back
Get SIGNAL/NOISE in your inbox daily

A new study from Carnegie Mellon University and Anthropic found large language models could simulate real world attacks