The target paper proposes six solutions to increase the reliability of infant research, with expected benefits for robustness and replicability. In this commentary, I highlight that signal detection theory can help the framing of reliability issues, and that reliability does not address the core problem of internal validity. The relative merits and challenges of the six solutions are reviewed. A seventh solution is proposed to address the validity issue that remained overlooked. The lack of general theories about the infant mind is probably the biggest obstacle to progress in the field. Instead of whether questions, asking how questions should be at the forefront of future research.