OpenAI’s president does ‘all the things,’ except answer a question

So far, the strongest witness in Elon Musk’s case against OpenAI has been Greg Brockman’s diary. Brockman himself comes in second.

Brockman was called to the stand in a rather unusual manner – he was first cross-examined, followed by direct examination – and he had serious high school debate club energy. There were a lot of “I wouldn’t characterize it that way,” “I wouldn’t say it that way,” and “That sounds like something I wrote.” Can I see it in context? When Musk’s lawyer, Steven Molo, read some evidence aloud, Brockman pedantically corrected him if he skipped a word, even if that word was “a” or “the.” When asked if Microsoft’s $10 billion investment was OpenAI’s biggest financial event, Brockman said it was the only $10 billion investment. Come on.

I’ve said before that if you can define the word “epistemology”, you shouldn’t testify in your own defense. So the lawyer skipped a word: Is it really worth taking the jury’s time to tell us all this? Except being the smartest boy in the world to your parents.

That would have been bad enough. But the journal entries—a series of text files from his computer—were worse, because they very clearly demonstrated Brockman’s greed and opportunism at least around 2017. Here’s one: “Besides, we also realize that it would be a mistake to steal the nonprofit from him…to convert to a b-corp without him…it would be morally bankrupt and he’s really not an idiot.” » Here’s another: “Maybe we should just go into a for-profit business. Making money for us sounds great and all.” There’s also this: “we can’t say we’re engaged in a non-profit organization. I don’t mean we’re committed. If three months later we do a b-corp, that’s a lie.”

“It would be a mistake to steal the nonprofit” is very close to Musk’s phrase “stealing from a charity,” I note.

We haven’t completed the direct review yet, so I’m sure we’ll hear something exculpatory about the events that inspired these entries. But between Brockman’s attitude toward the cross and the journal entries, I don’t think I’d trust him to watch my bag while I use the restroom.

Musk’s team tries to paint Brockman as greedy, which I accept. The famous “What will get me to $1 billion?” » from Brockman’s journal made an appearance. We established that Brockman’s stake in OpenAI’s for-profit venture was worth approximately $30 billion. Molo asked Brockman why he didn’t donate $29 billion to OpenAI’s nonprofit arm if $1 billion was enough for him.

Brockman could have said something like: “If I got rid of all my holdings at once, OpenAI would receive much less than $39 billion, because that’s how supply and demand works.” He could have said something like, “It’s an important signal to other investors that I can get in on the game.” Or maybe, “That’s just my net worth on paper. It’s not real.”

He didn’t do any of that. Brockman responded with nonsense about the value of the nonprofit’s stake in the for-profit enterprise. Molo said that didn’t answer his question and asked again. We discussed this topic for a while; The heads of the jurors moved back and forth as if they were watching a tennis match. Brockman never answered the question.

No detail was too small to discuss. Molo asked if purple boxes were something OpenAI typically used to draw attention to something important, and Brockman said no. Then we all read in the document that OpenAI typically used them in employee and investor documents to highlight important things. In my notes I wrote, “Why are we fighting about this damn purple box? »

Molo dealt another blow by discussing the various agreements that OpenAI had made with companies in which Brockman had a stake: Cerebras, CoreWeave, Stripe and Helion Energy. Given the sheer number of companies that use Stripe, its OpenAI deal seems paltry — but OpenAI commitments seriously matter to both Cerebras and CoreWeave.

Brockman also has direct financial ties to Altman due to a compensation package he was offered during the launch of OpenAI. He owns a 1% stake in the Altman family office, which Brockman got instead of Y Combinator shares because “we ran out of Y Combinator shares to meet other requirements.” [employees’] offers. » In a 2017 email, Musk’s body man Jared Birchall writes to Musk that Altman revealed this to him, and Musk forwards Birchall’s email to Brockman with a “?” » Apparently, Musk was unaware of the deal and Brockman had to explain it.

I dwell on the various ways Brockman made himself untrustworthy because it’s fun to watch powerful men squirm. But it also influenced my view of his direct testimony, which began afterwards. Brockman began by telling a story of OpenAI’s founding that seemed like it had been polished for a thousand podcasts and keynotes. When asked what he did as president of OpenAI, he replied: “I do everything.” If we hadn’t been in a courtroom, I would have screamed. Millennial vocabulary is a fucking tragedy.

In this story, OpenAI was the idea of ​​Brockman and Sam Altman. Brockman had told Altman of his interest in AI as he was leaving Stripe (“I’m thinking about doing an AI thing,” to which Altman apparently responded, “I’m thinking about doing an AI thing too.”). They stayed in touch. The original idea was allegedly to have a research arm of Y Combinator, which Musk destroyed because he did not want to be affiliated with Y Combinator.

Imagine a montage of cozy dinners, trips to Napa (“our van got stuck in traffic for an hour and a half and no one noticed” because the conversation was so good), conferences on AI. Damn it! It was so cool! Everyone got along so well and had such creative energy! We were treated to a very long story from Ilya Sutskever hesitating to leave Google, and then a photo Brockman took from the first day of OpenAI, with everyone working from their apartments. (Pictured: Altman. Missing: Musk.) I think you understand what I mean; I definitely have Brockman’s. It was Altman and Brockman’s baby. Only after Musk completed closing calls with the team Altman and Brockman had assembled did Musk tell them he wanted to become more involved.

Musk appeared in the testimony as a distant and sometimes threatening figure. At a dinner, he asked if Google’s Denis Hassabis was evil. In fact, Musk seemed “very consistent and obsessed” with Hassabis, and never mentioned Larry Page, who Musk said was the reason OpenAI came to be. In text messages from Sutskever to Brockman, Sutskever wrote: “Elon could spend a half day a week with us. I have imagined how this would go and am concerned that our work environment could become very stressful.”

Sutskever was right to worry; Musk is known for being difficult. I imagine we’ll know more tomorrow. But as things stand, the jury will have to decide which of the two not particularly trustworthy men he trusts more. I don’t envy them this task.