What Sam Altman’s surprise sacking means for the AI race

user avatar Author: Editors Desk Source: The Economist
November 18, 2023 at 07:29
It is a big setback for OpenAI, and could slow the industry as a whole
Image: Getty Images
Image: Getty Images

How quickly the mighty fall. Ever since the release of Chatgpt a year ago, Sam Altman has been the human face of the generative artificial-intelligence revolution. As recently as November 16th the co-founder and boss of Openai was touting the virtues of ai to executives and world leaders at the Asia-Pacific Economic Co-operation summit in San Francisco. The very next day he was out on his ear. A blog post on Openai’s website said the board “no longer has confidence” in Mr Altman’s leadership because “he was not consistently candid in his communications with the board”. Another shock came hours later. Greg Brockman, chairman of the firm’s board and another co-founder, resigned in response to Mr Altman’s sacking.

The defenestration was all the more surprising because Mr Altman seemed at the peak of his powers. He had recently completed a world tour where everyone from Narendra Modi to Emmanuel Macron jockeyed to meet him. On November 6th he had launched a suite of new AI tools at Openai’s developer day, drawing comparisons with Steve Jobs—a parallel that now seems ironic, considering that in 1985 Jobs too was booted out of the company he had founded. One startup boss says Mr Altman’s and Mr Brockman’s departures are as serious as if Larry Page and Sergey Brin had been kicked out of Google during its early years.

Openai’s employees and investors were blindsided by the move. Mr Brockman later tweeted that he and Mr Altman had not been aware of what was happening until minutes before the ousting. According to Axios, a news website, Microsoft, which has a 49% stake in the firm, was also in the dark until the last minute. Microsoft’s stock fell by 2% on the news, probably because the firm’s ai ambitions, including a hotly anticipated “copilot” for its Office suite, hinge on access to Openai’s technology.

The ousting raises three big issues: what led to the surprise sacking; what it means for the firm that has been at the frontier of generative ai; and what it means for the future of the technology itself. How Openai’s employees, the tech world and society writ large respond to the situation will be critical to what happens next.

The board has yet to offer a detailed explanation for its decision. The leading theory, put forward by Kara Swisher of New York Magazine, says that the rest of the board, assembled by chief scientist Ilya Sutskever, disagreed with Mr Altman on how the firm should balance making money with the safe release of its models. An employee at Openai flatly calls the situation a “coup d’état”. In a meeting with employees held shortly after the announcement, though, Mr Sutskever denied this characterisation, saying that the board was just doing its duty.

If it is true that Mr Altman’s defenestration resulted from disagreements over ai safety, it would be the most dramatic expression yet of a longstanding debate at the heart of Openai’s history and indeed the wider industry. Openai was founded as a non-profit in 2015 by Mr Altman, Mr Brockman and Mr Sutskever, a superstar AI researcher, among others. But in 2019, in need of cash to train models that were demanding ever more computer power, Mr Altman spearheaded the creation of a “capped-profit” company inside the non-profit and raised $1bn from Microsoft. In 2021 a handful of senior employees at the firm grew disillusioned with the firm’s more commercial focus and left to form a rival startup called Anthropic. The launch of Chatgpt was only one chapter in Mr Altman’s quest to turn what was once a tiny research lab into a nimble product-oriented company.

What does Mr Altman’s ousting mean for Openai? The immediate effect is chaos. In addition to Mr Brockman, three other senior engineers have left. Others could follow, especially if the ousting resulted purely from a strategic disagreement. There could also be financial repercussions. Speaking to your correspondent in August, an investor in Openai called Mr Altman the “only irreplaceable person” at the startup because of his top-notch recruiting and fundraising abilities. “Sam is the greatest fundraiser of all time…after Elon,” he said. But perhaps Openai no longer needs Mr Altman to hire staff and raise money, now that the firm is so well known, and has the backing of Microsoft. That is why the departure of Mr Brockman, widely considered the engineering brains of the startup, is even more stinging.

Openai is currently in talks to raise funds at a valuation of nearly $90bn, which would make it one of the most valuable private tech companies in the world. A private-markets broker says that before the announcement there had been “nothing but demand” for Openai’s shares. That valuation will now be tested, he says.

The impact on the wider industry is less clear. Mr Altman pushed Openai to “ship” new products into the world which gave it a first-mover advantage. Its competitors were forced to move faster to keep pace. A more safety-focused OpenAI will therefore slow down the whole industry, allowing competitors to catch up. AI startups that were building products with OpenAI’s technology may now think twice before tying themselves too closely to one company. And Mr Altman himself is a wildcard. Writing on X he promised he would “have more to say about what’s next later”. If what’s next is a new company, the drama could just be getting started.

Keywords
You did not use the site, Click here to remain logged. Timeout: 60 second