Completed
27:55 - Coherence in long video generation
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
How a Moonshot Led to Google DeepMind's Veo 3
Automatically move to the next video in the Classroom when playback concludes
- 1 0:00 - Intro
- 2 0:47 - Veo project's beginnings
- 3 3:02 - Veo's origins in Google Brain
- 4 5:07 - Video prediction and robotics applications
- 5 7:45 - Early progress and evaluation challenges
- 6 10:30 - Physics-based evaluations and their limitations
- 7 12:18 - The launch of the original Veo model
- 8 14:06 - Scaling challenges for video models
- 9 16:02 - The leap from Veo1 to Veo2
- 10 19:40 - Veo 3’s viral audio moment
- 11 21:17 - User trends shaping Veo's roadmap
- 12 23:49 - Image-to-video vs. text-to-video complexity
- 13 26:00 - New prompting methods and user control
- 14 27:55 - Coherence in long video generation
- 15 31:03 - Genie 3 and world models
- 16 35:54 - The steerability challenge
- 17 41:59 - Capability transfer and image data's role
- 18 47:25 - Closing
