🤖Dessa Launches World's First Real-Time Speech AI Model
Dessa's Speech AI sounds human, feels real
TL;DR
Dessa, a Block lab, launches the world's first full-duplex speech AI model, offering real-time conversations with human-like emotion, nuance, and pauses. It's a game-changer for voice AI.
Dessa, a Block lab, just launched the world's first full-duplex speech AI model, Dessa. This AI can hold real-time conversations, understand emotions, and mimic human speech patterns. If you're building voice assistants or working with speech recognition, this is a big deal. Dessa can learn language and facts from human speech alone, perform in-context learning, and even understand sarcasm and humor. It runs locally on a MacBook Pro M4 Max or better, making it accessible for developers to experiment with.
Key Points
Dessa is the first AI to hold real-time conversations with human-like emotion and nuance.
Dessa can learn language and facts from human speech alone, without text.
Dessa can perform in-context learning and understand sarcasm and humor.
Dessa runs locally on a MacBook Pro M4 Max or better, making it accessible for developers.
Dessa is giving early access to AI safety researchers to find risks.
Why It Matters
If you're building voice assistants or working with speech recognition, Dessa's full-duplex AI model changes the game. It offers real-time conversations with human-like emotion and nuance, making it a must-try for developers. The ability to learn from human speech alone and understand complex human behaviors like sarcasm and humor sets it apart from existing models.
Comments
Be the first to comment
Enjoyed this article?
Get it daily. 7am. Free. Reads in 5 minutes.
Join 3,567 builders reading daily.