AI Wiki · Mediocre Security
ProductModels & inference infrastructure

Block Dessa

Voice model from Block trained entirely on human speech.

Voice model from Block trained entirely on human speech; processes audio directly instead of converting to text first, aiming to catch cues like sarcasm and pauses. Research preview, reportedly runs locally on a MacBook Pro M4 Max (The Decoder; PYMNTS, Oct 2026).

Security and assurance angle

PYMNTS notes voice-impersonation and scam risks.

Connections

Block Dessa and everything one link away. Drag to move, scroll to zoom, click a node to open it.Open in the full graph →

What links here

Concept Multimodal

Publication PYMNTS · The Decoder

Sources

  1. Block, “Dessa - the Conversation Model trained only on speech”, Oct 23 2026
  2. Let's Data Science, “Block introduces Dessa, a speech-to-speech AI model trained on conversation”, Oct 23 2026
  3. PYMNTS, “Block's Voice AI Wins Praise and Triggers Scam Concerns”, Oct 25 2026
  4. TechCrunch, “Block's new voice model sounds eerily human”, Oct 24 2026