Recent studies reveal spontaneous implicit false‐belief understanding in infancy. But is this early ability genuine theory‐of‐mind? Spontaneous tasks may allow early success by eliminating the selection‐response bias thought to underlie later failure on standard (verbal) tasks. However, using anticipatory eye gaze, we find the same bias in non‐verbal tasks in both preschoolers and adults. We argue that the bias arises from theory‐of‐mind competence itself and takes the form of a rational prior to attribute one's own belief to others. Our discussion then draws attention to a number of other inferential hallmarks of early belief‐desire reasoning that together suggest it is the real deal.