Hacker News new | ask | show | jobs
by slipwalker 66 days ago
I set my "personal instructions"/customization for any LLM tool to highly adversarial mode. Explicitly confront and question every word i type.