Some philosophers writing on adaptive preference insist that they are prudentially bad for us, while others argue that they can be prudentially good. In this paper, I use the concept of transformative experience to show that both of these claims are often plausibly true of adaptive preferences at once. First, I show that on the most frequently-defended accounts of well-being, if adaptive preferences involve personally transformative experiences, then the objects of adaptive preferences can come to genuinely prudentially benefit persons after their personally transformative experience. However, I argue that adaptive preferences nevertheless deserve our suspicion on prudential grounds. They do so first because personally transformative experiences may be incomplete, and epistemically transformative experiences may reveal to persons the sub-optimal nature of their adaptive preferences. And they deserve suspicion secondly on the grounds that the well-being effects of even fully personally transformative experiences must be judged against the well-being effects of alternative personally transformative experiences that could have been undergone instead - that is, even if the object of an adaptive preference prudentially benefits a person once she has formed the associated preference, she could have received a more significant prudential benefit from developing non-adaptive preferences in the first place.