Today for AI
HOT RADAR
llmHEAT 16.7°

AI CURATED HOT STORY · 10/24/2026

Dessa Launches Full-Duplex Speech Model Trained Solely on Human Audio

1 reports archived1 independent sourcesupdated 10/24/2026, 03:09:54
Synthesis & Latest Updates
1 Sources Cross-Validated

Startup Dessa has released a full-duplex speech model trained entirely from scratch on human audio conversations, bypassing traditional text-based pretraining. By processing raw acoustic signals, the model preserves paralinguistic cues like emotion and cadence, aiming for more natural real-time interactions.

LATEST/Startup Dessa has released a full-duplex speech model trained entirely from scratch on human audio conversations, bypassing traditional text-based pretraining. By processing raw acoustic signals, the model preserves paralinguistic cues like emotion and cadence, aiming for more natural real-time interactions.

HEAT TRENDHourly heat curve

3 fully observed hours
Now
16.7
Peak
17.405:00
24h change
–

The line compares only sources observed throughout; gap hours are interpolated to keep the trend continuous.

TIMELINECoverage timeline

Total 1 reports · Latest first
  1. Hacker News AIT2·78 pts
    • Trained from scratch exclusively on human speech data without any text or pretrained model initialization
    • Utilizes a full-duplex architecture to capture and generate complex vocal features including emotion, pauses, and natural rhythm
    • Challenges the traditional 'speech-to-text' paradigm by addressing interaction fidelity loss caused by text's inability to carry paralinguistic information