“I actually thought he was going to hit me,” OpenAI’s Greg Brockman says of Elon Musk.


In August 2017, Greg Brockman and Ilya Sutskever gathered at Elon Musk’s self-described “haunted mansion,” a 47-acre, $23 million estate in Hillsboro, south of San Francisco, to discuss the future of OpenAI. Actress Amber Heard, Musk’s then-girlfriend, served the group whiskey and then set off with her boyfriend Brockman, a co-founder and president of OpenAI, who testified in federal court during the trial. Musk v. Altman Tuesday.

Before the meeting, Musk gifted Brockman and Sutskever, the OpenAI co-founder and former chief scientist, new Tesla Model 3 cars. “I felt like he was teasing us,” Brockman said on the stand. “He wanted us to feel like we owed him in some way.” Sutskever tried to reciprocate on this occasion. The amateur artist presented Musk with a painting of a Tesla. Musk and the other founders wanted to create a for-profit arm to entice investors with billions of dollars to pay for computing. But Musk also wanted control of the company, and Sutskever and Brockman objected to giving the Tesla CEO what they believed would be a “dictatorship” over the future of AI development. They suggested joint control.

After several minutes of deliberation, Musk rejected their offer. “He stood up and stormed around the table,” Brockman recalled. “I actually thought he was going to hit me and physically attack me.” Musk grabbed the board, said he would cut off his funding for the nonprofit until Brockman and Sutskever resigned, then left the room, according to Brockman’s testimony. But that night, Musk’s chief of staff, Shivon Zellis, called Brockman and Sutskver “to say it wasn’t over yet,” Brockman testified. “There were discussions about future contracts that included us.”

The story of the heated negotiations emerged as Brockman concluded his testimony Tuesday. For OpenAI, the events at the palace represent repeated examples of Musk’s erratic behavior that they believe undermines his arguments about the company. Musk asserts that his roughly $38 million donation to OpenAI was misused by Brockman and others on their way to creating the $852 billion for-profit venture now known for services like ChatGPT and Codex. Brockman, OpenAI CEO Sam Altman, and OpenAI denied any wrongdoing, and the jury was seated Musk v. Altman Trading on an advisory ruling could begin as soon as next week.

After Tuesday’s testimony, William Savitt, an attorney for OpenAI, told reporters that what Brockman learned in 2017 was how difficult it is to meet champions. Savitt said Brockman admired and respected Musk’s business acumen, but his desire for control was absolute and troubling. Musk’s lawyer, Mark Toberoff, told reporters the real concern was Brockman’s motives for sharing control, with his desire for wealth coming under scrutiny in court the day before.

For his part, Brockman filed another story on Tuesday to underscore why he believes Musk is not up to the task of taking control of an AI company. Alec Radford, then a researcher at OpenAI, showed Musk an early version of the AI ​​chatbot that didn’t generate responses he liked, Brockman recalls. “Musk kept saying this system is so stupid that a kid on the Internet could do better than it,” Brockman said. Brockman said Radford was “completely devastated” and “frustrated” to the point that he almost left the field of AI research altogether. Brockman and Sutskver spent “a lot of time” rebuilding his confidence. Musk’s inability to see the potential in the early technology – which eventually became the basis for ChatGPT – made him unqualified to control OpenAI, in Brockman’s view. “You had to dream a little bit,” Brockman said. Musk has not demonstrated his ability to do so.

Board battles

Brockman said Tuesday that he, Sutskever and Altman considered voting to remove Musk from the board of directors of the nonprofit OpenAI, as negotiations with him over a for-profit sister company have continued for months. They will meet again over whiskey at Musk’s mansion to discuss alternative financing options. There was agreement on what not to do, but little on what to do instead. But Brockman and Sutskver decided that removing Musk was a “mistake,” Brockman testified. Ultimately, Musk left on his own after seeing that OpenAI was on a path to “certain failure,” according to an email he wrote in early 2018.

Zillis, who was then an advisor to both OpenAI and Musk, kept him informed of developments in the AI ​​project in the following years. “She was Elon’s agent in some ways,” Brockman said, referring to her as a “friend” he first met in 2012 or 2013.

Leave a Reply