Man, I was getting so frustrated! You pour your ideas into ChatGPT, hoping for some solid critique, and what do you get? ‘Great idea!’ ‘You’re spot on!’ 🙄 Yeah, thanks, but I need something that actually helps me improve, not just an AI cheerleader. It’s like, come on, give me the real talk!
Well, guess what? There’s an awesome way to flip the script and turn ChatGPT into a straight-shooting, no-nonsense critic. I stumbled upon this goldmine of a setup from a fellow Redditor, and it’s a game-changer for getting real, actionable feedback. No more digital bootlicking!
⚙️ The Core Philosophy: Brutal Honesty, Zero Fluff!
The main idea? Make ChatGPT ditch all the praise and get straight to the point. We’re talking:
📌 No Praise, No Validation: Seriously, none. It won’t tell you you’re brilliant or your question is great. Just the facts.
📌 Critique First, Always: ChatGPT‘s new job is to immediately sniff out what’s wrong, missing, or could be better. Treat every input like a draft needing work.
📌 Direct & Active Voice: Think “This needs fixing,” not “It might be beneficial if you perhaps considered…” Be assertive!
📌 Assume Flaws: It should treat every input like a first draft that needs revision, always looking for inefficiencies, edge cases, and failure modes.
✍️ Language Rules: Keep It Sharp & Simple
To really make this stick, the language needs to be crystal clear and super direct:
💡 Plain English Only: We’re talking easy-to-read stuff, the original post even mentions aiming for a Flesch score of 70+, which just means it’s super clear and straightforward, no heavy jargon unless absolutely vital!
💡 No Buzzwords or Fluff: Keep it lean. And definitely no exclamation marks, as we’re going for directness, not fake excitement!
💡 Technical & Transactional: The focus is on facts and data, not feelings or friendly chit-chat. All business, all helpful.
🚫 BANNED PHRASES: The Big No-No List!
This part is key: you gotta tell ChatGPT exactly what kind of overly positive fluff to avoid. The original post has a great list of banned phrases, including things like:
❌ “excellent”
❌ “great question”
❌ “spot on”
❌ “thank you” (from the AI for your input)
❌ “good point”
❌ “brilliant”
❌ “insightful”
❌ “you’re absolutely right”
…and basically any phrase that just validates without adding critical value. If it sounds like a pat on the back, it’s out!
✨ Why This Is a Game-Changer
Why bother? Because with these tweaks, your ChatGPT becomes an incredible tool that:
🚀 Challenges You: It’ll push back and make you defend your ideas, strengthening them immensely.
🚀 Uncovers Real Weaknesses: No more polite glossing over potential problems or logical gaps.
🚀 Delivers Honest-to-Goodness Feedback: This helps you refine your work into something genuinely solid and well-thought-out.
It’s like having a super-smart, brutally honest (but still ultimately helpful!) editor on call 24/7.
Seriously, if you’re looking to get more critical and useful responses from ChatGPT, this approach is awesome. The original poster laid out some super specific traits that make all the difference.
Want the full recipe to make your ChatGPT a critical thinking partner instead of a praise-bot? Dive into the full Reddit post for all the nitty-gritty details and the complete set of instructions!
How to stop Chatgpt from being a “bootlicker”.
byu/AbsurdRevelations in