Playbook
Tasteful Skills
“Tasteful Skills” argues that the best agent skills are not documentation or best-practice lists.
Back to Podcast Digest
The Musk-OpenAI Trial Is Wilder Than Anyone Expected
TL;DR
Musk’s most explosive courtroom moment was admitting xAI used OpenAI distillation techniques “partly” — when OpenAI lawyer William Savitt asked whether Grok was trained using OpenAI model outputs, Musk said “partly,” prompting audible gasps and then defended it as standard AI validation practice.
The lawsuit now looks bigger than founder drama because it could disrupt OpenAI’s IPO path — the hosts argue even a partial win for Musk could slow OpenAI’s momentum, which matters because the company is burning cash fast and may need tens or hundreds of billions if an IPO slips.
Musk’s core jury message is brutally simple: OpenAI ‘stole a charity’ — he testified that he gave roughly $38 million to a nonprofit meant to build AI for humanity, not what he now frames as an executive-enriching for-profit machine, and kept returning to that line even after the judge told him to answer the actual question.
OpenAI’s counter is that Musk wanted control, not purity — Savitt pointed to 2017 discussions about a for-profit subsidiary, Musk’s desire for majority ownership and board control, and even a pitch for Tesla to acquire OpenAI, all to argue this is really a competitor attack.
The stakes reach far beyond Sam Altman and Greg Brockman personally — Musk is seeking their removal and a reversal of OpenAI’s restructuring, and the hosts say that if anything close to that happened it would ripple through every business built on GPT and through the broader AI-heavy capex economy.
The trial has already turned into public spectacle, not just legal process — the judge reportedly scolded both sides for litigating on X while Musk kept posting attacks like “Scam Altman” and “Greg Stockman,” making the whole thing feel, in the hosts’ words, too unbelievable even for a Hollywood script.
The Breakdown
A courtroom bombshell: Musk says xAI used OpenAI distillation “partly”
The clip opens on the line everyone will remember: William Savitt asks whether xAI used distillation techniques on OpenAI’s models to train Grok, and Musk answers, “partly.” That drew audible gasps in the courtroom. Musk tried to soften it by saying using other AIs to validate your AI is standard practice, but the moment landed like a thunderclap.
What Musk is actually asking the court to do to OpenAI
The trial opened in Oakland with a packed courthouse, OpenAI employees inside, protesters outside, and a case that could genuinely alter the AI industry. Musk wants Sam Altman and Greg Brockman removed and OpenAI’s restructuring into a for-profit subsidiary unwound. The hosts frame that not as tabloid drama, but as something that could derail OpenAI’s IPO ambitions and reshape the market around it.
Musk’s story: from idealistic donor to “I was a fool”
On the stand, Musk said he co-founded OpenAI in 2015 as a donation-backed nonprofit for humanity, not a startup to enrich executives. He testified that he gave about $38 million in “essentially free funding” and now sees that money as having helped create an $800 billion company. He broke the relationship into three phases: enthusiastic support, growing doubts, then the conclusion that OpenAI was “looting the nonprofit,” with Microsoft’s $10 billion investment in late 2022 as the turning point.
AI safety becomes part of the pitch to the jury
Musk also leaned hard on safety, saying OpenAI was meant to counterbalance Google. He repeated a striking story about Larry Page supposedly saying it would be “fine” if AI wiped out humanity as long as AI survived, then invoked the “Terminator situation where AI kills us all.” It’s not just legal argument — it’s Musk trying to remind the jury that this began, in his telling, as a mission with existential stakes.
OpenAI’s rebuttal: this is a competitor fight, not a moral crusade
Savitt’s response was to paint Musk as an opportunist trying to weaken a rival. The hosts cite evidence that in 2017 Musk discussed a for-profit OpenAI, wanted majority control and board control, and even pitched Tesla acquiring the company. Savitt also surfaced a Musk email after Andrej Karpathy was hired away to Tesla: “The OpenAI guys are going to want to kill me, but it had to be done.”
The judge, the X posts, and Musk’s “stole a charity” refrain
One of the wildest details is that the judge reportedly scolded both parties for litigating the case on X. Meanwhile Musk was posting lines like “Scam Altman” and “Greg Stockman” and boiling his argument down to one message: “They stole a charity.” According to courtroom reports, he kept returning to that point so often that the judge told him to stop repeating the talking point and answer the question.
Why the hosts think the real story is economic fallout
What fascinates the hosts most is not just the soap opera but the possibility of real downstream damage. Even if Musk doesn’t get the extreme result — removal of Altman and Brockman, a full unwind, or the mentioned $134 billion fine — smaller wins could still slow OpenAI’s path to IPO. Since OpenAI is burning money fast, they argue, any disruption could ripple into private fundraising, data center buildouts, and every company whose products depend on GPT.
Share
Keep Reading
Two Ways to Follow
The Weekly Echo. The inbox-shaped summary of what mattered.
New editorials announced here.
