Why did Musk and OpenAI part ways? Early emails reveal the truth

Source: Tencent Technology

According to a recent report by TheVerge, in 2016, OpenAI began to negotiate a cooperation agreement with Microsoft. The newly established startup hopes to obtain key computing resource support to build the chatbot ChatGPT. However, as the CEO of OpenAI, Sam Altman must first win the support of Elon Musk, the biggest "financial owner".

According to the latest disclosed emails, Altman sent Musk a message in September 2016: "We only need $10 million to get $60 million worth of computing resources, and we can also provide advice on Microsoft's deployment of technology in the cloud." Microsoft hopes that OpenAI can provide feedback and promote Microsoft's artificial intelligence tools, such as Azure Batch. However, Musk expressed disgust at the idea, saying that it made him "feel sick."

Altman then made another offer: "Microsoft is now willing to reach an agreement with us for $50 million and promises that 'OpenAI will evaluate whether it has done its best efforts according to the circumstances', and both parties have the right to terminate the agreement at any time. At the same time, the agreement does not contain any promotional content and there are no other additional conditions. In this way, we will not be used as a crappy marketing pawn by Microsoft. Do you think this proposal can be continued?"

Musk responded: "If they don't use this feature in unsolicited messages, I think it's fine. But if we can avoid making people feel like Microsoft's 'marketing bitch', then this agreement will be worth much more than $50 million."

Last week, Musk made these emails and other related emails public as part of his lawsuit against OpenAI and Microsoft. On the surface, Musk's move was to prove that there was an anti-competitive partnership between Microsoft and OpenAI. But more importantly, the emails reveal details of the early cooperation between Altman and Musk and the subsequent power struggle. Musk invested $50 million to $100 million in OpenAI in its early days.

The emails also trace the evolution of OpenAI from an open source nonprofit to a "de facto closed-source subsidiary" of Microsoft, during which the company abandoned its original intention to develop artificial intelligence for the benefit of all mankind. At the same time, they also witnessed the complete collapse of the partnership between Musk and Altman.

"Musk has tried to change his statement for the third time in less than a year, and this time his statement is even more unfounded and exaggerated than before," an OpenAI spokesperson said in a statement.

Musk co-founded OpenAI with Altman and others because they were worried that human-level intelligence technology would fall into the hands of tech giants like Google. However, they eventually found that OpenAI gradually evolved into the tech giant they worried about. After successfully winning Musk's coveted position as CEO of OpenAI, Altman decided to keep OpenAI's cutting-edge artificial intelligence technology strictly confidential. He claimed that the technology was too dangerous to be released publicly.

The decision angered Musk, who quit the board of OpenAI and founded his own artificial intelligence startup xAI last year. Nearly a decade after the two co-founded OpenAI, Musk is taking the competition to court, waging a fierce battle over "the unstoppable future of computing."

In an email to Musk in 2015, Altman expressed his deep thoughts on the future development of artificial intelligence. He wrote: "I have been thinking about whether it is possible to stop people from developing artificial intelligence. But if the development of this technology is inevitable anyway, it seems that it would be a good choice for companies other than Google to take the lead in this work."

1.Caught between two forces, lack of talent and funds

From the beginning, OpenAI was caught between two conflicting forces: on the one hand, its lofty ideal mission to benefit mankind, and on the other hand, the fierce competition with technology giants. Musk and Altman agreed that regardless of their respective motivations, it was crucial to ensure that they attracted top talent and raised a lot of cash. This early compromise laid the groundwork for Musk's later criticism of OpenAI for pursuing profits rather than sticking to its original intentions.

In 2015, OpenAI was still known as YC AI, a laboratory affiliated with YCR, the nonprofit research department of the startup incubator Y Combinator. Altman, then president of Y Combinator, used his extensive connections and resources to actively attract researchers and funds. Musk urged Altman and CTO Greg Brockman (current president) to seek more than $100 million in financing, reminding them that any amount below that would be insignificant compared to well-funded tech giants such as Google and Facebook.

"I think we should say we're going to start with a $1 billion commitment," Musk wrote in an email that year. "That's true, and I'll pay for any amount that nobody else has provided."

However, despite Musk's support and tens of millions of dollars in funding, the emerging organization still faced the same challenges that most startups would encounter in the early days: fierce competition for top talent. While OpenAI may be the hottest employer in Silicon Valley today, a decade ago (or before ChatGPT was launched), many top AI researchers were not interested in it.

To compete for the best AI researchers, Altman and his team designed a unique compensation plan: offering a base salary of $175,000, granting the title of YC "part-time partner", and 0.25% equity in each batch of YC startups. Today, OpenAI AI researchers are paid nearly $1 million a year. In an email to Musk, Altman called this plan the "Manhattan Project for Artificial Intelligence" and firmly believed that he could attract the top 50 of many top researchers to join, build this project in the form of a non-profit organization, let the technology benefit the world, and let the people doing this work get the same generous compensation as startups.

Altman's original intention was to build a founding team of 7 to 10 elite members, and he was determined to attract the best talents in the industry at all costs. However, they faced fierce competition from Google's artificial intelligence laboratory DeepMind.

In a December 2015 letter to Musk, Altman expressed concern: "DeepMind intends to make a generous offer to every member of OpenAI tomorrow in an attempt to strangle us. Do you object to increasing everyone's salary by $100k-$200k per year? While I think they are all passionate about the mission, a pay increase would be a positive sign that we are taking good care of them." He added: "It sounds like DeepMind is preparing to go to war with us."

Musk approved the salary increase. By February 2016, OpenAI's founding team received an annual salary of $275,000 and equity in YC, while other employees received an annual salary of $175,000 and a performance bonus of $125,000, or an equivalent amount of YC or SpaceX stock. But OpenAI co-founder Ilya Sutskever (former chief scientist) was a special case. It is reported that Sutskevi earned more than $1.9 million in 2016. He once told the media that he "rejected offers that were several times the salary offered by OpenAI."

Musk responded: "We need to do everything we can to attract top talent. We will continue to increase our investment. If we need to review the salaries of existing employees in the future, that's fine. Either we get the best talent in the world, or we will be defeated by DeepMind. As long as we can cultivate excellent talents, I am willing to support it." He also warned that DeepMind's victory brought him "great mental pressure" and "they are obviously making great progress, and they should make such progress considering the level of talent there."

2.Musk and Altman both want to be CEO,Motives questioned by other co-founders

Over time, Musk and Altman have a significant disagreement.

In August 2017, just as OpenAI was finalizing the details of its first round of financing of $200 million to $1 billion, some internal conflicts began to emerge. Shivon Zilis, a former OpenAI board member and the chief operating officer of Musk’s brain-computer interface company Neuralink, wrote to Musk about Brockman and Sutskevi’s concerns about the distribution of equity and control of the newly established for-profit OpenAI, and about whether Musk could devote enough time while serving as CEO. Musk responded: “This is annoying. Please encourage them to start a business. I’ve had enough.”

The following month, Sutskevi and Brockman jointly sent an email to Musk and Altman, further elaborating on their concerns. They pointed out that if Musk served as CEO, he might have “unilateral and absolute control” over the development of general artificial intelligence (AGI). At the same time, they also questioned Altman's motivations, asking why he valued "the title of CEO" so much and asking him "Is the development of AGI really his main motivation? How does this relate to your political goals? How has your thinking changed during this period?"

Although the email did not elaborate on the specific meaning of the word "political", Altman did become active in political campaigns in California earlier that year. Sutskevi and Brockman said they were blinded to these concerns during the early fundraising process.

DeepMind CEO Demis Hassabis

They continued: "OpenAI's goal is to create a beautiful future and avoid letting AGI (artificial general intelligence) take control of everything. We are also worried that DeepMind's Demis Hassabis (DeepMind CEO) may create an AGI hegemony. So, if you plan to become a ruler, or create a system that allows you to become a ruler, that is definitely not a good idea, especially given the efforts we are making."

However, Musk is not buying it. He responded, "I'm not going to fund OpenAI again unless you guys make a commitment. This makes me feel like a fool that I'm basically giving you free money to build a startup. End of discussion." Altman responded by saying he was still "passionate about the nonprofit structure," a statement that ultimately convinced Sutskevi and Brockman to back down.

Following this confrontation, the situation became even more complicated. Ziris relayed to Musk a conversation she had with Altman, in which Altman expressed that his trust in Brockman and Sutskevi had dropped significantly, and that he felt their messaging was inconsistent and sometimes naive. As a result, Altman decided to take a 10-day leave of absence to deal with the incident, and he needed to reassess his trust in the two co-founders and whether he was still willing to work with them.

However, just five months after Sutskevi and Brockman expressed their concerns about the power struggle, the situation once again reached a critical inflection point. During a heated argument, Musk became convinced that OpenAI had fallen far behind Google and proposed that he take control of the company himself. This is exactly the kind of situation that Brockman and Sutskevi had warned about before: that Musk might try to unilaterally control the direction of the company.

Musk said in 2018 that he was pessimistic about OpenAI's prospects of competing with DeepMind and Google. After OpenAI's management rejected his proposal, Musk left the board and cut off funding in February 2018, but he still provides consulting services to the company as an advisor.

Musk's departure posed a serious threat to OpenAI's nonprofit model. Since he had previously invested $100 million, his departure left OpenAI in a desperate need to find new sources of funding. In 2019, in order to reduce its dependence on Musk and fund training data centers, OpenAI designed a unique structure: a for-profit company controlled by a nonprofit organization. This structure attracted the attention of investors such as LinkedIn co-founder Reid Hoffman and venture capitalist Vinod Khosla, who participated in the first round of financing and pledged to provide nearly $1 billion in funds, but the initial funds were only $130 million.

During this period, Musk sent an email to Altman, suggesting that he was involved in the adjustment of the company's structure and asked Altman to make it clear that he had no financial interest in the for-profit department of OpenAI. Altman simply responded: "No problem." However, this email later became part of the lawsuit.

3.OpenAI's power struggle is not due to personal grudges

OpenAI has a huge influence in the artificial intelligence industry, and its internal battle for control has always attracted much attention. Although the struggle between Musk and Altman was fierce, Altman eventually won, giving him almost complete control of OpenAI.

The legal merits of Musk’s allegations against OpenAI and Microsoft may be debatable, as much of his lawsuit focuses on accusations of hypocrisy by Altman, which are rarely punished in court. However, the lawsuit provides Musk with a rare opportunity to publicly reveal some details about the origins of OpenAI and his own role. The evidence reveals that Altman already had a degree of power at the company’s inception, which may not be entirely consistent with the original intentions of his other co-founders.

In addition, the lawsuit provides a public view of the complex relationships and power struggles within OpenAI. Altman’s willingness to confront for-profit competitors head-on from the beginning highlights his strong position in the company. At the same time, this incident also allows us to see that Musk’s desire for power has not diminished, and he may hope to use this lawsuit to destroy Altman’s reputation and consolidate his position as the rightful steward of AGI.

However, it is also necessary to recognize that the narrative presented by this lawsuit may not be complete. Many important conversations and decisions may have taken place in private and have not been made public or included in the litigation materials. Therefore, we cannot fully understand all the details and background of this palace fight.

In general, the battle for control of OpenAI is not only a contest of personal grudges, but also concerns the future development direction and strategic positioning of the field of artificial intelligence. The outcome of this struggle will have a profound impact on OpenAI, and will also have a certain inspiration and reference for the entire industry.