Humans possess a capacity for what the ancient Greeks called logos — speech, language, rationality. In the words of the philosopher Charles Taylor, we are “the language animal.” Recent advances in AI invite us to consider anew the nature and significance of our human form of logos, and to ask whether and how such a capacity might be instantiated in a machine. Contemporary large language models (LLMs) are amazingly adept with language. How should we think about what these systems are doing with words? Do they possess genuine understanding of themselves or the world? What do they reveal to us about our own abilities for speech and thought? What do they suggest about the connections between life, agency, embodiment, and language? Can we envision machines with their own form of logos? What would those machines be like in their constitution and mode of functioning?