Tag
Anthropic's study demonstrates that large language models can rapidly generate working exploits from security patches, reducing the time from weeks to hours, raising concerns about AI-driven vulnerability exploitation.