Overview
Dessa is a block research-lab full-duplex speech-to-speech model trained only on human conversation audio (no text pretraining), positioned against cascaded ASR→LLM→TTS voice stacks.
Recent Developments
- 2026-10-24: Public research preview with waitlist; early access limited to AI safety researchers; runs locally on MacBook Pro M4 Max+; trained on one NVIDIA GB200 NVL72 rack (2026-10-24-block-dessa-official)
"Human-like" demos are company-recorded; no public weights, API, independent benchmarks, or release date as of launch week.