Clones of CD8⁺ T cells that have been selected in the primary response must have a mechanism by which they can continuously or intermittently generate new effector cells. Several years ago, this mechanism was proposed to involve a self‐renewing, stem cell‐like subset that could avoid the differentiating effects of interleukin‐2 (IL‐2). The model considered the stem cell subset to be contained within the central memory population of CD8⁺ T cells (T_CM). This proposal was inconsistent with subsequent findings suggesting that all antigen‐activated CD8⁺ T cells differentiated to effector cells (T_EFF) during the primary response and that T_CM developed during the memory phase by de‐differentiating from effector memory cells (T_EM). However, findings have since been reported that support the stem cell model. First, studies indicate that T_EM do not serve as the precursors of T_CM. Second, transcriptional repressors of IL‐2 signaling do enhance the memory response. Third, memory cells lacking effector functions and with a capacity to replicate in a secondary response develop in the absence of signaling through the IL‐2/IL‐15 receptor. Taken together, these findings suggest that antigen‐activated CD8⁺ T cells with a stem cell‐like capability for maintaining proliferative potential develop by an unknown IL‐2‐independent process. The challenge is now to identify this unknown pathway of clonal expansion.