I search the Internet and find the book The Machine by Georges Perec. The book, from 1968, was originally written as radio play, revolving around three computers analysing the Goethe poem “The Ramblers Lullaby II”. The computers process the poem in a number of different ways. In other words it deals with computers as creators and interpreters of text. The computer as an author of text, a common enough idea today, but not back then. Georges didn’t live to experience the breakthrough of personal computers. His only personal experience was likely the encounter with an ancient talking computer, owned by the Michigan State University, that he visited while working on the text. Thus he wrote about a future he would never see for himself, about the predicted singularity – the computer capable of self-improvement, the ultra intelligent machine. The Machine was broadcast on German radio in 1968 and later translated for an American art journal of which I have forgotten the name. Someone scanned this translation and made it available online as a PDF-download. That’s where I found it. I read about programs capable of interpreting texts. The development is fast and within a few years they predict that the computer will be able to author texts. I enter Goerges’s text into a program that I wrote, and then I wait. Together with Georges and the computer I create the text that Goerges once dreamt about. I write the code and observe the text evolving in a process governed by chance, as a game between me and the computer, but also in a way between Goethe and Perec. The project is a media archeological art work about the history and dreams of computers and it is using code and chance to create a remake of Perecs’s book, a text that later was assembled into an artist book. The project resulted in knowledge about how to use code to create texts and use Internet as resource for writing and create chance. The project also reflects on the future of writing in relation to AI and Machine learning. In the software I am using algorithms that are common in machine written poems. I am interested in mistakes from texts created from Machine learning and AI and how the mistakes generate new content and language. How can computers write stories and what happen if you combine AI with your own ideas and (human) writing? In the software I combine theses algorithms with my own written code and text. I use the text as a visual element and combine it with my own language and the language of Gothe and Perec. It all became 400 different books with random content. Since I use chance to create the content I didn’t know what was going to happen and I became the observer of my own artworks being created by the computer. The project was exhibited in the exhibition, I want to believe, (Gibca Extended) and as a performance at Linköpings Universitet (ReprecDigit)