For a time this weekend, it looked like Sam Altman might return as a conquering hero to OpenAI, the company whose board had fired him as CEO on Friday.
It would have been another shocking twist in a saga that was already riddled with them. And Altman had a lot of influence. OpenAI employees had supported him since his firing, and OpenAI investors were pressuring the board of directors to bring him back. Billions of dollars (and possibly the trajectory of the entire AI industry) hung on the fate of the board’s decision, and many expected them to bow to pressure and back down.
Instead, the board stood firm, rejecting the return of Mr. Altman and stated in a late-night memo to employees Sunday that removing him was “necessary to preserve the board’s ability to execute its responsibilities and advance the mission of this organization.” He named Emmett Shear, the former head of Twitch, as interim boss.
Hours later, Microsoft CEO Satya Nadella announced that Altman and his top lieutenant, Greg Brockman, would join the tech giant to lead a new AI research division.
The OpenAI saga is far from over. Things are changing quickly and there is a lot we still don’t know, including why the board decided to fire Altman in the first place. (In Sunday’s memo, the board said there had been no specific incident that led to the firing, but that Mr. Altman had simply lost its trust.)
But even without knowing much about the incident that caused it, we can begin to assess the damage.
The most obvious loser in all of this is OpenAI itself.
Before Friday, the company was the hottest name in tech, with a famous leader, a well-known product in ChatGPT, and a killer lineup of AI talent that was the envy of Silicon Valley giants. It was in the middle of a takeover bid that would have allowed employees to cash out their shares at a staggering valuation, and its cutting-edge AI language model, GPT-4, was the best in its class. Now, the company is in chaos. Its top leaders are gone. Morale is destroyed. The takeover bid may fail. The new CEO has said that he wants slow down AI. And the company is still heavily dependent on Microsoft, whose enormous computing power it needs to run its models and which, starting Monday, will have a mini-OpenAI growing within it, led by Altman and made up of former OpenAI employees. .
The OpenAI board of directors can be satisfied with this result; after all, they chose him, even after having the chance to back out. But they seem foolish for not explaining why Mr. Altman was fired, and until they share more information, it’s hard to imagine the bases aligning.
No one’s weekend had a bigger turnaround than Mr. Nadella’s.
On Friday. When Altman was fired, it looked like Nadella might lose one of her most powerful allies. Microsoft invested $13 billion in OpenAI, and under Altman’s leadership, the company had become a key partner of Microsoft. Its technology is the backbone of many AI services, such as the company’s Copilot AI suite of products, on which Microsoft is betting the future of its business.
Mr. Nadella clearly would have preferred that Altman be reinstated. But when it became clear that wasn’t happening, he did the next best thing: he launched into offering jobs to Altman, Brockman, and their loyalists.
Strategically, it was a masterstroke. Now, Microsoft will be able to continue using OpenAI models to power its products in the near term, while also giving a new team led by Altman the money and computing power it needs to build new Microsoft-owned models in the long term. You’ll get a pool of talented AI researchers from OpenAI, and Microsoft now effectively owns 100 percent of a new AI lab that any Silicon Valley venture capitalist would have lined up to fund.
Winners: AI Doomers and Effective Altruists
For years, a community of AI researchers and activists (many of them affiliated with the effective altruism movement, whose followers think reason and data should determine how to do the most good, have warned that AI systems were becoming too powerful and that runaway AI could pose an existential threat to humanity.
People with these fears, sometimes derided by their critics as “doomers” or “slowers,” were once considered fringes. But in recent years, they have been moving closer to the mainstream, gathering signatures for open letters and warn regulators to take AI safety seriously. And on Friday they took down the CEO of the world’s leading artificial intelligence company.
Ilya Sutskever, the chief scientist at OpenAI, who led the coup against Altman, is not an effective altruist, but seems to have been motivated for similar fears. And two of the board members who supported the coup, Tasha McCauley and Helen Toner, have ties to Effective Altruists groups.
If OpenAI ends up being irreparably harmed by Altman’s firing, people will blame the board for ruining one of Silicon Valley’s most promising young companies and destroying billions of dollars in shareholder value.
But the board has clearly succeeded on its own terms. They were concerned that Altman was moving too quickly to build powerful and potentially harmful AI systems, and they stopped him. This is a victory for the cause, even if it comes at the company’s expense.
No one was more supportive of Altman’s return to OpenAI than the investors and venture capitalists who backed him and who risked losing their money if he left.
Many of these investors are techno-optimists who believe that AI will be a pure good for society, and they loved Altman’s essentially optimistic view of the future of AI. (They also loved that it made them a lot of money.)
These investors now have stakes in a company with an interim CEO, a workforce in rebellion, and an unclear path forward. What’s worse, the only way they can invest in Mr. Altman’s new company is by buying Microsoft stock.
Unclear: OpenAI’s rivals
It is not yet clear whether rival AI companies will benefit from Altman’s ouster.
On the one hand, companies like Google, Anthropic, and Meta could benefit from a weakened OpenAI if it allows them to catch up with the company’s AI progress or divert key employees. (Recruiters I didn’t waste time trying to poach disgruntled OpenAI workers on Friday).
But it also means they will compete with a stronger Microsoft. And it means Altman’s new AI efforts won’t be limited by the same intricate nonprofit governance structure that OpenAI was, meaning it could move even faster.