OpenAI fiasco: Experts warn of potential brain drain

Everyone from tech industry insiders to average bystanders marvelled over the past week at how some of the brightest minds in Silicon Valley managed to briefly turn the world’s leading AI company into a laughing stock.

The drama that unfolded at OpenAI nearly a year to the day after the institute first captured people’s imagination with the release of ChatGPT is now over—or at least paused—with co-founder Sam Altman reinstated as CEO.

But experts now believe OpenAI crossed the Rubicon and there is no going back to the way things were, even if Altman’s in-house rivals, among them Ilya Sutskever, wish for nothing more.

A brain drain now threatens to derail OpenAI, warns Georgetown University assistant professor Jason Schloetzer. As the erstwhile non-profit establishes itself first and foremost as a commercial entity under the de-facto control of Microsoft, its talent faces a fateful choice.

“This could be the time you have to pick one side or the other,” says Schloetzer, whose focuses are corporate governance and artificial intelligence. “If you believe that having all of the OpenAI talent together in the same spot has tremendous value, then splitting them up reduces the pace of innovation.”

Each employee will now have to question whether they support “a lot of people getting rich”—as Altman phrased it earlier this year—or OpenAI’s original mission, which foresees ensuring everyone benefits from their work, and not just a few.

In a worst-case scenario, some of their best staff could follow the lead of Dario and Daniela Amodei, a brother and sister team at OpenAI who left in December 2020 to form their own AI startup, Anthropic.

Veteran venture capitalist Chamath Palihapitiya has even wondered whether OpenAI will still exist.

“OpenAI failure, if that is what ultimately happens, will scatter teams everywhere. The net result of that is more competition,” he warned.

The problem is the organization has been diminished after five dizzying days during which the board first sacked Altman with barely any explanation, only to hire him back after an equally abrupt change of heart.

Sutskever, an OpenAI board director and its chief scientist, even turned on the very plotters he had first sided with and expressed deep regret for his actions, saying: “I will do everything I can to reunite the company.”

The drama became so gripping, it was compared to Game of Thrones, the HBO fantasy drama best known for its perpetual palace intrigue and backstabbing. According to Vinod Khosla, the first VC investor to put money into OpenAI, the ensuing damage reflected poorly on the entire field and “set back the promise of artificial intelligence”.

It was ultimately the threat of a brain drain that brought the coup to an end. Microsoft CEO Satya Nadella shrewdly forced the board’s hand by giving a blanket job guarantee to anyone who wanted to leave.

He wasn’t the only one interested in poaching OpenAI’s talent, either.

Salesforce founder and CEO Marc Benioff wasted no time in appealing personally to prospective deserters, making a standing offer to match the full cash and salary benefits of any OpenAI staffer on the spot.

“Sent me your CV directly to [email protected],” he offered.

OpenAI couldn’t afford that kind of exodus, bringing the brief revolt against Altman to a close.

Sutskever, along with two independent directors, researchers Helen Toner and Tasha McCauley, were replaced with ex-Twitter chairman Bret Taylor and economist Lawrence Summers, both considered more investor-friendly.

“OpenAI is nothing without its people,” chief technology officer Mira Murati acknowledged, echoing a widely held sentiment within the company.

With the situation now apparently resolved in the interests of OpenAI’s shareholders, co-founder Greg Brockman, who had briefly resigned as chairman in protest of Altman’s sacking, attempted to put on a brave face.

“We are so back” he tweeted, posting a group photo with smiling staff members.

VC backer Khosla agreed, hoping to turn the page on the fiasco: “Time to forgive, forget and move on everyone.”

But the episode exposed an underlying rift that cannot be papered over with photos.

The stark choice employees now face between profits and purpose harkens back to OpenAI’s very roots.

The company/nonprofit was founded in late 2015 as a counterweight to Google, with a goal to ensure advanced AI would not “harm humanity or unduly concentrate power,” as its founding charter stipulates.

Just prior to OpenAI’s founding, Elon Musk—one of OpenAI’s original co-chairs, alongside Altman—became concerned, he says, by the “cavalier” attitude of Google co-founder Larry Page toward the dangers posed by AI, especially since the latter had just bought DeepMind.

Eventually, however, OpenAI needed to compromise on its ideals in order to scale its ambitions faster than the $1 billion in committed funds allowed. As a non-profit, it didn’t have those kinds of resources, so it chose to fundraise from the private sector.

To remain true to its primary mission—developing AI that benefits all of humanity—a solution was found in 2019 to create a hybrid governance structure. This would open it up to investors that would own equity in a new company called OpenAI LP, while ensuring board governance remained with the non-profit.

All investors to the new limited partnership were asked to sign the same obligation as employees—that the founding charter “always comes first, even at the expense of some or all of their financial stake”.

But the private sector ultimately has a fiduciary duty to shareholders to protect their investment. So even if the OpenAI board still technically called the shots, the halcyon days of being unaccountable ended. The chaos surrounding Altman’s brief sacking made that obvious.

“There is no OpenAI without Microsoft leaning in in a deep way to partner with this company on their mission,” Nadella warned this week.

Christiaan Hetzner

Copyright: 2023 Fortune Media IP Limited