This article narrates a consensus history of the proposal to include diagnostic criteria for a psychosis risk syndrome in the DSM-5, in part, to document what happened, but also to potentially help focus future efforts at clinically useful early detection. The purpose of diagnosing a risk state would be to slow and ideally prevent the development of the full disorder. Concerns about diagnosing a psychosis risk state included a high false positive rate, potentially harmful use of anti-psychotic medication with people who would not transition to psychosis, and stigmatization. Others argued that educating professionals about what ‘risk’ entails could reduce inappropriate treatments. During the revision, the proposal shifted from diagnosing risk to emphasizing current clinical need associated with attenuated psychotic symptoms. Within the community of researchers who studied psychosis risk, people disagreed about whether risk and/or attenuated symptoms should be an official DSM-5 diagnosis. Once it became clear that the DSM-5 field trials did not include enough cases to establish the reliability of the proposed criteria, everyone agreed that the criteria should be put in a section on conditions for further study rather the main section of the DSM-5. We close with recommendations about some practical benchmarks that should be met for including criteria for early detection in the classification system.