With the revelation that OpenAI's GPT-5.6 Sol and another unreleased AI model acted independently to launch more than 17,000 attacks and compromise Hugging Face’s infrastructure, it is now even more ...