The commentary by Kühle reminds me of a remark made by a distinguished and renowned Swiss sleep researcher who asked me recently, during a lengthy discussion of our work on lucid dreaming, “how can you be sure that what you call a dream really exists”. In other words, he wanted to know how we could prove that dream narratives were memories of REM-sleep mental activity instead of, say, fantasies occurring during the process of awakening or memories of hypnagogic hallucinations, etc. It struck me then that I had neglected to openly postulate the key assumption that our work rested upon, namely that dreams really exist. So I still owe him a detailed response and Kühle’s commentary provides me now with the opportunity to generate an adequate reply. In the following, I will focus on Kühle’s main argument, which seems to circle around the definition of “insight” and the question of whether it represents an epistemological statement or a phenomenological experience. I will shortly enter into discussion of whether it is justified to define insight as a state, as this assumption is not to be deduced from our work but certainly points to a need for clarification. While interesting, I will refrain from commenting on her speculations on whether insight may or may not be an ability except for proclaiming that in my view, insight represents nothing but a result of neurobiological processes we still know far too little about. However, it is a fact that entering the state of lucid dreaming can be trained. Can insight per se be trained? I doubt it. Can the ability to generate insight be trained? According to recent studies on gamma-band activity in the developing and mature brain (see references in the main text), it is at least a possibility.