AI is moving beyond chatbots.

Block recently introduced Dessa, a speech model designed for full-duplex interaction, pushing toward more natural, real-time voice conversations.

The interesting challenge isn’t just generating a realistic voice. It’s handling interruptions, listening while someone speaks, responding naturally, and maintaining conversational flow.

Imagine AI assistants that can handle customer support, help developers debug code through voice, or act as real-time personal assistants.

But there’s a catch.

If an AI can sound completely human, how will we know whether we’re speaking to a person or a machine?

I’d love to hear your thoughts:

🔹 Will voice become the primary interface for AI assistants?

🔹 What matters more: natural speech, low latency, or intelligence?

🔹 Should AI-generated voices always identify themselves as AI?

Would you trust an AI voice assistant to handle an important conversation without human supervision?

Source: https://block.xyz/

i.redd.it