Jack introduces Dessa, a speech model meant to talk like a person
A full-duplex voice model that handles emotion, pauses and natural cadence.
Jack (@jack) introduced Dessa on Friday, describing it as a full-duplex speech model built so talking to AI feels like talking to a person. The post says it handles emotion, nuance, pauses and natural cadence, and it drew about 9,200 likes.
Full-duplex means the model can listen and speak at the same time, rather than waiting for you to finish before it replies. That is what lets a voice assistant be interrupted or react mid-sentence, the way people do on a call. The post gives no benchmarks or release details beyond that description.
Why it mattersVoice is where AI assistants still feel the most robotic. If Dessa works as described, it points toward voice chat that feels less like a turn-taking menu. Treat it as a claim until you can try it yourself.
