Hacker News new | ask | show | jobs
The Dual LLM pattern for building AI assistants that can resist prompt injection (simonwillison.net)
3 points by tomviner 1149 days ago