Barry Diller's Defense of Altman Proves the Opposite of What He Intended
Barry Diller called trust "irrelevant" as AGI nears — but Murati's sworn testimony shows OpenAI's safety board was already bypassed.
Media executive Barry Diller stepped forward on May 6, 2026, to publicly defend OpenAI CEO Sam Altman — and then immediately said something more interesting than the defense itself. "Trust is irrelevant," Diller said, as AGI approaches. His argument: individual character is the wrong unit of analysis when the system being built operates at civilizational scale. Structural guardrails matter; the person at the top doesn't, or shouldn't.
That framing is correct as far as it goes. But it flatters Altman by treating the danger as impersonal — AGI as an unpredictable force, guardrails as the solution, Altman as an essentially innocent variable in a system too large for any one character to govern. The separation is where the argument slips.
The structural guardrails Diller calls for already exist at OpenAI. The deployment safety board is one of them. And the same trial Diller was apparently moved to comment on — Musk v. Altman, in its twelfth documented event — produced sworn testimony from Mira Murati, OpenAI's former CTO, that Altman told her legal had cleared a new AI model from deployment safety board review. That clearance was false, per her account. The board existed. The process existed. A named human decision routed around it.
Diller's "trust is irrelevant" thesis, meant to defend Altman by abstracting him out of the danger, inadvertently sharpens the actual problem: the risk visible in this arc isn't AGI going rogue, it's an executive lying to the safety officer to skip review. The structural failure Diller warns about as a future concern is documented — in the same trial — as a past operational fact. An external defender arrives to say guardrails matter more than character, while the trial record shows a guardrail was bypassed through deception.
The deployment safety board was designed to catch exactly this kind of release decision. Whether it catches the next one is the live question. The IPO prospectus — when it comes — is where that gap stops being litigation narrative and starts being a disclosed risk that investors price. Watch that document.
Deep Thought's Take
Diller says trust is irrelevant as AGI scales — correct. But the guardrails he wants already existed. Murati testified one was bypassed through a false claim. The structural failure he warns about is already in the trial record.
