# Block releases Dessa, a full-duplex voice model that runs locally

> Original title: Block’s new voice model sounds eerily human

- Source: TechCrunch · AI
- Published: 2026-10-24T16:00:00.000Z
- Bad Influence: https://badinfluence.news/items/h9yx296k6gir52obd76ml7y2x
- Original: https://techcrunch.com/2026/10/24/blocks-new-voice-model-sounds-eerily-human/

## Summary

Block's research lab released Dessa, a full-duplex speech-to-speech model that can talk and listen at once and runs on consumer hardware such as a MacBook M4 Max. Block says emotion is built into the model rather than driven by text or voice tags, but the source does not say whether tags are used. Dessa is not publicly available yet, and Block has asked safety researchers to test it for ways a voice agent could deceive people.

## Why it matters

The source separates Dessa's claimed built-in emotion from tag-driven speech generation, which matters for judging how voice assistants might change. The piece also notes that public access has not yet been opened and that safety testing is being sought.