I return to Break It Down Show with Pete A. Turner and Brad Hutchings to look more closely at what a large language model actually does. We discuss the idea of AI as a language coprocessor: powerful at transforming signs and patterns, but not a mind, a colleague, or a source of judgment.
That framing opens a broader conversation about SkillSaige, semiotic processing, and the systems we build around these models. We ask how technical boundaries can protect human agency, reduce waste, and make AI products more useful by refusing roles the underlying technology cannot responsibly fill.