I saw the author speak on a panel about AI and writing, arguing that writers should be able to use any tools they see fit to bring their creative work to life. But I didn’t get particularly curious about the book until I learned that he trained a large language model on the poetry of Marianne Moore to produce the lyrical stylings of Charlotte, a fictional generative AI platform. In this book, aging poet Marian Ffarmer is hired by a major tech company to collaborate with Charlotte on a long poem. Ffarmer’s character is inspired by more than Moore’s poetry: Like Moore, Ffarmer is an amiable recluse in a codependent relationship with her mother, an odd duck who is known as much for her cape and tricorne hat as for her poems. I’m no Moore scholar, but I worked in a museum that houses her faithfully reconstructed study, and it was both enjoyable and unsettling to see flashes of Moore’s familiar-to-me history made unfamiliar in the character of Ffarmer–perhaps similar to the way it can be enjoyable and unsettling to see an AI program write poetry in real time.
I thought the novel would be gimmicky. It is actually lovely, with Ffarmer’s late-life whims and worries rendered with dignity and humor, and all the questions raised by her AI co-poet/co-pilot treated seriously and thoughtfully. They are, this book demonstrates, the same questions raised about poetry even without AI: what is the work of poetry? what is it for? who is it for? why do we read it, or don’t we? why are some poets elevated by fame while others are forgotten? when we are driven to create art, what do we owe our art and what do we owe to others? I may regret it, but I think this book leapt onto this year’s list of Books I Love.