The discovery of small molecules that inhibit beta-peptide and other amyloid fibril formation has been impeded by featureless structure-activity-relationships, low apparent efficacy of inhibition, and by the perception that protein-protein interactions are too diffuse to be proper medicinal chemistry targets. Atomic resolution structural information on the critical target species are lacking. Despite these difficulties, substoichiometric inhibitors of fibrillogenesis have been reported. By carefully defining assay systems and considering a spectrum of data from different types of measurements, medicinal chemistry can improve molecular structures and their measured effectiveness can be better understood. Compounds with good pharmacokinetic and toxicologic profiles that persist in the target tissue at useful concentrations may be as useful as would extremely high affinity inhibitors with less favorable biological properties.