Cells proliferate or differentiate. Conventional wisdom holds that these two dominant forces regulating cell state display a yin-yang relationship, but like all reductionist conclusions, the absolute dichotomization of cell behaviors must be considered with caution, especially in the context of cancer, wherein rules are frequently broken. Cytotoxic cancer therapies target proliferating cells but are limited by toxicities to normal tissues and resistance mechanisms in tumors. Inducing differentiation offers potential tumor control with reduced toxicity. Glioblastoma (GBM) is a highly lethal, heterogeneous cancer characterized by unrestrained growth accompanied by aberrant lineage differentiation with a desperate need for more effective therapies. GBM intratumoral heterogeneity is derived from the complex interplay between genetic and nongenetic forces that evolve temporally and spatially to create cellular hierarchies with self-renewing, tumorigenic stem-like cells at the apex. These so-called cancer stem cells (CSCs) remain in a hyperproliferative, nondifferentiated state and contribute to tumorigenesis (1). Over the years, numerous cellular regulatory networks involved in development, wound responses, epigenetic modification, and autophagy have been identified and targeted to induce differentiation of GBM CSCs (2). In many of these studies, differentiation is an afterthought or the presumed natural, indirect consequence of interfering with an oncogenic pathway and is assessed solely by marker expression (of neuronal and astrocytic lineages) along with decreased capacity to proliferate. A unidirectional, irreversible march toward terminal differentiation may not be relevant in cancer because tumor cells often display simultaneous expression of markers of stem/progenitor cells and differentiated progeny (sometimes with more than one lineage at a time). The plasticity of the cellular hierarchy may manifest in normal tissues, with reporter mouse lines suggesting that colon progenitor cells may be able to replenish the stem cell compartment under stress (3). It is uncertain whether a terminal differentiation status exists in GBM, as some studies argue that GBMs can arise from differentiated glia and neurons (4, 5). Perhaps a better definition of terminal differentiation in the context of GBM could be based on functional differentiation that involves the interaction of tumor cells with the normal brain. Studies have yet to go beyond marker analysis to clearly detail or trace that a putative terminally differentiated GBM cell of the neuronal lineage, for example, meets the criteria of being classified as a differentiated neuron, such as showing the ability to establish action potentials, synapses, relay signals, and produce neurotransmitters. Furthermore, differentiation is not just restricted to neural cells because GBM CSCs have been shown to give rise to supporting vasculature, including endothelial cells and pericytes (6, 7). In PNAS, Hu et al. use a creative discovery approach to prospectively identify and test the activity of genes known to influence terminal differentiation of GBM CSCs (8).