Oct4, a class V POU-domain protein that is encoded by the Pou5f1 gene, is thought to be a key transcription factor in the early development of mammals. This transcription factor plays indispensable roles in pluripotent stem cells as well as in the acquisition of pluripotency during somatic cell reprogramming. Oct4 has also been shown to play a role as a pioneer transcription factor during zygotic genome activation (ZGA) from zebrafish to human. However, during the past decade, several studies have brought these conclusions into question. It was clearly shown that the first steps in mouse development are not affected by the loss of Oct4. Subsequently, the role of Oct4 as a genome activator was brought into doubt. It was also found that the reprogramming of somatic cells into induced pluripotent stem cells (iPSCs) could proceed without Oct4. In this review, we summarize recent findings, reassess the role of Oct4 in reprogramming and ZGA, and point to structural features that may underlie this role. We speculate that pluripotent stem cells resemble neural stem cells more closely than previously thought. Oct4 orthologs within the POUV class hold key roles in genome activation during early development of species with late ZGA. However, in Placentalia, eutherian-specific proteins such as Dux overtake Oct4 in ZGA and endow them with the formation of an evolutionary new tissue-the placenta.