Gradium CEO: every voice agent today is a half-duplex walkie talkie
AI Engineer · youtube · 2026-09-15
A talk by Gradium co-founder and CEO Neil Zeghidour argues that every real-time voice model shipping today is half duplex — either listening or speaking — while real phone calls involve up to 20% overlapping speech. That gap is the case for full-duplex voice models.
Key points:
- Traces the arc from a 2011 phone assistant mapping transcripts to app actions, through open-ended voice chat with zero agency, to cascaded voice agents that call tools but inherit text-pipeline latency and flattened emotion
- Live demo: back-channeling ("mm hmm", "yeah") makes current models stop dead, exposing broken turn-taking
- Technical core: 8 words ≈ 3 seconds = 72,000 timesteps at 24kHz; attention cost grows with the square of sequence length, so raw audio can't fit in an LM. Neural codecs compress audio into learnable tokens; full duplex requires modeling two token streams at once
- Candid tradeoff: every naturalness gain costs intelligence under a fixed weight budget. Two paths forward: scale speech-to-speech models, or split interface from intelligence
More from Multimodal
- Open-source MoneyPrinterTurbo turns one idea into finished short videos, voice and subtitles included — Roger_M_Taylor · 2026-09-15
- Inside ElevenLabs' Internal Design Tooling: 40M Unique Voice Orbs Generated at Scale — NevFlynn · 2026-09-15
- OpenShorts: open-source UGC video studio hits 4,000+ GitHub stars, turns product links into AI ads — Roger_M_Taylor · 2026-09-15
- Flam's 26B MoE Falcon model returns first token in 30ms, specialized for Indic languages — testingcatalog · 2026-09-15
- 4DAnyone Falters on Coats and Odd Poses but Remains a Great Layout Tool — mickmumpitz · 2026-09-15
- Creator builds a template library for Gemini + HyperFrames to clone reels on demand — toolstelegraph · 2026-09-15