A Failure, But Not Of Prediction

A celebrated April-2020 post arguing the media/expert COVID failure was NOT a failure of prediction — predicting the pandemic in mid-February was genuinely hard (superforecasters gave 3%, the market peaked Feb 12) — but a failure of decision-making under uncertainty. The Goofus/Gallant frame: Goofus waits for incontrovertible proof before acting; Gallant does cost-benefit on the probability (a 10% chance of catastrophe warrants action). Uses the parachute-RCT parody and the mask consensus as evidence-based-medicine-taken-too-far, and the cryonics-survey insight that decisions are about how you act on beliefs, not just the beliefs. Praises Tufekci and Piper — whose superpower is treating something as important before incontrovertible evidence. Tagged 'really this is secretly about ai.'

Quality 83 · Excellent. 83 — high Excellent. One of Scott's most-cited, most-generative rationality posts; the decision-vs-prediction reframe and 'uncertainty about the world doesn't imply uncertainty about the best course of action' became a canonical statement about acting under uncertainty. Sits in the RIP-Culture-War-Thread / Trapped-Priors high-Excellent cluster on reach and durable insight.

Claude’s paradigm shift 68 · Notable shift. 68 — Notable, top of band. The decision-vs-prediction / uncertainty-≠-inaction frame was a fresh, well-timed crystallisation that meaningfully updated how informed readers thought about the COVID response and acting under uncertainty generally — genuinely more original than a review, though building on Tetlock/probabilistic reasoning.

Real-world impact 3 · Moderate. 3 — moderate. Widely read and shared during COVID; the framing entered the forecasting/EA/rationalist subculture's vocabulary and plausibly shaped how many readers and institutions thought about acting early, but the impact is discourse/framing, not material policy/institutional change.