Sam Altman says it’s relentless and all-consuming to run OpenAI teams after shock CTO departure

On Wednesday, chief technology officer Mira Murati said she was stepping down from her role, alongside vice president of research Barrett Zoph and chief research officer, Bob McGrew—leading to a leadership changeover Altman characterized as amicable and uncoordinated.

“I obviously won’t pretend it’s natural for this one to be so abrupt, but we are not a normal company,” he told staff in a letter he posted on social media, calling their jobs “relentless” and “all-consuming.” 

The departures come at a historic juncture in OpenAI’s nine-year history. Instead of celebrating the one-year anniversary of its November 2022 ChatGPT launch that sparked an AI gold rush, last November was instead the scene of a failed coup against Altman by the non-profit board that controls the company. 

This led the CEO to admit its hybrid structure was a clumsy compromise to keep the team intact as it shifted focus from pure research to commercializing its intellectual property. 

Now it appears that time has come to complete OpenAi’s metamorphosis, with plans in the works to shed the non-profit shell that still controls it on paper as it seeks to raise fresh equity capital at a reported $150 billion valuation. (A spokesman for OpenAI told Fortune earlier this month the non-profit will continue to exist since it is “core” to its mission, but did not explain what role it would play.)

Altman himself is set to reap a windfall gain in the process, potentially being awarded a 7% stake in the new entity, according to Bloomberg.

‘Nothing without its people’

Companies in fast-paced industries are only as good as their human capital, and Altman initially managed to hold on to staff split between researchers still dedicated to its original non-profit mission and those looking to quickly commercialize the technology.

“OpenAI is nothing without its people,” became the unifying motto of OpenAI at the time, a sentiment Murati herself shared in November 2023. 

That now appears to be aging poorly.

Chief scientist Ilya Sutskever, who regretted his role plunging OpenAI into crisis last year, departed in May to found his own AI startup (it recently raised $1 billion from investors, or $100 million per employee). 

Fellow co-founder John Schulman works at Anthropic as of last month, leaving just three of the original team of eleven still around. One of those, president Greg Brockman, has gone on a sabbatical through the end of this year.  

Recently former researcher Daniel Kokotajlo told Fortune half the company’s AI safety staff have left the company in the past several months, while his former colleague Jan Leike skewered Altman’s leadership after quitting.

Competition heating up in the form of Elon Musk’s xAI

Meanwhile, the competitive landscape is heating up. Elon Musk, still sore that he lost control of OpenAI, is pressing hard to recruit the best talent to his new startup xAI—even if that means poaching them from Tesla. 

Earlier this month he completed the acquisition of 100,000 Nvidia H100 processors for AI training— amassing 10 times the computing power he needed to train Grok 2—and more are on their way.

“Our fundamental competitiveness depends on being faster than any other AI company,” Musk argued in July. “This is the only way to catch up.”

Grok 3 is currently in training, and should Musk’s team at xAI hit his ambitious December release schedule, it could potentially eclipse OpenAI’s GPT-4 omni model, its most advanced yet. 

AI industry critic Gary Marcus has subsequently warned OpenAI investors to rethink whether the company in its current state is worth what they are claiming, with so many top minds leaving.

“People are valuing this company at $150 billion dollars? Absolutely insane,” he posted on Wednesday. “Investors shouldn’t be pouring [in] more money at higher valuations, they should be asking what is going on.”