Eleven v4 Turbo Gets Voice Agents to Full DuplexOriginalElevenLabs shipped Eleven v4 Turbo today: 150ms to first speech, streaming both ways, a turn-taking model around it. The full duplex setup I wrote about in August.Voice AI
Full-Duplex Voice AI Needs a New ArchitectureOriginalThe PACE paper exposes a flaw at the heart of full-duplex voice AI: the model answers from context the user never actually heard.Voice AI