The past 30 years have seen multiple attempts at demonstrating the safety and efficacy of drugs for Alzheimer's disease (AD), predominantly to improve symptoms. Only five drugs (tacrine, donepezil, rivastigmine, galantamine, memantine) have obtained regulatory approval in most countries. Their cost-effectiveness from a societal perspective has not been universally recognized, and anybody who thinks these drugs are useful for individual patients will have to agree that the improvement above the starting point of treatment is moderate. Most of the benefit has been in slowing down progression of symptoms rather than a readily detectable improvement above baseline. There have also been attempts at arresting progression of AD, but all have failed until now. Should we change our approach to developing new drugs for AD so as to move forward? This review will highlight some options to consider in the development of future drugs for AD, with emphasis on strategies to prevent AD or arrest its progression.