After a checkered weekend for OpenAI, its co-founders Sam Altman and Greg Brockman, one fired and the other resigned, have just been recruited by Microsoft. They will lead a “new AI research team”.
Hours after this announcement, nearly 500 OpenAI employees sent an open letter to their board of directors, declaring that they could resign and join the duo at Microsoft if the two co-founders were not immediately reinstated.
As a reminder, the ongoing saga began Friday evening when OpenAI released a statement announcing Altman’s ouster, due to what the board said was a lack of confidence in his ability to continue to lead the ‘business.
Mira Murati sought to rehire Sam Altman and Greg Brockman
“Mr. Altman’s departure follows a review process by the Board of Directors, which concluded that he was not always candid in his communications with the Board of Directors, which hindered the ability of the latter to exercise its responsibilities”, declared the board of directors. “OpenAI was structured to advance our mission: to ensure that artificial general intelligence (AGI) benefits all of humanity. The Board of Directors remains fully committed to serving this mission.”
The board expressed gratitude for Mr. Altman’s contributions to the founding and growth of OpenAI, but noted that new leadership was “necessary” for the company to move forward.
So he named CTO Mira Murati as interim CEO, describing her as “exceptionally qualified” as she has led the company’s research, product and security functions.
On Sunday, numerous news reports suggested that Mira Murati was looking to rehire Sam Altman as well as Greg Brockman. The board also reportedly reconsidered its decision, under pressure from investors and company staff who demanded Mr. Altman’s return. But none of this ultimately happened.
Twitch co-founder Emmett Sear at the helm of OpenAI
Instead, Twitch co-founder and former CEO, Emmett Shear, has been named interim CEOand replaced Mira Murati in this position which she also held on an interim basis for a few hours.
Amid speculation about what might happen next, Microsoft Chairman and CEO Satya Nadella announced Monday that he had recruited Altman and Brockman, along with “colleagues” from OpenAI, to form “a new research team on AI. Microsoft, however, seems not to put its eggs in one basket.
“We remain committed to our partnership with OpenAI and are confident in our product roadmap, our ability to continue to innovate with everything we announced at Microsoft Ignite, and to continue supporting our customers and partners,” said Nadella in a post on X. “We look forward to getting to know and working with Emmett Shear and the new OpenAI leadership team.”
Microsoft does not put its eggs in one basket
The Microsoft CEO added that his company would move “quickly” to provide the resources its new AI research team would need to succeed.
In response to his appointment at Microsoft, Mr. Altman said simply: “The mission continues.”
He will lead Microsoft’s new team as CEO, according to Nadella, who added that this team will join other teams that have built “independent identities and cultures” within Microsoft, including GitHub and LinkedIn.
Exodus of OpenAI employees to Microsoft announced
Shortly after Mr. Nadella hired Mr. Altman, nearly 500 OpenAI employees sent a letter to their board saying they risked heading in the same direction if Mr. Altman and Mr. Brockman were not reinstated at OpenAI. They also demanded that all board members resign and two new independent directors be appointed.
“Your actions have exposed your inability to oversee OpenAI. We cannot work for or with people who lack skill, judgment, and care for our mission and our employees,” it says. read in the letter. “Microsoft assured us that there were positions for all OpenAI employees in this new structure, should we decide to join.”
Among the signatories to the letter are Mr. Murati and OpenAI co-founder and chief scientific officer Ilya Sutskever, who is believed to have played a role in the decision to remove Mr. Altman. Mr. Sutskever serves on the board of directors of OpenAI.
Ilya Sutskever publishes his regrets
Mr Sutskever, who had remained silent since the news broke on Friday, posted a message on X just before the letter surfaced, in which he said: “I deeply regret participating in the board’s actions. I never intended to harm OpenAI. I love everything that we built together and I will do everything I can to reunite the company.”
In an X post on his own nomination, Emmet Shear presented a 30 day plan for OpenAI, which includes launching an independent investigation into “the entire process that led to this situation” and producing a full report. Also on his to-do list is “reforming” the management and leadership team, following recent departures, with the aim of delivering results for clients.
Adding that he had “verified” the reasons for Mr. Altman’s ouster, Mr. Shear said: “The board did not remove Sam from his position because of a lack of political will. The board did not terminate Sam due to a specific disagreement over safety. Their reasoning was completely different. I’m not crazy enough to take this position without the support of the board to bring our models to market “.
Divergences on how to ramp up to AI “superintelligence”
During his visit to Singapore last June, Mr Altman said it was important for the public to discover and experience AI, even as the technology continues to evolve. This would be more effective than building and testing a piece of technology behind closed doors and releasing it to the public assuming all possible risks have been identified and eliminated, he noted. “You can’t learn everything in a laboratory,” he added.
In a July 2023 post, Sutskever wrote alongside machine learning researcher Jan Leike: “Superintelligence will be the most important technology humanity has ever invented and could help us solve many of the world’s most difficult problems. important people in the world. But the immense power of superintelligence could also be very dangerous and lead to the disempowerment of humanity, or even its extinction.”
“Currently, we do not have a solution to direct or control a superintelligent AI and prevent it from behaving in a disordered manner,” also said Mr. Sutskever, who also sits on the board of directors of OpenAI. “Our current AI alignment techniques, such as reinforcement learning from human feedback, rely on the ability of humans to supervise AI, but humans will not be able to reliably supervise AI systems that are much smarter than us. So our current alignment techniques won’t be able to accommodate superintelligence. We need new scientific and technical breakthroughs.”
In his post, Mr. Sutskever announced the creation of a new team of researchers and engineers specializing in machine learning, responsible for working on this problem. Their efforts include developing a scalable learning method, validating the resulting model, and stress testing the entire pipeline.
To go further on the impact of the Sam Altman / OpenAI affair on the corporate AI sector
Source: “ZDNet.com”