Sam Altman, who is now the former CEO of OpenAI © Presse-citron
It's the theatrical twist of the weekend: while we thought Sam Altman was absolutely essential at the head of OpenAI, to whom we owe ChatGPT, the board of directors of the firm surprised everyone by convening the board of directors of OpenAI. emergency the co-founder of the firm Friday November 17, 2023 – to inform him that he is being dismissed from his position, with immediate effect. Since then, the announcements have been coming one after the other and it is not easy to easily find oneself in this somewhat chaotic situation.
The group's technical director, Mira Murati, was initially named interim president of OpenAI (according to the open letter from employees included in the last part, the directors subsequently changed their minds). The board of directors is still looking for a definitive successor to lead the firm. For his part, Sam Altman is fine, thank you – he soon begins new adventures at Microsoft with another great figure of the firm. And he is not the only one to change or threaten to change companies.
In fact, of the 770 OpenAI employees, 743 are now threatening to leave to follow Sam Altman… Here is a complete summary of the situation:
The official OpenAI press release, published immediately on Friday, is not sparing of details on the precise reasons for the dismissal of Sam Altman. We can simply read:
“Mr. Altman leaves office after extensive board review. The latter determined that his frankness in communications was not consistent, which compromised the board's ability to fulfill its responsibilities. As a result, the board no longer believes in its ability to effectively lead OpenAI”.
Sam Altman did nothing illegal or commit professional misconduct
So what is the famous “franchise problem” which the press release mentions? This Monday, November 20, new details on the preliminary meeting of the board of directors which led to the dismissal emerge.
As far as we know, there is no evidence to indicate that the leader is accused of anything illegal. The company's COO, Brad Lightcap, explains in an internal memo intended for OpenAI employees (taken up by Axios):
“We can say with certainty that < strong>the board's decision was not in response to any professional misconductor anything that might concern our finances, our business, our security or our privacy practices. This is actually a communication problem between Sam and the board of directors.
The proponents of a non-profit and exclusively scientific OpenAI have (almost) won…
The New York Times also reveals in another article the < strong>central role of Ilya Sutskever, scientific director of the organization. He would have defended the idea of landing Sam Altman during this quasi-plenary meeting, accusing him of giving too commercial a turn to OpenAI.
From then on, the scenario of 8217;a“putting OpenAI back on track” a prudent organization, with a scientific and above all non-profit aim, is essential. The official OpenAI press release specifies further:
“OpenAI was designed with a precise objective: ensuring that intelligence general artificial is truly for the benefit of all humanity. Our board of directors is entirely dedicated to this mission. We are deeply grateful for everything Sam has contributed since the creation of OpenAI and during its development. However, we are convinced that a change is necessary in management…
…but the decision also caused the departure of Greg Brockman and sparked a game of dominoes
On the sidelines of this meeting (which he was not allowed to attend) the board of directors also decided to remove its president Greg Brockman< /strong>. If he remains in his other functions within OpenAI, the latter was involved in many of Sam Altman's actions to find new funding.
However, he is also the keystone of many success of the firm, notably the GPT-4 model, or the training system of the OpenAI Gym models. This personality was seen as central to developing artificial general intelligence (AGI) such as OpenAI's vision.
In a tweet this Saturday, November 18, 2023, Greg Brockman explains that he and Sam Altman “are shocked and saddened by the decision of the board of directors”:
Sam and I are shocked and saddened by what the board did today.
Let us first say thank you to all the incredible people who we have worked with at OpenAI, our customers, our investors, and all of those who have been reaching out.
We too are still trying to figure out exactly…
— Greg Brockman (@gdb) November 18, 2023
He continues by providing a chronology of events:
<em”>”Here is what happened as we experienced it:
- [Thursday evening, editor's note], Sam receives a message from Ilya asking him to talk Friday at noon. When he joins the Google Meet meeting, he finds himself in front of almost the entire board, except Greg. Ilya then announced his dismissal and that the information would soon be public.
- Greg, for his part, received a message from Ilya at 12:19 p.m. asking him to call quickly. Four minutes later, he receives a Google Meet link. He is told that he is being removed from his position on the board, while remaining essential to the company, and that Sam has been fired. Almost simultaneously, OpenAI published an article on its blog.
- To our knowledge, OpenAI management was only informed shortly after [these two meetings], exception made of Mira Murati who had known about it since the day before.”
Greg Brockman concludes: “We are touched by your support; THANKS. But don't worry about us. We will get through this. New and great perspectives are coming.”
Moreover, we learn that Mira Murati has obviously not accepted was replaced as interim president of OpenAI. She simply posted on 𝕏 a message of support for Sam Altman and Greg Brockman: “OpenAI is nothing without its employees”.
< p>Instead, this Monday, November 20, 2023, Emmett Shear, former CEO of the Amazon Twitch service, was appointed to ensure the transition. The atmosphere promises to be frosty in the firm's offices in the coming days.
OpenAI was founded in December 2015 by Ilya Sutskever, Greg Brockman, Trevor Blackwell, Vicki Cheung, Andrej Karpathy, Durk Kingma, Jessica Livingston, John Schulman, Pamela Vagata, Wojciech Zaremba, Elon Musk and Sam Altman. And until 2018, the firm did indeed remain 'non-profit'.
But very quickly internal struggles appeared. Particularly on the side of Elon Musk who began to scold Sam Altman (visibly trying to take the direction of the organization by force), before being disembarked by the board of directors. Elon Musk accused the leader of bringing OpenAI under the control of big Big Tech names like Google.
However, from 2019, OpenAI is changing its structure to allow limited profits (100 times the initial investment). It was then that OpenAI accepted an initial investment from Microsoft of $1 billion – while using Microsoft Azure data centers. OpenAI Global LLC then announces its intention to commercialize its technologies.
Sam Altman then explains his intention to spend this investment “over 5 years, or even faster” ; and to add that OpenAI will likely need an endowment “of an unprecedented level for a non-profit organization” in order to achieve this. reach the IAG.
This increasingly commercial orientation sows internal trouble. Which gets worse between 2020 and 2023, when Microsoft announces a new investment of $10 billion over several years.
Microsoft has until now been banking heavily on its partnership with OpenAI and its massive investment in the company. Despite its strong position within OpenAI, Microsoft did not really know of the dismissal in advance. Which casts a certain chill, from the outset, on the long-term relations between OpenAI and Microsoft.
Microsoft therefore tried to reverse the decision of the board to 8217;administrationthis weekend, alongside other investors. And here again, the outcome of this story is not at all good news for OpenAI. On Sunday, a new press release from the board of directors announced that the firm is maintaining its decision and will not bend under pressure from its financiers.
However, as of Monday, the boss of Microsoft is making a announcement on 𝕏 that some would call somewhat “passive-aggressive”:
We remain committed to our partnership with OpenAI and have confidence in our product roadmap, our ability to continue to innovate with everything we announced at Microsoft Ignite, and in continuing to support our customers and partners. We look forward to getting to know Emmett…
— Satya Nadella (@satyanadella) November 20, 2023
“Our commitment to our partnership with OpenAI remains strong and we are confident in our product development plan, our ability to innovate, as demonstrated at Microsoft Ignite, and our continued support of our customers and partners. We look forward to collaborating with Emmett Shear and the new OpenAI leadership team. In addition, we are delighted to announce that Sam Altman and Greg Brockman, along with their colleagues, will join Microsoft to lead a cutting-edge AI research team. We are ready to quickly provide them with the resources they need to realize their ambitions.”
You read that right: Sam Altman, Greg Brockman, and “their colleagues” simply join Microsoft, where their action is unlikely to cause any friction. We note that Microsoft suggests that it has managed to recruit an unknown number of other OpenAI employees. On social networks, many collaborators, like Mira Murati, posted the same message of support, which gives an idea of the potential hemorrhage.
Seen from the outside, OpenAI seems to not only endanger its immediate future but also its very missionto lay the foundations for a controlled and responsible IAG. The firm is in fact losing real talent, and alienating practically all of its investors; an instability which should have a lasting impact on OpenAI's operations and slow down the development of its projects.
On the other hand, Microsoft is recovering some of the best elements of OpenAI, while emerging from this story with an attractive image, which could convince many other employees of the firm to join him, whether immediately or in the coming months and years. In our opinion, Microsoft is therefore the big winner in this story– and OpenAI, such as the need to more effectively supervise the development of AI, the two big losers.
By making such an abrupt decision, by alienating investors, and by separating from talents, OpenAI has in fact compromised its chances of remaining a major player in the sector, and therefore to have the slightest influence allowing it to impose its vision in the immediate future. It’s as if the firm had brought down new guardrails, trying to return to its fundamentals.
De facto, this surprise decision profoundly reshuffles the cards in the AI sector. Firstly because, in addition to Microsoft, other players can take advantage of this period of unrest to continue to drain the company's current talents. Which means that teams like those from Bard, Grok or other AI could more easily gain the advantage.
Companies that depend on OpenAI AI for their services will have comparatively undoubtedly an impression of stagnation and many of them will be tempted to use the competition. The emergence of more efficient models from various players should also complicate the establishment of rules to govern the development of AI.
Which increases their potential dangerousness. Enough to encourage States to take up the subject to impose a framework – if possible global – to the entire sector. But, history gives more and more reasons to sigh that the table now looks like a lost game.
The situation is still under development and there is a good chance that new surprises will emerge in the coming days. to not miss any new news, follow our liveblog dedicated to the situation at OpenAI.