In the summer of 2017, OpenAI built an artificial intelligence system that could play a popular video game called Defense of the Ancients, known as Dota.
Inside a stadium in Seattle with more than 20,000 spectators, this A.I. system won an international Dota tournament, beating many of the world’s best players. When OpenAI’s president, Greg Brockman, emailed Elon Musk about the win, the tech mogul was elated. “Time to make the next step for OpenAI. This is the triggering event,” said Mr. Musk, who was backing OpenAI financially.
The next day, Mr. Brockman met Mr. Musk and his chief of staff, Shivon Zilis, and several others at a party house Mr. Musk had recently purchased just south of the city. There, they began discussing ways of transforming OpenAI into a for-profit company, according to testimony by Mr. Brockman and other evidence presented on Tuesday in the blockbuster trial that pits Mr. Musk against the maker of ChatGPT.
Mr. Musk has sued OpenAI, accusing Mr. Brockman and its chief executive, Sam Altman, of breaching the A.I. lab’s founding contract by putting commercial gain over the public good. He is asking for $150 billion in damages and a court order that would unwind the for-profit company that OpenAI created last year. He also wants an order removing Mr. Altman from the OpenAI board of directors.
Mr. Musk founded OpenAI as a nonprofit in 2015 alongside Mr. Brockman, Mr. Altman and a group of A.I. researchers, before leaving the organization. Mr. Altman and the other founders then attached a for-profit company to the A.I. lab and began raising billions of dollars from the tech giant Microsoft and other investors.
With these commercial efforts, Mr. Musk argues, OpenAI has abandoned the original mission of the nonprofit. But during his second day of testimony, Mr. Brockman described how Mr. Musk spent several months in 2017 working with the other OpenAI founders to turn the operation into a for-profit company.
Under questioning from Sarah Eddy, one of OpenAI’s lawyers, Mr. Brockman read a 2017 text he sent to Ms. Zilis in which he detailed a meeting he had with Mr. Musk that July. “He said nonprofit was def the right one early on, may not be the right one now,” the text read, referring to Mr. Musk.
Mr. Brockman also detailed a meeting in August 2017 when he and another OpenAI co-founder, Ilya Sutskever, discussed a possible for-profit company with Mr. Musk. Mr. Musk had recently given Tesla electric cars to the other founders of OpenAI. To show his gratitude, Dr. Sutskever brought Mr. Musk a painting he had made of a Tesla.
During the meeting, Mr. Brockman and Dr. Sutskever said they would not agree to give Mr. Musk full control over the proposed for-profit, Mr. Brockman said from the stand. Mr. Musk then paused to think for several moments, before standing up, walking around the table and moving in Mr. Brockman’s direction.
“I thought we was going to hit me. I thought he was going to physically attack me,” Mr. Brockman told the court.
Before storming out of the room, Mr. Brockman said, Mr. Musk told them: “When will you be departing OpenAI? I will withhold funding until you decide what you are going to do.”
When Mr. Musk’s lead counsel, Steven Molo, began another round of questions for Mr. Brockman, he showed a email from later in 2017 in which Mr. Musk said he “would unequivocally have initial control” of the proposed for-profit company “but this will change quickly.”
Mr. Molo then suggested to Mr. Brockman that he did not have the business sense to understand that Mr. Musk did not want total control of the company, Mr. Brockman pushed back at the characterization. The two sparred for several minutes, with Mr. Molo pointedly raising his voice and Mr. Brockman responding firmly but calmly.
When Mr. Brockman began his testimony on Monday, Mr. Musk’s lawyers challenged his credibility and tried to show that he was driven by greed. He continued to say that he was primarily motivated by OpenAI’s original mission to build A.I. for the good of humanity.
After Mr. Brockman finished his testimony on Tuesday, Mr. Musk’s legal team showed part of a video deposition by OpenAI’s deputy general counsel, Robert Wu, in which he detailed the agreement from 2019 in which OpenAI created its first for-profit company. This was a “capped profit” company that limited the profits shared with its investors.
Mr. Wu said that the OpenAI nonprofit transferred employees and intellectual property to the capped profit company and retained few employees of its own. He then explained that Microsoft invested a total of $13 billion in the capped profit company.
In the deposition, Mr. Wu also explained that the nonprofit would not have received any funds from the venture until Microsoft and other partners received more than $250 billion in compensation.
(The New York Times has sued OpenAI and Microsoft, claiming copyright infringement of news content related to A.I. systems. The two companies have denied the suit’s claims.)
Cade Metz is a Times reporter who writes about artificial intelligence, driverless cars, robotics, virtual reality and other emerging areas of technology.
The post Elon Musk Wanted OpenAI to Go Commercial, Company President Says appeared first on New York Times.
