AI expert Gary Marcus has been following the OpenAI turmoil this weekend with interest. And, as he wrote on Sunday, his “stomach hurts.”
OpenAI’s board shocked investors and employees on Friday by firing CEO Sam Altman. But it now seems likely that not only will the board’s decision be overturned and Altman returned to his position, but the board members will also be expelled, according to Bloomberg.
Mark wrote about the situation on his Substack, sharing a analysis written by FortuneIt’s Jeremy Kahn earlier today. Whatever the board’s reasons, the reasons are stated were vague– it’s not a good sign if it’s easily controlled, says Marcus, professor emeritus of psychology and neural sciences at New York University and host of the show Humans versus machines podcast.
While OpenAI began as a nonprofit in 2015, four years later Altman, shortly after becoming CEO, created a commercial arm, governed by the nonprofit parent company. Altman, unusually, held no stake in the company. This diminished his influence with the board of directors, which, as he often pointed out, had the power to fire him.
“No one should trust anyone here,” he told Bloomberg this summer. “The board of directors can fire me. I think it’s important.
In OpenAI’s unusual structure, a board of directors “with no financial interest was supposed to look after humanity,” Marcus wrote. “The spirit of the original agreement was that everything the for-profit did was supposed to be in service of the nonprofit. »
Indeed, the board was supposed to oversee the profit-capped company, with an eye on the broader mission: ensuring that safe artificial general intelligence (AGI) “is developed and benefits the entire world.” humanity”. AGI refers to a system that can match humans faced with an unknown task.
So while it’s Microsoft’s large sums of money and computing resources that keep OpenAI running – the software giant has committed at least $13 billion to OpenAI but so far has none provided only one part – the nonprofit board was clearly still in control.
But as Kahn wrote, “the structure was basically a ticking time bomb. By addressing a single legal entity, Microsoftfor the majority of the cash and computing power OpenAI needed to accomplish its mission, it essentially handed over control to Microsoft, even though that control was not codified in any formal governance mechanism.
Faced with the potential financial fallout from Altman’s removal, “the nominally subordinate for-profit organizations (both employees and investors) quickly set to work to expel the board and overturn its decisions,” a writes Marcus. “All signs indicate that financially interested stakeholders will quickly emerge victorious. »
Altman had told investors that if he returned to OpenAI, he wanted a new board and governance structure, according to THE Wall Street Journal.
“The tail therefore appears to have wagged the dog, potentially jeopardizing the original mission, if the Board’s concerns were well-founded,” Marcus wrote. “If you think OpenAI has a chance, eventually, at AGI, none of this bodes particularly well.”