According to the amyloid's cascade hypothesis the wrong cut of the Ab protein and its aggregation into ordered fibrils is at the origin of the Alzheimer’s disease. Following this aggregation process with microscopic details is far to be possible experimentally. However, in silico modelling can help. Recently we have started working on a multi-scale simulation scheme to challenge the problem, and we could follow the aggregation of a simple but insightful amyloid peptide, the Ab(16-22), into prefibrillar states. You can enjoy the details here. However, too much focus on the aggregation of Ab seems to be the wrong thing to do. In fact, it is a news of recent days that another potential drug intended to remove Ab plaques from the brain of Alzheimer's affected patients just does not work, and the clinical trial was stopped, read more here. This news failure questions the amyloid cascade, or just indicates that when the symptoms are visible, treating Ab plaques is just to late, some interesting points are debated here.