4 ideas
Isolated facts don't stick or generalize; you need the big models from many disciplines (psychology, economics, physics, biology, math, engineering) hung on a latticework so experience attaches to them. A person who knows only one discipline is 'a man with a hammer' to whom every problem looks like a nail, so fluency across roughly 80–90 core models is what prevents systematically distorted judgment.
Instead of asking how to achieve a goal, ask what would guarantee failure, then avoid those things. Many hard problems become tractable when approached backward, as when Munger explains how to achieve a miserable life (unreliability, envy, resentment, ignoring others' lessons) to show how to achieve a good one. Avoiding stupidity is often easier and more reliable than trying to be brilliant.
Human misjudgment comes from a set of distinct psychological tendencies (incentive-caused bias, social proof, reciprocation, commitment and consistency, deprival superreaction, and others). When several of them push in the same direction at once, they combine non-linearly into extreme outcomes, which explains phenomena like cult behavior, auction frenzies, and open-outcry fraud that no single bias accounts for.
Define the narrow area where you genuinely have an edge and refuse to act outside it. Within it, wait patiently, like a batter with no called strikes, until an obvious mispriced opportunity appears, then bet heavily, because a few large, concentrated, well-understood bets outperform many small, poorly understood ones.