Full Duplex

A full-duplex model listens and speaks simultaneously, completely removing the turn-taking assumption: overlapping speech and arbitrary interruption are natural behavior that doesn’t require detection. Pioneer – Moshi (parallel user & model audio streams + textual “inner monologue”, ~200ms latency). Thesis: interactivity should be built into the model, not bolted on (TML Interaction Models, micro-rounds ~200ms). GPT-Live brought to industrial scale: continuous processing of input and output, tens of decisions per second, plus separation of fast and slow thinking via delegation to a background model.

Related: [Voice Paradigms], [Fast and Slow Thinking Architecture]