Think
- ..through the problem yourself before you start typing.
- decide what ***context*** is relevant -
- What **background**?
- What **constraints**?
- What **data**?
- ….the AI needs to give a useful answer.
- Short Prompts from https://open.substack.com/pub/aiblewmymind/p/i-dont-want-ai-to-replace-my-mind
collapsed:: true
-
17 short prompts for deeper thinking
Usually, I advocate for longer, more carefully crafted prompts because they’re excellent predictors of output quality, and the very process of crafting them brings clarity of mind. But starting is more important than doing nothing, so these short questions offer a bridge between knowing better and doing better.
The prompts here are based on well-established thinking frameworks you might already know. They can help you catch problems before they become costly mistakes and turn AI from a risky gamble into a dependable thinking partner.
But remember: none of these replace the crucial first step of thinking through the problem yourself before you start typing. That step also forces you to decide what context is relevant - what background, constraints, or data the AI needs to give a useful answer.
-
1. 5 Whys: uncover the real root cause instead of fixing surface symptoms
This method helps you find the root cause of a problem. You start with the problem, then keep asking “why?” until you get past surface-level answers and uncover what’s really driving it. It keeps you from fixing symptoms instead of causes.
Prompt:
*“Here’s the problem. Ask me ‘why’ until we find the root cause.”* -
2. Problem Restatement: see fresh angles by reframing the problem in new ways
The first way you describe a problem isn’t always the best. Restating it in different ways opens up new angles and can completely change how you solve it.
Prompt:
*“Here’s the problem. Give me 3–5 alternative framings that shift perspective, not just wording.”* -
3. Inversion & Guardrails: spot failure modes early and build safeguards before they appear
Most plans don’t collapse because of bad luck, they fail because of blind spots we didn’t prepare for. Instead of only asking how to succeed, flip the question: how could this fail, and how can you protect against it? Looking at failure modes first helps you build guardrails before they show up.
Prompt:
*“Here’s my goal. Show me the biggest ways this plan could fail, what would cause them, and the guardrails I can put in place now to avoid sabotaging it later.”* -
4. Socratic Questioning: sharpen your thinking by exposing hidden assumptions and weak logic
This is about digging deeper with questions instead of accepting things as given. You test assumptions by asking things like “What do you mean by that? What evidence supports it? What if the opposite were true?” It sharpens your ideas.
Prompt:
*"Question me like Socrates about this idea until I hit the core principle or expose a contradiction."* -
5. Outside View: stay realistic by looking at similar cases
We often get trapped in our own perspective. The outside view asks: how did similar projects turn out? This gives you a reality check from past patterns.
Prompt:
*“Here’s my plan. Based on what usually happens in comparable cases, how likely is it to succeed or fail, and why?”* -
6. Devil’s Advocate: strengthen ideas by forcing them to withstand tough criticism
This forces your idea to stand up against criticism. By having the AI argue against it, you expose weaknesses you might ignore. If your idea survives, it’s stronger.
Prompt:
*“Here’s my idea. Argue against it as strongly as possible.”* -
7. Red Team / Blue Team: stress-test plans with a structured debate of strengths vs. weaknesses
This is like a structured debate. One side defends your idea (blue team), the other side attacks it (red team). It shows you both strengths and weaknesses.
Prompt:
*“Here’s my plan. Run a Red Team vs Blue Team debate with multiple arguments on both sides. Summarize where the defense is strongest, where the attack is strongest, and what that means for me.”* -
8. Unintended Consequences: anticipate ripple effects and side impacts before they catch you off guard
Every action creates ripple effects, many of them invisible at first. By asking where things could backfire or produce side effects, you avoid short-term wins that turn into long-term problems. This lens pushes you to look beyond the obvious outcome.
Prompt:
*“Here’s my plan. What edge cases might I be missing now that could lead to unintended consequences later?“* -
9. Steelman: understand the other side fully by building its strongest possible case
Instead of making the other side look weak, you make the strongest possible version of the opposing view. This helps you understand it better and strengthens your own position.
Prompt:
*“Here’s my position. Build the strongest possible version of the opposing view, as if its smartest defender wrote it."* -
10. Triangulation: cross-check claims with multiple perspectives for a clearer picture
No single source is perfect. By looking at the problem from multiple perspectives (data, expert opinion, case studies), you get a clearer and more reliable picture.
Prompt:
*“Here’s my claim. Test it from three independent angles: 1) what the data suggests, 2) what experts or research say, and 3) what past real-world cases show. Where do they align, and where do they conflict?"* -
11. Barbell Strategy: balance safety and upside by testing extremes in parallel
Borrowed from investing, the barbell strategy means exploring extremes instead of just the middle. One path is ultra-conservative, another is high-risk/high-reward. Running both in parallel can balance safety and upside.
*“Show me one ultra-conservative way to do this, and one high-risk/high-reward way. How could I run both in parallel?”* -
12. Multiple Models: see the problem more fully by layering different mental lenses
No single lens explains everything. Looking at a problem through different models (e.g. economics, psychology, systems thinking) gives you a fuller picture.
Prompt:
“Here’s the situation. Apply multiple mental models to it (psychology, economics, strategy). Compare how each would frame the problem.” -
13. Falsification: replace wishful thinking with ideas that survive real evidence tests.
Most ideas sound good until you ask the harder question: what evidence would prove this wrong? Falsifiability forces you (and AI) to stop defending an idea and start testing it.
Prompt:
*"Here’s my claim/idea. Check current evidence with sources that could prove it wrong.”* -
14. Pros/Cons/Alternatives (PCA): avoid tunnel vision by weighing trade-offs and alternatives together.
Pros and cons alone can be limiting. By forcing yourself to add at least one alternative option, you make better comparisons and avoid tunnel vision.
Prompt:
*“Here’s the answer. Give me pros, cons, and one alternative.”* -
15. Reframing Setbacks: Turn obstacles into opportunities and keep momentum moving forward
How we interpret setbacks shapes how we respond. Reframing in the moment is difficult, but AI can help us step back, question limiting beliefs, and see opportunities we might otherwise miss.
Prompt:
*“Here’s the setback I’m facing. Help me reframe it as an opportunity. What lessons, strengths, or advantages could come from it?”* -
16. Hanlon’s Razor: save mental energy by looking for non-malicious explanations first
Not everything goes wrong because someone’s out to get you. More often, mistakes, miscommunication, or chance are to blame. By deliberately asking for non-malicious explanations, you stop yourself from jumping to the worst conclusion in work, relationships, or your own analysis.
Prompt:
*“Here’s what happened. Give me 3 non-malicious explanations for it.” *
-
17. Incentives: reveal what truly drives behavior by mapping hidden motivators
Using this lens helps you see why deals stall, why systems keep producing the “wrong” outcome, or what’s really driving behavior in news and politics, opening your eyes to dynamics you might otherwise miss.
Prompt:
*“Map out the hidden incentives here. Who benefits, who loses, and how does that shape behavior?”*
-