Deepgram CPO: voice AI is now continuous listening/thinking/speaking on a real-time context bus
schwentker · x · 2026-11-01
Deepgram CPO Ed Anuff describes a shift in voice AI architecture: no longer "listen, think, speak" in turns, but continuous listening, thinking and speaking connected by a real-time context bus — "speech-to-speech without the black box."
CEO Scott also shared a north star of "complaint-driven development": you should be able to complain at any moment and the product should get better in seconds — an aspiration, not yet a product promise.
Related event: Deepgram CPO on Real-Time Context Bus; Pipecat Co-creator on Staying Open(2 posts)→
More from coding & agent
- HauntBnB: a multimodal vision agent that inspects your room for 'paranormal hazards' — mrdzyn · 2026-11-01
- traits.sh turns agent code-review feedback into reusable traits, enforcing Redacted for secrets — samgoodwin89 · 2026-11-01
- Google's rakyll: agent quality hinges on locked-down, reproducible environments — IaC principles still apply — rakyll · 2026-11-01
- Pre-call guardrails that stop a runaway AI agent before the bill arrives — MaverikSh · 2026-11-01
- Browser-Based Video Mapping Built with Claude Code, Open for Agent Remixing — ismaelfaro · 2026-11-01
- Dev runs Codex-driven Blender 3D modeling agent through 8 iterations, hits prompt limits — ZeYanjie · 2026-11-01