Creator demos mid-take body transformation with Luma agents as ultra-realistic home-video style replaces AI slop
SimplyAnnisa · x · 2026-09-04
Creator @mrjonfinger shares an experiment using Luma AI agents to transform a subject's body type mid-take. Alongside it circulates a full prompt for a 30-second 1080p ultra-realistic home video: a young woman in a quiet coastal neighborhood, with identity/outfit locked throughout and deliberate handheld artifacts—shake, autofocus hunting, exposure shifts, noise, compression—to shed the AI-slop look. Realism is clearly the direction video generation is heading.
More from Multimodal
- Clapper to launch macOS version this weekend, expanding open-source AI video creation — flngr · 2026-09-04
- Astra beats Fable 5.1 in video gen duel: samurai strike and ghost-story mood win — eyishazyer · 2026-09-04
- A reusable 'Data Mosh Surrealist' prompt template for glitch-art AI images — LudovicCreator · 2026-09-04
- FastH3 Ref2V: open-source controller for continuous character-driven AI video streams in ComfyUI — EarthDefenceForces · 2026-09-04
- RadixArk's Miles uses a shared post-training design for VLMs and diffusion models — ying11231 · 2026-09-04
- Cat Tales: Whiskerhold — a 25-minute AI-generated short film shared on Reddit — MosskeepForest · 2026-09-04