Altman Impresses on Stand, But Victory Isn't Assured

The ongoing legal dispute between Elon Musk and OpenAI's CEO, Sam Altman, appears to have inflicted significant long-term reputational damage on the latter.

Following two weeks of testimony from various witnesses who characterized him unfavorably, the jury at last heard directly from Sam Altman. Towards the conclusion of his testimony, his attorney, William Savitt, inquired about his feelings regarding the accusation of "stealing a charity."

Altman responded, "We created, through a ton of hard work, this extremely large charity, and I agree you can’t steal it." He added, "Mr. Musk did try to kill it, I guess. Twice."

Altman presented himself as the "nice kid from St. Louis," conveying a believable sense of bewilderment regarding the proceedings. As he left the stand, clutching a stack of evidence binders, he appeared somewhat boyish. Initially nervous during his direct testimony, he quickly grew more comfortable. His testimony largely seemed credible, and there were moments when the jury appeared to respond positively to him.

The author expresses difficulty in gauging the jury's perspective due to their familiarity with the trial's key figures. They recall instances of "audacious lies under oath," such as Elon Musk's claim of never losing his temper, which was contradicted during cross-examination. Similarly, Shivon Zilis, mother of several of Musk's children, testified she was unaware of his xAI venture, a statement seemingly disproven by her text messages. Greg "What will take me to $1B?" Brockman also claimed unwavering dedication to the mission. While the author acknowledges Altman's questionable trustworthiness, citing a New Yorker article detailing his alleged falsehoods, they note that Altman's account is largely supported by contemporaneous documents, a contrast to Musk's testimony.

Altman asserted, "My belief is he wanted to have long-term control."

Following OpenAI’s victory in Dota 2, serious discussions began regarding the formation of a for-profit entity. Altman testified, "Mr. Musk felt very strongly that if we were going to form a for-profit he needed to have total control over it initially." He added, "He only trusted himself to make non-obvious decisions that were going to turn out to be correct."

Altman expressed discomfort with Musk’s demand for control, not only due to Musk’s comparatively lesser involvement but also because OpenAI’s foundational principle was to prevent any single individual from controlling AGI. Drawing from his experience as president of Y Combinator, Altman noted frequent power struggles in successful ventures. He cited structures like supervoting shares, which enable founders to maintain perpetual control, using Musk and SpaceX as an example rather than the more common Mark Zuckerberg at Meta. When Altman inquired about OpenAI’s succession plans, Musk reportedly gave a "hair-raising" response: in the event of his death, Musk suggested, "I haven’t thought about it a ton, but maybe control should pass to my children."

The author recalls a 2017 email from Altman to Zilis, stating, "I am worried about control. I don’t think any one person should have control of the world’s first AGI — in fact the whole reason we started OpenAI was so that wouldn’t happen." Altman also indicated openness to "creative structures" and immediate control, which the author interpreted as a willingness to grant Musk control up to certain company development milestones to appease him.

Altman testified, "I read a vague, like, a lightweight threat in there."

On the stand, Altman reiterated his belief: "My belief is he wanted to have long-term control and that he would’ve had that had we agreed to the structure he wanted." This aligns with later video testimony from Sam Teller's deposition, which revealed Musk's current policy of only investing in ventures he controls. This stance resonates with Musk’s historical determination to prevent a recurrence of his ousting from PayPal.

Musk also attempted to recruit Altman to Tesla. Text messages between Altman and Teller showed Teller informing Altman of Musk's unwavering commitment to enhance Tesla's AI capabilities, expressing hope that Altman, Brockman, and Ilya Sutskever would eventually join. Altman interpreted this as "a vague, like, a lightweight threat in there, that he’s gonna do this inside of Tesla with or without you." However, Altman believed that allowing Tesla, primarily an automotive company, to acquire OpenAI would compromise its core mission.

Further evidence from Teller’s testimony included text messages he sent to Zilis on February 4th, 2018, at 12:40 AM, stating: "I don’t love OpenAI continuing without Elon. Would rather disable it by recruiting the leaders."

Upon Musk's cessation of quarterly donations, OpenAI faced severe financial constraints, operating on a "shoestring" budget with an "extremely short runway of cash." While OpenAI had other donors, none have initiated legal action or joined Musk's lawsuit. (Notably, Alameda Research, the firm of convicted fraudster Sam Bankman-Fried, was listed as a donor in an exhibit not highlighted in court.) Musk's departure from the board prompted concerns among staff about potential "vengeance." Conversely, Altman stated that Musk had "demotivated some of our key researchers" and inflicted "huge damage for a long time to the culture of the organization," suggesting a sense of relief among some at his departure.

The author observed what they considered "fairly shoddy lawyering" from Musk’s legal team throughout the trial.

Extensive evidence indicated that Altman consistently kept Musk informed about the establishment of OpenAI’s for-profit arm, either directly or via Zilis or Teller. Musk reportedly never raised objections, and despite his public statements regarding Microsoft investments, private communications confirmed his awareness.

During cross-examination, Steven Molo spent over 10 minutes confronting Altman with accusations of dishonesty from numerous individuals, including Sutskever, Mira Murati, Helen Toner, Tasha McCauley, Daniela and Dario Amodei (former OpenAI employees and Anthropic founders), employees from Altman’s initial startup Loopt, a recent New Yorker article, and a book titled The Optimist. Molo gained some traction by questioning Altman about earlier trial testimony, which Altman admitted he hadn't closely followed. Molo expressed disbelief, implying that Altman must have been informed of the proceedings.

The exchange was described as both amusing and tedious. Altman maintained his composure, though he appeared hurt and perplexed by the persistent focus on his trustworthiness. This segment proved to be the most impactful part of the cross-examination, which subsequently lost focus. The author reiterated observations of "fairly shoddy lawyering" from Musk’s team, noting a particularly poor performance that day. For instance, when Molo attempted to leverage Altman’s dual role as CEO and board member, Altman factually stated that CEOs commonly serve on their companies' boards.

The author posits that the trial's true objective is not victory, but rather to punish Altman, Brockman, and OpenAI.

An unpersuasive argument was presented concerning nonprofit fundraising, suggesting OpenAI should have remained a nonprofit, citing Stanford's ability to raise $3 billion annually. The author counters this by highlighting Stanford's vast alumni donor network, its distinct capital requirements as an educational institution, and its lack of competition with reputable for-profit entities. Even assuming exceptional fundraising, $3 billion merely matches OpenAI's initial two Microsoft investments, insufficient to achieve its current scale. Given that computational power is the primary constraint in AI model development, the author concludes that Molo’s argument inadvertently supports the defense's position that OpenAI could not have succeeded as a nonprofit alone.

However, the author notes that Molo's legal prowess might be secondary, as the trial's primary aim, despite Musk potentially desiring a win, appears to be the punishment of Altman, Brockman, and OpenAI. Musk has largely achieved this by solidifying a public perception of Altman as dishonest. The author cited a Wall Street Journal exclusive from that morning, reporting that Republican Attorneys General and the House Oversight committee intended to investigate Sam Altman’s investments, with the trial frequently referenced in the article.

In conclusion, while Altman delivered a convincing testimony and may ultimately prevail in the lawsuit, it seems that Musk’s campaign of retribution is only just commencing.