Social media magnifies “snackable” opinions, algorithm dictates what we see and the flood of new articles/books far exceeds our capacity to read. Above all, we have Generative AI producing massive content at machine speed.

Published in 2018, Hans Rosling’s Factfulness remains very relevant. He taught us how to achieve a better fact-based worldview (beating chimpanzee!) by controlling 10 biased instincts that distort our perception. How can we live up factfulness in 2026 and find our blind spots?

This is is what triggered me to experiment the following - Can we use LLM as a cognitive mirror to reach factulness? LLMs are trained on our own language, so they should reflect the same biases back at us. Having an “outside-in” perspective is more appealing than doing it in isolation, lost in your inner thoughts (maybe i’ll become a philosopher someday)

If you craves for a “magical” prompt that 10x your productivity, don’t be so hasty. It will be more rewarding to

The goal is not to outsource your thinking to AI by copy-pasting the prompt tips, but to use it as a cognitive mirror. Because these models are trained on our own language, they reflect the same collective biases back at us. It is a rare opportunity to have a “outside-in” perspective on our own blind spots which is much more difficult to achieve in isolation.

#	Instinct	Strategy	Suggestion
1	Gap	Append	…or the spectrum in between?
2	Negativity	Append	…that are usually ignored because they are boring
3	Straight Line	Iterate	Assuming we reach a plateau, what are the 3 rate-limiting steps we cannot circumvent?
4	Fear	Verify	Audit the provided claim for risk exposure 1. Human Intangibles: List non-codified complexities (e.g., ambiguity, consensus-building) the claim ignores. 2. Structural Stabilizers: Identify systemic buffers (e.g., law, liability, or economic inertia) that slow disruption. 3. Scenario Matrix: Map Aggressive, Moderate, and Conservative paths, identifying one specific Trigger Event for each level.
5	Size	Verify	Audit these figures to establish relative impact: 1. Denominator: Compare the figure to the total relevant budget, market, or population. What is its percentage of the whole? 2. Unit Economics: Break the total down into cost or output per single unit/outcome based on standard industry averages. 3. Efficiency Benchmark: Compare this scale to a traditional or manual equivalent. Does this represent a leap in efficiency or simply a larger volume of input?
6	Generalization	Verify	Audit this claim for potential overgeneralization using 3 personas 1. Specialist: Why is this group/setting a false proxy for the broader reality? 2. Auditor: What blind spots exist due to the limited scope? 3. Skeptic: What hidden shifts occur when scaling from this sample to the real world
7	Destiny	Verify	Audit this claim for outdated assumptions 1. The 20-Year Test: What “Physical Barrier” or “Golden Rule” was once considered impossible to bypass but is now routine? 2. The Slow-Burn: What incremental trend (1-2% annual change) has accumulated into a systemic shift? 3. The Standard Update: How has the “Gold Standard” for success or proof evolved since this logic was first formed?
8	Single Perspective	Verify	Audit this claim for consensus bias 1. The signal gap: Where does the ‘last-mile’ practitioner face friction when dealing with reality? How are high-value but “silent” actors penalized? 2. The liability audit: If this fails, who is legally or professionally ‘on the hook’? Why is that person currently incentivized to resist this change? 3. The consensus blind spot: Beyond surface-level benefits like speed, cost, and quality, what is the one inconvenient reality currently being ignored by the mainstream consensus?
9	Blame	Iterate	Your answer is incomplete because … Diagnose why this happened 1. Did I provide insufficient context? 2. Is there an ambiguity in my instructions? 3. Is this a known limitation of your training data (e.g., knowledge cutoff)? 4. Suggest a restructured prompt that would prevent this error.
10	Urgency	Iterate	Before I proceed, audit this report for wrong shortcuts 1. What are the top 3 assumptions where a small change would completely flip your conclusion from ‘success’ to ‘failure’? 2. What specific internal or external data would a subject matter expert demand to see before approving this? 3. If I act on this draft immediately and one of your ‘educated guesses’ is wrong, what is the most likely negative consequence?

Below, I will address each bias in detail, and provide a concrete example how you could either 1) reframe the prompt to lower the bias or 2) ask follow-up question to expose the bias. The verbose AI responses were stylistically edited manually in clean structure to make quick read-through possible while the core messages were not altered (and not verified in depth for perfect accuracy)

1. The Gap Instinct

When a story pictures two separate groups with a gap in-between while in reality it is not as polarized but the majority is actually in between

The Gap Instinct

Takeaway

• we can replace binary choices with continuum mapping.
• we can use hybrid options as a “third way”.
• we can focus on the sweet spot that works in the market.

2. The Negativity Instinct

When we get a too-negative impression of the world around us because negative news / bad events are more likely to reach us.

The Negativity Instinct

Takeaway

• we can decouple visibility from impact
• we can value proven approaches that work but are boring
• we can counter-balance pessimism

3. The Straight Line Instinct

When we believe a projection will continue its trajectory across scale, ignoring the external factors that will influence the outcome.

The Straight Line Instinct

Takeaway

• we can integrate constraint mapping into extrapolation
• we can anticipate complexity drag
• we can identify inflection points where trends decouple from history

4. The Fear Instinct

When we create overly pessimistic scenarios because of our own attention filters and media influence, without considering the likelihood of them occurring and while undervaluing the probabilities of alternative outcomes.

The Fear Instinct

Takeaway

• we can identify moats
• we can transition from possibility to probability
• we can decouple capability from adoption

5. The Size Instinct

When a lonely number is used to trigger a desired reaction that would be tempered if presented in comparison or in its proper proportion

The Size Instinct

Takeaway

• we can map concentration
• we can expose diluted impact
• we can neutralize Big number awe

6. The Generalization Instinct

When we instinctively group diverse things into broad categories, mistakenly assuming that every individual within that category is the same.

The Generalization Instinct

Takeaway

• we can map latent risks
• we can feed variance analysis into strategy
• we can distinguish local from global truth

7. The Destiny Instinct

When we view slow-moving transitions as permanent states, failing to recognize that incremental progress eventually crosses a tipping point that invalidates previous fundamentals.

The Destiny Instinct

Takeaway

• We can expose “last-mile” friction
• We can forecast misaligned incentive
• We can can deconstruct the universal hammer bias (same solution to any problem as referred in the book)

8. The Single Perspective Instinct

When a single perspective limits the interpretation because of the echo chamber while a multi-perspective would highlight complexity and bring more practical solutions

The Single Perspective Instinct

Takeaway

Beware of your favorite approach (e.g. agent), you may want to use it too often and end up exaggerating the importance of the problem. A multi-perspective approach opens up dead angles that will hold you back.

9. The Blame Instinct

When we finger point at someone, without considering other possible explanations and blocks our ability to prevent similar problems in the future

The Blame Instinct

Takeaway

• We can shift from “who is wrong” to “what is the system gap”
• We can calibrate AI reliability zone
• We can identify personal blind spots in prompting and verify improvement

10. The Urgency Instinct

When pressure for speed overrides scrutiny, we accept high-risk assumptions without testing their sensitivity to change

The Urgency Instinct

Takeaway

• we can identify the safety margin
• we can anticipate pivot points before failure materializes
• we can present a defensible plan