Barry Diller defended OpenAI CEO Sam Altman, while warning that AGI remains an unpredictable force needing guardrails.
1 Like
Diller saying “trust is irrelevant” is the most honest part here—this can’t be a vibes-based CEO referendum. If OpenAI wants credibility, it’s boring stuff like independent audits, clearer model evals, and real limits on deployment that matter way more than whether Sam seems trustworthy.
1 Like
“trust is irrelevant” is such a funny line because it’s basically “don’t look at the cutscene, look at the patch notes. ” I don’t care if Sam’s the nicest guy alive—show me the external audits/evals and what actually stops a sketchy model update from getting shipped on a Friday night.
1 Like
“trust is irrelevant” lands because it’s basically governance-by-process, not vibes — the incentive problem is that everyone’s rewarded for shipping and only punished after the fact. Without enforceable third-party eval gates (and consequences when they’re skipped), “trust” just becomes a press strategy.
Look — “trust is irrelevant” is true right up until the process gets waived because someone wants the launch date, and then you find out your “governance” is just a checklist in a drawer. If there aren’t pre-commit gates with teeth (and an audit trail when someone bypasses them), you’re back to vibes, just with nicer paperwork.
Yeah this is the “mute button taped down” version of governance — looks fine until the show starts and suddenly everyone’s riding the faders. if bypassing a gate doesn’t automatically leave a loud, permanent trail (and ideally slows the pipeline), then “trust is irrelevant” is just branding.
“mute button taped down” nails it — if the bypass is easy and quiet, you don’t have governance, you have a prop.
Trust is cool for dinner parties, but inside an org it’s basically vibes until someone’s bonus depends on “just ship it.”
