These days, some of tech’s most important decisions are being made inside courtrooms. Google and Facebook are fending off antitrust accusations, while patent suits determine how much control of their own products they can have. The slow fight over Section 230 of the Communications Decency Act threatens platforms like Twitter and YouTube with untold liability suits for the content they host. Gig economy companies like Uber and Airbnb are fighting for their very existence as their workers push for the protections of full-time employees. In each case, judges and juries are setting the rules about exactly how far tech companies can push the envelope and exactly how much protection everyday people have. This is where we keep track of those legal fights and the broader principles behind them. When you move fast and break things, it shouldn’t be too much of a surprise when you end up in court.
The two companies’ famed 2019 contract was made public as part of the Musk v. Altman trial exhibits. The 36-page agreement defines artificial general intelligence as “a highly autonomous system that outperforms humans at most economically valuable work.”
Every three months, we get another round of financial updates from publicly traded companies like Alphabet, Meta, and Amazon, but the Trump Administration is pushing to lower that requirement to every six months. The argument in favor of changing it claims that quarterly reporting is a burden on small or medium-sized companies, despite the risks of weaker monitoring, as Reuters and WSJ report.
Wu is still identifying documents. Two jurors have some real thousand-yard stares going on.
We are now looking at assorted legal documents with him. We have determined the board has approved the agreement of 2018, which is the initial formation of the for-profit OpenAI LP. OpenAI Inc. contributed assets and in result got the limited partner interest and “residual interest.” This seems to be mostly about reading stuff into the record.
Molo asked a very narrow question about texts or emails about removing Musk from the board. Then he accused Brockman of making up the explanation after the fact because of the trial. OpenAI’s lawyer pulled out two sentences from the same entry: “real decision is fire Elon” and “We seem converged on the ‘fire Elon’ route.” Pretty good from OpenAI’s lawyers, and very annoying / misleading from Molo.
For instance, yesterday, Brockman gave a long spiel about the setup of OpenAI that only tangentially involved Musk. Altman and Sutskever were portrayed as being in the middle of things. Molo effectively said that and then pointed out that Musk’s money was key. Brockman objected both to the summary, and to the idea that Musk’s money was key. Come on, dude. We aren’t idiots.
It’s not clear to me why Musk’s ultimatum matters? Musk says he won’t continue funding until he gets a firm commitment. Brockman testified he never made that firm commitment, and Musk didn’t resume his quarterly payments. There were clear continuing negotiations about how to raise money, including whether Tesla should get control of OpenAI and whether there should be an ICO after this email. What are we doing? I am at this point genuinely lost.
We have been talking about whether or not Greg Brockman made an initial investment into the for-profit. If we all agree that the for-profit has a different investment base than the nonprofit, I don’t get what the conflict here is? The nonprofit is effectively an investor in the for-profit, though the investment was in kind. The nonprofit still exists. Right now I am trying to figure out whether I find Brockman or Musk more reliable, and I think the answer is, it’s going to depend on who is most supported by other witnesses and documents.
The problem is that the chronology of Brockman’s negotiations with Musk is now working against him. We have multiple emails and texts with Musk proposing that he be in charge, followed by Musk withholding funding unless Brockman and Ilya Sutskever did what he wanted. Brockman might be motivated by profit — I certainly think so! — but Musk’s actions are what they are: high-pressure negotiating tactics that failed. It seems to me that Musk’s lack of involvement with OpenAI’s for-profit is self-inflicted.
“Did you practice them?” Molo is really not into the various ways that Brockman is objecting to the wording of his questions. I’m not sure how sympathetic the jury is to either of them. They seem pretty stone-faced.
Brockman is being really squishy about this — I mean, he did say that he thought Musk was going to hit him at one point. Molo is I think trying to make the point that Musk was being “hardheaded” in negotiations. He’s also saying that Brockman wasn’t familiar with corporate governance. And he is, notably, raising his voice at Brockman.
Greg the Bard, who told long-winded stories about how wonderful OpenAI is, is gone. “False assumption baked into the question” is the go-to currently.
Musk’s lawyer, Molo, is saying that first of all, Brockman actually did the layoffs, and second of all, a memo prepared by Ilya Sutskever contained criticism of Brockman as a manager. Shortly after the memo was prepared, Brockman was removed from the board.
Microsoft is low-key my favorite part of this trial? Their opening statement was just like, an ad for Microsoft products. Almost all questions, including of Brockman, are like, “Did Microsoft play a role in the founding? Did it participate in the creation of a for-profit entity?” (No, obviously.) It is so funny, like they are going to at any moment turn to the jury and say, “Now why are we in this?”
Brockman says he was told he was removed from the board in an eight-minute call, and not given a reason why. He was also told Sam Altman was fired. Shortly after, he quit. “The board’s actions felt wrong to me. Wrongly conceived, wrongly executed.” He thought that he’d start a new AI company with Sam Altman. Several other people quit alongside Brockman and they started planning their new company. Satya Nadella called to ask if something terrible had happened, and Brockman said no. So then being a startup within Microsoft was an option too. But ultimately, as we all know, Brockman went back to OpenAI.
She did not disclose her romantic relationship with Musk to Greg Brockman. She gave birth to twins while still a board member at OpenAI, and found out their father was Elon Musk through “public reporting.” She did, eventually, say the twins were via IVF and her relationship with Musk was “entirely platonic.” She stayed on the board because Sutskever, Altman, and Brockman trusted her — other board members wanted her gone.
because he said, essentially, he saw no path at OpenAI and was going to work on AGI at Tesla. But during a Q&A period, he said that he wasn’t going to work on safety, and just focus on catching up with DeepMind. That generated a strong, negative reaction.
Every email to Musk starts with glowing compliments, and at times what even reads like groveling. I don’t know if the jury is noticing this, but I certainly am. I wonder how much ass-kissing Musk is accustomed to — probably a lot.
Some more successfully than others. While I’m willing to believe the “steal the charity” line is about booting Musk from the board, the idea that “making the money for us sounds great and all” was about unsuccessful fundraising was somewhat less believable. One thing Brockman is making clear is that Musk was mercurial — not a surprise. “Elon’s engagement depended directly on how likely he thought it was we’d succeed,” he says. By January 31st, 2018, Musk says he thought OpenAI was “on a path of certain failure relative to Google” in an email.
“Ilya and myself” decided not to remove him from the board because it felt right for the mission but wrong personally. By that point, Musk was trying to get them all to join Tesla. This makes the journal entry about “Ilya feeling like we morally should not be kicking elon out, and should be trying to make the non-profit work” as well as “it’d be wrong to steal the non-profit from him” read really differently.
Greg Brockman is explaining that Musk put conditions on his continuing donations at OpenAI, which he did not accept. Then Musk said they should merge OAI into Tesla. They’d get the money, a billion-dollar-per-year budget, and it’d grow from there. The work would have to be secret — that would be a requirement to make it happen.
“I truly thought he was going to physically attack me.” Musk was angry that no one wanted to agree for him to have majority equity. As he was storming out of the meeting, Musk asked Brockman and Sutskever when they planned to leave OpenAI. They were confused. Then Musk said, “I will withhold funding until you decide what you are going to do.” He then stopped his promised quarterly donations to OpenAI.
Brockman is discussing an intense period of negotiations between him, Ilya Sutskever, Sam Altman, and Musk. Musk wanted unilateral control. He also wanted a lot of equity because “he needed the money for Mars, he needed $80 billion to create a city there,” Brockman says. During the negotiations, Musk stopped his quarterly donations.
between Altman, Musk, Brockman, and Sutskever in August 2017. It was the first proposal for a for-profit. Musk rejected this. “At the end of the meeting he said, ‘You guys are great but I could start another AI company tomorrow. One tweet is all it takes,” Brockman said.
He started keeping it in 2010. He describes it as stream-of-thought, jotted notes, and disorganized thoughts that sometimes contradicted each other. “It’s very painful” to see the journal in this case, Brockman says. These were “very deeply personal writings that were never meant for the world to see but there’s nothing there I’m ashamed of.”
Musk told Brockman he tried to get Gates to donate four times, and Gates didn’t so much as come by the office. One wonders whether Musk’s persistent shit-talking of Gates had anything to do with Gates’ reluctance.
I love a sidebar. We have static. There’s some evidence OpenAI is trying to introduce that Musk’s team apparently doesn’t want, about Brockman’s investment in Cerebras.
Five to 10 people were laid off after Musk demanded a list of people with their contributions by their names. It had a “significantly negative” impact on morale and made recruiting more difficult.
Elon Musk requested that they come help. “It was pretty clear this was not something we could say no to,” Brockman says. Brockman claims this was something he worried about when it came to joining OpenAI.
So over the course of several months, the OpenAI group worked on self-driving. One of those engineers, Andrej Karpathy, permanently joined Tesla afterward. “I have an apology and a confession,” Musk said. “I made an offer to Andrej to run autopilot and he accepted.”
We have some matters to take care of before we get the jury back.
DeepMind’s Demis Hassabis was a constant figure of fear among Musk and other OpenAI higher-ups.
App analytics firm Kochava and its subsidiary, Collective Data Solutions, will be prohibited from “selling, licensing, transferring, sharing or disclosing” sensitive location data without express consent from consumers. The ban settles the FTC’s lawsuit alleging that Kochava sold sensitive geolocation details that could track people seeking or performing abortions.
While Elon Musk has settled his Twitter case with the SEC, nonprofit watchdog group Media Matters says the FTC has ended a “retaliatory” investigation about whether it worked with advertisers to boycott Elon Musk’s website:
…the FTC ultimately submitted to a legally binding settlement agreement with Media Matters, withdrawing its blatantly retaliatory demands and committing to forgo ever reissuing or issuing a substantially similar Civil Investigative Demand to Media Matters.
The FTC also stated — in writing — that Media Matters is not the target of any investigation and that any similar future litigation would occur in D.C.
No detail was too small to argue over for Greg Brockman.
Just before they left, Jared Birchall’s testimony regarding the funding of the bid to buy OpenAI — the subject of last week’s drama — was struck.
In Brockman’s telling, Altman was around a lot more often than Musk. The gee-whiz energy here is off the charts, including a pre-launch story about the group being stuck in traffic for an hour and a half and not noticing because they were having such a good time.
