Since Friday evening, have we been witnessing the self-destruction of OpenAI, the most prominent startup of the moment, just a few days away of the first anniversary of its flagship service, ChatGPT? Friday evening, to everyone’s surprise, the board of directors of the gem of generative artificial intelligence fired its founder andDirector General, Sam Altman, effective immediately. The reason given: a lack of “ franchise ” and of ” transparency » from the entrepreneur towards the board of directors… but no more details.
OpenAI board putsch
The news had the effect of an explosion. Because it is nothing more and nothing less than a putsch by the board of directors: neither the employees, nor Microsoft, which nevertheless holds 49% of the capital of OpenAI and has injected more than 13 billion dollars since 2019, nor the startup’s other investors were consulted before the board decided to activate the guillotine.
This anomaly is explained by the fact that OpenAI was launched in 2015 as a non-profit research laboratory. In fact, its board of directors brings together only the three co-founders and three researchers very attached to the original mission of OpenAI, that is to say “ AI research serving the common good “. The composition of board has barely changed over time, in part because of the desire for total control of Sam Altman, Silicon Valley’s new charismatic guru. Thus, despite the transformation of OpenAI into a company in 2018, and the enormous investments by Microsoft and the American fund Sequoia to develop it, the latter obtained neither a seat on the board nor the right to consult or veto its decisions.
“ Partnering with the most cutting-edge startup in generative AI was so important for Microsoft that they agreed to inject $13 billion without having the slightest control in terms of governance. This is a major error, an aberration in the economic world. The overthrow of Sam Altman would not have been possible in a normal company », Estimates Stéphane Roder, CEO and founder of AI Builders, an independent consulting firm specializing in AI.
Stunned and weakened – Microsoft’s stock price plunged 2% in post-closing trading on Friday evening – Satya Nadella, the CEO of the software and cloud giant, tried everything during the day on Saturday to save his head by Sam Altman. He was helped by the dozens of messages from OpenAI employees supporting their former boss, and especially by the resounding resignation of the company’s number 2, co-founder Greg Brockman, followed by that of certain key OpenAI engineers and executives. .
Without success: Sunday evening, OpenAI confirmed the ousting of Sam Altman and announced the appointment of Emmett Shear, former general manager of the video game streaming platform Twitch, known for his desire to slow down the development of artificial intelligence , which he considers to be a threat to humanity.
The revolt of OpenAI employees
But the game does not seem to be over yet. Monday morning, Microsoft regained control by announcing the recruitment of Sam Altman and Greg Brockman. The first is named CEO of a new AI research team, and everything indicates that Microsoft has rolled out the red carpet to ensure that the father of ChatGPT, whose technology irrigates all of Microsoft’s products, is not poached by the competition, Google and Amazon in the lead.
“ We learned how to give space to startup founders and innovators to create independent entities and cultures within Microsoft », declared Satya Nadella on X, citing the success stories of GitHub, Mojang Studios and LinkedIn, all three acquired and strongly developed by Microsoft within independent structures.
Since then, other leading AI researchers have left the OpenAI ship. Among them is Jakub Pachocki, who had been among the main minds behind the development of GPT4, OpenAI’s latest language processing model and currently the most advanced on the market. But also Szymon Sidor, who worked on artificial general intelligence (AGI), and Aleksander Madry, director of a research unit dedicated to the risks of AI. The technical director of OpenAI, Mira Murati, announced Friday evening as interim CEO before the appointment of Emmett Shear, for her part publicly displayed her support for Sam Altman, like dozens of other executives.
At the same time at OpenAI, the debacle continues to get worse. Monday afternoon, no less than 500 employees of the company sent an ultimatum to the board of directors, demanding its resignation and the reinstallation of Sam Altman at the head of the startup under penalty of a massive resignation which could downright cause the downfall of OpenAI.
The crisis is such that Ilya Sutskever, the scientific head of OpenAI considered to be the architect of the dismissal of Sam Altman, confirmed to the online media Techcrunch that he is also joining Microsoft! In the process, he published his regrets on I deeply regret my role in the actions of the board of directors. I never intended to harm OpenAI and I will do everything to reunite the company “, he tweeted. However, according to several American media, this researcher member of the company’s board of directors is the one who convinced his colleagues that leadership ” reckless and greedy » by Sam Altman threatened the raison d’être of OpenAI and that the race for artificial general intelligence could not be conducted under these conditions…
Microsoft’s “stroke of genius”
According to the American press, Microsoft is trying to recover as many employees as possible who leave OpenAI, regardless of the cost. The site The Information estimates that dozens of people announced their departure on Monday. “ Basically, the crisis at OpenAI is disastrous for Microsoft because this crisis shows that the startup is completely beyond its control and this is extremely worrying in terms of management. Today, the collapse of OpenAI or the poaching of its talents by the competition would cause it a lot of harm. On the other hand, Satya Nadella’s spectacular reaction to directly recover the talents is a true stroke of genius », continues Stéphane Roder.
Indeed, OpenAI currently enjoys a technological lead over the competition, and a first-place premium over the market. But “ talent is key in the battle for artificial intelligence », confirms Ghislain de Pierrefeu, analyst at Wavestone. “ The market is in its infancy and technology is evolving so quickly that having the best talent is essential » he adds.
Could the crisis at OpenAI be a blessing in disguise for Microsoft? “ In any case, it allows him to reinternalize part of the AI skills he needs, even if this is done in chaos. », replies the analyst. And it costs less than buying OpenAI, all without needing validation from competition authorities… Obviously,The market seems to approve: in trading before the stock market opened on Monday, the value of Microsoft shares rebounded by almost 5 points.
Microsoft well placed whatever happens?
Given the frenetic pace of developments since Friday evening, it is impossible to predict how the situation will evolve or rebound in the coming hours or days.
Two scenarios remain. If the crisis at OpenAI is resolved with the reinstatement of Sam Altman and the resignation of the board of directors, Microsoft will be a winner because as a shareholder with 49% of the capital, it will certainly be able, this time, to obtain real power in the new governance of the company.
On the contrary, if OpenAI loses its workforce and becomes an empty shell, Microsoft will have lost a lot of money, but nothing dramatic for a group valued at $2.8 trillion and endowed with almost unparalleled financial power in the world. The chaos caused, however, would slow down innovation in generative artificial intelligence and provide momentum to competitors with the opportunity to catch up, but Microsoft would not exit the game, as analyst Xiaodong Bao of Edmond firm explains. of Rothschild Asset Management:
“ Microsoft already provides OpenAI’s cloud infrastructure which is entirely hosted on Azure, and they have signed a long-term partnership to operate the technology. Additionally, the majority of OpenAI customers come from Microsoft because they subscribe via the Azure cloud and via services like Copilot. In this context, if Microsoft also recovers the best of OpenAI’s technical and R&D team, they are strengthening », he deciphers.
For the analyst, the big loser in this scenario could be… Nvidia, the world number one in artificial intelligence chips. “ Sam Altman was seeking Saudi funding for his new AI chip startup, which is also one of the reasons for the discord with the OpenAI board », he deciphers. However, its deal with Microsoft certainly includes, according to the American press, the financing of this project, which would contribute to the independence and technological sovereignty of the United States, in accordance with the objective set by the White House of centralize the development of artificial intelligence.