Nuseir Yassin shares a prompt that makes AI give you an answer, then argue against it. He appends "Give me the answer, then destroy your own answer" to force the model to show both sides instead of just echoing what you want to hear. He tested it on an argument with his girlfriend and got the AI to flip positions entirely.
This is the most important AI prompt. If you put it in ChatGPT, your life will change. It's just one line. Right now, AI has a big problem. It only agrees with you. It confirms what you already believe. So here is the fix. Every time you ask AI something, add this line at the end of your prompt: Give me the answer, then destroy your own answer. This Changes everything. Now you get two answers, one that agrees with you and another that completely disagrees. I tried this during an argument with my girlfriend. I asked AI for advice. It agreed with me and it said I was right. Then I said, "Destroy your own answer," and suddenly it took her side. And the truth was in the middle between the first and the second answer. So the next time you use AI, remember: ask for an answer, then ask it to destroy its own answer. See you next week.
