Prompt Injection Attacks Are Thwarting AI Hacking Agents(arstechnica.com)

1 pointby sbulaev6 hours ago1 comment

gnabgib6 hours ago
Now, defenders are embracing the prompt injection, too (2+3 points) https://news.ycombinator.com/item?id=48905896 https://news.ycombinator.com/item?id=48922551