Author: Lan Xi Source: Weibo

Because Musk is suing OpenAI, many originally confidential private emails have been disclosed. Based on the email conversation records, The Verge has compiled a timeline of OpenAI's early top-level design. Although there is a lot of jargon, it is very interesting to connect them together. This is how real business battles are, plain and simple:

• In 2015, before OpenAI existed, its predecessor was Y Combinator's AI Lab. Sam Altman, who was the president of YC at the time, personally invested $10 million and started recruiting everywhere.
• At this time, Musk suggested that he should raise at least $100 million, because any capital below this number would be easily crushed by Google and Facebook.
• Sam Altman sent an email to Musk, pouring his heart out, wondering if he should choose to stop the development of AI for the sake of humanity. If he can't stop it, then it might be right for the leadership of AI to fall into the hands of a company other than Google.
• Then OpenAI spun off independently, Musk set a small goal of raising $1 billion first, let Sam Altman raise money as much as possible, and promised to make up the final shortfall;
• Because OpenAI was inconspicuous in Silicon Valley at the time and couldn't attract top talents, Sam Altman designed a radical compensation plan: a base salary of $175,000 per year, plus a batch of 0.25% YC equity grants.
• Sam Altman said to Musk that this would be a 'Manhattan Project' for the artificial intelligence industry, hoping to attract 50 top researchers to join. Although the company remains non-profit, employees who contribute labor can still receive returns similar to entrepreneurship.
• In 2016, DeepMind, Google's AI laboratory, noticed OpenAI's actions and began to call everyone contacted by OpenAI, trying to lure them away with double the salary;
• Sam Altman panicked and quickly wrote an email to Musk, hoping to approve a salary increase of $100,000 to $200,000, otherwise DeepMind would win the war ahead of time;
• Musk agreed, and the early team members received salaries close to $300,000 plus stock options. Musk even contributed SpaceX's equivalent stock to fill the gap.
• With the exception of three people, Greg (the subsequent president), Ilya (the subsequent chief scientist), and Trevor (the subsequent researcher), who received higher salaries, it seems that Sam Altman did not arrange for himself to be included.
• Ilya's income exceeded $1.9 million in 2016, which helped him to some extent resist the recruitment from Google;
• Musk agrees with OpenAI's generous salary payment when there is nothing, "We continue to do our best to attract top talent, and if money can solve the problem, then nothing could be better, as long as we are not defeated by DeepMind, I can accept any cost";
• Musk had an extreme aversion to Google at the time, telling Sam Altman that the mere thought of DeepMind becoming the winner caused him to have a mental breakdown; he couldn't allow Google to dominate the world.
• In 2016, OpenAI was still short of money, especially unable to afford chips. Sam Altman went to find Microsoft and negotiated a plan to exchange $10 million for the equivalent of $60 million in Computing Power resources. The condition was that OpenAI would use Microsoft's Azure cloud services and help promote it in the tech community.
• Musk told Sam Altman, this tm is disgusting;
• There's no way, Sam Altman went to talk to Microsoft again, and the new offer is for Microsoft to invest 50 million US dollars, without any additional conditions, and the cooperation can be terminated at any time. Sam Altman, with resentment, said to Musk, 'It doesn't look like Microsoft's lapdog anymore, are you satisfied?';
• Musk's merciless reply email stated that he also had to request Microsoft not to issue a press release on this matter. Just not being Microsoft's lapdog is worth several 50 million dollars; In 2017, OpenAI was preparing for a new round of financing. A woman who was later to have three children for Musk and was on the board of directors told Musk that the founding team was not very willing to have Musk as CEO because they felt that Musk was managing too many companies and could not devote 100% of his time to OpenAI.
• At that time, it was the period when TSL was just a few steps away from bankruptcy. Musk, visibly frustrated, had no interest in OpenAI's tricks and replied to the insider's email, saying, 'I'm tired of it, let them handle the company.'
• Not long after, Ilya and Greg relentlessly pushed forward the ice-breaking plan, writing long emails while copying Sam Altman and Musk, expressing concerns about Sam Altman's insistence on the CEO title and doubting his loyalty to the ideal of AGI. At the same time, they explained to Musk that the reason for rejecting him as CEO is because they know his personality would definitely seek absolute control over AGI;
• In other words, their core concern is that while the risks of AI are enormous, there may be even greater risks in appointing a single individual to control AI. They want to conceive of a more decentralized governance structure;
• Both Musk and Sam Altman are now in a state of collapse, without doing anything, they put on a big analysis, all kinds of grand narratives about how to determine the fate of humanity. Musk directly responded, saying the discussion is over. If you treat me as just a cash cow, I will Rug Pull.
• Sam Altman appeased his colleagues on the surface, expressing confidence in the non-profit nature of the company. But privately, he mentioned that Greg and Ilya can be immature at times. He needs to think carefully before taking a vacation to determine whether he can trust them not to stab him in the back.
• In 2018, Musk judged that OpenAI's progress was completely behind DeepMind, once again proposed to be the CEO himself, otherwise it would not be possible to beat Google. However, the founding team of OpenAI still refused him, leading to Musk's withdrawal from the board of directors and stopping the capital donation;
• Suddenly losing Musk's money briefly put OpenAI at a disadvantage. In order to obtain financing, OpenAI, for the first time, changed the company and designed a structure controlled by a non-profit organization, a limited-profit company.
• Musk had already begun to realize the need to preserve evidence for future litigation. He emailed Sam Altman, asking him to clearly prove that he (Musk) had no financial interest in the new company structure, and Sam Altman promptly replied, "Received";
• The Verge's conclusion is that the power struggle around OpenAI essentially stems from two vastly different visions colliding. At least from the information displayed in the emails, it is clear that OpenAI's ability to turn the tide and reach the current heights is inseparable from the impetus provided by Sam Altman. He always manages to obtain what he wants at all costs, while Musk, not lacking in power, seeks to retaliate against any behavior that belittles him and firmly believes that he is the chief steward of AGI.