Explaining the Upheaval at OpenAI

Hello! We’re back with another bonus edition of On Tech: A.I., a pop-up newsletter that teaches you about artificial intelligence, how it works and how to use it.

The artificial intelligence landscape will never be the same after the extraordinary upheaval at OpenAI, the start-up that set off a technology arms race by releasing ChatGPT nearly one year ago.

The OpenAI board ousted Sam Altman as chief executive on Friday, shocking employees and investors. His exit set off a series of head-spinning developments, as the board briefly considered and then rejected a proposal to bring him back.

Microsoft, the company’s biggest investor, announced on Sunday that it would hire Altman and his co-founder, Greg Brockman, to run a new research lab — an apparent rupture in the tight relationship between OpenAI and the tech giant, which invested $13 billion in the start-up. The majority of OpenAI employees have `threatened to jump ship to Microsoft.

The weekend’s turmoil also highlighted an unresolved debate at OpenAI and in the larger tech community: Is artificial intelligence the most important new technology since web browsers, or is it potentially dangerous to humanity — or both?

Today, with help from Cade Metz, Kevin Roose and their colleagues on the Times tech team, we’ll bring you up to speed on where this fast-moving story stands, and on where it might go. Warning: There may be more plot twists.

What just happened?

• On Friday, Altman was abruptly dismissed as OpenAI’s chief executive for reasons that are still not clear. Some tech observers compared the shock to when Steve Jobs was forced out of Apple in 1985.

• “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,” OpenAI’s board said in a memo.

• Mira Murati, the company’s chief technology officer, was named interim chief executive.

• Greg Brockman, another co-founder, was stripped of his chairmanship and quit.

• Investors in OpenAI — who have little power because of the company’s quirky corporate governance structure (more on that below) — began plotting a way for Altman to return.

• Talks to bring Altman back faltered, and OpenAI’s board named its second interim chief in two days. Emmett Shear, the former chief executive of the streaming service Twitch, replaced Murati.

• Hours later, Microsoft said that it would hire Altman and Brockman to lead an advanced research lab at the tech giant. Altman wrote on the X platform, formerly Twitter, that “the mission continues.”

• By Monday morning, almost all of OpenAI’s nearly 800 employees had signed a letter saying they might quit to join Altman’s new project at Microsoft unless the start-up’s board resigned, three people who viewed the letter told Cade.

What really happened?

Ilya Sutskever, OpenAI’s chief scientist who is also a co-founder and board member, was increasingly worried that OpenAI’s technology could be dangerous and that Altman was not paying enough attention to that risk, three people familiar with his thinking told Cade.

Kevin wrote that the board “was worried that Altman was moving too fast to build powerful, potentially harmful A.I. systems, and they stopped him.”

In yet another plot twist, Sutskever wrote on X early on Monday morning: “I deeply regret my participation in the board’s actions. I never intended to harm OpenAI. I love everything we’ve built together and I will do everything I can to reunite the company.”

In short, we still don’t know exactly what went down this weekend or the ultimate outcome of all the turmoil.

OpenAI’s ‘messy’ history

Altman, Brockman and Sutskever created OpenAI in 2015 alongside nine others, including Elon Musk. The group founded the A.I. lab as a nonprofit, saying that unlike a traditional tech company — say, Microsoft — it would not be driven by commercial incentives.

In 2018, after Musk parted ways with OpenAI, Altman transformed the lab into a for-profit company controlled by the nonprofit and its board. Over the next several years, he raised the billions of dollars the company needed to build things like ChatGPT.

“OpenAI has just been a messy company always,” said Casey Newton, Kevin’s co-host on the “Hard Fork” podcast. Musk fell out with the company and ended up walking away; he founded an A.I. company called xAI this year. Another group of people who left OpenAI went on to start Anthropic, another competitor.

“In the A.I. world, there are a lot of disputes,” Casey said, “and they often end up with people slamming doors and often going to start their own A.I. companies.”

OpenAI’s unusual corporate structure also appears to have played a role in Altman’s ouster. OpenAI is controlled by the board of a nonprofit that can decide the company’s leadership. Investors like Microsoft have no formal way of influencing decisions, and many of the top leaders, including Altman, do not own any shares in the company.

“That scenario makes this kind of drama more likely,” Casey said.

The effective altruism movement

For years, a community of A.I. researchers and activists — many affiliated with the effective altruism movement, whose adherents think that reason and data can be used to determine how to do the most good — have warned that A.I. systems are becoming too powerful, and that out-of-control A.I. could pose an existential threat to humanity.

People with these fears — sometimes mocked as “doomers” — were once considered fringe. But over the past several years, they’ve been moving toward the mainstream, gathering signatures on open letters and warning regulators to take A.I. safety seriously.

Ilya Sutskever, OpenAI’s chief scientist, who led the coup against Altman, is not an effective altruist, but he appears to have been motivated by similar fears. And two of the board members who supported ousting Altman, Tasha McCauley and Helen Toner, have ties to effective altruist groups.

And if this movement sounds familiar, it may be because of the travails of Sam Bankman-Fried, the disgraced crypto mogul who also supported effective altruism.

What does Microsoft get from this?

Microsoft was said to be particularly alarmed by Altman’s sudden dismissal, and led the failed campaign to have him reinstated. The tech giant, along with other OpenAI investors like Thrive Capital and Sequoia Capital, found out about Altman’s firing a mere minute before the announcement.

Satya Nadella, Microsoft’s chief executive, was reportedly deeply involved in the talks. On Sunday night, he said Microsoft remained “committed” to OpenAI, but stressed that the new unit Altman and Brockman would run within Microsoft would be “setting a new pace for innovation,” in an apparent contrast with the OpenAI board’s desire for caution in developing A.I. technology.

Kevin said that Nadella ended the weekend a winner:

“On Friday, when Altman was fired, it looked like Nadella might lose one of his most powerful allies,” he wrote. “Microsoft invested $13 billion in OpenAI, and under Mr. Altman’s leadership, the company had become a key partner of Microsoft’s. Its technology is the backbone of many of the A.I. services, such as the company’s suite of Copilot A.I. products, that Microsoft is betting the future of its business on.”

Nadella “would have clearly preferred to see Altman reinstated,” Kevin concluded. “But when it was clear that wasn’t happening, he did the next best thing: swooping in to offer jobs to Altman, Brockman and their loyalists.”

Microsoft stock, which plummeted after news of Altman’s firing on Friday, recovered its value on Monday and set a new record high.

Now what?

Casey and Kevin discussed on this weekend’s edition of “Hard Fork” how Altman’s stature in Silicon Valley allowed him to recruit lots of top-flight talent to OpenAI. The flip side: His absence could hamper the company’s fortunes.

“There were a lot of people who went to work because they worked for Sam Altman,” Casey said. “On Monday, they’re going to go in to work for someone else.”

The letter from employees who threatened to join Altman’s new project at Microsoft if the OpenAI board did not resign was, curiously, also signed by Sutskever.

“Before Friday, the company was the hottest name in tech, with a celebrity leader, a household-name product in ChatGPT, and a murderers’ row of A.I. talent that was the envy of Silicon Valley giants,” Kevin wrote.

But now, “the company is in chaos. Its top leaders are gone. Morale is shattered.”

The company also remains highly dependent on Microsoft for its computing power. Starting today, Kevin noted, Microsoft “will have a mini-OpenAI growing inside of it, led by Altman and staffed by former OpenAI employees.”

“OpenAI’s board may be satisfied with this outcome — after all, they chose it, even after being given a chance to backtrack. But they look silly for not explaining why they fired Altman, and until they share more information, it’s hard to imagine the rank-and-file falling in line.”

— Reporting by Cade Metz, Kevin Roose, Mike Isaac, Jason Karaian, John Koblin and Kevin Granville.