Literary theory for robots How computers learned to write

In this thought-provoking treatise, Tenen (Plain Text), a Columbia University English professor and former software engineer, examines the forebears of text-generating artificial intelligence. Highlighting how contemporary concerns about AI echo centuries-old debates, Tenen notes that 17th-century poet Quirinus Kuhlmann objected to German polymath Athanasius Kirchner's Mathematical Organ--a box-shaped device with a complex system of wooden slates that, when properly arranged, could "compose music, write poetry... and even do advanced math"--because Kuhlmann believed it reduced users to parroting information, instead of producing genuine knowledge. Elsewhere, Tenen covers such experiments as William Cook's 1928 Plotto manual for generating story ideas ("imagine a really complicated Choose Your Own Adventure story," which allowed authors to chart a plot from start to finish) and linguist Noam Chomsky's semisuccessful attempts to teach English grammar to a primitive computer in the 1960s. The history provides crucial perspective on the contemporary AI boom, and Tenen's incisive analysis offers cautious optimism about the future, suggesting that while AI will upend the jobs of legal professionals, writers, and other "intellectual laborers," those workers will also be "freed" from regurgitating facts and can instead "challenge themselves with more creative tasks." Timely and original, this is an essential resource on the history of text-generating AI, and its future. Photos. (Feb.)

An intriguing glimpse into how the secret machinery that makes our technology work has deep roots in philosophy, poetry, and linguistics. We've become so used to computers that can understand what we put on a screen--checking our spelling and correcting our grammar--that we forget that they use complex, clever processes to do so. In this brief, pithy book, Yi Tenen, a former software engineer at Microsoft and current affiliate at Columbia University's Data Science Institute, suggests that giving computers literacy should be seen as one of the most essential technological feats of the 20th century. Locating the beginning of the story is difficult, since Arabic philosophy, Chinese numerology, and other intellectual traditions from around the world and across centuries all have antecedents. Mechanical cypher systems also played a role, as did early attempts at creating a math-based language. Babbage, Bacon, and Leibniz were all interested in giving machines an interactive element, and Turing made a huge contribution with his theories of coding. For decades, people had contrasting ideas on how to get computers to "read," but all ran into the problems of defining intelligence, communication, and understanding. As Yi Tenen shows, we tend to see the human mind as a metaphor, but the author believes that this is not really an appropriate comparison: The repetitive, algorithmic pattern of machine learning does not reflect the conceptual, intuitive nature of human development. The advent of personal computers started tech on the path to artificial intelligence, although getting everything to work together requires extensive collaboration and creativity. Yi Tenen, stirring some wit and anecdotes into the story, sets out the material in non-technical terms, making for an entertaining, informative read. An eclectic and erudite tale of how wide-eyed visions become smart, interactive tools. Copyright (c) Kirkus Reviews, used with permission.