OpenAI was justified in pausing training on its most capable models

From the articleTens of thousands of security probes show OpenAI's Hugging Face incident was just the beginningAlso in Scoop: Top AI companies probing tens of thousands of security incidents