The machine learning installation "Blissful" deals with the failure of human and machine pattern recognition. A computer with a camera analyzes cgi clouds passing by and tells short stories about the cloud figures it recognizes. It is located on a hill which reminds of the standard background image "Bliss" of Windows XP. Machine algorithms like ChatGPT produce linguistically plausible sounding outputs, where the degree of truth is secondary. The algorithms are programmed to rather give a wrong and human sounding answer than none at all. This behavioral design is also reinforced by the Turing test fascination, i.e. the moment in which we can hardly distinguish machine behavior from human behavior. The installation aims to explore these moments of failure: the computer's interpretation of random noise patterns (the clouds) as meaningful, and the visitors' interpretation of these misinterpretations as meaningful because of their human-like nature.