Contra Stone On EA

Point-by-point rebuttal of Lyman Stone's 'Why Effective Altruism Is Bad', across seven sections. The standout is methodological: Stone infers EA's charitable impact from city-level Google-Trends/giving data, and Scott shows the analysis is hopelessly underpowered -- ~2,500 EAs in a 10M Bay Area population shift average giving from 3% to 3.0025%, far below Stone's 0.5pp resolution; you'd need 500,000 EAs in the Bay to register, so just measure the rare event directly (SSC-survey EAs give ~6% vs non-EAs ~1.5%). Then the recurring 'EA is more than IPA++' thread: EA is a philosophy plus community plus scale, not just RCTs (GiveWell solved the overhead-metric problem IPA didn't); the broad cause list reflects diminishing returns on $10B, not incoherence; the terrorism digression is dismantled (terrorism is wrong, doesn't work, and massively backfires -- the SBF cautionary tale); the animal-welfare 'evolutionary error' objection is met with 'so are art, loving your kids, and caring about distant strangers'; and EA as social technology / Alcoholics-Anonymous-for-doing-good.

Quality 70 · Strong. Strong floor. A characteristic, effective contra-piece; the power-calculation point ('don't escalate a rare event to the population level when you can check it directly') is a genuinely useful, citable nugget. Held to the Strong floor because it is reactive and bounded -- a competent demolition of a weak target rather than a standalone contribution.

Claude’s paradigm shift 45 · Moderate. Moderate, low. Mostly redeploys Scott's existing EA defences (social technology, tower of assumptions); little new in its moment beyond the sharpening of the power-calc critique.

Real-world impact 2 · Minor. Minor. Within the EA-discourse defence literature; conceptual, no material reach. 2.