The most compelling witness for Elon Musk's legal case against OpenAI has turned out to be a journal. Greg Brockman's journal. And Brockman himself, now on the stand, is doing his best to make everyone forget that.
OpenAI's president was called to testify in an unusual order — cross-examination before direct — and he came prepared with the energy of a high school debater who read the rulebook twice. Every other answer was some variation of "I wouldn't characterize it that way" or "That sounds like something I wrote, but can I see it in context?" When Musk's attorney Steven Molo read passages from evidence aloud, Brockman would stop him to correct a skipped word. Even if that word was "a." Even if that word was "the."
It was a performance that said a lot without saying much at all. Brockman clearly came prepared to give as little ground as possible. Whether that strategy is working is another question.
The trial centers on Musk's lawsuit against OpenAI, the company he co-founded and later departed. Musk has argued that OpenAI abandoned its original nonprofit mission in pursuit of profit, and that its close relationship with Microsoft represents a fundamental betrayal of what the organization was built to be.
Brockman's journal entries have surfaced as some of the more telling pieces of evidence in the case, offering a contemporaneous window into early decisions at the company. That the journal's author is now on the stand trying to re-contextualize those same entries adds a certain tension to the proceedings.
The trial is still ongoing. What's clear already is that whatever OpenAI's legal team had planned for Brockman's testimony, his own past writing keeps getting in the way.
Full coverage of the trial is available at The Verge.
