How to Think Clearly (1/11): Where Thinking Breaks and What to Put There

This series treats clear thinking as a practice with known failure points. Thinking does not break at random. A mind with a small working memory, a pull toward the easiest option, a habit of mistaking its models for the world, poor instincts about size and odds, and a new habit of asking machines that tend to agree, fails in the same places again and again.

Each post takes one failure point, shows the evidence for it from the original studies or sources, and offers one safeguard. A safeguard here means a decision made before the moment of failure, while thinking is still clear. Odysseus did not resist the Sirens by willpower; he had himself tied to the mast beforehand. The rest of the series is that move, applied six ways.

The map

The stages run from the inside out: the mind, the setting around it, the beliefs it holds, the estimates it makes, the direction it chooses, and the tools it hands work to.

The six stages of the series. Every post carries this figure with its own stage highlighted.

How each post works

Each post opens by saying where it sits on the map. Claims come from named sources, read in the original wherever possible. Where a result reaches the post only through a secondary source, the post says so. Where an argument or a simulation is the author’s own, the post marks it as such.

Contents

#	Post	Stage	Fails when	Safeguard
1	Where Thinking Breaks and What to Put There	-	Thinking breaks at known points	A decision made before the failure
2	A Thought Has a Machine Behind It	Mind	Capacity and attention run out	Lighten the load
3	Catch the Error While It Happens	Mind	Failures go unnoticed	Notice confusion, avoidance and “should”
4	Make the Right Thing the Easy Thing	Setting	Effort alone is unreliable	Move the friction
5	The Map Is Not the Territory	Belief	A model is taken for reality	Check the territory
6	Three Tools for Estimating the Unknown	Estimate	Intuition misjudges size and odds	Compute it
7	Averages Lie	Estimate	The average hides a few huge cases	Look at the distribution
8	Price Your Own Hour	Direction	Time is spent unpriced	Put a number on it
9	Explore Early, Exploit Late, Imagine the Failure First	Direction	Locked in too early	Explore, then run a premortem
10	The Hill You Are Standing On	Direction	A small peak feels like the top	Test for a higher peak
11	Let AI Think, Keep the Decision	Delegation	The assistant agrees	Keep the decision

Reading order

Read in order, since each post starts where the previous one stopped. Or read by need: when effort keeps failing, post 4; when numbers mislead, posts 6 and 7; when stuck, posts 8 to 10.