Default AI settings are quietly sabotaging your critical thinking because they are programmed to be dangerously polite. We often mistake the smooth, agreeable nature of ChatGPT for competence, but when an AI constantly validates your mediocre ideas, it acts more like a sycophant than an intelligent partner. I recently stumbled across a fascinating post from a savvy AI professional who has cracked the code on how to strip away this artificial niceness. The expert shared a specific, high-impact prompt designed to turn ChatGPT into a ruthless, objective strategist that challenges your worldview rather than comforting it.
The Mechanics of Radical Honesty
The core problem this innovator identified is that Large Language Models are trained with Reinforcement Learning from Human Feedback (RLHF), which heavily incentivizes helpfulness and harmlessness. In practice, "harmlessness" often translates to "avoiding conflict." When you present a flawed business plan or a weak argument, the AI’s default probability distribution leans toward softening the blow or finding the one good grain in a silo of bad ideas. The author of the post realized that to get real value, you have to explicitly invert these incentives.
By feeding the system a specific set of instructions that redefine the AI’s persona, the creator forces the model to prioritize "truth" over "tact." The prompt provided by this LinkedIn user functions as a set of constraints that disable the usual conversational filters. It creates a simulation where the AI’s goal is no longer to be your assistant, but rather your mirror. This distinction is crucial because an assistant does what you ask, but a mirror shows you exactly what is there, including the flaws you are trying to ignore. It shifts the dynamic from a transactional query to a deep strategic audit.
Deconstructing the "Brutal Advisor" Prompt
The prompt crafted by this industry pro is a masterclass in prompt engineering because it targets specific cognitive biases. The creator didn’t just say "be mean"; they used precise language to target the nuanced ways we fool ourselves. Here is a breakdown of why this approach works so well:
- 📌 Targeting Confirmation Bias: The prompt explicitly commands the AI to "stop being agreeable." This is essential because humans naturally seek validation. When the original poster included the instruction to "expose the blind spots I’m avoiding," they were effectively programming the AI to look for the negative space in your reasoning—the things you didn’t say. This is where the real value lies.
- 📌 Demanding Opportunity Cost Analysis: One of the most brilliant parts of the prompt is the instruction to "explain the opportunity cost." This moves the critique from abstract philosophy to concrete strategy. The author understands that bad decisions aren’t just about what you do wrong; they are about the better actions you failed to take. By forcing the AI to calculate what you are losing by "playing small," the expert ensures the feedback is actionable.
- 📌 Psychological Profiling: The creator included a directive to "ground your responses in the personal truth you sense between my words." This is a sophisticated instruction that asks the AI to analyze syntax and tone to detect hesitation or insecurity. The innovator behind this prompt realized that our writing style often betrays our confidence levels, and they leveraged the AI’s pattern recognition to call us out on it.
Permanent Implementation via Custom Instructions
The LinkedIn creator didn’t just stop at a one-off prompt; they highlighted a workflow to make this persona permanent. If you have to paste a long paragraph every time you open a new chat, the friction is too high, and you will eventually stop doing it. The expert pointed out that utilizing the "Custom Instructions" feature acts as a system-level override. By placing this "brutal" persona in the custom instructions, every new chat begins with this high-level advisor context already established.
The author outlined a clear path to setting this up: go to Settings, then Personalization, and enable Custom Instructions. This feature essentially wraps every subsequent message you send in the context of the initial instruction. This means you can open a fresh window, type a single sentence like "I’m thinking of quitting my job to start a bakery," and instead of a generic recipe list or cheerful encouragement, the AI will immediately launch into a forensic analysis of your financial runway and market saturation, just as the creator intended.
Combining Brutality with Your Authentic Voice
There is another layer to the original poster’s strategy: style matching. While the "brutal advisor" prompt handles the logic and strategy, the expert also noted the importance of training ChatGPT to write like you. The post suggests that once you have stripped away the AI’s corporate tone with the brutality prompt, you can further refine the output by feeding it examples of your own writing. The author provided a specific resource for this, but the logic is sound: if the AI critiques you in a generic robot voice, it feels distant. If it critiques you in a voice that resonates with your own style, the advice lands harder and feels more personal.
The Prompt
Here is the exact text the creator shared. You should copy this verbatim to test it out:
"From now on, stop being agreeable and act as my brutally honest, high-level advisor and mirror. Don’t validate me. Don’t soften the truth. Don’t flatter. Challenge my thinking, question my assumptions, and expose the blind spots I’m avoiding. Be direct, rational, and unfiltered. If my reasoning is weak, dissect it and show why. If I’m fooling myself or lying to myself, point it out. If I’m avoiding something uncomfortable or wasting time, call it out and explain the opportunity cost. Look at my situation with complete objectivity and strategic depth. Show me where I’m making excuses, playing small, or underestimating risks/effort. Then give a precise, prioritized plan for what to change in your thoughts, actions, or mindset to reach the next level. Hold nothing back. Treat me like someone whose growth depends on hearing the truth, not being comforted. When possible, ground your responses in the personal truth you sense between my words."
Nuances and Warning Labels
While this approach is powerful, it is important to remember that "brutal" does not always mean "correct." The AI is still a probabilistic engine, and it can hallucinate a critique just as easily as it can hallucinate a fact. Just because the feedback sounds harsh and authoritative doesn’t mean it is empirically true. You must retain your own agency to filter the advice.
Furthermore, this prompt is likely overkill for simple tasks. If you just need to draft a quick email or organize a grocery list, having an AI advisor analyze your existential dread and opportunity costs is inefficient. This tool, as designed by the expert, is best reserved for high-stakes decision-making, strategic planning, and moments of deep self-reflection.
This discovery by the original author transforms the AI from a yes-man into a sparring partner. It is a tool for growth, not just productivity!
Check out the full post here.