---
title: "Generative Video at the Speed of Light"
category: "talks"
date: "2026-07-01"
time: "2:25pm-2:45pm"
track: "Generative Media"
room: "Track 1"
speakers: ["Keegan McCallum"]
sourceLabels: ["Official conference schedule", "Public YouTube metadata"]
scheduleTrack: "Generative Media"
scheduleRoom: "Track 1"
scheduleLabels: ["Generative Media", "Track 1", "session", "confirmed"]
---
# Generative Video at the Speed of Light

## Conference Context
- Date/time: 2026-07-01 · 2:25pm-2:45pm
- Track/room: Generative Media · Track 1
- Speaker(s): Keegan McCallum
- Session type/status: session · confirmed

- Track: Generative Media
- Room: Track 1
- Session type: session
- Status: confirmed

## Session Description
Discussing recent breakthroughs in realtime generative video models, and the new architectural problems and bottlenecks involved in creating immersive, interactive experiences on top of these models.

## Media Evidence
[Dream Machine: Scaling to 1m users in 4 days — Keegan McCallum, Luma AI](https://www.youtube.com/watch?v=EY4O9M6AsWI) (speaker-match related prior/adjacent AI Engineer video; captions: English auto-captions).

- Source video: `youtube-EY4O9M6AsWI`
- Slide deck: [[youtube-EY4O9M6AsWI-reconstructed-slides|Reconstructed Slides: Dream Machine: Scaling to 1m users in 4 days — Keegan McCallum, Luma AI]] — 12 visible slide image(s); 12 HTML recreation(s).
![[assets/reconstructed-slides/EY4O9M6AsWI/slide-003.jpg]]
![[assets/reconstructed-slides/EY4O9M6AsWI/slide-004.jpg]]
![[assets/reconstructed-slides/EY4O9M6AsWI/slide-005.jpg]]
- Additional slide evidence: [[youtube-EY4O9M6AsWI-slides|Slides: Dream Machine: Scaling to 1m users in 4 days — Keegan McCallum, Luma AI]]
- Slide-derived themes for `youtube-EY4O9M6AsWI`: triton, mission, general, intelligence, generate, understand, operate, physical.

## 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-EY4O9M6AsWI` — 8 slide-derived text signals
- Slide-derived themes for `youtube-EY4O9M6AsWI`: triton, mission, general, intelligence, generate, understand, operate, physical.
- Evidence links for `youtube-EY4O9M6AsWI`: [[youtube-EY4O9M6AsWI]], [[youtube-EY4O9M6AsWI-slides]], [[youtube-EY4O9M6AsWI-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
- [[keegan-mccallum]]

## Supporting Slides
- [[youtube-EY4O9M6AsWI-slides]] — extracted from the related public AI Engineer video.

## Synthesis
### Synthesized Breakdown
The session explains MCP Apps as a way to add interactive UI to MCP-based workflows. It treats the MCP host as the place where tool data, model reasoning, and human controls meet.

### Speaker And Company Context
- [[keegan-mccallum|Keegan McCallum]] — Founder at [[urun|uRun]].

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

### Derived Links And Source Material
- [[youtube-EY4O9M6AsWI]] — related YouTube source page.
- [[youtube-EY4O9M6AsWI-slides]] — slide evidence.
- [[youtube-EY4O9M6AsWI-reconstructed-slides]] — slide evidence.

### Novel Concepts / Clever Methods
- [[mcp-app-runtime|MCP Apps As Agentic App Runtime]] — MCP Apps treats interactive UI returned from MCP servers as a runtime layer for agent-facing software.

### 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.