NobodyWho Publishes 25-Line Local Python Sketch of Jev-Style Decision Modeling

NobodyWho has published a deliberately minimal, local Python sketch of Jev-style decision modeling, showing how a tiny Qwen3 model can classify an input by reading option-token logits instead of calling an API. The demonstration arrives a week after TypeSafe AI launched Jev, an early-access System One model built to return typed probabilities rather than generated text, and it has sharpened a central question: how much of the idea is a new model class, and how much is a disciplined interface around constrained choices?