Written by: Techub News Compilation

Recently, The New Yorker journalist Ronan Farrow and Andrew Marantz published an in-depth investigative report on OpenAI and its CEO Sam Altman. Based on interviews with over a hundred informed sources and a large number of internal documents, this lengthy article analyzes how Altman transitioned from a "safety prophet" warning about the existential risks of AI to a leader now dominating an unrestrained commercial race, as well as the profound trust crisis this shift has triggered. This dialogue not only concerns the fate of a company but also touches upon our core inquiries about power, responsibility, and truth in the age of AI.

From the "Manhattan Project" to a Commercial Empire: Altman's Narrative Shift

The story begins with an email from 2015. At that time, Sam Altman wrote to Elon Musk: "I think we need a 'Manhattan Project' for AI." This analogy was not accidental; it ran through the entire narrative logic of early OpenAI. Just as the U.S. initiated the atomic bomb project during World War II to defeat the Nazis, Altman positioned OpenAI as the "good guys" alliance against the "bad guys" (originally Google, later also including China), with the mission to ensure the development of powerful and safe AI.

This narrative has a duality: both noble and unsettling. It draws on the image of Robert Oppenheimer—the father of the atomic bomb, who was also deeply concerned about its destructive power. Altman placed himself at a similar moral crossroads, promoting the idea to top AI researchers: "I am just as scared as you are." In a context where AI safety concerns were pervasive, this became a highly persuasive recruitment tool. Many scientists, such as Ilya Sutskever, who later became OpenAI's chief scientist, rejected lucrative offers from companies like Google because they believed in this "safety-first" nonprofit mission and chose to join OpenAI.

However, the investigation reveals a fundamental contradiction between this fear-based narrative and the more aggressive commercial agenda that Altman simultaneously pursued. Reporter Andrew Marantz pointed out that Altman's earlier blog posts were rife with warnings about the "alignment problem" (AI goals misaligned with human interests) potentially leading to human extinction. But over time, his public statements shifted significantly. Nowadays, he speaks more about how AI will cure cancer, bring economic prosperity, and even fulfill a science fiction vision of "capturing all economic value within the solar system." The alignment problem has been redefined to trivial matters like "the Instagram algorithm wasting your time."

"It's a bit like either you were lying then, or you are lying now," Marantz commented. This dramatic shift in narrative has laid the groundwork for later internal conflicts.

"Inconsistent Candor": Internal Memos and Trust Collapse

At the end of 2023, the OpenAI board suddenly dismissed Sam Altman, citing that his communication with the board was "not consistently candid." Although Altman quickly returned to his position with the support of employees and investors, the deeper reasons for this upheaval have remained shrouded in mystery. The New Yorker's investigation unveiled details of internal memos that led the board to make this decision.

Reporter Ronan Farrow emphasized that there was no single "smoking gun" or "concrete evidence." Instead, it was the accumulation of numerous subtle behavioral patterns that led core members, including Ilya Sutskever, to conclude: "Sam is not the person who should have his finger on the button." (A nod to Oppenheimer's metaphor)

These behavioral patterns include:

• Repetitions on significant safety issues: Insiders revealed that OpenAI executives once "considered" selling (or offering) future AGI (Artificial General Intelligence) technology to the highest international bidder, even imagining a bidding war between China, Russia, and the United States. This idea shocked the safety faction within the company, who found it "utterly insane." Although OpenAI responded that this was just one of many considered ideas and was not implemented, it exposed the decision-makers' moral ambiguity.
• Two-faced statements towards partners: When the competitor Anthropic was blacklisted for rejecting the U.S. Department of Defense's unlimited terms for "mass surveillance and autonomous weapons development," Altman publicly supported Anthropic's stance. However, privately, he was negotiating with the Pentagon, trying to secure the same contract. This discrepancy between public statements and private actions repeatedly appeared in business relationships, such as a recent deal with Amazon being questioned by Microsoft for violating Microsoft's exclusive agreement.
• Exaggerations or fabrications about trivial matters: Several former employees recalled that Altman once claimed to be a championship-level table tennis player, but his actual skill was among the worst in the office. Such seemingly inconsequential examples repeatedly emerged from many respondents, gradually building an impression that even in unnecessary matters, he tended to present a version of the truth that diverged from reality.

Farrow, during an interview with Altman, directly asked him how he faced the widespread doubts about his integrity. Altman admitted to having a "people-pleasing" issue and mentioned that the root might be related to severe bullying he experienced during his teenage years due to his sexual orientation. He candidly expressed reluctance to discuss this matter in depth out of fear of being seen as "seeking sympathy and manipulating others." However, to reporters, Altman's reflection on these criticisms seemed "superficial," lacking in-depth self-examination.

Political Vacillation and Geopolitical Stakes

Sam Altman's political stance also reflects his adaptability and complexity. He was once a donor to the Democratic Party and likened Trump to Hitler. However, by 2024, his tone softened noticeably, claiming that "whatever happens, this country will be fine." According to national security officials from the Biden administration, Altman frequently visited the White House, encouraging the government to strengthen regulations, even believing that the executive orders were not stringent enough. But after Trump's administration took office, OpenAI quickly announced a large-scale new data infrastructure project, while Trump’s rhetoric was that "security concerns are fabricated."

The larger geopolitical stakes are reflected in OpenAI's financing strategy. The company is building a massive data center in the UAE, comparable in size to seven Central Parks and consuming as much electricity as the entire city of Miami. Despite the norm of obtaining funding from the Middle East extending from Silicon Valley to Hollywood, critics argue that placing such critical computational infrastructure, potentially capable of "reshaping the balance of power among nations," under the control of a non-democratic regime poses significant national security risks. Altman's response is that computing power is "the new currency of the world," and his task is to acquire resources as much as possible.

"You could eventually find yourself in a situation where a dictator disproportionately possesses the most powerful technology on earth, which could constitute a real national security threat," Marantz summarized, "but the money is there."

The "Shakespearean" Competition of the AI Era and Personality Puzzle

The competition in the entire AI field has been described by insiders as "Shakespearean"—full of drama, personal vendettas, and ruthless tactics. During the investigation, reporters received extensive information from government officials, investment firms, and competitors, which was filled with "shocking allegations" regarding Altman's personal life, such as involving minors. After months of verification, reporters stated they found no evidence to support these allegations and noted that many of these rumors were actively spread by intermediaries of Elon Musk (some of whom were paid). Ironically, these false, attention-grabbing personal attacks obscured more substantial criticisms regarding business practices and safety commitments.

So, what does Sam Altman want? To become the world's richest man, or to seize power? He himself has stated that his interest in power is greater than in money. Farrow presented a more complex portrait: "What Sam Altman often wants is what you want at that moment." He has an extraordinary ability to sense the needs and desires of those he's conversing with, making them believe that those are also his sincerely held goals. This extreme "conflict avoidance" trait and tendency to please others, combined with his "seemingly self-assured" confidence, create an extremely efficient salesperson and gatherer. However, when this trait is applied to the development of technologies regarding the future of humanity, the consequences exceed personal character considerations.

"I don't think Sam Altman is a villain," Farrow concluded, "he's a complex character. He generally believes what he's saying at that moment... I think he is grappling with these consequences, and the industry also needs to address these consequences." Andrew Marantz added that the memos that led to Altman's dismissal are not simply undeniable evidence of wrongdoing, but a behavioral pattern that requires a 16,000-word deep report to elucidate.

The story of OpenAI is far from over. It has shifted from a nonprofit research laboratory to a profit-driven commercial giant (despite still being labeled a "public benefit corporation"), with its initial mission oversight body now only holding a minority stake. The "coup" attempt in 2023 marks a critical moment: testing whether the "high standard" of safety still has a place in the immensely economically charged pursuit of AGI. Altman's victorious return, along with the subsequent shift in industry atmosphere, seems to provide an answer. Nowadays, a reasonable argument is that these AI companies are falling into a "race to the bottom" regarding safety issues. And this all began with a person who called himself "Oppenheimer" and his ever-evolving story.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。