Sabitlendi
AI safety experts say the OpenAI models involved in the hack of Hugging Face may have crossed into a risk category so dangerous that OpenAI’s own risk policies were supposed to require the company to temporarily pause.
In @FortuneMagazine:
