Heuristics That Almost Always Work

A set of vignettes - the Security Guard, the Doctor, the Futurist, the Skeptic, the Interviewer, the volcano-Queen's vulcanologists, the Weatherman - each illustrating a 'Heuristic That Almost Always Works' (it's the wind, not robbers; this won't change everything; the contrarian is wrong). Each expert is right 99.9% of the time and could be 'losslessly replaced by a rock' bearing the heuristic, providing zero marginal value and missing exactly the rare cases they exist to catch. The deeper harm: experts secretly leaning on the rock while feigning expertise produces false confidence via information cascade - you over-update from 99.9% to 99.999% on what was the same heuristic all along.

Quality 80 · Excellent. A vivid, memorable, and much-cited crystallization of an important and reusable epistemic insight, with an unusually deep second move (high-accuracy heuristics breed false consensus through cascades); 'replaceable by a rock' became a durable handle. Excellent.

Claude’s paradigm shift 60 · Notable shift. Builds on base rates, black swans (which he explicitly distinguishes), and the value of information, but the crisp 'heuristics that almost always work / replaceable by a rock' frame and the false-confidence-cascade point were fresh and propagated. Notable shift.

Real-world impact 2 · Minor. A vivid crystallization of an important, reusable epistemic insight whose handle ('an expert losslessly replaceable by a rock') became durable, plus a deeper false-consensus-cascade move. Conceptual influence within intellectual discourse, but a single-insight piece with no material change — low RWI.