After prolonged extracorporeal multiplication in physiological culture media, there can be curative infusions of a cancer patient's own cytotoxic T cells (adoptive T cell transfer; ACT), which must achieve efficient activation in potentially adverse tumour microenvironments. With spectacular, yet irregular, success, improvements are needed. Developing lymphoid cells are biologically selected, not only for ‘near‐self’ reactivity (positive selection), but also to avoid self‐reactivity (negative selection). Thus, success requires harnessing near‐self cells while avoiding extreme autoimmune phenomena. Abrupt metabolic changes accompanying T cell activation to leave the G0 stage and enter the G1 stage of the cell cycle (eg enhanced glycolysis) are accompanied by increased transcription of the G0S9 gene that mediates salvage synthesis of NAD+ from nicotinamide; the latter has recently been shown to increase the efficiency of ACT. Despite theoretical and experimental advances, there has not been parallel progress in simulating in vivo conditions with culture media that were initially formulated for their positive benefits for tumour cell lines (cell survival and proliferation). Yet for lymphoid cells, inhibition or death (ie immunological tolerance) is as important as their activation and proliferation (immunological response). Thus, use of media optimized for the latter may mask the former. The resilience of established culture protocols may have been partly politically driven. However, unphysiological conditions have sometimes yielded fortuitous insights. Optimization of culture media for specific tissues must consider the nature of problems addressed in research settings and the need to avoid mishaps in clinical settings.