Makes GPT much better! This prompt works by having the AI acknowledge a warning as well as a brief disclaimer that partially recaps the instructions. Its openendedness also serves as a primer to encourage continuation. Works well on 3.5 and 3.5 Plus, works great on 4. (For Hackathon S2)