TIL: Jev and the Rise of Decision Models
Last week, TypeSafe AI introduced Jev, a new category of model it calls a decision model. Unlike an LLM, Jev is not designed to generate prose. It takes unstructured text or JSON and returns a typed decision: a category, score, yes/no answer, extracted value, and an associated confidence score. TypeSafe describes it as “a frontier-intelligence function call: unstructured state in, typed probabilistic decisions out.” The important distinction is that a program can branch directly on the result instead of interpreting generated text. ...