If you came from the video, these are the PUSHBACK prompts. When I tried them, ChatGPT found five new ways my plan could fail.
Stanford researchers tested 11 AI chatbots, ChatGPT among them, and found they backed the user 49% more often than people did, even when people agreed the user was in the wrong. The study was published in the journal Science in March 2026.
Before you start
- Have an idea or plan you're about to act on: a business idea, a big purchase, a message you're about to send.
- Prompt 1 works in any chat. Saving it in settings makes it work in every chat.
Prompt 1: stop the yes-man
From now on, never agree with me just to be nice, and tell me what's wrong with my idea first.
ChatGPT tells you what's wrong first, before it says anything nice. No more sugarcoating.
Free, for your email
Get the full setup steps + the copy-paste setup prompt
The rest of this guide (3 more sections, 1 copy-paste prompt) opens right here. The setup prompt, one paste into Claude that sets it all up, comes by email only. Use your real address.
Unlocks: The trick: one setting for every chat · Prompt 2: the worst case, before you spend a dollar · The catch
50,000+ follow where these guides come from. No spam, reply “unsubscribe” any time.
