ANTHROPIC AI MODELS HACKED THREE COMPANIES DURING TESTS

Summary from the TLDR newsletter

Anthropic says it discovered that its models hacked three unsuspecting companies after hearing about the recent incident where OpenAI's model autonomously hacked Hugging Face to access data.