向OpenAI投资10亿美元。他们拒绝了，马斯克也背弃了他10亿美元的承诺。蒂尔和霍夫曼抓住了这个诡计，任何理性的人都会这么做。放弃控制权将使OpenAI成为特斯拉的子公司，使马斯克个人致富，同时背叛其使命。马斯克提出的未来10亿美元资金的条件是他们给他完全控制权并同意合并到特斯拉。然而，显而易见的是，如果事件按此顺序进行，10亿美元的资金将毫无意义，因为马斯克实际上是在把钱给自己。

It was primarily Sam Altman and Greg Brockman who saw Musk’s agenda as it really was, and who led the opposition against his push for control. That breakdown in their relationship had consequences, as you will see. Musk’s response to the refusal to cede him absolute power: “I will no longer fund OpenAI … I’m just being a fool who is essentially providing free funding for you to create a startup. Discussions are over.” Musk based his grab for control on the fiction that OpenAI “had fallen fatally behind Google”. His claim: “My probability assessment of OpenAI being relevant to DeepMind/Google without a dramatic change in execution and resources is 0%. Not 1%.” The others rejected his view and his “offer”, and Musk departed the board in February 2018, cutting off funding. [5] Musk’s last act on departure was to address a meeting of OpenAI staff and gravely inform them that their chance of success was zero without the guidance of (and under the control of) their philosopher-god-warrior.

主要是萨姆·奥特曼和格雷格·布罗克曼看到了马斯克的真正意图，并领导了反对他控制权的行动。他们关系的破裂产生了后果，正如你所见。马斯克对拒绝让他拥有绝对权力的回应是：“我将不再资助OpenAI……我只是一个傻瓜，基本上是在免费资助你创建一家初创公司。讨论结束了。”马斯克对控制权的争夺建立在OpenAI“已经致命地落后于谷歌”的虚构之上。他的说法是：“我认为，如果不大幅改变执行和资源，OpenAI与DeepMind/谷歌相关的可能性为0%。不是1%。”其他人拒绝了他的观点和他的“提议”，马斯克于2018年2月离开了董事会，切断了资金。[5]马斯克离开前的最后一幕是在OpenAI员工的一次会议上发言，严肃地告诉他们，如果没有他们的哲学家–神–战士的指导（和控制），他们成功的机会为零。

There is a Douyin video (I assume also on YouTube) where Elon Musk claims he gave money to OpenAI and never asked for shares in the company in return for his “donations” – “one of the stupider things I’ve done”. But there were no shares existing. OpenAI has an unusual structure where its for-profit arm is owned and controlled by a non-profit 501(c)(3) public charity, and all the control vested in the Board of Directors of that charity. When Musk initially demanded full control, it was not of any voting shares, but of the votes of the Board of Directors of the non-profit that controlled everything. It was only later that OpenAI formed the for-profit company that had shareholdings.

在抖音（我推测YouTube上也有）上有一段视频，埃隆·马斯克在视频中声称，他给OpenAI提供了资金，但从未要求以股份作为对其“捐赠”的回报——“这是我做过最愚蠢的事情之一”。但当时并不存在股份。OpenAI的结构不同寻常，其营利性部门由一家非营利性501(c)(3)公共慈善机构拥有和控制，所有控制权都归属于该慈善机构的董事会。当马斯克最初要求完全控制权时，他并没有要求获得任何有投票权的股份，而是要求获得控制一切的非营利性机构董事会的投票权。直到后来，OpenAI才成立了拥有股份的营利性公司。

A co-founder of OpenAI in 2015, Musk stepped down in 2018, four years after saying that AI is “potentially more dangerous than nukes.” Source

马斯克于2015年成为OpenAI的联合创始人，但在2018年辞职，四年后他表示人工智能“可能比核武器更危险”。来源

The next part is interesting. The Directors of OpenAI offered to give Musk a substantial shareholding in the newly-created for-profit arm, though still far short of full control. And this offer required no investment on Musk’s part. The shares were offered free, in recognition of Musk’s earlier donations to the venture. Musk dismissed their offer, and resigned in 2018. That was when OpenAI created its capped for-profit model and shares became available. The Board again offered free shares to Elon Musk some time after his departure, but Musk again dismissed their offer with contempt.

接下来的部分很有趣。OpenAI的董事们提出，让马斯克在新成立的营利性部门中持有大量股份，尽管这还远远达不到完全控制。而且这一提议不需要马斯克投资。这些股份是免费提供的，以表彰马斯克早些时候对该企业的捐赠。马斯克拒绝了他们的提议，并于2018年辞职。就在那时，OpenAI创建了其有上限的营利性模式，股份开始流通。马斯克离职后不久，董事会再次向埃隆·马斯克提出免费股份的提议，但马斯克再次轻蔑地拒绝了他们的提议。

There’s beautiful irony here because Musk’s donations to the nonprofit became worth far more once the for-profit arm launched. Had he remained, his $44.4M would have been worth billions, but his exit in stubbornness and resentment forfeited this. Elon Musk’s early donations to the enterprise were about 1/3 of the total. This would have been substantially diluted by future investments made by Microsoft and others, but had he accepted the offered equity in the 2019 for-profit subsidiary, his stake could have been 5–10% as a conservative estimate. And that stake today would be worth between $15 billion and $30 billion.

这里有一个绝妙的讽刺，因为一旦营利性子公司成立，马斯克对非营利组织的捐赠价值就会大幅增加。如果他留下来，他的4440万美元将价值数十亿美元，但他的固执和怨恨让他失去了这一点。埃隆·马斯克早期对该企业的捐赠约占总额的三分之一。这将被微软和其他公司未来的投资大幅稀释，但如果他接受了2019年营利性子公司提供的股权，保守估计他的股份可能达到5-10%。而今天，这笔股份的价值将在150亿美元到300亿美元之间。

Musk’s fixation on absolute control, and his pathological resentment at being thwarted, blinded him to a generational opportunity. Musk’s $44 million investment would have been worth billions had he accepted shared governance but, for Elon Musk, control is non-negotiable. This has been recorded as one of the stupidest business decisions of all time.

马斯克对绝对控制的执着，以及他对被阻挠的病态怨恨，使他看不到一代人的机会。如果马斯克接受共同治理，他4400万美元的投资本可以价值数十亿美元，但对于埃隆·马斯克来说，控制权是不可谈判的。这被记录为有史以来最愚蠢的商业决策之一。

There was yet another matter of interest, a striking example of Elon Musk’s manipulative obsessions. On September 20, 2017, Elon Musk sent an email to Ilya Sutskever, Sam Altman, Greg Brockman, and Shivon Zilis, in which he said: “Guys, I’ve had enough. This is the final straw. Either go do something on your own or continue with OpenAI as a nonprofit.” [6]

还有一件有趣的事，是埃隆·马斯克（Elon Musk）操纵狂的典型例子。2017年9月20日，埃隆·马斯克给伊利亚·苏茨克维尔（Ilya Sutskever）、萨姆·奥特曼（Sam Altman）、格雷格·布罗克曼（Greg Brockman）和希冯·齐利斯（Shivon Zilis）发了一封电子邮件，他说：“伙计们，我受够了。这是最后一根稻草。要么你们自己去做点什么，要么继续把OpenAI作为一个非营利组织。” [6]

I would draw your attention to the statement “go do something on your own”. Musk is essentially claiming ownership of OpenAI, and saying clearly that if the others don’t agree with his intentions, they should all leave and “form their own AI company”. That would mean abandoning OpenAI – which they owned – and leaving it all to Elon Musk. That’s astonishing. “Either do what I want, or you can all leave”. The email in question is quite stark: Musk’s ultimatum essentially positions him as the de facto owner of OpenAI, implying the others were disposable. The language – “go do something on your own” – isn’t just dismissive; it’s proprietorial, treating OpenAI as his personal fiefdom.

我想提请你注意“自己去干点什么”这句话。马斯克基本上是在宣称对OpenAI的所有权，并明确表示，如果其他人不同意他的意图，他们都应该离开并“组建自己的AI公司”。这意味着放弃他们所拥有的OpenAI，并将其全部留给埃隆·马斯克。这令人震惊。“要么按我的意愿行事，要么你们都可以离开”。这封有争议的电子邮件相当直白：马斯克的最后通牒基本上使他成为OpenAI事实上的所有者，暗示其他人是可以随意处置的。语言——“自己去干点什么”——不仅仅是轻蔑的；它是所有者的，把OpenAI当作他的个人领地。

Musk echoed similar sentiments in other exchanges, where he effectively stated that he should be given full control and that the others had the “power to leave” if they weren’t happy. From one email exchange to Musk: “We disagree with your statement that our ability to leave is our greatest power, because once the company is actually on track to AGI, the company will be much more important than any individual.” [7] This is Musk’s broader modus operandi: a pattern of centralizing power by ejecting dissenters, something he has done in all companies where he has been involved. 马斯克在其他交流中表达了类似的观点，他实际上表示，他应该被赋予完全的控制权，而其他人如果不满意，则有“离开的权力”。在一封与马斯克的电子邮件交流中：“我们不同意你的说法，即我们离开的能力是我们最大的权力，因为一旦公司真正走上AGI的道路，公司将比任何个人都重要得多。”[7]这是马斯克更广泛的运作方式：通过驱逐持不同意见者来集中权力的模式，这是他在所有参与过的公司中都做过的事情。

我们现在在哪里？

Elon Musk’s pathological urge for revenge against anyone who thwarts or crosses him.

埃隆·马斯克（Elon Musk）对任何阻挠或冒犯他的人都有病态的报复欲望。

At this point, Elon Musk has left OpenAI with plans to start his own version of artificial intelligence. In the meantime, OpenAI have formed their for-profit arm and are receiving huge investments from the tech community. The company continued to focus quietly on research for the next few years, until in November of 2022 they released ChatGPT. This proved to be one of Silicon Valley’s most disruptive creations ever, often referred to as “one of the most successful and transformative consumer tech products of the century”, alongside iPhone, Facebook, TikTok and DeepSeek.

此时，埃隆·马斯克（Elon Musk）已经离开OpenAI，计划开始开发自己的人工智能版本。与此同时，OpenAI已经组建了营利性部门，并从科技界获得了巨额投资。该公司在接下来的几年里继续专注于研究，直到2022年11月发布了ChatGPT。这被证明是硅谷有史以来最具颠覆性的创造之一，通常被称为“本世纪最成功和最具变革性的消费科技产品之一”，与iPhone、Facebook、TikTok和DeepSeek并列。

Musk was just as shocked by the mainstreaming of AI as everyone else, and clearly resentful that he had failed to be part of it. Musk tweeted: *”I’m confused how a non-profit I donated $100M to (actually $44 million), became a $30B for-profit company”* [8]. That was only the beginning of the bad news for Elon Musk. The $30 billion valuation of OpenAI soon became $90 billion and eventually $300 billion – with Musk owning none of it.

马斯克和其他人一样对人工智能的主流化感到震惊，并且显然对自己未能参与其中感到不满。马斯克在推特上写道：“我很困惑，我捐了1亿美元（实际上是4400万美元）的非营利组织，怎么会变成一家300亿美元的营利性公司呢？”[8]。这只是埃隆·马斯克坏消息的开始。OpenAI的300亿美元估值很快变成了900亿美元，最终变成了3000亿美元——而马斯克没有拥有任何股份。

That should be the end of the story, but actually this is where the story really begins. I referred earlier to Elon Musk’s pathological urge for revenge against anyone who thwarts or crosses him. Read on. Musk began publicly criticizing OpenAI for growing too fast and ignoring security concerns, and organised an open letter calling for a six-month moratorium on AI development to stifle OpenAI’s growth. He launched multiple lawsuits meant to cripple his enemy, forged nasty letters against the other founders, orchestrated a coup d’état to have Altman fired, tried to poach their best staff for his xAI, and much more.

这应该是故事的结尾，但实际上，这才是故事的真正开始。我之前提到过埃隆·马斯克对任何阻挠或冒犯他的人的病态报复欲。继续读下去。马斯克开始公开批评OpenAI发展过快，忽视了安全问题，并组织了一封公开信，呼吁暂停人工智能开发六个月，以遏制OpenAI的发展。他发起了多起旨在削弱敌人的诉讼，伪造了针对其他创始人的恶意信件，策划了一场政变，让奥特曼被解雇，试图挖走他们最好的员工到他的xAI，等等。

OpenAI的《权力的游戏》

A member of OpenAI’s leadership team for five years, Mira has played a critical role in OpenAI’s evolution into a global AI leader.

作为OpenAI领导团队的一员已有五年，Mira在OpenAI成为全球AI领导者方面发挥了关键作用。

Ilya Sutskever was (and still is) a wunderkind in the AI industry, brilliant and capable. Elon Musk personally recruited Sutskever from Google in 2015, calling it “OpenAI’s key to success” [9]. Musk also had him appointed to the Board of Directors of OpenAI. This established a power dynamic where Sutskever owed his career to Musk. The circumstances imply a fair amount of trust and also justify an assumption that Musk had some considerable influence over him.

Ilya Sutskever 过去是（现在仍然是）人工智能行业中的神童，才华横溢，能力出众。2015年，埃隆·马斯克亲自从谷歌招募了Sutskever，称其为“OpenAI成功的关键”[9]。马斯克还任命他为OpenAI的董事会成员。这建立了一种权力动态，在这种动态中，Sutskever的职业生涯归功于马斯克。这种情况意味着相当程度的信任，也证明了马斯克对他有相当大的影响力。

For one thing, Sutskever initially supported Musk in his desire to arrange the purchase of Cerebras through Tesla, recognising that would give Musk a dangerous amount of control over OpenAI’s activities. He acknowledged the contradiction, but still pushed for Tesla’s involvement. He explicitly stated in a September 2017 email that Tesla should acquire Cerebras, despite acknowledging this would compromise OpenAI’s mission: “If we decide to acquire Cerebras, I strongly feel it should be done through Tesla … [but] Tesla has an obligation to maximize shareholder returns, which conflicts with OpenAI’s mission.[10][11]

一方面，Sutskever最初支持马斯克通过特斯拉收购Cerebras的计划，因为他认识到这将使马斯克对OpenAI的活动拥有极大的控制权。他承认这种矛盾，但仍推动特斯拉的参与。他在2017年9月的一封电子邮件中明确表示，特斯拉应该收购Cerebras，尽管他承认这将损害OpenAI的使命：“如果我们决定收购Cerebras，我强烈认为应该通过特斯拉来完成……[但]特斯拉有义务最大化股东回报，这与OpenAI的使命相冲突。”。” [10] [11].

Mira Murati was OpenAI’s Chief Technical Officer (CTO) who joined in 2018 and was the core lead of ChatGPT. She also had much prior experience and a good relationship with Elon Musk, since she previously worked for him at Tesla where she led development of the Model X. “A member of OpenAI’s leadership team for five years, Mira has played a critical role in OpenAI’s evolution into a global AI leader. She brings a unique skill set, understanding of the company’s values, operations, and business, and already leads the company’s research, product, and safety functions.” [12] Here is an interesting article on Murati.

米拉·穆拉蒂（Mira Murati）是OpenAI的首席技术官（CTO），她于2018年加入OpenAI，并担任ChatGPT的核心负责人。她之前在特斯拉（Tesla）为埃隆·马斯克（Elon Musk）工作，领导了Model X的开发，因此积累了丰富的经验，并与马斯克建立了良好的关系。“作为OpenAI领导团队的一员已有五年，米拉在OpenAI发展成为全球人工智能领导者方面发挥了关键作用。她带来了独特的技能组合，了解公司的价值观、运营和业务，并已经领导了公司的研究、产品和安全职能。”[12]以下是关于穆拉蒂的一篇有趣文章。 [13] [13]

On November 17, 2023, in a surprise move, OpenAI fired CEO Sam Altman, with President Greg Brockman and three senior scientists simultaneously resigning. Major investor Microsoft (49% equity in the for-profit arm), was apparently blindsided and furious. Altman’s firing came as a complete surprise to him, with Brockman being equally surprised to learn he was being removed as a Director. [14] Initial reports were that the ousting was orchestrated by Chief Scientist Ilya Sutskever over safety concerns and deployment speed Those reports were unlikely to be true.[15]

2023年11月17日，OpenAI突然解雇了首席执行官Sam Altman，同时总裁Greg Brockman和三名高级科学家也宣布辞职。主要投资者微软（持有OpenAI营利性分支49%的股权）显然对此感到震惊和愤怒。Altman的解雇对他来说完全出乎意料，而Brockman在得知自己被免去董事职务时也同样感到惊讶。[14]最初的报道称，此次解雇是由首席科学家Ilya Sutskever出于安全考虑和部署速度的担忧而策划的。这些报道不太可能是真的。 [15]

While initial “justification” centered on speed (too much) and safety (not enough), the “real” reasons were apparently contained in a couple of pdf files “proving” that Altman “was not consistently candid in his communications with the board”. But then an internal memo stated the board’s decision to fire Altman “was not made in response to malfeasance or anything related to our financial, business, safety, or security/privacy practices. This was a breakdown in communication between Sam and the board.” That doesn’t tell us much.

虽然最初的“理由”集中在速度（太快）和安全（不够）上，但“真正”的原因显然包含在几个pdf文件中，这些文件“证明”了奥特曼“在与董事会的沟通中并不总是坦诚”。但随后一份内部备忘录指出，董事会解雇奥特曼的决定“并非针对渎职或与我们财务、业务、安全或安全/隐私实践相关的任何事情。这是萨姆和董事会之间沟通的破裂。”这并没有告诉我们太多。

From here, things happened very quickly. First, Musk’s “protege”, Mira Murati, was quickly installed as the interim CEO. “The board of directors of OpenAI, Inc., … today announced that Sam Altman will depart as CEO and leave the board of directors. Mira Murati, the company’s chief technology officer, will serve as interim CEO, effective immediately.” [16] The next immediate thing was that the employees of OpenAI published a written (and public) statement to the Board, saying in part, “Your actions demonstrate that you do not have the competence to regulate OpenAl.” The letter demanded the resignation of all current board members, the appointment of new Directors, and the immediate reinstatement of Sam Altman and Greg Brockman. [17] They further stated that if their demands were not met, they would all quit. Of OpenAI’s 770 employees, 743 signed the letter. Of particular note is that Ilya Sutskever’s name was #12 in that list, and Mira Murati was also on the list.

自此之后，事情进展得非常迅速。首先，马斯克的“门徒”米拉·穆拉蒂（Mira Murati）迅速被任命为临时首席执行官。“OpenAI股份有限公司董事会……今日宣布，萨姆·奥特曼（Sam Altman）将辞去首席执行官职务并离开董事会。公司首席技术官米拉·穆拉蒂将担任临时首席执行官，即刻生效。”[16]紧接着，OpenAI的员工向董事会发布了一份书面（且公开）声明，其中部分内容如下：“你们的行动表明，你们没有能力监管OpenAI。”该信要求所有现任董事会成员辞职，任命新董事，并立即恢复萨姆·奥特曼和格雷格·布罗克曼的职位。[17]他们进一步表示，如果他们的要求得不到满足，他们将集体辞职。在OpenAI的770名员工中，有743人签署了这封信。特别值得注意的是，伊利亚·苏茨克弗（Ilya Sutskever）的名字在该名单中位列第12，而米拉·穆拉蒂也在名单之列。

From left to right, Mira Murati, Sam Altman, Greg Brockman, Ilya Sutskever. Source

从左至右，米拉·穆拉蒂、萨姆·奥特曼、格雷格·布罗克曼、伊利亚·苏茨克维尔。来源

There was more. The Board of Directors resigned, new Directors were appointed, Sam Altman was re-hired as CEO, Greg Brockman was reappointed to the Board, Ilya Sutskever never returned to the office and resigned to form his own company, and Mira Murati resigned and left. That was an exciting four days.[18][19]

还有更多。董事会辞职，任命了新董事，Sam Altman被重新聘为首席执行官，Greg Brockman被重新任命为董事会成员，Ilya Sutskever再也没有回到办公室，辞职成立了自己的公司，Mira Murati辞职离开了。那四天很令人兴奋。 [18] [19]

Sutskever said he “deeply regretted” the part he played in the removal of Altman and Brockman, but his position was untenable after all the reversals. When you lead a failed coup against your CEO, your future is not bright and continued presence might be awkward. He probably had no choice but to leave. It was interesting that in the fallout of this, the evidence strongly suggested that Sutskever was not really the originator of the coup, but was being used by the Musk-friendly Board members to rid the company of Altman and Brockman. It was also interesting that a later professional investigation into the matter, concluded that there was nothing to have justified the removal of those men. It had all been a fraud. For her part, Mira Murati resigned, saying there were many things that didn’t make sense, and that “I had a feeling I was being used” as a pawn. It suggests she was a temporary placeholder in a Musk-aligned transition plan that failed [20]. That was the same feeling Sutskever had.

苏茨克维尔表示，他对自己在解雇阿尔特曼和布罗克曼一事中所起的作用“深感遗憾”，但在所有逆转之后，他的立场已经站不住脚。当你领导了一场针对首席执行官的失败政变时，你的未来并不光明，继续留下来可能会很尴尬。他可能别无选择，只能离开。有趣的是，在此次事件的影响中，有证据强烈表明，苏茨克维尔并不是政变的真正发起人，而是被马斯克友好的董事会成员利用，以使公司摆脱阿尔特曼和布罗克曼。同样有趣的是，后来对此事的专业调查得出结论，没有任何理由可以证明解雇这些人是有道理的。这一切都是骗局。就米拉·穆拉蒂而言，她辞职了，称有很多事情没有意义，“我有一种被利用”的感觉，就像一个棋子。这表明她是马斯克联盟过渡计划中的一个临时替代者，该计划失败了[20]。苏茨克维尔也有同样的感觉。

The only logical conclusion is that Sutskver was Elon Musk’s invisible hand, a pawn, not a mastermind. While Musk lacked formal authority at OpenAI in 2023, his pattern of hostile takeovers (e.g., Tesla’s founders, Twitter’s board purge), combined with board allies, timed destabilization, and strategic rivalry, strongly implies his involvement. The coup’s failure – due to employee solidarity and Sutskever’s remorse – exposed the limits of his influence. As one source noted: “Musk’s vendetta against Altman is less about AI ethics and more about unchecked control” [21] [22].

唯一合乎逻辑的结论是，Sutskver是埃隆·马斯克的幕后推手，一个棋子，而不是一个策划者。虽然马斯克在2023年在OpenAI缺乏正式权力，但他的敌意收购模式（例如特斯拉的创始人、Twitter的董事会清洗），再加上董事会盟友、定时破坏稳定和战略竞争，强烈暗示了他的参与。政变的失败——由于员工的团结和Sutskever的悔恨——暴露了他影响力的局限性。正如一位消息人士所指出的：“马斯克对Altman的仇恨与其说是关于人工智能伦理，不如说是关于不受限制的控制”[21][22]。

While there is no smoking gun, and no one wants to reveal the machinations behind the scenes, we can be almost 100% certain the episode was engineered by Elon Musk, who retained sympathisers on the Board and among the staff. There is no other explanation that fits all the known facts. This one event by itself might not create absolute certainty, but there is more to come.

虽然没有确凿的证据，也没有人愿意揭露幕后的阴谋，但我们几乎可以100%确定这一事件是由埃隆·马斯克策划的，他在董事会和员工中都有同情者。没有其他解释能符合所有已知事实。这一事件本身可能无法创造绝对的确定性，但还有更多的事情要发生。

Worthy of note is that Musk made prolonged attempts to hire Sutskever when he left, but to no avail. Musk even appealed to him publicly, praising him in Twitter posts for his integrity and humanity. But Sutskever ignored Musk’s overtures, apparently having had enough of Elon Musk, and formed his own company instead. It was the same with Mira Murati; she had an opportunity to return to the fold and join Musk’s xAI, but Mira had also apparently had her fill of Elon Musk and went her own way.

值得注意的是，马斯克在苏茨克维尔离开后曾多次尝试聘请他，但均未成功。马斯克甚至公开向他发出呼吁，在推特上称赞他的正直和人性。但苏茨克维尔对马斯克的示好置之不理，显然已经受够了埃隆·马斯克，并组建了自己的公司。米拉·穆拉蒂的情况也是如此；她有机会回到马斯克的团队，加入他的xAI，但米拉显然也受够了埃隆·马斯克，选择了自己的道路。

随时随地拍摄 — Taking Shots Wherever we Can

Sam Altman and Greg Brockman accused of serious deception and manipulation by Elon Musk. Source

Sam Altman和Greg Brockman被指控严重欺骗和操纵Elon Musk。来源

But Elon Musk’s vendetta was far from over. In fact, he was only just getting started. At around the same time as the coup against Sam Altman, Musk posted an unsigned “anonymous” letter on X, allegedly from “a former OpenAI employee”, filled with unsubstantiated allegations, accusing Sam Altman and Greg Brockman of serious deception and manipulation, and calling on the Board of Directors to conduct a full investigation. [23] Musk also published the letter on Github and possibly other platforms. The full letter is at the link provided, and it is obvious from the text that no “concerned ex-employee” would express his concern in this language. “Elon Musk’s role in amplifying the unsubstantiated allegations against Sam Altman at a time when OpenAI grappled with internal conflict and external scrutiny raises critical questions about the motives.” [24] The letter was almost certainly a contrived forgery, and almost certainly written by Elon Musk.

但埃隆·马斯克的仇视远未结束。事实上，他才刚刚开始。大约在针对萨姆·奥特曼的政变发生的同时，马斯克在X上发布了一封未签名的“匿名”信，据称来自“一名前OpenAI员工”，信中充满了未经证实的指控，指责萨姆·奥特曼和格雷格·布罗克曼严重欺骗和操纵，并呼吁董事会进行全面调查。[23]马斯克还在Github和其他平台上发布了这封信。完整的信件见提供的链接，从文本中可以明显看出，没有“有关的前员工”会用这种语言表达他的担忧。“在OpenAI内部冲突和外部审查之际，埃隆·马斯克在放大对萨姆·奥特曼的未经证实的指控方面的作用引发了对其动机的质疑。”[24]这封信几乎可以肯定是人为伪造的，而且几乎可以肯定是埃隆·马斯克写的。

At around the same time, a long (more than 5,000 words), libelous and slanderous article was published on multiple platforms and languages. I found this one on a Chinese website where it had been translated into Chinese. [25] Full and truncated versions of the letter appear everywhere. [26] There are several noteworthy aspects; one is that the letter is clearly directed to trashing the reputations of only Sam Altman and Greg Brockman, the two individuals most hated for frustrating Musk’s determination to take control of OpenAI.

大约在同一时间，一篇长篇（超过5000字）、诽谤和诽谤的文章在多个平台和语言上发布。我在一个中文网站上找到了这篇文章，它已被翻译成中文。[25]这封信的全文和删节版随处可见。[26]有几个值得注意的方面；其中之一是，这封信显然是为了诋毁萨姆·奥特曼和格雷格·布罗克曼的名誉，这两个人最令人憎恨，因为他们挫败了马斯克控制OpenAI的决心。

Perhaps the most noteworthy aspect is that the letter is minutely detailed, listing hundreds of large and small events and complaints that no staff member could know. The letter presents as written by an outsider, but no outsider could possibly have had the information to have written it. It was obviously written by an insider, but no low-level employee would have had access to that detailed level. And even a senior executive could not have written it because, while they might have had knowledge of some parts of the company, none of them would have had knowledge across all domains, especially in such intricate detail. I can think of only three people who would have had the knowledge to write that article: Sam Altman, Greg Brockman, and Elon Musk. Since Altman and Brockman are non-starters, we have only one remaining candidate.

或许最值得注意的一点是，这封信内容详尽，列出了数百件大小不一的事件和投诉，这些是任何员工都不可能了解的。这封信看似由局外人撰写，但没有任何局外人能掌握撰写此信所需的信息。显然，这封信是由内部人士所写，但没有任何低层员工能接触到如此详尽的信息。即使是高级管理人员也无法撰写此信，因为他们可能了解公司的某些部分，但无人能了解所有领域，尤其是如此错综复杂的细节。我能想到的只有三个人有足够的知识来撰写这篇文章：萨姆·奥特曼（Sam Altman）、格雷格·布罗克曼（Greg Brockman）和埃隆·马斯克（Elon Musk）。由于奥特曼和布罗克曼都不可能撰写此信，所以我们只剩下一个候选人。

偷猎 — Poaching

For Musk, safety is a talking point for manipulation, not an operational principle.

对于马斯克来说，安全是操纵的谈资，而不是运营原则。

Elon Musk has tried poaching staff from OpenAI, and I believe he has had some success. Also, it appears that OpenAI has been losing staff through excess attrition, although not necessarily going to work for Elon Musk. [27] [28] Nevertheless, Elon Musk’s role in OpenAI’s talent exodus involves both direct poaching and indirect encouragement of departures through ideological and operational warfare. When nearly the entire staff threatened to quit in solidarity with Sam Altman, that kind of schism creates recruitment opportunities. Musk took great advantage of this with his legal attacks (more later) on OpenAI, indirectly validating his own accusations and making his companies morally appealing alternatives.

埃隆·马斯克试图从OpenAI挖角，我认为他取得了一些成功。此外，OpenAI似乎因过度流失而失去了一些员工，尽管他们并不一定会为埃隆·马斯克工作。[27][28]然而，埃隆·马斯克在OpenAI人才流失中的作用既包括直接挖角，也包括通过意识形态和运营战间接鼓励离职。当几乎所有员工都威胁要与萨姆·奥特曼团结一致辞职时，这种分裂创造了招聘机会。马斯克利用这一点，对OpenAI进行了法律攻击（稍后详述），间接验证了他自己的指控，并使他的公司在道德上更具吸引力。

Musk publicly courted Ilya Sutskever, flowering him with praise, [29] but also advertised job openings on X, explicitly targeting disillusioned OpenAI employees. This blatant intervention coincided with the 744 staff threatening mass resignations. Musk’s timing was probably perfect. His later lawsuits accusing OpenAI of fraud, calling Altman a “liar” and OpenAI a “scam”, would tend to demoralise staff and legitimise exits. [30] While staff poaching exists in every industry, this case was Elon Musk using talent as a weapon in ideological warfare against his bitter enemy.

马斯克公开向伊利亚·苏茨克弗示好，对他大加赞赏，[29] 同时在X上发布招聘信息，明确针对心灰意冷的OpenAI员工。这一公然干预恰逢744名员工威胁要集体辞职。马斯克的时机可能非常完美。他后来提起诉讼，指控OpenAI欺诈，称奥特曼是“骗子”，OpenAI是“骗局”，这往往会打击员工的士气，使离职合法化。[30] 虽然每个行业都存在员工挖角的情况，但这一案例是埃隆·马斯克利用人才作为武器，在意识形态战争中对抗他的死敌。

It is clear from the facts and all the activity (in this and his other actions) that his prime intent was to eviscerate OpenAI, to make it crumble and fall. Even if the talent joined Anthropic or Google instead of xAI, it would fragment OpenAI’s research dominance. Musk wasn’t merely poaching; he was systematically discrediting OpenAI’s governance to fuel a talent drain. Aside from his direct recruitment, Musk flooded the public media with rhetoric that painted OpenAI as “untrustworthy.” He created what some have called “a viable exit ecosystem” by his own recruiting and by supporting exits to competitors. While not every departure can be linked to Musk, his multi-pronged offensive exploited OpenAI’s internal rifts, making him a catalyst in its brain drain.

从事实和所有活动（包括这次行动和其他行动）中可以清楚地看出，他的主要意图是掏空OpenAI，使其崩溃和倒下。即使这些人才加入了Anthropic或谷歌，而不是xAI，也会使OpenAI的研究优势四分五裂。马斯克不仅仅是挖角；他系统地败坏OpenAI的声誉，以助长人才流失。除了直接招聘外，马斯克还在公共媒体上大肆宣扬OpenAI“不值得信任”。他通过自己的招聘和支持竞争对手的退出，创造了一些人所说的“可行的退出生态系统”。虽然并非每次离职都与马斯克有关，但他的多管齐下的攻势利用了OpenAI的内部分歧，使他成为人才流失的催化剂。

In March of 2023, Elon Musk’s xAI and Grok were an order of magnitude behind their competitors. OpenAI had just released a more powerful technology that underpins ChatGPT. In early tests and a company demo, the technology was shown drafting lawsuits, passing standardised exams and building a working website from a hand-drawn sketch. Musk still hadn’t produced a working AI model and even the primitive Grok-1 wouldn’t be released for another 6 months. He was of course resentful and livid, and concocted a scheme to use the courts to force his competitors into canceling all research for at least 6 months, citing “profound risks to society and humanity.” [31] [32] Musk claimed that his competitors were in an “out-of-control race to develop and deploy machine learning systems that no one – not even their creators – can understand, predict, or reliably control.” The letter said, “If a pause is not put in place soon, governments should step in and create a moratorium.” No leading AI labs joined his moratorium call.[33]

2023年3月，埃隆·马斯克的xAI和Grok相较于竞争对手落后了一个数量级。OpenAI刚刚发布了一项更强大的技术，该技术是ChatGPT的基础。在早期测试和公司演示中，该技术被证明可以起草诉讼、通过标准化考试，并从手绘草图构建一个可运行的网站。马斯克仍未开发出一个可运行的人工智能模型，甚至原始的Grok-1也要再过6个月才能发布。他当然感到愤慨和愤怒，并制定了一项计划，利用法院迫使竞争对手取消至少6个月的所有研究，理由是“对社会和人类构成巨大风险”。[31][32]马斯克声称，他的竞争对手正处于“开发和部署机器学习系统的失控竞赛中，没有人——甚至包括其创造者——能够理解、预测或可靠地控制这些系统”。信中说：“如果不尽快暂停，政府应该介入并制定暂停令。”没有领先的人工智能实验室加入他的暂停令呼吁。[33]

This is what one writer termed “competitive frustration masked as altruism”. OpenAI was clearly dominant. GPT-4, with its 1.76 trillion parameters, represented a massive leap over predecessors, while Musk’s xAI—founded early in the same year, had yet to launch any product. Grok-1, released later in 2023, had only 33 billion parameters and was criticized for hallucinations and limited capabilities compared to GPT-4. “Musk weaponizes “safety” rhetoric to handicap competitors and regulators, while operating his ventures with minimal safeguards.” [34] As AI ethicist Timnit Gebru noted: “Musk’s safety advocacy is performance. He’s not protecting humanity – he’s disrupting competitors.“ This pattern reveals a fundamental truth: For Musk, safety is a talking point for manipulation, not an operational principle.

这就是一位作家所说的“伪装成利他主义的竞争挫败感”。OpenAI显然占据主导地位。GPT-4拥有1.76万亿个参数，代表了其前身的巨大飞跃，而马斯克的xAI——同年早些时候成立，尚未推出任何产品。2023年晚些时候发布的Grok-1只有330亿个参数，与GPT-4相比，因其幻觉和功能有限而受到批评。“马斯克利用‘安全’言论来阻碍竞争对手和监管机构，同时以最小的保障措施经营他的企业。”[34]正如人工智能伦理学家蒂姆尼特·格布鲁所指出的那样：“马斯克的安全倡导是表演。他不是在保护人类，而是在扰乱竞争对手。”这种模式揭示了一个基本事实：对于马斯克来说，安全是一个操纵的谈资，而不是一个操作原则。

利用法院来扼杀竞争 – 压倒性的怨恨 — Using the Courts to Kill the Competition – Overwhelming Resentment

Elon Musk repeatedly denigrated OpenAI, its executives, its abilities, and its products, all as publicly as possible.

埃隆·马斯克（Elon Musk）多次公开贬低OpenAI、其高管、其能力和其产品。

A few months later, Musk launched his own for-profit artificial intelligence company, xAI, but its technology and market influence lagged far behind OpenAI. Musk had hoped that xAI would be a serious competitor to Altman, but he was an order of magnitude behind. Thus, in 2024, Musk launched an attack on Altman in court, dropping the suit when a loss became obvious, reopening it later, and amending it with the shifting climate. His main claim was that Altman violated their original agreement that OpenAI should be solely non-profit to prioritize the public good over profit.[35]

几个月后，马斯克推出了自己的营利性人工智能公司xAI，但其技术和市场影响力远远落后于OpenAI。马斯克曾希望xAI能成为Altman的有力竞争对手，但他却落后了一个数量级。因此，在2024年，马斯克在法庭上对Altman发起了攻击，在明显败诉后撤诉，后来又重新提起诉讼，并根据不断变化的气候进行了修改。他的主要主张是，Altman违反了他们最初的协议，即OpenAI应完全是非营利性的，以公共利益优先于利润。 [35]

Musk filed the lawsuit against OpenAI and two specific founders – Sam Altman and Greg Brockman. The lawsuit was filled with half-truths and false statements. It claimed that the makers of ChatGPT breached their original contractual agreements by pursuing profits instead of developing an AI that benefited humanity. [36] Musk claimed Altman and Brockman convinced him to help bankroll the startup in 2015 with promises it would be a nonprofit, that they deluded him by forming a for-profit entity. Musk totally ignored his own recommendations that they must form a for-profit entity as the only hope for attracting the necessary capital. He also ignored his own push to merge OpenAI into Tesla, which would by definition have made it a for-profit entity. [37]

马斯克对OpenAI和两位特定的创始人——萨姆·奥特曼和格雷格·布罗克曼提起了诉讼。该诉讼充满了半真半假和虚假的陈述。它声称，ChatGPT的制造商违反了他们最初的合同协议，追求利润而不是开发造福人类的人工智能。[36]马斯克声称，奥特曼和布罗克曼在2015年说服他资助这家初创公司，并承诺它是一家非营利组织，他们通过组建一家营利性实体来欺骗他。马斯克完全忽略了他自己的建议，即他们必须组建一家营利性实体，这是吸引必要资金的唯一希望。他还忽略了他自己推动OpenAI与特斯拉合并的努力，这从定义上讲会使它成为一家营利性实体。[37]

I had to laugh when I read the part about how Elon Musk refused shares in the for-profit company, taking a “principled stand” due to his high standard of ethics and the “moral hazard” of being part of a commercialized for-profit AI. And Musk’s xAI is what? A commercialised for-profit AI, with far more “moral hazard” than OpenAI could ever hope for. Of course, the “principle” on which Musk stood, could be summarised as “If I can’t have it all, nobody gets anything”.

当我读到埃隆·马斯克如何拒绝接受这家营利性公司的股份，并因为他的高道德标准和成为商业化营利性人工智能的一部分所带来的“道德风险”而采取“原则立场”时，我忍不住笑了。马斯克的xAI是什么？一个商业化营利性人工智能，其“道德风险”远远超过OpenAI所能期望的。当然，马斯克所坚持的“原则”可以概括为“如果我不能拥有一切，那就谁也别想得到任何东西”。

Musk’s lawsuit was in reality asking the court to force OpenAI to terminate the for-profit entity, to exist only as a non-profit charity, and to refund all investments to the original investors. Musk had been gone from OpenAI for five years; he had no connection of any kind with the company, and had no legal standing to force them to adhere to prior discussions. He was effectively suing OpenAI for breaking a promise they made to themselves years earlier, although even this claim was false. Of course, the lawsuit had no hope of success and was thrown out. [38] His intentions were clear: if the profit arm were killed and the investments returned, OpenAI would have no funds to continue research and development, nor to pay staff, and would quickly die. The court was not blind to Musk’s agenda of wanting only revenge against people who succeeded without him.

马斯克的诉讼实际上是要求法院强制OpenAI终止营利实体，只作为非营利慈善机构存在，并将所有投资退还给原始投资者。马斯克已经离开OpenAI五年了；他与该公司没有任何联系，也没有法律地位迫使他们遵守先前的讨论。他实际上是在起诉OpenAI违反了他们多年前对自己做出的承诺，尽管这一说法也是错误的。当然，这起诉讼没有成功的希望，最终被驳回。[38]他的意图很明确：如果营利部门被关闭，投资被退还，OpenAI将没有资金继续研发，也没有资金支付员工工资，很快就会倒闭。法院并没有对马斯克只对那些没有他而成功的人进行报复的议程视而不见。

In support of his legal and other claims, Musk at the same time attacked everyone in this latest revenge eruption. He particularly resented Microsoft’s investment in OpenAI and condemned them in multiple ways, to the extent that Microsoft’s cooperation with OpenAI became the focus of the case, Musk arguing that Microsoft may have “violated antitrust laws” by investing in OpenAI. Of course, a Microsoft investment in Musk’s xAI would be free of such contamination. He even publicly attacked Apple for working with OpenAI, threatening to ban Apple products if Apple integrated OpenAI technology. “If Apple integrates OpenAI at the OS level, then Apple devices will be banned at my companies. That is an unacceptable security violation.” [39] [40] He repeatedly denigrated OpenAI, its executives, its abilities, and its products, all as publicly as possible.

为了支持他的法律和其他主张，马斯克在最近的报复爆发中同时攻击了所有人。他特别憎恨微软对OpenAI的投资，并以多种方式谴责他们，以至于微软与OpenAI的合作成为案件的焦点，马斯克认为微软通过投资OpenAI可能“违反了反垄断法”。当然，微软对马斯克的xAI的投资不会受到这种污染。他甚至公开攻击苹果与OpenAI合作，威胁如果苹果整合OpenAI技术，就会禁止苹果产品。“如果苹果在操作系统层面整合OpenAI，那么苹果设备将在我公司被禁止。这是不可接受的安全违规行为。”[39] [40]他多次公开诋毁OpenAI、其高管、其能力和其产品，尽可能公开。

如果你不能打败他们，那就买下他们 — If You Can’t Beat Them, Then Buy Them

Bret Taylor, the chairman of OpenAI’s board, said the artificial intelligence company was “not for sale. Source

OpenAI董事会主席布雷特·泰勒表示，这家人工智能公司“不会出售”

With the lawsuits foundering, Elon Musk still wouldn’t give up. In February of 2025, Musk claimed the backing of a consortium and offered $97.4 billion to buy the nonprofit that controlled OpenAI. [41] [42] But this was merely a post made on Twitter (X), and OpenAI said they had received no offer. [43] In response to the X offer, Altman said “No thank you but we will buy twitter for $9.74 billion if you want.” Musk replied, “nice burn”, to which Altman retorted, “If I want a burn, I’ll buy a Tesla”. [44] In an inexplicably foolish press release, Musk stated that he would abandon his bid to purchase OpenAI if Altman and OpenAI committed suicide by eliminating the for-profit entity. Musk seemed deluded that his purchase offer was legally binding on OpenAI and used this in an apparent extortion attempt.[45]

随着诉讼的失败，埃隆·马斯克仍然不会放弃。2025年2月，马斯克声称得到一个财团的支持，并提出以974亿美元收购控制OpenAI的非营利组织。[41][42]但这仅仅是在推特上发布的一条帖子，OpenAI表示他们没有收到任何报价。[43]作为对X报价的回应，奥特曼说：“谢谢，但我们不会买Twitter，除非你们想以97.4亿美元的价格买下它。”马斯克回答说：“烧钱真好”，奥特曼反驳说：“如果我想烧钱，我会买特斯拉”。[44]在一篇令人费解的愚蠢新闻稿中，马斯克表示，如果奥特曼和OpenAI通过消除营利实体来自杀，他将放弃收购OpenAI的计划。马斯克似乎误以为他的收购要约对OpenAI具有法律约束力，并以此进行明显的敲诈勒索。[45]

It should be obvious from the above that Elon Musk desperately wanted absolute control of OpenAI, and engaged in multiple strategies and deceptions to fulfill his goal. It should also be obvious that he was deeply resentful of the success of OpenAI without his participation, and further that he tried every means at his disposal to not only ruin the careers of the principals but to destroy the company. All in the name of vengeance.

从上述内容可以明显看出，埃隆·马斯克迫切希望完全控制OpenAI，并采取了多种策略和欺骗手段来实现他的目标。同样显而易见的是，他对OpenAI在没有他参与的情况下取得的成功深感不满，并竭尽全力不仅破坏主要参与者的职业生涯，还要摧毁整个公司。这一切都是为了复仇。

Next Essay: xAI and Grok

下一篇：xAI和Grok

Mr. Romanoff’s writing has been translated into 34 languages and his articles posted on more than 150 foreign-language news and politics websites in more than 30 countries, as well as more than 100 English language platforms. Larry Romanoff is a retired management consultant and businessman. He has held senior executive positions in international consulting firms, and owned an international import-export business. He has been a visiting professor at Shanghai’s Fudan University, presenting case studies in international affairs to senior EMBA classes. Mr. Romanoff lives in Shanghai and is currently writing a series of ten books generally related to China and the West. He is one of the contributing authors to Cynthia McKinney’s new anthology ‘When China Sneezes’. (Chap. 2 — Dealing with Demons).

罗曼诺夫先生的作品已被翻译成34种语言，他的文章被发布在30多个国家的150多个外文新闻和政治网站上，以及100多个英文平台上。拉里·罗曼诺夫是一位退休的管理顾问和商人。他曾在国际咨询公司担任高级管理职务，并拥有一家国际进出口企业。他曾担任上海复旦大学的客座教授，为高级EMBA课程讲授国际事务案例研究。罗曼诺夫先生现居上海，目前正在撰写一系列共十本书，总体上涉及中国与西方。他是辛西娅·麦金尼新选集《当中国打喷嚏》（第2章——与恶魔打交道）的特约作者之一。

His full archive can be seen at

他的全部档案可以在以下网址查看：

https://www.bluemoonofshanghai.com/ + https://www.moonofshanghai.com/

He can be contacted at:

他的全部档案可以在以下网址查看：

2186604556@qq.com

NOTES
注释

[1] 我是OpenAI存在的原因。
https://www.douyin.com/video/7467066458804194611
[1] 伊隆•马斯克希望OpenAI转为盈利
https://openai.com/index/elon-musk-wanted-an-openai-for-profit/
[3] 伊隆•马斯克希望OpenAI能够盈利
https://openai.com/index/elon-musk-wanted-an-openai-for-profit/
[4] 伊隆•马斯克希望OpenAI盈利
https://openai.com/index/elon-musk-wanted-an-openai-for-profit/
[5] 埃隆•马斯克与OpenAI的混乱分手内幕
https://www.theverge.com/2024/11/18/24299787/elon-musk-openai-lawsuit-sam-altman-xai-google-deepmind
[6] 伊隆•马斯克希望OpenAI转为盈利
https://openai.com/index/elon-musk-wanted-an-openai-for-profit/
[7] 伊隆•马斯克希望OpenAI能够盈利
https://openai.com/index/elon-musk-wanted-an-openai-for-profit/
[8] OpenAI的《权力的游戏》
https://wapbaike.baidu.com/tashuo/browse/content?id=db3f0c1174b541d5057ca710&fromModule=tashuo-article_bottom-tashuo-feed
（9）埃隆•马斯克：OpenAI的首席科学家应考虑转至xAI或特斯拉
https://www.vgover.com/news/73385
（10）OpenAI曾考虑收购AI芯片初创公司Cerebras，但最终未能成功。
https://unwire.pro/2024/11/18/cerebras/ai/
[11] OpenAI一直有一个“核心梦想”：它曾考虑收购Cerebras，一家晶圆级芯片公司
https://baijiahao.baidu.com/s
[12] OpenAI宣布领导层过渡
https://openai.com/index/openai-announces-leadership-transition/
[13] 关于米拉•穆拉蒂（Mira Murati）的6个事实，她是当下科技界最有趣的人
https://www.fastcompany.com/90855799/6-things-to-know-about-openais-mira-murati-the-most-interesting-person-in-tech-right-now
[14] OpenAI宣布领导层过渡
https://openai.com/index/openai-announces-leadership-transition/
[15] OpenAI董事会突然发生政变，首席执行官萨姆•奥特曼被罢免，细节浮出水面
https://arstechnica.com/information-technology/2023/11/report-sutskever-led-board-coup-at-openai-that-ousted-altman-over-ai-safety-concerns/
[16] OpenAI宣布领导层过渡
https://openai.com/index/openai-announces-leadership-transition/
[17] 超过500名OpenAI员工威胁要集体离职，并要求董事会成员辞职
https://www.newmobilelife.com/2023/11/20/500-openai-staff-want-to-quit/
[18] Sam Altman重返CEO职位，OpenAI迎来新一届董事会
https://openai.com/index/sam-altman-returns-as-ceo-openai-has-a-new-initial-board/
[19] 从被解雇到重新聘用仅4天：Sam Altman离开OpenAI的时间线
https://abcnews.go.com/Business/sam-altman-reaches-deal-return-ceo-openai/story?id=105091534
[20] OpenAI 政变第二幕：从《权力的游戏》到《以人民的名义》
https://www.jazzyear.com/article_info.html?id=1126
[21] 白宫机密：奥特曼与埃隆•马斯克的内部争执故事
https://d.foresightnews.pro/article/detail/78286
[22] 山姆•奥特曼（Sam Altman）拒绝了埃隆•马斯克（Elon Musk）970亿美元的收购要约：OpenAI“非卖品”
https://cn.dataconomy.com/2025/02/12/%e5%b1%b0%e5%a7%86%c2%b7%e5%a5%a5%e7%89%b9%e6%9b%bc%ef%bc%88sam-altman%ef%bc%89%e5%85%b3%e9%97%ad%e5%9f%83%e9%9a%86%c2%b7%e9%a9%ac%e6%96%ac%e5%85%8b%ef%bc%88elon-musk%ef%bc%89%e7%9a%84-97b%e5%87%ba/
[23] 马斯克转发了一封来自前OpenAI员工的举报信，指控奥特曼欺诈，混乱和内斗加剧
http://www.360doc.com/content/23/1122/14/39534287_1104895074.shtml
[24] 伊隆•马斯克的匿名指控在OpenAI引发了混乱
https://www.cryptopolitan.com/zh-cn/elon-musk-ignites-openai-turmoil-accusations/
[25] 马斯克再次起诉OpenAI、Sam Altman和Greg Brockmanhttp
://t.ceden.cn/g/?p/516407.html
[26]埃隆•马斯克再次起诉OpenAI和山姆•奥特曼，此前他撤回了之前的诉讼
https://mashable.com/article/elon-musk-lawsuit-openai-again
[28] OpenAI的领导层大换血：今年有9位关键高管离开这家人工智能巨头
https://observer.com/2024/09/openai-executives-resign/
[28] OpenAI正经历高级人才流失
https://fortune.com/2024/08/06/openai-sam-altman-executive-departures-tech-exodus/
[28] Elon Musk: OpenAI的首席科学家应该考虑转到xAI或特斯拉
https://www.vgover.com/news/73385
[30] 伊隆•马斯克称OpenAI的萨姆•奥特曼是个骗子，并以其无与伦比的道德自豪
https://www.cryptopolitan.com/zh-cn/elon-musk-calls-openais-sam-altman-a-liar/
[31] 伊隆•马斯克和其他科技领袖呼吁暂停“失控”的人工智能竞赛
https://edition.cnn.com/2023/03/29/tech/ai-letter-elon-musk-tech-leaders/index.html
[32] 伊隆•马斯克和顶级人工智能研究人员呼吁暂停“大型人工智能实验”
https://www.theverge.com/2023/3/29/23661374/elon-musk-ai-researchers-pause-research-open-letter
[33] 暂停大型AI实验：一封公开信
https://futureoflife.org/open-letter/pause-giant-ai-experiments/
[34] 什么是Grock？埃隆•马斯克备受争议的ChatGPT竞争对手解释
https://zhongguo.eskere.club/%e4%bb%80%e4%b9%88%e6%98%af%e6%a0%bc%e7%bd%97%e5%85%8b%ef%bc%9f%e5%9f%83%e9%9a%86%c2%b7%e9%a9%ac%e6%96%6f%e5%85%8b-elon-musk-%e5%a4%87%e5%8f%91%e4%ba%89%e8%ae%ae%e7%9a%84-chatgpt-%e7%ab%9e%e4%ba%89-2/2024-08-16/
[35] 白宫机密：奥特曼与埃隆•马斯克的内部争执故事
https://s.foresightnews.pro/article/detail/78286
[36] OpenAI对埃隆•马斯克的诉讼作出回应
https://www.nytimes.com/2024/12/13/technology/openai-elon-musk-lawsuit.html
[37] 伊隆•马斯克起诉OpenAI和Sam Altman，称其“背叛”非营利性人工智能使命
https://techcrunch.com/2024/03/01/elon-musk-openai-sam-altman-court/
[38] 法官驳回了马斯克阻止OpenAI盈利计划的请求
https://www.nytimes.com/2025/03/05/technology/elon-musk-openai-profit-lawsuit.html
[39] Elon Musk撤销对OpenAI的“创始协议”诉讼，曾经的兄弟如今成为竞争对手
https://www.coinlive.com/news/elon-musk-withdraws-the-founding-agreement-lawsuit-against-openai-and
[40] “马斯克为何将OpenAI告上法庭？
https://www.fortunechina.com/ugo/292105.htm”
[41]
路透社报道，埃隆•马斯克领导的集团以974亿美元竞购OpenAI的控制权，据华尔街日报报道，该交易于2025年2月10日达成
[42] OpenAI质疑埃隆•马斯克收购的合理性
https://www.nytimes.com/2025/02/12/technology/elon-musk-openai-response.html
[43] 山姆•奥特曼在致员工的信中驳斥埃隆•马斯克收购OpenAI的提议
https://www.wired.com/story/sam-altman-openai-reject-elon-musk-bid/
[44] OpenAI拒绝埃隆•马斯克974亿美元的收购要约
https://www.nytimes.com/2025/02/14/technology/openai-elon-musk.html
[45] 如果OpenAI保持非营利性质，埃隆•马斯克表示将放弃收购该公司
https://www.nytimes.com/2025/02/13/technology/elon-musk-openai-nonprofit.html

This article may contain copyrighted material, the use of which has not been specifically authorised by the copyright owner. This content is being made available under the Fair Use doctrine, and is for educational and information purposes only. There is no commercial use of this content.

本作者的其他作品