Live updates from Elon Musk and Sam Altman’s court battle over the future of OpenAI

About 200 people work on safety at OpenAI.

Kolter laid out OpenAI’s different safety groups: the safety systems team, which works on guardrails and evaluations; the preparedness team, which deals with OpenAI’s preparedness framework; the alignment team, which helps train models on ways that “align with human values”; the model policy team, which develops the model spec; and other teams focusing on investigations. When speaking about the controversial dissolution of OpenAI’s superalignment team and AGI readiness team, he said some of that research is being done by other teams.

The chair of OpenAI’s safety and security committee said they’ve formally delayed its model releases.

Dr. Jeremy “Zico” Kolter said that so far, there have been two times when the committee “formally requested a delay of models.”

Irritatingly, no one has asked him why he’s called “Zico.”

Anyway, he’s a member of the nonprofit board of the OpenAI foundation, but not the for-profit. He’s a safety expert.

Microsoft establishes that OpenAI has other investors...

And that Musk hasn’t sued any of them. Softbank’s investment is bigger — so is Nvidia’s and so is Amazon’s. That’s all from Microsoft.

We see the Musk “bait and switch” texts again.

In response, Altman texted, “I agree this feels bad — we offered you equity when we established the cap profit, which you didn’t want at the time but we are still very happy to do any time you like.” Molo tries to ask if this is a bribe — if Altman is trying to say that if Musk lets him get away with stealing from a charity, he’ll split the loot. Savitt objects, YGR sustains.

What if we had a drinking game for this trial?

Granted we’d all be dead as a result because who can keep up with this but... Drink every time:

Musk says, “This is a bait and switch” in a October 2022 text chain.

Molo says, isn’t he saying you stole a charity? Altman looks confused. “No?”

Molo is not doing especially impressive lawyering here.

It’s funny that he’s on the team alleging money overrode ethics, because, well, I guess it takes one to know one? Throughout the case, we’ve seen a number of own goals from the Musk team. This morning, when the Musk team was requesting to ask questions about safety concerns with AI, YGR snapped, “What else do you think you want to do? Because you do not want to be held in contempt I guarantee you.” She’s tough with lawyers as a general rule, but woof.

Molo asked Altman if he’d ever fire himself as CEO of the OpenAI for-profit.

“I have no current plans to do so,” Altman replied, adding, “I’ve never thought about it before.”

“The blip” again.

Y’all I am so sick of this. We are literally litigating a week in 2023.

Well, I do love a long inquiry into the linear nature of time.

Molo is asking about a series of terms on the Microsoft deal that were developed post 2020. None of them are on the 2018 term sheet because they hadn’t been negotiated yet. Fascinating stuff from Molo — is this really the best use of his time?

The difference between Musk and Altman on cross is really stark.

While Musk was ready to get into a fight over anything and everything, Altman has rather mildly answered every insulting question Molo has asked him. Molo has just accused Altman of lying on his direct examination about Musk trying to tuck OpenAI into Tesla. In his deposition, Altman says he’s not sure one way or the other about that and then asks to scroll to the top of the deposition for context. Molo says no. A juror smiles. I expect I know what we’re going to see on the redirect.

Ronan Farrow’s article is brought up.

OpenAI’s lawyers have objected to the question, which is about the New Yorker article which “basically calls you dishonest,” Molo says. YGR would like us to move on from all the people who have called Altman a liar. It’s been more than 10 minutes of this. I think everyone got the point.

This cross is spicy!

Sam Altman is responding to being repeatedly called a liar, both in this trial and elsewhere. We’re just hearing a list of people who have called Altman a liar or a schemer, including the Anthropic cofounders, Dario and Daniela Amodei. Altman just seems confused.

Mr. Molo is going directly in at Altman: “Do you always tell the truth?”

“I believe I am an honest and trustworthy businessperson,” Altman says. We are now hearing about Ilya Sutskever’s testimony that he thought Altman was dishonest, and Mira Mirati’s as well. Altman is responding to this with confusion, seems hurt, and is speaking very softly.

“If I knew how difficult and painful this was going to be, I never would have tried,” Altman said.

He also says it’s the most meaningful thing in his life besides his family. It’s “awesome and fulfilling,” Altman said. He also testifies he is still enthusiastic about the nonprofit structure because it’s now one of the largest nonprofits in the world. “Mr Musk did try to kill it, at least twice.” We are now about to witness the cross.

We are now talking about Altman’s investments.

He says he never received any money from OpenAI’s startup fund. “I temporarily held the gp position because as the only person on the executive team without OpenAI equity, if anyone else had that… it would’ve caused adverse tax consequences.” He says he’s recused from any related-party transactions and let the boards of the relevant companies decide what to do. This is followed by some PR talk about the OpenAI foundation’s Alzheimer’s work. Okay.

“I had poured the last years of my life, and I was watching it be destroyed,” Altman said.

He was tempted to go to Microsoft to work on pure research effort, and he felt very angry. “I’m sure I could have made a ton of money and had a much easier life at Microsoft but I cared about the mission and the people,” he said. So he returned. As for the board, he says “I feel badly for the misunderstandings” with the board.

“I was in this like fog of war, I didn’t know what was going on,” Altman says of what happened next.

People started quitting OpenAI. Altman went to Brockman’s house and wanted to figure out a way to stabilize OpenAI. He had calls with board members about coming back to OpenAI. “Although I was still very angry” and thought he’d have to “clean up a mess I didn’t make,” he said he was willing to come back under other conditions.

We are now onto “the Blip.”

All the board would tell him was that he wasn’t consistently candid and they weren’t going to get into why. Altman was completely shocked, he says. He also told them that their plan to announce it via a blog post would throw things into chaos. “If this is the decision, this is a terrible way to execute it,” Altman said he told them. The board told him it was too late to stop it.

OpenAI has raised “approximately $175 billion” in investment, Altman says.

2023 was “the beginning of the inflection.” ChatGPT had been introduced, and “it became clear to us we would need a lot more compute.” They needed it for both research and for the models being used by the public. Around then, Shivon Zilis resigned from the OpenAI board. Shortly after that, Musk announced xAI, and Altman says there were”a lot of efforts to recruit our employees” and “negative tactics from Mr. Musk toward us.” Musk’s lawyers don’t like this but over their objections, Altman “started to hear rumblings” about litigation.

Altman seems to be getting into his testimony...

...and it’s a little catty. (I live!) At one point, Altman says that Zilis told him Musk had “front-runner-itis” — but there’s an objection that stops Altman from telling the rest of this story. Altman looks slightly disappointed. We then hear that Zilis advised Altman on how to engage Musk so that Musk wouldn’t “bash us on Twitter.”

Musk didn’t invest in the OpenAI for-profit because “he was no longer going to invest in any startups he did not control.”

Musk didn’t raise any objections, though. And then he sent the infamous message where he rated OpenAI’s chances as zero. Altman appears to be concentrating hard on his testimony but is coming across as being a little bewildered about why he is here at all — but maybe that’s just how his eyebrows look at all times.