Seedance 2.5: Where to Cut
A pause after a spilled drink, and what I want from longer clips and editing a few seconds at a time.
Lab
Notes from working through scenes, references, and the tools used to make them.
MiniMax H3: Before Generation
How an old voiceover mistake shapes the way I read H3's open release.
Text Is Not Replaced by Reference
Why reference images do not make scene writing obsolete; they make the text layer more specific.
Scene Memory Is Not Mood
Why time, light, floor, furniture, material, and zone anchors should be treated as continuity tokens in AI video production.
Seedance Is Not a Storyboard Generator
A production note on treating Seedance as a local interpreter, not a memory-bearing storyboard system.
Structural Pass Is Not Production Pass
A field note on semantic QA, production readiness, and render-feedback routing for AI video prompt systems.
Seedance 2.0: Multi-shot Becomes Director Control
A note on Seedance as a move from single-clip generation toward multimodal shot-flow control.
Kling 3.0: Reference, Audio, and Storyboard Control
A note on Kling 3.0 as a shift from isolated clips toward reference-driven sequence control.
Hailuo: Performance, Motion, and Cost
A note on why practical video models compete on repeatable performance, not only peak demos.
HunyuanVideo and Wan: Open Models Become Infrastructure
A note on open video stacks, first-last-frame control, and why inspectability matters.
Veo 2: Cinematography Becomes Promptable
A note on why Veo 2's real shift was camera behavior, physics, and film grammar.
Luma Dream Machine: Video as a Creative Instrument
A note on keyframes, camera control, extension, and why iteration surface changes model value.
Character Placement and Continuity Tests
A practical note on identity, placement, and relationship anchors across shots.
Prompt Structure and Scene Rhythm
How character assets become shot prompts without overloading the model.
Building Reference Sets for Characters
Reference curation as a system for identity, silhouette, costume, posture, and emotional range.
Spatial Anchors: Terrain, Light, and Time
Using place, direction, light, and repeated objects as scaffolding for generated scenes.
From Script to Shot Flow
Turning narrative beats into ordered visual states, not just a list of pretty shots.
Camera Language in AI Video
Framing, lens behavior, and movement as continuity tools, not decorative cinema words.
Sora: The World-Simulator Frame
A note on why Sora changed the question from clip generation to visual state simulation.
Lumiere: Generating the Whole Motion at Once
A note on whole-clip generation as an answer to global temporal consistency.
Stable Video Diffusion: The Open-Weights Pivot
A note on why open weights changed video generation from a demo into material for inspection.
Runway Gen-2: The Browser Studio Moment
A note on why Gen-2 mattered as a creative interface, not only as a model.
Imagen Video and Phenaki: Fidelity or Time
A note on the early split between sharper video and longer narrative continuity.
Make-A-Video: Learning Motion Without Paired Captions
A note on the data problem behind early text-to-video generation.
CogVideo: Inheriting Image Models for Motion
A note on the early strategy of borrowing image-model semantics before learning motion.
Video Diffusion Models: Motion Enters Diffusion
A note on why video generation first had to become a problem of time, not image quality.