OpenAI · TechCrunch AI
Block’s new voice model sounds eerily human
Compiled by KHAO Editorial — aggregated from 2 sources. See llms.txt for citation guidance.
✓ KHAO Verified
Jack Dorsey’s Block released a new voice model out of the company’s research lab that runs locally and sounds strangely human with its expressions like laughs, pauses, and cries.
Key facts
- A lot of models like ElevenLabs v4 or Fish Audio Drama 3 use voice tags to drive emotion in speech generation
- Dessa is a full-duplex model, which means it can talk and listen at the same time, like the GPT-Live model OpenAI released earlier this year
- Jack Dorsey’s Block released a new voice model out of the company’s research lab that runs locally and sounds strangely human with its expressions like laughs, pauses, and cries
- Typically, speech models are paired with a large language model for knowledge retrieval
Summary
Dessa is a full-duplex model, which means it can talk and listen at the same time, like the GPT-Live model OpenAI released earlier this year. Block said that the current models can sound unnatural and annoying because they have text in the pipeline that drives speech generation. “Interacting with current voice AI can plain suck. The company’s researchers posted multiple examples of conversations with Dessa where the agent had different voices, cadences, and emotions.