The concept of schizophrenia only covers the 30% poor outcome fraction of a much broader multidimensional psychotic syndrome, yet paradoxically has become the dominant prism through which everything ‘psychotic’ is observed, even affective states with mild psychosis labelled ‘ultra-high risk’ (for schizophrenia). The inability of psychiatry to frame psychosis as multidimensional syndromal variation of largely unpredictable course and outcome – within and between individuals – hampers research and recovery-oriented practice. ‘Psychosis’ remains firmly associated with ‘schizophrenia’, as evidenced by a vigorous stream of high-impact but non-replicable attempts to ‘reverse-engineer’ the hypothesized biological disease entity, using case–control paradigms that cannot distinguish between risk for illness onset and risk for poor outcome. In this paper, the main issues surrounding the concept of schizophrenia are described. We tentatively conclude that with the advent of broad spectrum phenotypes covering autism and addiction in DSM5, the prospect for introducing a psychosis spectrum disorder – and modernizing psychiatry – appears to be within reach.