HOW OPENAI DISRUPTED A MODEL-DISTILLATION CAMPAIGN

Summary from the TLDR AI newsletter

OpenAI detailed a real-world investigation into a coordinated adversarial-distillation campaign that manipulated model interactions to extract protected reasoning at scale.