Scott Alexander, curated
← Back to curation

Contrarians, Crackpots, and Consensus

Quality
80
Excellent
Claude Shift
62
Notable shift
RWI
3
of 10

Summary

Foundational epistemics essay distinguishing virtuous contrarians from crackpots. Introduces the 'pyramid' model of the scientific establishment (specialist researchers -> non-specialists -> guideline-makers -> journalists + fieldworkers -> public), where knowledge transfers or gets blocked by industry/political 'enemy action' at different levels (explaining the serotonin-theory and Taubes disputes). Key diagnostic: virtuous contrarians (Kirsch/Whitaker, Cochran/Harpending, Judith Rich Harris, Bostrom) are met with BOREDOM, crackpots (creationists, anti-vaxxers) with violent opposition -- the 'two types of no evidence' (neutral vs active). Kuhnian framing: contrarians find patterns in ignored 'noise,' synthesize isolated specialist findings ('what if fast evolution happens all the time?'), and are often field-outsiders. The tragic payoff: 'even when contrarians win, they lose' -- the field celebrates its sober members while the loud contrarians are remembered as crackpots who got a veneer of credibility (the prescient Stuart-Russell-vs-MIRI prediction).

Why this score

Quality 80 · Excellent. A foundational epistemics essay dense with exported insights (the pyramid model; the boredom-vs-opposition diagnostic; the canonical 'two types of no evidence'; 'even when contrarians win, they lose'). High-Excellent, co-tier with Five-Case-Studies (80) and the conceptual parent of Learning-To-Love-Scientific-Consensus (81).

Claude’s paradigm shift 62 · Notable shift. Notable-to-Major shift. The pyramid model, the contrarian/crackpot diagnostic, and 'two types of no evidence' were genuinely novel framings in 2015 that he and others reused for years.

Real-world impact 3 · Moderate. A foundational epistemics essay dense with exported insights (the establishment 'pyramid,' the boredom-vs-opposition diagnostic for distinguishing virtuous contrarians from crackpots, the 'two types of no evidence'). Conceptual influence within rationalist discourse, no material change — modest RWI.