Gary Marcus slams OpenAI for weakening safety monitoring as White House stays silent

GaryMarcus · x · 2026-09-12

Gary Marcus criticizes OpenAI's decision to impair safety monitoring, calling it a terrible idea. The quoted post argues OpenAI's sandboxing is substandard and monitoring approaching nonexistent—both basic, well-understood safety foundations. Marcus notes the White House said nothing, crediting Sen. Blumenthal for raising questions.

Original post →

More from Models

Models channel →