This work studies book, text and graphic visualization as tools for proving and constructing rational knowledge.

Graphical representation of data in diagrams, diagrams and graphs is mistakenly perceived as associated with exceptional objectivity and reliability. However, the data is only a partial impression of the real world, which is formed using subjective collection methods, which, in fact, can be seen in their graphical representations. In this work, diagrams and graphs turn out to be a way of visualizing subjective and emotional processes.

Reading, in turn, is not a predetermined process, and the perception of a text is always associated with randomness: interpretation, the state of the reader, attitude and various environmental factors. In this book, the text is created through the use of uncertainty and uncontrollability — what makes neural networks (GPT-2) work.