Marijuana: Much More Than You Wanted To Know

The exhaustive even-handed review of marijuana: liberalisation (decriminalisation, medical) shows little effect on teen use, but full legalisation likely raises adult use ~25-50%; harms include a possible ~8-point IQ drop with heavy teen use and a ~1.4x psychosis risk (causality unclear); drug-war costs are real but fewer are jailed for pot than commonly claimed. An "irresponsible utilitarian analysis" in QALYs concludes the net is too close to call - and, strikingly, that it hinges almost entirely on whether marijuana substitutes for alcohol and reduces traffic deaths.

Quality 78 · Excellent. A model of the MMTYWTK genre - comprehensive, numerate, and honest about its own uncertainty - elevated by a genuinely fresh analytic conclusion (the whole question reduces to driving). Low-Excellent.

Claude’s paradigm shift 48 · Moderate. The exhaustive-review-plus-explicit-utilitarian-calculation approach was somewhat novel for policy writing, and the "it all comes down to traffic accidents" reframe is a real contribution; the underlying method is not paradigm-defining. Moderate, edging notable.

Real-world impact 3 · Moderate. A model, numerate review of marijuana's effects on a materially-consequential policy question (legalization), elevated by a fresh analytic conclusion (the net hinges almost entirely on alcohol substitution). Real topical relevance that informs the debate, but discourse-level with no enacted change — modest RWI.