Everything You Need to Know About the Ridiculously Chaotic ‘Coup,’ Implosion, and Counter-Revolution at OpenAI

Late Friday, OpenAI began to implode in a fashion quite unlike anything in the history of Silicon Valley. The organization, which has become famous around the world for its development of the AI chatbot ChatGPT and text-to-image program DALL-E, suddenly found itself in the middle of a Succession-like drama so chaotic, confounding, and fast-moving that it was hard to keep the story straight. The situation was probably best summarized by The New York Times' Mike Isaac, who described the entire mess as a “fucking shitshow” on Sunday night. 

That all goes to say that if you are having trouble making sense of what just transpired, you are not alone. To help you make sense of what just happened—and why—we have decided to break down exactly how we got here, who is fighting who, and where things will likely go from here.

OK, let’s start with the basics: Who got fired and why?  

It all began with a strange press release on Friday afternoon, released by four members of the artificial intelligence company’s board, stating that they were kicking CEO and co-founder Sam Altman out of the company due to unspecified “communications” issues. 

Altman has been one of the most well-connected people in Silicon Valley for a while now. From 2014 to 2019, he was president of Y Combinator, an influential startup accelerator, during which time he also helped start the initially nonprofit OpenAI, for which he and Elon Musk served as the first co-chairs of the board. In 2019, he became the company’s full-time CEO.

Have information about OpenAI? We want to hear from you. From a non-work device, contact our reporter at maxwell.strachan@vice.com or via Signal at 310-614-3752 for extra security. 

In the year since, Altman has become not only the face of OpenAI, but in many ways of artificial intelligence more broadly. For him to have been booted so suddenly—he had been onstage at OpenAI’s first developers day just this month—led to intense speculation about if he had committed some unknown heinous act. Altman avoided diving into the mess on Friday, only saying on X: “i loved my time at openai. it was transformative for me personally, and hopefully the world a little bit. most of all i loved working with such talented people. will have more to say about what’s next later.” 

But Altman’s own brother, Lattice CEO Jack Altman, was more aggressive. “To all the people gleefully hating today, please know you’re betting against the wrong guy,” he wrote. 

Who did the firing? Maybe that can offer some clues? 

Technically, that would be the four members of the board that are not Altman and OpenAI president Greg Brockman, another co-founder and Altman ally. One of them is Quora CEO Adam D’Angelo, and two others are Tasha McCauley, an entrepreneur who is married to actor Joseph Gordon-Levitt, and Helen Toner, director of strategy at Georgetown Center for Security and Emerging Technology. McCauley and Toner have been linked to the Effective Altruism movement made famous by disgraced former FTX CEO Sam Bankman-Fried, a group that has been known, among other things, for fearing the apocalyptic potential of AI. 

The fourth is OpenAI’s chief scientist and co-founder Ilya Sutskever, who is now the sole OpenAI co-founder left on the board. Highly respected in the AI field, Sutskever has been personally recruited to join OpenAI by people including Musk after a stint at Google’s Brain Lab.  While he was interested in developing Artificial General Intelligence, or AGI, Sutskever also represented the company’s original focus on AI safety—so much so that he reportedly once commissioned a wooden effigy or a rogue AGI to burn at an offsite meeting for company leadership.

Over time, Sutskever became increasingly concerned about Altman, who was cavorting with Middle East sovereign wealth funds and SoftBank’s Masayoshi Son to potentially fund outside projects, according to a lengthy report by Bloomberg this weekend. While Sutskever was focused on safety, some feared Altman was too focused on growth.  Adding to his frustration, Altman tried to “reduce Sutskever’s role at the company” in recent months according to Bloomberg. 

Then, at OpenAI’s Developers Day this month, Altman announced plans to make customizable versions of ChatGPT—a concern for advocates obsessed with AI safety, Sutskever included. Soon after, Sutskever went to the board to reportedly express his frustrations and concerns and “ultimately got the board on his side,” according to longtime tech reporter Kara Swisher. 

As part of the reshuffling, the board announced that president Brockman also would step down from his role as chairman of the board, though he would remain at the company under interim CEO Mira Murati, who had been OpenAI’s chief technology officer. 

On Friday, Brockman began to piece together the series of events as he and Altman had experienced them. According to his account, Sutskever had contacted Altman on Thursday asking to talk the next day at noon. When Altman joined a Google Meet, everyone on the board, minus Brockman, was on the call. Sutskever then informed Altman he was fired. Nineteen minutes later, Sutskever texted Brockman asking to jump on a call, where Sutskever informed he was being kicked off the board but remained “vital to the company.” 

The blog post announcing the changes was published “around the same time.”  

Understandably pissed, Brockman announced Friday to staff (and then on X) that he was quitting. Soon after, news broke that three senior researchers had also resigned.

This all seems pretty weird. Is there anything particularly strange about OpenAI that made something like this more likely to occur? 

The chaos is, at least in part, a result of OpenAI’s odd organizational structure, which is a manifestation of its atypical history and the competing ideological factions within the company. The organization was first formed in 2015 as a nonprofit by a collection of people that included Altman, Brockman, and Sutskever, to reflect its self-stated high-minded goals and desire to not bow to typical corporate desires. 

But in 2019, the nonprofit announced that it was transforming itself into a new type of organizational structure, which it described as “a hybrid of a for-profit and nonprofit.” The idea was to make it easier for OpenAI to raise more capital and, in return, distribute a capped percentage of profits to investors. The idea was to “put our overall mission—ensuring the creation and adoption of safe and beneficial AGI—ahead of generating returns for investors.”

But critically, as OpenAI announced at the time of the reorganization, the small nonprofit board would continue to remain in control, meaning that outside investors could exert no power over major decisions and the board did not have the typical fiduciary duty to maximize shareholder value. More importantly, it meant a few people held major power. 

Despite attempts to create a new type of organization that could raise capital without the typical pressures of capitalism, tensions spilled over at times. In 2020, some OpenAI employees left over concerns the company had veered from its original mission and created the AI startup Anthropic. 

What does Microsoft have to do with all of this? 

OpenAI’s innovative new structure was successful in allowing the organization to take in billions in investment, namely from Microsoft which ultimately invested $13 billion in the company, meaning it owned a 49 percent stake in the company. 

Considering it owned nearly half of the company, you might think Microsoft knew Altman was about to get shit-canned before it happened. But that didn’t really happen. People over there were only informed moments before the release. Because of its close ties and investment, Microsoft’s share price plunged after the news broke, reducing the company’s value by tens of billions of dollars. Whoops. Microsoft released a statement saying it remained “committed” to the partnership, CEO Satya Nadella was reportedly “livid.”

Were employees OK with this? 

No. OpenAI convened for a Friday all-hands meeting after the firing, during which employees asked Sutskever if the action could be considered a “coup” or “hostile takeover.” Sutskever reportedly “dismissed” such ideas, arguing that it was necessary to protect the site’s core mission of making sure artificial intelligence was developed to humanity’s benefit. 

The employees absolutely did not buy it. Saturday morning, those still at OpenAI were left searching for answers to what had happened. COO Brian Lightcap circulated a memo internally, which was obtained by Axios, letting the team know management had been as surprised as anyone else and that they were talking with the board to “try to better understand the reasons and process behind their decision.” Employees weren’t willing to wait to better understand the reasons and process and started to demand they bring Altman and Brockman back.

“Dozens” of people—including some OpenAI employees—joined Altman at his $27 million mansion to try and learn what they could about the latest developments. “i love the openai team so much,” Altman wrote on X. Online, OpenAI employees started to tweet heart emojis over the weekend, which The Verge’s Alex Heath reported were a covert way to signal “who would leave and follow Altman and Brockman to a new company if the board coup succeeds.”

OK, so the employees wanted to bring Altman back. But surely the board had made up its mind? 

You would think that. But nothing about this situation is normal. Investors also wanted Altman back, and by Saturday, The Verge was reporting that Altman was already back in discussions with the board to rejoin OpenAI as CEO, though he was “ambivalent” about doing so. Amazingly, it soon came out that the board had “agreed in principle to resign” and allow Altman and Brockman to come back, but then “waffled.” Nevertheless, chief strategy officer Jason Kwon told employees Saturday night he was “optimistic” a deal would get done.

By Sunday, Altman and Brockman were back at company headquarters on Murati’s invitation, where Altman posted a photo of himself with a guest badge. “first and last time i ever wear one of these,” he wrote. Negotiations continued throughout the day, as delivery men reportedly arrived with Boba and, later, McDonald’s. A noon deadline became a 5 p.m deadline after Sutskever asked for more time, according to Kara Swisher. Microsoft’s Nadella reportedly got “quite involved.” A new board construction was discussed, with names like Marisa Mayer, Brian Chesky, and Sheryl Sandberg floated. By Sunday evening, all signs seemed to indicate that Murati was on track to bring back Altman and Brockman as she negotiated the terms with board member Adam D’Angelo. 

“What a fucking special company,” wrote Jack Altman, Sam’s brother. 

Wow! So did it happen??

No, it did not! Late Sunday, Sutskever himself delivered news to the staff that Altman would not be rejoining. Within hours, dozens of additional OpenAI employees had quit. Instead, the four remaining members of the board announced in an internal memo that the company had hired former Twitch CEO Emmett Shear as its new interim CEO, ending Murati’s few days in charge.

In the memo, the board once again defended its decision without providing much in the way of specific information, saying, “Put simply, Sam’s behavior and lack of transparency in his interactions with the board undermined the board’s ability to effectively supervise the company in the manner it was mandated to do.” 

Shortly thereafter, Shear published a long tweet about his hiring, saying he had reflected on the decision for a “few hours” before accepting the “once-in-a-lifetime opportunity.” 

In his message, he said it was “clear that the process and communications around Altman’s removal has been handled very badly” and said the board “did not remove Sam over any specific disagreement on safety,” though he did not clarify why it did remove him, only saying “their reasoning was completely different from that.” T

To get to the truth of what happened, he planned to hire an independent investigator, among other right-the-ship actions. “Significant governance changes” were on the table “if necessary.”

Why did they hire the Twitch guy…

It seems like Shear represented a desire to develop OpenAI’s products more safely and carefully. Motherboard dug through his public comments on social media about AI and found that he described himself as a "doomer" who is "in favor of slowing down" AI development. A source told Bloomberg more plainly that his belief that AI presented an existential threat was a key reason for his hiring. (It’s worth noting that Altman has been prone to make similarly apocalyptic predictions about the potential power of AI, although his tenure was defined by pushing new AI products to market.)

Why did they get so close to rehiring Altman if they were also preparing to hire the Twitch guy?

It would seem like the reason is that the board itself was struggling with what the correct course of action was moving forward. Again, this board is now just four people. 

Sounds weird, but OK. Did the employees buy this? 

Again, no. OpenAI employees have been tweeting the same phrase over and over: “OpenAI is nothing without its people.” More concretely, on Monday morning, 550 of OpenAI’s 700 employees signed a letter demanding the resignation of the entire board of directors, saying they were all ready to quit and join Microsoft unless they reinstated Altman and Brockman, appointed two lead independent directors, and then quit. 

“Despite many requests for specific facts for your allegations, you have never provided any written evidence,” the letter read. Even more startling, according to the letter, the board told leadership that allowing the company to be “destroyed ‘would be consistent to the mission.’”

“Your actions have made it obvious that you are incapable of overseeing OpenAI,” the signees wrote, adding elsewhere. 

Among them were Murati and, shockingly, Sutskever, who had just days before had reportedly played a key role in firing Altman and demoting Brockman. 

Wait, what? The guy who was furious at Altman and played a large role in his firing is now signing a letter in protest of the firing? 

Not only that, he also posted about what a big mistake he made. 

“I deeply regret my participation in the board's actions,” Sutskever wrote in a post of his own. “I never intended to harm OpenAI. I love everything we've built together and I will do everything I can to reunite the company.” Altman responded with two heart emojis. 

So maybe Altman will come back? 

It seems pretty unlikely, but never say never. Microsoft’s Nadella announced late Sunday that he was bringing on Altman, Brockman, and their colleagues to run a “new advanced AI research team” at Microsoft while carefully noting that the company remained committed to the OpenAI partnership. Observers have noted that Microsoft hiring OpenAI's key leadership and the majority of its talent would be a coup of its own, effectively acquiring all of OpenAI's expertise without actually buying the company or facing potential regulatory scrutiny. 

On Monday afternoon, reports emerged that Altman and Brockman are still thinking of returning to OpenAI and that the Microsoft deal is not finalized. Altman tweeted, "we are all going to work together some way or other, and i’m so excited," and followed up by saying that "satya and my top priority remains to ensure openai continues to thrive."

Of course, given how this story has played out already, who knows? Right now, nothing is a done deal.