Michael Burry 抨击 OpenAI、Anthropic 呼吁放缓 AI 是“利己之举”

知名投资人 Michael Burry 抨击 OpenAI、Anthropic 及其他科技巨头的领导者，称他们组织呼吁放缓 AI 发展步伐是“利己之举”——他加入了科技界大佬、特朗普总统乃至中国政府的行列，共同批评所谓的 AI 恐慌论。

Burry 因在约 2008 年次贷危机中成功预判并从中获利，而被好莱坞大片《大空头》（“The Big Short”）所推崇。这位做空投资者在 X 上写道：“让我们都花一点时间理解一下，OpenAI、Anthropic 和其他大型超大规模云服务商的高管谈论放缓发展，是多么利己。”

Burry 表示，AI 巨头呼吁踩刹车，是为“真实且无法控制的增长放缓”提供掩护；AI 实验室正难以维持 AI 热潮早期所经历的高速增长。他指出，随着“IPO 看起来将被推迟”，一段降温期将有所帮助。

Michael Burry 加入了 David Sacks 和 Bill Gurley 等科技界大佬的行列，批评席卷硅谷的所谓 AI 恐慌论。Bloomberg via Getty Images

他还将 AI 领导者持续发出的负面论调批评为一种监管俘获：即某一领域的领军者利用监管和放缓措施，阻碍新兴竞争对手。他将放缓呼吁描述为 AI 巨头在大规模 IPO 前的营销手段。

“IPO 需要炒作和夸大；‘我们太强大了，可能会变得危险’就是炒作和夸大，”Burry 写道。Burry 此前曾警告，当前向 AI 投入数十亿美元的热潮，与 2000 年代初的互联网泡沫相似。

OpenAI 和 Anthropic 均在今年夏初向美国证券交易委员会提交了 IPO 文件。Altman 表示 OpenAI 将推迟 IPO，并于周五告诉 Fortune，在当前的 AI 担忧氛围下上市将是一个“不合时宜的时刻”。

Burry 在帖子中称，大语言模型并非通用人工智能，因此“没有什么 AI 可以放缓”；他补充说，“LLM 不是 AI，也不会成为 AGI。”针对 AI 的担忧主要聚焦于 AGI，即机器拥有与人类同等智力能力的技术进展状态。理论上，这种系统可能会在重大决策中将人类排除在外，或因其可能比人类更聪明而对人类构成风险。

Burry 的言论发布前，Anthropic CEO Dario Amodei、OpenAI 的 Sam Altman、Google DeepMind 负责人 Demis Hassabis 和 Elon Musk 均曾呼吁政府加强对 AI 的监管，并要求以更慢的速度推进这项新技术的发展。这些 AI 领导者在近日发布相关文章和社交媒体帖子，相关讨论在周末达到高潮，并导致 AI 股票在本周开盘时承压。

这些公开表态在 AI 领域的主要声音中引发了强烈反响。

OpenAI 首席执行官 Sam Altman 表示，他认同有必要放缓 AI 发展。REUTERS

“不要假装放缓发展的动机完全是利他的，”特朗普总统的科技顾问、白宫前 AI 与加密货币主管 David Sacks 在周六的一则社交媒体帖子中表示。

顶级风险投资人、且经常批评顶尖 AI 公司的 Bill Gurley 告诉《纽约时报》：“如果我们要进行监管，监管者不能是他们认识的人，而且他们不能自己制定规则。”

就连副总统 JD Vance 也发表了看法：“我必须说，仅从个人角度而言，看到这么多前沿 AI 科技公司跑到政府面前，恳求政府监管它们，我感到有点奇怪。这在我看来有一点像特洛伊木马。”

Anthropic CEO Dario Amodei 周六在其个人网站发表的一篇文章中表示，提升 AI 模型能力的进展应当放缓。OpenAI CEO Sam Altman 在 X 上回应 Amodei 的文章：“我同意 Dario 的看法，我们需要控制前沿发展的节奏。”

SpaceXAI（此前为 xAI）创始人 Elon Musk 在自己的 X 帖子中写道：“Dario 是对的。”

微软 CEO Satya Nadella 也支持放缓，称需要“为将对齐作为设计目标而进行的研究、专注和审慎的节奏控制”。

Anthropic CEO Dario Amodei 表示，提升 AI 模型能力的进展应当放缓。Getty Images

Google DeepMind CEO Demis Hassabis 此前曾支持一项协调一致的行动：如果风险升级，应放缓前沿 AI 的发展。今年 7 月，他呼吁建立由美国主导的全球 AI 监管机构，对最强大的模型进行审查。

这些放缓呼吁，是围绕 AI 安全以及护栏机制应在多大程度上控制开发速度这一时常自相矛盾的事件脉络中的最新一环。尤其是 Anthropic，一方面数月来不断加剧人们对潜在 AI 末日情景的担忧，另一方面又筹集了大量资本，以在竞争中领先于对手。

特朗普政府于 8 月最终确定了一项用于评估新 AI 模型的自愿框架，但该计划的具体细节尚未公开发布。

Famed investor Michael Burry slammed leaders of OpenAI, Anthropic and other tech giants for orchestrating a “self-serving” call to slow the pace of AI development – joining tech bigwigs, President Trump and even the Chinese government in panning so-called AI fear-mongering.

Burry – the short-seller lionized in Hollywood blockbuster “The Big Short” for calling and profiting from the subprime mortgage crisis circa 2008 – wrote on X: “Let’s all take a moment to understand how self-serving it is for OpenAI, Anthropic and other execs of big hyperscalers to talk of slowing things down.”

Burry said AI juggernauts’s calls to hit the brakes were “cover for real uncontrollable slowing growth” as AI labs struggle to maintain the stratospheric growth they experienced earlier in the AI boom. A cooling period would help “as IPOs look to be pushed out,” Burry noted.

Michael Burry has joined a chorus of tech bigwigs including David Sacks and Bill Gurley panning the so-called AI fear-mongering sweeping Silicon Valley. Bloomberg via Getty Images

He also panned the negative drumbeat coming from AI leaders as regulatory capture – where the leaders in a space use regulations and slowdowns to thwart up and coming competition. He framed the calls for a slowdown as a marketing ploy ahead of massive IPOs for the AI giants.

“IPOs need hype & puffery; ‘we are so awesome it could become dangerous’ is hype & puffery,” Burry wrote. Burry previously warned that the current rush to pour billions of dollars into AI resembles the dot-com bubble of the early 2000s.

OpenAI and Anthropic both submitted IPO filings with the Securities and Exchange Commission earlier this summer. Altman said OpenAI would delay its IPO, telling Fortune on Friday that a public listing amid the current AI concerns would come at an “ill-advised moment.”

Burry in his post argued that large language models are not artificial general intelligence and thus there is “nothing AI to slow down,” adding “LLMs are not AI and won’t be AGI.” AI fears have largely focused on AGI, a state of technological progress in which machines will posses human-level intellectual capacity and could theoretically cut humans out of the loop in consequential decisions or pose a risk to humanity because they could outsmart humans.

Burry’s statements come after Anthropic CEO Dario Amodei, OpenAI’s Sam Altman, Google DeepMind chief Demis Hassabis and Elon Musk have called for greater government oversight of AI and for the development of the new technology to happen slower. The AI leaders posted essays and social media posts on the subject in recent days, culminating over the weekend, and have hammered AI stocks to start the week.

The missives have caused an uproar by leading voices in the AI space.

OpenAI Chief Executive Officer Sam Altman said he agrees with the need for a slowdown in AI development. REUTERS

“Stop pretending the motivation to slow down is purely altruistic,” said David Sacks, a technology adviser for President Trump and the White House’s former A.I. and crypto czar, in a Saturday social media post.

Top venture capitalist Bill Gurley and frequent critic of top AI firms, told the New York Times: “If we’re going to regulate, it can’t be anyone they know, and they can’t write the rules.”

Even Vice President JD Vance weighed in: “I have to say, just personally, I feel a little bit weird about the fact that you have so many frontier AI tech companies kind of coming to the government and begging the government to regulate them. It feels a little bit to me like a bit of a trojan horse.”

In an essay on his personal website on Saturday, Anthropic CEO Dario Amodei said that the progress of improving AI model capabilities should be slowed down. OpenAI CEO Sam Altman responded to Amodei’s essay on X. “I agree with Dario that we need to pace the frontier.

In his own X post, Elon Musk, founder of SpaceXAI (previously xAI), wrote, “Dario is right.”

Microsoft CEO Satya Nadella has also endorsed a slowdown, saying “the research, focus, and deliberate pacing needed to get alignment right as the design goal.”

Anthropic CEO Dario Amodei said that the progress of improving AI model capabilities should be slowed down. Getty Images

Google DeepMind CEO Demis Hassabis has previously backed a coordinated effort to slow frontier AI development if risks escalate. In July, he called for a US-led global AI watchdog to review the most powerful models.

The calls for a slowdown are the latest in the at times contradictory saga about AI safety and to what extent guardrails should control the pace of development. Anthropic, in particular, has spent months stoking fears about potential AI doomsday scenarios while at the same time raising torrents of capital to race ahead of competitors.

The Trump administration in August finalized a voluntary framework for evaluating new AI models, though details of the plan have yet to be publicly released.