Are you curious about AI’s ability to examine a human produced poem and how the internal process works? “Sonnet 4.6 Tries, Fails, and Tries Again to Write a Human Poem” is a dialog between a human poet and Claude, Sonnet 4.6, in which Claude looks a short poem, critiques it, then tries to write a poem on the same topic. It fails miserably, then uses its own process to critique its work than tries again. In the process of trying to improve its own writing the LLM eventually discloses what’s missing in AI that prevents it from learning to write a truly human poem.

Read the full dialog by author and editor, David Jibson HERE for free.