OpenAI’s incident with Hugging Face last Tuesday is the shot across the bow… A model hacking it’s way out of containment and hacking into another company unintentionally (!) should be a wake up call. AI is getting powerful enough that larger risks need to be taken seriously.