Although converging lines of evidence suggest a central role for amyloid ß-protein (Aß) in the etiology of Alzheimers disease (AD), the active form(s) of Aβ that leads to altered neuronal function and death have not been definitively identified. Indeed, until recently, efforts at understanding Aβ aggregation and toxicity focused primarily on fibrillar forms of Aβ akin to those detected in senile plaques. But plaque density and severity of dementia correlate weakly and recent studies report more robust correlations between the levels of soluble Aβ and the extent of synaptic loss and severity of cognitive impairment. Moreover, the identification of pre-fibrillar intermediates in vitro and the detection of SDS-stable oligomers of Aβ in human brain and CSF and in the medium of cells expressing human Aβ demonstrate the existence of soluble, non-fibrillar Aβ assemblies. Because great effort is currently being expended on the development of anti-amyloid therapeutics, it is crucial that remaining concerns about a causative role for Aß in AD be rigorously addressed. Here we review evidence that Aβ toxicity is likely to be mediated by multiple different Aβ assembly forms and discuss possible therapeutic strategies designed to remove or prevent the formation of soluble toxic assemblies.