Full body waifus, AI dreams, realtime AI music, open-source Gemini Omni: AI NEWS
Open-source frontier models like ByteDance's Bernini and NVIDIA's Deja View are expanding local capabilities for video editing and 3D reconstruction.
What this lesson covers
This news roundup covers major AI releases, including ByteDance's video-editing model Bernini, NVIDIA's 3D reconstruction model Deja View, and Google's low-latency real-time music generator. The video details how these open-source tools can be run locally or integrated into production pipelines.
Key takeaways from the creator
AI-extracted notes, not independently verified product claims. Timestamp links let you check each point in the original video.
- 0:57 ↗
ByteDance's Bernini is a unified open-source video editing model capable of modifying videos using text prompts, images, or video references.
- 3:46 ↗
NVIDIA's Deja View is an efficient 3D reconstruction model that transforms multiple images into 3D Gaussian splats using a recurrent transformer block.
- 5:27 ↗
Google and Meta's Pager model handles panoramic geometry reconstruction to predict 3D depth and surface normals from single 360-degree images.
- 7:13 ↗
Google's Magenta Real-Time 2 brings music generation latency down to 200 milliseconds, allowing live control via MIDI and audio inputs.
Workflow outlined in the video
- Review the Hugging Face and GitHub repository pages for Bernini to understand hardware requirements and local setup instructions.
- Test NVIDIA's Deja View for efficient 3D scene reconstruction from multi-view image sets.
- Explore Magenta Real-Time 2 for low-latency live music generation within your audio production workflow.
Before you use this workflow
These notes describe the source video at its publication date. Model access, pricing, connectors and interfaces may have changed. Check the original source and the provider’s current documentation before installing an add-on, connecting an account or spending credits.
ReelStack has not independently tested this workflow. Preview one representative shot and check motion, continuity and output quality before applying it to a full production. No result, cost saving or model capability is guaranteed.
Explore the tools
How to interpret the numbers
152,938 views captured 23 October 2026. ReelStack recommendation score: 44/100. Formula: radar-v5. This is ReelStack’s calculation, not a YouTube rating or a measure of factual accuracy.
Show the calculation
- 45% performance: 1.26× views vs 11 other channel videos observed at a similar age. View-evidence factor 99% (views / (views + 1,000))
- 25% freshness: 0/100 with a 14-day half-life
- 20% momentum: 17 views/day vs 11 channel baseline (6 comparable uploads), with the same view-evidence factor
- 10% engagement: 75/100; likes + 4× comments, smoothed with a 500-view neutral prior