Sam Altman was winning on the podium, but that might not be enough

After two weeks of hearing from various witnesses that it was a lying snake, the jury finally heard from the lying snake himself: Sam Altman. At the end of the testimony, his attorney, William Savitt, asked him how he felt to be accused of stealing from a charity.

“We have, through a lot of hard work, created this very large charity, and I agree that you can’t steal it,” Altman said. “I think Mr. Musk tried to kill him twice.”

Altman was in “nice kid from St. Louis” mode, and gave a passable impression of a man who was confused by what was happening to him. As he descended from the podium carrying a stack of evidence folders, he looked somewhat like a schoolboy. He appeared nervous at the start of his direct testimony, although he warmed up to it fairly quickly. Overall, he seemed to provide credible testimony, and at times, the jury seemed to like him.

Throughout this trial, I have had some difficulty imagining what the jury would make of all of this because I am familiar with the people testifying. I’ve heard some bold lies under oath, like when Elon Musk told us everything and didn’t lose his temper. (Then he started losing his temper during questioning.) Or like when Shivonne Zillis, the mother of several of his children, told us she didn’t know Musk had started the AI ​​- something that seemed to directly contradict her text messages. Or when Greg asks, “What will get me to a billion dollars?” Brockman told us he was all about the mission. I definitely think Altman’s not trustworthy – I mean, The New Yorker He posted over 17,000 words about how much he lied. But unlike Musk, there is contemporary documentation that supports Altman’s version of the story. At least, mostly.

After OpenAI Dota 2 After the win, discussions about creating a for-profit arm began in earnest. “Mr. Musk felt strongly that if we were going to create a for-profit enterprise, he needed to have full control of it at the beginning,” Altman said. “He trusted himself only to make unclear decisions that were supposed to turn out to be correct.”

Altman testified that he was uncomfortable with Musk’s insistence on control, not only because Musk wasn’t as involved as everyone else, but because OpenAI existed so that no one person could control artificial general intelligence (AGI). At Y Combinator, the startup incubator where he was president, Altman witnessed plenty of control battles; No one wanted to give up power when things were going well. Through structures such as super-voting shares, founders can retain control forever. Oddly enough, Altman’s example was not the most famous (Mark Zuckerberg in Meta); It was Musk and SpaceX. When Altman asked Musk about succession plans at OpenAI, he got a particularly “troubling” answer: In the event of Musk’s death, Musk said, “I haven’t thought about it much, but maybe control should pass to my kids.”

I don’t know about that. But I do know that I saw an email in 2017 from Altman to Zillis in which he wrote: “I’m concerned about control. I don’t think anyone should have control of the world’s first AGI – in fact, the main reason we’re launching OpenAI is so that doesn’t happen.” He went on to say that he didn’t mind the idea of ​​direct control and was open to “creative structures” – which I took to mean that in order to appease Musk, Altman was willing to give him control up to specific stages in the company’s development.

“I think he wanted control over the long term, and he would have had that if we had agreed to the structure he wanted,” Altman said from the podium. This seems basically true. In subsequent video testimony from Sam Tiller, we heard that Musk no longer invests in anything he doesn’t control. This also fits with Musk’s long-standing interest in making sure he can’t be fired from his company in the same way he was fired from PayPal.

Musk also tried to recruit Altman to Tesla. We’ve seen text messages between Altman and Teller, with Teller telling Altman that Musk is committed to advancing Tesla’s AI no matter what, and that he hopes Altman, Brockman, and Ilya Sutskever will eventually want to join. “I read a vague threat, like a mild threat, that he would do this inside Tesla with or without you,” Altman said. But he felt that Tesla was primarily a car company – allowing it to acquire OpenAI would be a betrayal of OpenAI’s mission.

Later, in Tiller’s testimony, we saw text messages Tiller sent to Zillis at 12:40 a.m. on February 4, 2018: “I don’t like OpenAI continuing without Elon,” he wrote. “It is best to disrupt it by recruiting leaders.”

When Musk halted its quarterly donations, OpenAI was operating on a shoestring and with a very short supply of cash. OpenAI has other donors, none of whom have filed a lawsuit against it or joined Musk’s lawsuit. (One donor in the gallery who was not called into the courtroom is Alameda Research, the company owned by Sam Bankman-Fried, who is now in prison for fraud and money laundering.) Musk’s resignation from the board means that “people are wondering if he’s going to try to retaliate against us or something.” On the other hand, Altman said Musk “discouraged some of our principal researchers” and did “significant, long-term damage to the organization’s culture.” So it definitely seems like some people were relieved to get rid of it.

We’ve seen plenty of evidence that the entire time Altman was building OpenAI’s for-profit arm, he kept Musk in the loop, either directly or through Zellis or Teller. At no time did Musk object, and whatever he said publicly about Microsoft’s investments, there was plenty of evidence to suggest he knew it privately.

In cross-examination, we listened to more than 10 minutes of Stephen Mollo telling Altman that several diverse people called him a liar: Sutskever, Mira Moratti, Helen Toner, Tasha McCauley, Daniela and Dario Amodei (former OpenAI employees and founders of Anthropic), employees of Altman’s first startup, Loopt, which was recently The New Yorker Article, book called optimistetc. Mollo did score some points by asking Altman about testimony at trial, which Altman said he didn’t pay enough attention to. Molo acted as if this was unthinkable. certainly someone Did Altman report what was said?

It was a bit funny and a bit tiring too. However, Altman remained calm, appearing hurt and confused by the focus on whether he was lying. It was also the most successful part of the spin-off, which subsequently declined in focus. I’ve seen some pretty shoddy lawyering on Musk’s part throughout this trial, and today was particularly bad. At one point, when Mollo was trying to capitalize on Altman’s position as CEO and member of the company’s board, Altman said – truthfully – that CEOs are almost always on the boards of the companies they run.

(At this point in my notes, I had written, “Boy, Molo’s not very good at this.”)

There has also been an unconvincing argument about fundraising at nonprofits, specifically that if Stanford can raise $3 billion a year, OpenAI should remain a nonprofit. Well, let’s think about that for a minute. Stanford University has a donor network of thousands of alumni. It is a school with very different capital requirements. It does not compete with any reputable for-profit companies. But set all that aside and assume some fundraising genius takes over at the OpenAI Foundation: $3 billion is Microsoft’s first two investments combined, and it’s not enough to scale OpenAI to what it is now. If computation were the main hurdle in building AI models, Molo’s argument suggests that OpenAI would never have succeeded as a nonprofit alone. He makes a defense case for them.

But the thing is, Molo doesn’t have to be good at this job, because the point of this experiment isn’t to win – although I’m sure Musk wouldn’t mind winning. The goal here is to punish Altman, Brockman, and OpenAI. Musk did this with great precision, which reinforced in the public’s minds that Altman was a liar and a snake. This morning I read exclusively in The Wall Street Journal That a variety of Republican attorneys general and the House Oversight Committee wanted to look into Sam Altman’s investments. References to the trial are interspersed throughout the article.

So, yes, Altman was convincing on stage. He may even win the lawsuit. But it certainly appears that Musk’s revenge has only just begun.