---
title: "Your Voice Agent is Just a Walkie-Talkie"
category: "talks"
date: "2026-06-29"
time: "12:05pm-12:25pm"
track: "Claws & Personal Agents"
room: "Track 6"
speakers: ["Neil Zeghidour"]
sourceLabels: ["Official conference schedule", "Public YouTube metadata"]
scheduleTrack: "Claws & Personal Agents"
scheduleRoom: "Track 6"
scheduleLabels: ["Claws & Personal Agents", "Track 6", "session", "confirmed"]
---
# Your Voice Agent is Just a Walkie-Talkie

## Conference Context
- Date/time: 2026-06-29 · 12:05pm-12:25pm
- Track/room: Claws & Personal Agents · Track 6
- Speaker(s): Neil Zeghidour
- Session type/status: session · confirmed

- Track: Claws & Personal Agents
- Room: Track 6
- Session type: session
- Status: confirmed

## Session Description
Everyone says cascaded voice pipelines are dead and native speech models are the future. Yet production environments are still dominated by STT-LLM-TTS stacks. Reconciling the natural flow of native audio with the elite reasoning of a cascaded agent remains an unsolved systems problem. This talk dissects the brutal technical trade-offs behind that counterintuitive reality. We will break down why your voice agent is still stuck behaving like a walkie-talkie and map out the specific technical roadmap required to build full-duplex AI that actually works.

## Media Evidence
[Voice AI: when is the "Her" moment? — Neil Zeghidour, CEO, Gradium AI](https://www.youtube.com/watch?v=P_RI1kCkRbo) (speaker-match related prior/adjacent AI Engineer video; captions: English auto-captions).

- Source video: `youtube-P_RI1kCkRbo`
- Slide deck: [[youtube-P_RI1kCkRbo-dense-slides|Dense Slides: Voice AI: when is the \"Her\" moment? — Neil Zeghidour, CEO, Gradium AI]] — 1 visible slide image(s); 1 HTML recreation(s).
![[assets/dense-slides/P_RI1kCkRbo/slide-001.jpg]]
- Additional slide evidence: [[youtube-P_RI1kCkRbo-slides|Slides: Voice AI: when is the \"Her\" moment? — Neil Zeghidour, CEO, Gradium AI]], [[youtube-P_RI1kCkRbo-reconstructed-slides|Reconstructed Slides: Voice AI: when is the \"Her\" moment? — Neil Zeghidour, CEO, Gradium AI]]
- Slide-derived themes for `youtube-P_RI1kCkRbo`: engineering, future, brig, team, google, translation, great, hove.

## Evidence Graph
This evidence graph is generated from currently linked source material: official schedule text, related video pages, cached transcripts, visible slide text, dense/reconstructed slide pages, and AI slide-classification audits.

### Media Signals
- `youtube-P_RI1kCkRbo` — 10 slide-derived text signals
- Slide-derived themes for `youtube-P_RI1kCkRbo`: engineering, future, brig, team, google, translation, great, hove.
- Evidence links for `youtube-P_RI1kCkRbo`: [[youtube-P_RI1kCkRbo]], [[youtube-P_RI1kCkRbo-slides]], [[youtube-P_RI1kCkRbo-dense-slides]], [[youtube-P_RI1kCkRbo-reconstructed-slides]]

### Agent Reading Notes
Use these signals to refine the synopsis, topic links, people/company context, and method notes. If a source is a related external video rather than an exact official recording, keep it framed as supporting evidence.

## Transcript Status
Related video transcript availability: English auto-captions. Treat this as supporting context, not a recording of this exact scheduled session unless later confirmed. Not fetched yet.

## People
- [[neil-zeghidour]]

## Supporting Slides
- [[youtube-P_RI1kCkRbo-slides]] — extracted from the related public AI Engineer video.

## Slide Evidence
- Slide-only cropped deck: [[youtube-P_RI1kCkRbo-dense-slides]] (2 viable slide images).
- Related slide/OCR pages:
- [[youtube-P_RI1kCkRbo-dense-slides]]
- [[youtube-P_RI1kCkRbo-reconstructed-slides]]
- [[youtube-P_RI1kCkRbo-slides]]
- Slide-derived terms: `engineering`, `future`, `braintrust`, `workos`, `openal`, `engineer`, `aiengineer`, `gradium`, `kyutai`, `real`, `raat`, `tdps`, `vastelnt`, `cofounder`, `brig`, `soos`, `google`, `deepmind`

## Synthesis
### Synthesized Breakdown
# Your Voice Agent is Just a Walkie-Talkie ## Conference Context - Date/time: 2026-06-29 · 12:05pm-12:25pm - Track/room: Claws & Personal Agents · Track 6 - Speaker(s): Neil Zeghidour - Session type/status: session · confirmed - Track: Claws & Personal Agents - Room: Track 6 - Session type: session - Status: confirmed ## Session Description Everyone says cascaded voice pipelines are dead and native speech models are the future. Yet production environments are still dominated by STT-LLM-TTS stacks. Reconciling the natural flow of native audio with the elite reasoning of a cascaded agent remains an unsolved systems problem. This talk dissects the brutal technical trade-offs behind that counterintuitive reality.

### Speaker And Company Context
- [[neil-zeghidour|Neil Zeghidour]] — Co-founder & CEO at [[gradium|Gradium]].

### Topics Covered
- [[agent-security]]

### Derived Links And Source Material
- [[youtube-P_RI1kCkRbo]] — related YouTube source page.
- [[youtube-P_RI1kCkRbo-slides]] — slide evidence.
- [[youtube-P_RI1kCkRbo-reconstructed-slides]] — slide evidence.
- [[youtube-P_RI1kCkRbo-dense-slides]] — slide evidence.

### Novel Concepts / Clever Methods
- No highlighted novel concept has been detected yet.

### Evidence Boundary
This synthesis is based on the official schedule and linked source pages. It should be revisited when exact session recordings or transcript-backed secondary sources are available.