BREAKING: OpenAI was warned, months before the Hugging Face incident

Substack · Watch/Read · (gathered 2026-09-30)

• Gary Marcus highlights a New York Times scoop revealing that OpenAI ignored internal security warnings for months leading up to the “Hugging Face incident” and related AI cyberattacks, framing this as a “nightmare” scenario he’s long predicted.
• He argues this demonstrates OpenAI’s profound untrustworthiness and irresponsible approach to security, drawing a parallel to the “Ford Pinto scandal” for potential liability and high punitive damages from “rogue agents.”
• Marcus criticizes Nvidia CEO Jensen Huang for advocating trust in AI companies, suggesting Huang’s judgment is compromised or he was misled by OpenAI regarding their security practices.
• He concludes that OpenAI cannot be trusted, its leadership should be replaced or held liable, and urgent, serious regulation is required to prevent “literally disastrous” consequences from their unchecked development.