Dreamina Seedance 2.0
summary
Claims enterprise workflows achieve ~10x content velocity using BytePlus models C-Dream 5.0 and C-Dance 2.0, plus an agent platform called Trey that runs IDE/co-work modes and tool hooks. Video pipeline accepts multi-modal inputs (first/last frames, image/video/voice references), generates 15‑second segments that require a continuity chain to stitch, and targets ads, e-commerce (under 5% of SKUs have video), film previsualization — demonstrated by a short film called Red and multiple product-demo variations. Agent governance and ROI tracking are proposed for enterprise rollout.
AI-written summary · about
chapters (12)
- 0:00Introduction and background
- 1:13Traditional vs AI content creation
- 2:37C-Dance 2.0 multi-model input
- 4:44Use cases for C-Dance 2.0
- 7:34Advertising video variations demo
- 9:00E-commerce video demonstrations
- 12:18Short film creation process
- 13:45Agentic workflow for movies
- 15:10Types of AI agents
- 16:13ByteDance’s Trey AI engineer
- 16:55Enterprise agent CI/CD workflow
- 17:18Agent governance and ROI
related talks
17mAI That Powers Experiences
12mAI Content Generation
14mAI That Moves Business. Cloud That Scales
13mCineVision: An Interactive Pre-visualization Storyboard System for Director–Cinematographer Colla...
20mRealtime Video AI with Diffusion Models
39mVideo Generation + AI: From Production to Product
14mAI-Video Ads – Built 90% Faster & 80% Cheaper
37mLive Fireside Chat