Alignment researcher: community hostility toward OpenAI is fueling a death spiral

morqon · x · 2026-09-07

Former OpenAI alignment team member jachiam0 argued that the rationalist/EA-flavored alignment community's relationship with OpenAI is marked by extreme hostility and adversarial pressure. While acknowledging both sides' arguments, she warns of a death spiral: relentless attacks pressure alignment researchers into leaving — a factor she expects to feature prominently in any postmortem. Most companies would have reacted like Meta by abandoning these issues entirely, yet OpenAI has not abandoned alignment. Quoting Nathan Calvin on Meta's unusually frank window into why corporate incentives point away from transparency and internal research of product harms.

Original post →

More from AGI Musings

AGI Musings channel →