#full-duplex

Block releases Dessa voice model for local GPU use

Jack Dorsey's company claims its new full-duplex speech model eliminates the text-to-speech pipeline to capture human nuance and emotion on consumer hardware.