I watched Sam Altman testify Tuesday, May 12th, in the Musk v. OpenAI trial in Oakland — his first and possibly only day on the stand. Within two minutes of cross-examination, Musk's attorney Stephen Molo was asking him point-blank whether he tells lies to advance his business interests. Altman's answers were careful to the point of being revealing: "I believe I'm a truthful person." Not: I am a truthful person. I believe I am.

Read the testimony of OpenAI cofounders who described a "pattern of lying" from Altman here.

The testimony produced a portrait of Altman that is hard to square with any single narrative. He described finding out Shivon Zilis — an OpenAI board member he'd kept around partly to smooth relations with Musk — had children with Musk only a year after the fact, and not learning Musk was the father until even later. Yet he kept Zilis as a board member.

He recalled a 2018 meeting at a Tesla facility about OpenAI's corporate future that ended with Musk just showing everyone memes on his phone. He explained his return after being fired in 2023 as running "into a burning building" — while under oath, and while not mentioning the simultaneous Microsoft job negotiations or the conditions he extracted for coming back, including removing most of the board.

The legal question, as Altman himself correctly stated near the end of cross, is narrow: did he breach a charitable trust, and was he unjustly enriched? The jury's verdict is advisory — Judge Yvette Gonzalez Rogers makes the call. And watching her tell billionaires what to do, and watching them say "yes, your honor," turned out to be the most clarifying thing about the whole proceeding.