For AutoSummarize, Jason Huff took the 100 most downloaded copyright-free books from Project Gutenberg at the time and condensed them to a few lines each using the Autosummarize feature in Microsoft Word, which was marketed as automatically summarizing documents to a desired length. Microsoft Word confidently concluded the creation of a summary with the message, “Word has examined the document and picked the sentences most relevant to the main theme.” However, this feature was discontinued by Microsoft the same year Huff published his book, probably due to its high inaccuracy which often resulted in humorous fallacies, especially when dealing with rhetorical or poetic genres.

This shortcoming is clearly evident in Huff’s collection as well. Already the very first summary of Henrik Ibsen’s A Doll’s House gives erroneously the proper names or speaker indications as the most relevant information of the text: “Nora. Nora. Nora. Nora. Nora. Nora. Nora! Nora. Nora. Nora” (p. 1). But in a way it is of course just as true as that All About Coffee is primarily about “Coffee_ / Coffee. Coffee. Coffee. Coffee. Coffee. Coffee. Coffee. Coffee. / COFFEE” (p. 14). At least the algorithm seems to be sure of itself, as shown by the quite similar result for Alice’s Adventures in Wonderland, of which there are three different versions in Project Gutenberg and thus also three summaries in Huff's book (cf. pp. 11–13). Again and again, however, page numbers, captions, chapter headings, index entries, column titles, footnotes slip into the summaries.

Thus, the result of Huff’s automatic digestion is not only a poetic reading of a classical canon through the lens of failing algorithmic optimization, but also a documentation of an outdated tool, making this book an object of potential media-archeological investigation into the imaginaries and dead ends of writing programs.