Unveiling Altman's Other Side: Lies, Power Plays, and a Trillion-Dollar Gamble

When a company mastering civilization-altering technology rushes toward a trillion-dollar valuation, the market cares about growth and computing power, but the central issue surrounding Sam Altman is just one: trust. A survey spanning interviews with hundreds of people raises the question—Is this key figure of the AI era worthy of entrusting the future?

According to an in-depth report published by The New Yorker on April 6, doubts about Altman do not come from external sources, but rather from his closest coworkers and board members. The first accusation in an internal memo is “lying.” In their eyes, Altman is accustomed to telling different versions of reality to different people, and this man is steering the direction of AI.

The five-day power storm in 2023, in which Altman was fired and promptly reinstated, did not clarify the issues but instead deepened doubts: When capital, employees, and technology all rally behind him, has a CEO already gone beyond any real constraints?

Meanwhile, OpenAI is venturing deep into military, Middle Eastern capital, and sensitive national infrastructure fields, while internal commitments to safety have been weakened, and “safety” has even disappeared from the core business. The idealist shell is being reshaped by practical interests.

The question thus escalates—this is no longer just about personal integrity, but structural risk: When the keys to AGI are concentrated in the hands of a few, and their credibility is in doubt, does humanity still hold the power of a “brake”?

Altman once said AI needs a touch of “fictional” magic. But the real uncertainty is: As the boundaries between reality and narrative start to blur, are we facing a world-changing genius, or an immense story that cannot be verified?

“Lies”—Secret Memo from Former Colleagues

In fall 2023, OpenAI Chief Scientist Ilya Sutskever secretly relayed about 70 pages of internal documents—Slack chat logs, HR files, explanatory texts—to three other board members. Some content was preserved via cellphone photos, apparently to evade company device monitoring.

These internal documents had never been fully disclosed before. One memo dedicated to Altman opened with a list, the heading “Sam consistently exhibits…,” the first item: “Lying.”

Sutskever’s accumulating concerns were rooted in his thinking. He was once close friends with Altman, in 2019 even hosting OpenAI co-founder Greg Brockman’s wedding at the office.

But as he increasingly believed the company was approaching its ultimate goal—creating AI comparable or superior to human cognitive abilities—his doubts about Altman grew. At the time, he told another board member:

“I don’t think Sam is the one who should hold the button.”

AI policy expert Helen Toner and entrepreneur Tasha McCauley received the files and saw them as confirmation: both women had already formed similar judgments. “He’s been entrusted with humanity’s future,” a board member told The New Yorker, “but he’s not trustworthy.”

From Y Combinator to OpenAI: A Controversial Rise

Altman did not appear out of nowhere. His professional patterns began long before OpenAI.

After a year and a half at Stanford, he joined the famous startup incubator Y Combinator (Y.C.), founded by Paul Graham.

His first company, Loopt, a location-based social network, reflected a long-standing trait: interpreting vague rules in ways advantageous to himself.

During Loopt, employees were already puzzled by his exaggerations. One ex-employee recalls Altman bragging about being a ping-pong champion, “Missouri High School Ping-Pong Champion or something,” but he was actually one of the worst players in the company.

A senior employee, Mark Jacobstein, assigned as Altman’s “guardian” by investors, later described this habit as:

“The boundary between ‘I think I can do it’ and ‘I have already done it’ is blurred, and at its most extreme, it veers toward a Theranos-like scam (Silicon Valley’s famous fraud).”

In 2014, Graham promoted Altman to Y.C. President before retiring. By 2018, several Y.C. partners were deeply dissatisfied, and, according to those present, Graham and his wife had a candid conversation with Altman, effectively forcing him to leave.

Graham privately told Y.C. colleagues, before Altman was required to leave, “Sam kept deceiving us.”

Altman insists he was never fired from Y.C., and his departure was entirely voluntary. Graham’s public statements are more vague; he told The New Yorker: “We have no legal power to fire anyone. All we can do is apply moral pressure.”

OpenAI’s Founding: Using Doomsday Narratives to Rally Idealism

In 2015, Altman emailed Elon Musk, laying the foundation for OpenAI’s creation.

Previously, Altman was consistently a technology optimist, but his AI discourse quickly turned doomsday. He repeatedly warned, “AI technologies shouldn’t be dominated by profit-driven corporations.” “If this is going to happen,” he wrote to Musk, “it’s best done by someone other than Google.”

Thus, OpenAI was founded as a nonprofit, with a board legally obligated to put humanity’s safety above corporate interest—even above company survival.

Founders publicly declared that AI might be the most dangerous invention in human history, and thus required an extraordinary corporate governance structure. This narrative was persuasive, successfully recruiting top researchers from Google and academia to a cash-strapped organization.

Sutskever recalls the recruiting slogan: “You are going to save the world.”

However, internal conflicts simmered from the start. According to over 200 pages of private notes collected by Dario Amodei (who later founded Anthropic)—never publicly released—Altman was sending different signals early on.

During the 2019 negotiations for a $1 billion Microsoft investment, Amodei made a company charter clause on “merging and assisting” (if another company safely develops AGI first, OpenAI will assist not compete) a negotiating bottom line, and received Altman’s verbal assurance.

But as the deal neared signing, Amodei found a clause quietly added to allow Microsoft to block any OpenAI merger. He confronted Altman, who initially denied its existence; Amodei had to read the contract aloud, with another colleague present to confirm.

Amodei later wrote in his notes: “His response was almost certainly a lie.”

“Coup” and Restoration: A Five-day Drama

One weekend in November 2023, while the board read out the statement over videocall informing Altman he was no longer an OpenAI employee, he was in Las Vegas watching F1 racing.

The news shook the tech world. Microsoft, OpenAI’s strategic partner having invested over $13 billion, was informed only minutes before the firing. “I was very shocked,” Microsoft CEO Satya Nadella said.

LinkedIn co-founder and OpenAI investor Reid Hoffman launched calls, trying to find out what Altman had done wrong. “I had no idea what the hell happened,” Hoffman told The New Yorker. “We were searching for misuse of funds, or sexual harassment, but found nothing.”

Altman returned to his $27 million San Francisco mansion, quickly forming an informal “exile government,” including Airbnb co-founder Brian Chesky, investor Ron Conway and crisis PR expert Chris Lehane. His phone records show he spent over 12 hours a day in calls.

Meanwhile, he coordinated with Nadella and signaled to acting CEO Mira Murati: his allies were “going all out,” and “digging up negative information” to damage the reputations of those who opposed him (Altman says he doesn’t recall such a conversation).

Under intense external pressure, most employees signed a public letter demanding Altman’s reinstatement; Sutskever eventually signed too.

The board was forced to compromise; Helen Toner and Tasha McCauley lost their seats, while the two new board members (former Harvard president Lawrence Summers and former Facebook CTO Bret Taylor) were both chosen after deep communication with Altman.

“Could you do this: make Bret, Larry Summers, Adam the board, have me as CEO, and Bret lead the investigation?” Altman texted Nadella. In less than five days, he was back.

Employees refer to this episode as “the Blip” (short disappearance), borrowing from Marvel films: characters vanish and return unscathed, but the world is upheaved.

The Investigator’s Report That Was Never Written

The dismissed board members made independent investigation a condition of their departure. OpenAI hired WilmerHale, which led Enron and WorldCom investigations, for this “review.” But according to six insiders, the process seemed designed to limit transparency.

Investigators initially didn’t contact several key figures; some employees feared lack of anonymity and didn’t dare to speak frankly. “They only focused on the board turmoil’s narrow time window, not his long-term integrity,” one employee said.

In March 2024, OpenAI announced Altman was cleared but released no report. Only an 800-word statement was posted on OpenAI’s website, admitting “trust was broken.”

Insiders reveal there was no written report because none was ever drafted; findings were presented orally to Summers and Taylor. “The review’s conclusion was not that Sam was honest,” someone said, “but that he could continue as CEO.”

Several current and former employees at the scene felt shocked by this lack of transparency. Shortly after his reinstatement, Altman rejoined the board.

From “Human Safety First” Nonprofit Promise to Profitization and Recapitalization

OpenAI was originally set up to restrain commercial impulses with a nonprofit structure, the board obliged to put “human safety ahead of company success, even survival.” Internal records show this promise was always being eroded.

In a December 2022 board meeting, Altman assured members several GPT-4 features were approved by the safety review committee; Toner requested documents and found controversial features—allowing “fine-tuning” and deployment as personal assistants—were not approved.

Additionally, she discovered Microsoft had launched early ChatGPT versions in India, bypassing safety checks, a fact Altman never mentioned in lengthy board presentations.

When safety researcher Jan Leike was appointed head of the “Superalignment team,” the official statement promised “20%” of OpenAI’s compute. But four eyewitnesses say the actual allocation was 1%-2%, and “most Superalignment compute used the oldest, worst chip clusters.”

Leike complained to Murati; she told him not to bring it up again, saying “that promise was never realistic.”

In 2024, Sutskever and Leike both left. Leike wrote on social media: “Safety culture and processes have yielded to shiny products.” OpenAI dissolved the Superalignment team, and in its latest IRS filings, the word “safety” disappeared from descriptions of its “most important business activities.”

Trillion-dollar Gamble: Betting the Future of AI on Authoritarian Powers

Altman’s most ambitious business gamble—and the most worrying to outsiders—is extending AI infrastructure to authoritarian regimes in the Middle East.

As early as 2016, he befriended Saudi Crown Prince Mohammed bin Salman, privately calling him “friend.” In 2018, a week after journalist Jamal Khashoggi was murdered, Altman announced joining the Saudi-led Neom city advisory board, but later withdrew under colleague pressure.

But, a policy adviser recalls, Altman began asking if there was a way to circumvent moral pressure to keep receiving Saudi funding: “The question wasn’t ‘Is this bad,’ but ‘If we do this, what’s the consequence? Export control issue? Sanctions? Can I not be found out?’”

He then turned to the UAE. The president’s brother, Sheikh Tahnoon bin Zayed al-Nahyan, oversees $1.5 trillion in sovereign wealth and runs G42, the state AI group.

Altman formed a deep personal relationship with him, calling him “dear friend” on social media, and in 2024 was invited aboard his $250 million super yacht “Maryah.”

In 2023, Altman quietly pushed a plan dubbed “ChipCo”: Gulf states would invest billions to build large chip foundries and data centers in the Middle East. “To my understanding, the board was totally unaware when this happened,” a board member said.

The Biden administration was wary; Commerce officials told Altman clearly: “We don’t intend to build advanced chips in the UAE.”

However, when Trump returned to the White House, the situation changed completely. On the first week, Altman stood in the White House Roosevelt Room and announced the “Stargate” plan—a $500 billion joint venture to build massive AI infrastructure in the US.

Soon after, Biden’s AI export restrictions were lifted, Altman accompanied Trump to visit the Saudi Royal Family, and Saudi announced the creation of a national AI company. Meanwhile, OpenAI announced a data center in Abu Dhabi seven times the size of Central Park, with power usage comparable to Miami.

“We are building portals, through which we are really summoning the aliens,” a former OpenAI executive said. “Sam added another portal in the Middle East. This is one of the most irresponsible things ever.”

Regulation and Lobbying: From “Welcoming Regulation” to Strong Opposition to Safety Laws

For public image, Altman long posed as an outlier among tech giants—he actively called for government regulation of AI, and in 2023 testified to the US Senate, even proposing a new federal regulatory body.

Senator John Kennedy of Louisiana seemed moved, leaning on his hand, saying maybe Altman himself should enforce the rules.

But, as Time magazine reports, OpenAI in 2022-2023 successfully lobbied to weaken EU legislation to regulate large AI companies. In California, when an AI safety bill—very similar to Altman’s Senate testimony—was introduced, OpenAI publicly opposed it and privately issued threats.

“This year, we’ve seen increasingly cunning and deceptive tactics from OpenAI,” a legislative aide said to The New Yorker. Investor Conway lobbied leaders including Nancy Pelosi and Gavin Newsom, and the bill was vetoed after passing.

This year, a pro-AI super PAC called “Leading the Future” began funding candidates opposed to AI regulation.

OpenAI officially claims it won’t fund such groups, but one main donor, Greg Brockman, pledged $50 million. That year, Brockman and his wife donated $25 million to pro-Trump “MAGA Inc.”

Anatomy of Power: A Contradiction Between Pleasing Others and Manipulating Them

Among dozens of interviewees for The New Yorker, most shared a common psychological portrait of Altman. One board member said:

“He is not constrained by truth. He has two traits that almost never coexist in one person: First, a strong desire to please others, to be liked in every interaction; second, almost antisocial indifference to the consequences of deceiving people.”

This board member was not the only one spontaneously using the phrase “antisocial.” Altman’s peer, Aaron Swartz—a legend in tech circles, before his death—had expressed concerns to friends about Altman, in very strong terms.

Several senior Microsoft executives state that, though Nadella has long supported Altman, Microsoft’s relationship with him has grown tense. One says:

“He misleads, misrepresents, renegotiates, goes back on his word. I think there’s a real chance—not large, but real—that he’ll ultimately be remembered as a Bernie Madoff or Sam Bankman-Fried-level fraudster.”

Altman and his supporters’ explanation for these accusations is his behavior is simply “normal competitive business practices,” while critics are sci-fi “doomsayers” or jealous failures.

“He is too immersed in self-belief,” former board member Sue Yoon says, “so he sometimes does things that make no sense in the real world. But he doesn’t live in the real world.”

The Trillion-dollar Question: Should He Rule It All

OpenAI is reportedly preparing for an IPO, with a possible valuation near $1 trillion. Altman steers one of the fastest-growing startups ever, with fundraising—its latest round over $120 billion—four times the largest private fundraising in history.

For the foreseeable future, this company’s decisions will affect hundreds of millions: when AI should stop or go forward; whom it should help or deny service; whether it can make decisions autonomously in battle, without human oversight.

The issue isn’t just that Altman is morally flawed—any big company operator knows this isn’t unusual.

The issue is: OpenAI’s rationale rests on one premise—it differs from other companies, its leaders must have “exceptional integrity,” because they control a civilization-changing technology.

Sutskever originally believed deeply in this mission; it’s why he refused Google’s multimillion-dollar salary.

In an interview, a reporter asked Altman if running an AI company required “higher standards of integrity.”

It ought to be an easy question. He hesitated, then said, “I think a lot of industries can have a huge impact on society, for better or worse.” After a moment, he added, “Yes, it requires higher integrity, and I feel the weight of that responsibility every day.”

However, in OpenAI’s latest IRS filings, “safety” has disappeared as a core business description. The original recruiting slogan—“You are going to save the world”—has been replaced internally with: “Feel the AGI.”

As Altman once told the media, almost philosophically, he allowed his AI models to retain a certain “fictional” ability, not pursuing 100% accuracy, because “that kind of magic is what people like most.” That, perhaps, best illustrates who he is. ▪

Risk warning and disclaimerThe market entails risks, investments should be made cautiously. This article does not constitute personal investment advice and has not taken into account the specific investment objectives, financial conditions, or needs of individual users. Users should consider whether any opinions, viewpoints, or conclusions in this article fit their particular circumstances. Investing accordingly is at your own risk.